Proceedings of 2004 CANBERRA FORUM - Introduction to the RMA

RMA DVA ESO

10th YEAR 10th YEAR 10th YEAR

CANBERRA

Tuesday & Wednesday, 30-31 March 2004

Venue:Rydges Lakeside Hotel

London CircuitCanberra City

Proceedings of

2004 CANBERRA FORUM

1Repatriation Medical Authority

Proceedings of a joint

RMA, DVA & ESO Forum.

All correspondence to:

The Registrar

Repatriation Medical Authority

GPO Box 1014

BRISBANE QLD 4001

2 Repatriation Medical Authority

© Repatriation Medical Authority 2004

This work is copyright. Apart from any use permitted under the Copyright Act 1968, no partof this work may be reproduced or copied by any process without written permission of thepublisher. Requests and inquiries concerning reproduction and rights should be addressed tothe Repatriation Medical Authority, Brisbane.

The Repatriation Medical Authority shall not be responsible for the results of any actionsarising out of the use of any information in this publication nor for any errors or omissionscontained therein. The publisher, the Repatriation Medical Authority, and the authorsexpressly disclaim all liability to any person in respect of anything and of the consequencesof anything done or omitted to be done by any such person in reliance, whether whole orpartial upon the whole or any part of the contents of this publication.Published by the Repatriation Medical Authority, Brisbane, 2004

ISBN 1 920720 32 4

Publication Number: P1070


Table of Contents:Foreword...................................................................................................................................4Abbreviations and Acronyms..................................................................................................5Ministerial Opening Address ..................................................................................................6

The Hon Danna Vale MPThe RMA – Ten Years On & Its Evolving Role ....................................................................9

Professor John Kearsley, RMA MemberStress and Stressors ...............................................................................................................17

Professor Beverley Raphael, RMA Member“Normal” Population Abnormalities Versus Risk Factors.................................................23

Professor Andrew Wilson, RMA MemberCritical Appraisal and Causal Inference .............................................................................26

Professor John Kaldor, RMA MemberThe Expectations from Health Studies................................................................................32

Professor Ken Donald, RMA ChairmanProtective Effects of Exposures ............................................................................................37

Professor John Kaldor, RMA MemberSyndromes and Causes – Illness and Disease .....................................................................40

Dr Justine Ward, RMA SecretariatSummary of Issues Raised by ESOs ....................................................................................48

RMA Processes ..............................................................................................................49Issues Relating To Particular Factors .............................................................................53Stress and Stressors ........................................................................................................54Health Studies.................................................................................................................56

APPENDICES

1. Determination of SOPs Flowchart – 2004 Forum ..............................................................592. Background Terminology: Introduction To Some Epidemiological Terms.........................61

Dr Alex Bordujenko, RMA Secretariat3. Workshop Summary ............................................................................................................724. Bibliography/References .....................................................................................................755. Organisations Represented at the Forum ............................................................................78

Technical/scientific papers presented at the forum by invited speakers but notreproduced in this publication

Military Rehabilitation & Compensation (MRC) BillMr Bill Maxwell. Division Head, Compensation and Support, DVAMr John Douglas, Director, Compensation and Support, DVAMr Andrew Moorhead, Director, Compensation and Support, DVA

Korean War Cancer and Mortality StudyDr Keith Horsley, Director of Research Studies, DVA

Recent Initiatives in the ADFAir Commodore Tony Austin, Director-General of Defence Health Service, ADFColonel Stephan Rudzki, Director of Preventive Health, Defence Health ServicesBranch, ADF


ForewordThe 2004 RMA/DVA/ESO forum was held in Canberra on the 30 and 31st March. Theobjectives of the forum were to: restate RMA processes; inform representatives of the Veterancommunity; address issues of concern to the Veteran community; provide a forum forDepartment of Veterans’ Affairs (DVA) and Department of Defence (DoD) representatives toaddress current issues; and to provide the opportunity to identify future challenges.

The forum was attended by around 45 representatives of Ex-Service Organisations (ESOs) andrepresentatives of DoD, DVA, the Repatriation Commission, the Veterans’ Review Board andthe Specialist Medical Review Council, as well as the members of the RMA and the RMAsecretariat. It was opened by the Hon. Danna Vale, MP, Minister for Veterans’ Affairs.

The format of the forum involved presentations, a workshop on critical appraisal and causalinference, and question and answer sessions. Opportunity for the latter was provided at the endof each presentation as well as at specific sessions dedicated to answering questions submittedprior to the forum. There were thus opportunities for active participation by ESOrepresentatives.

The topics of the various presentations were chosen to explain RMA processes, elaborate onthe legislative and scientific constraints under which the RMA operates and to reflect currentissues of concern.

The purpose of this document is both to record the content of the proceedings and to provideveterans or their advocates with a helpful reference for matters which arise when dealing withthe Statements of Principles (SOPs). The principles and processes by which the RMAoperates, as outlined in the proceedings of the 1998 forum, are essentially reconfirmed in thisdocument.

All presentations given by members of the RMA are included, either as a paper or as editedversions of the audio transcript. Presentations given by guest speakers are listed but notincluded because they provided background information of a topical nature which was onlyindirectly related to RMA functions. All questions and answers from throughout the forumhave been brought together from the audio transcripts, grouped in broad categories and editedfor clarity. Additional useful general material is provided in the Appendix.

Formal and informal feedback received by the RMA about the forum has been very positive.Delegates felt that the forum met its objectives and was relevant and useful. We hope thereforethat this document will be a relevant and useful resource, both to those who were present andthose who, in the future, need to gain a greater understanding of the RMA and the SOPsystem.

Professor Ken DonaldChairpersonRepatriation Medical Authority


Abbreviations and AcronymsAPPVA Australian Peacekeepers & Peacemakers Veterans Association

ASASA Australian Special Air Service Association

AVADSC Australian Veterans & Defence Services Council

BMI Body Mass Index = W/H2 (W = weight in kg, H = height in metres)

DDT Dichlorodiphenyltrichloroethane

DoD Department of Defence

DVA Department of Veterans’ Affairs

ESO Ex-Service Organisation

GARP Guide to the assessment of rates of veterans’ pensions

MET A unit of measurement of the level of physical exertion. 1 MET = 3.5 ml ofoxygen/kg of body weight per minute or, 1.0 kcal/kg of body weight per hour,or resting metabolic rate

MRCA Military Rehabilitation and Compensation Act

PTSD Post-traumatic stress disorder

RDFWA Regular Defence Force Welfare Association

RMA Repatriation Medical Authority (the Authority)

RSL Returned & Services League of Australia Limited

SMRC Specialist Medical Review Council

SMSE Sound medical-scientific evidence

SOP Statement of Principles

TPI Australian Federation of Totally & Permanently Incapacitated Ex-Servicemenand Women

VEA Veterans’ Entitlements Act

VRB Veterans’ Review Board

VVAA Vietnam Veterans Association of Australia

VVFA Vietnam Veterans’ Federation of Australia


Ministerial Opening Address

The Hon Danna Vale MP

Minister for Veterans’ Affairs

Repatriation Medical Authority Forum

30 March 2004

Canberra


Good morning, everyone. I’d like toacknowledge Professor Ken Donald andmembers of the Repatriation MedicalAuthority and the RMA Secretariat staff, DrJonathan Phillips and members of theSpecialist Medical Review Council, Dr NeilJohnston and members of the RepatriationCommission, Brigadier Bill Rolfe, PrincipalMember of the Veterans’ Review Board, Ex-Service Organisation leaders, and ladies andgentlemen. It is a great privilege for me tojoin you today, to celebrate the achievementsof the Repatriation Medical Authority.

This year marks the 10th anniversary of theestablishment of the RMA. In just a decadethe Authority has established itself as anintegral and respected agency in our worldclass repatriation system.

The RMA was established in 1994 inresponse to audit findings that the veterandisability compensation system wassuffering from a lack of consistency in itsdecision-making.

The introduction of Statements of Principlesin 1994 by the RMA, was intended toprovide a firm foundation to ensure not onlythat payments were generous, but thatveterans were treated consistently and fairly.

The RMA has developed SOPs based onsound scientific evidence and has worked toensure the SOPs evolve as new studiesemerge on the health factors affecting ourservicemen and women.

Ten years on, the combination of the SOPs,expert computer systems developed by DVA– and a commitment to training staff and ex-service pensions officers – has delivered anenviable system for delivering compensationthat meets the needs of disabled veterans.

Processing times for primary claims havefallen and processing costs are down – bothan indication that the job is being done moreefficiently.

There has been a significant improvement inthe quality of decision-making and a notablereduction in rates of appeal for review ofdecisions – meaning that the job is beingdone more effectively.

Importantly, the partnership between theRMA, DVA and Ex-Service Organisationsalso has strengthened veteran confidence inthe integrity and fairness of thecompensation claims system.

When caring for veterans it is necessary tostrike a balance between the generoussupport that the community expects forthose who have served our nation, and acertain amount of rigour in applyinglegislative provisions.

While no system can be perfect, thearrangements we have in place today meanthat veterans can have high levels ofconfidence that they are being treatedconsistently.

A great deal of credit for the RMA’s successmust go to Professor Ken Donald.

Professor Donald has chaired the authoritysince its inception and his expertise,experience and dedication to the RMA’swork are greatly valued by the Government.

I would also like to express my appreciationto the RMA members and their Secretariatstaff, many of whom are also celebratingtheir 10th anniversary with the authority thisyear.

But while we are celebrating a significantmilestone, that doesn’t mean that we will beresting on our laurels.

In 2004, the repatriation system is poised toenter a new stage of its evolution, with theestablishment of the Military Rehabilitationand Compensation Scheme.

This will be a landmark moment in thehistory of Australian repatriation – the


creation of a single scheme designed to meetthe care and compensation needs of bothserving members of the Australian DefenceForce and the next generation of veterans.

I am looking forward to its speedy passageso that we can carry forward our intention tocommence the new scheme from the

1 July 2004.

The new scheme will bring together the bestelements of the Veterans’ Entitlements Actand the Safety, Rehabilitation andCompensation Act.

One of the most important elements of thescheme will be the inclusion of the RMA’sStatements of Principles as the determininginstrument for compensation claims.

The inclusion of SOPs in the new schemefollowed extensive consultation withveterans and serving members and reflectstheir successful use under the VEA.

While this decision has generally beenwelcomed by those who have played anactive role in the development of thelegislation, I would like to address concernsraised recently in the legal profession thatthe use of SOPs will make it more difficultfor ADF members to obtain compensationfor injuries suffered on peacetime service.

This quite clearly is not the case.

Statements of Principles already deal withall forms of service, from peacetime towarlike service. They could even be used incivilian compensation cases, as the RMAinvestigates all possible causal factors fordiseases and conditions that are covered bySOPs.

The legal concerns raised also overlook thefact that Statements of Principals are notstatic documents. They are subject to reviewto ensure they are kept up to date with thelatest scientific evidence.

And, where a condition is not covered by anexisting SOP, the connection betweenservice, of whatever kind, and the injury ordisease is made on the basis of the medicalevidence available.

Obviously, some changes may be necessaryin applying Statements of Principles to thehealth and compensation needs of servingpersonnel.

We have no intention of just superimposingthe existing system onto the ADF profile,without testing the adequacy of the SOPs forthat purpose and satisfying serving membersof their comprehensiveness.

The RMA has an excellent track record ofconsultation and of responding to thegenuine needs of ex-servicemen and women.

Professor Donald has already indicated tome the authority’s intention to consultwidely with the defence force communityonce the legislation has passed Parliament,to ensure that the Statements of Principleseffectively meet the needs of ADF membersunder the new scheme.

I might say that this is not the first time thatlegal arguments have been raised claimingthat Statements of Principles would make itharder for veterans to access theirentitlements.

I expect those involved in the establishmentof the Repatriation Medical Authority willrecall that similar concerns were raisedabout the move to SOPs to determineveterans’ claims under the VEA.

In this 10th anniversary year we can lookback and see the proof to the contrary.

I look forward to the Repatriation MedicalAuthority continuing its dedicated work onbehalf of all those who serve in the defenceof Australia.


The RMA – Ten Years On & Its Evolving Role

Professor John Kearsley, RMA Member

Paper prepared from edited transcripts of the RMA Forum

March 2004


The RMA – Ten Years On & 1998Canberra Forum Publication

My role, as the first RMA speaker, is reallyto set the scene for this conference, to offeran overview, to review our achievements andto review the limitations under which wealso work as the Repatriation MedicalAuthority. I’d also like to provide a sense ofthe journey that has led us to theseachievements.

When the RMA was first created 10 yearsago, the SOP concept was a largely foreignarea. There was no template, there was noblueprint and there was no history that thefledging RMA could reflect upon. What wehad was the Veterans’ Entitlements Act andsome very skilled people.

At this point I would like to pay tribute toour Chairman, Professor Ken Donald,because I think he has been an amazingexample of how to make an organisationforge ahead. Where there was nothing, therenow is something substantial, and ProfessorDonald has been, in significant part,responsible for that achievement. One of thethings Professor Donald has alwaysimplored to RMA members is that we needto act lawfully. It is important to understandthe framework in which we’ve had to work.

The first part of my talk is really an overviewor a review of our achievements, and I willhighlight seven achievements. The secondpart of my talk will be to review some of thelimitations that really are, to a great extent,not in the RMA’s direct control. It isimportant because sometimes individuals,and members of Ex-Service Organisations,do not agree with us or our views. It’simportant to understand that a lot of thistension is actually not our fault. We are oftenrestricted by either the poor quality of thescience, the poor quality of the publishedpeer reviewed literature, or sometimes the

legislation itself in respect of occasionalunrealistic expectations.

The first RMA achievement is theestablishment and refinement of the SOPsystem. Back in 1994, we didn’t really havea history and we didn’t know what a SOPwould look like. By the end of the year 2003,the RMA had produced nearly 1200instruments, covering a total of 276individual conditions and the percentage ofconditions covered by SOPs at the primarylevel in 2003 was 93.2%. The RMA hascovered the vast majority of the importantSOPs.

Initially, of course, the first priority was todetermine as many SOPs as possible. Wechose those SOPs for which we thoughtthere were likely to be many claims and forwhich there was good and reliable peer-reviewed published evidence. In the firstthree years after the inception of the RMA,229 conditions were covered, and 667instruments were determined. The emphasisnow has changed in that, while we are stillcreating SOPs, we now spend most our timereviewing SOPs because the nature of thepeer reviewed literature is that it is alwayschanging and being refined.

In each investigation the evidence for allpotential factors is reviewed. We havebecome very particular about how we writethe factors because, as you will be aware,there are a large number of verysophisticated SOP users. These include theveteran community, Ex-ServiceOrganisations, the courts, the AAT, the VRB,the SMRC and the wider community.

The SOPs are now readily available on theRMA Website, and so there is a lot morescrutiny as to what we write and the factorsthat are created. I will come back to thatissue of interpretation afterwards. Currently,of course, there’s a steady and increasingflow of requests for investigations, or review.


As I have said, new information accumulatesrapidly, so that most requests require areview of the whole evidence and thereforean investigation is advertised. Requests forinvestigations can also be received from theSMRC. There is now an expanding role forSOPs for serving personnel because of thenew MRC Act.

I might just go back to this point aboutreviews and the thoroughness of the SOPs.The current review of Malignant Neoplasmof the Prostate gland for instance will take inexcess of eight months to fully and properlyre-examine. So these reviews do take time.We review the whole SOP and all the factorsjust to ensure that no new factor has emergedor existing factors have changed.

Now, why do SOPs change? I want to clarifyfirst of all that, in fact, existing SOPs do not ofthemselves change. If we want to change aSOP, the SOP needs to be revoked and thenreissued. SOPs change for a number ofreasons. One, as I’ve said, is that there is newsound medical-scientific evidence. Uponreview a change may or may not be required,depending on the evolution of that informationin peer reviewed journals. There may beadministrative reasons for change- someonemay have left out a semi-colon, there might bean alteration in the format, or at times theremight be just too many amendments.

Sometimes also, there are problems ininterpreting the factors, or what we mean bya disease entity. I recall that there was greatargument at one stage of what we meant bythe “large intestine”. Now, for most peoplethe large intestine is the large intestine, butwhen we have to define it so carefully inSOPs we have to make sure that we knowexactly where the ileo-caecal valve fits. Is itpart of the large or small intestine? And doesthe recto-sigmoid include the largeintestine? These are issues about clarity ofdisease definitions.

Sometimes, we know what we think andwhat we meant, but other people think thatwe meant something different. The courts,ESOs and veterans are extremely helpful inhelping to clarify factors, becausesometimes what we write as a factor in termsof the exposure or the amount of theexposure is open to interpretation. Wecertainly thank the ESOs and other peoplewho have pointed out variableinterpretations of our factors to us.

The RMA flowchart on Determination ofSOPs shows the step by step process bywhich the RMA acts to produce a SOP (seeAppendix 1). When we receive a requestfrom an ESO, the Repatriation Commission,other organisations, or veterans, the RMAgoes through a process. The first step is; isthe condition under consideration a disease?If it’s not then we are unable, under thelegislation, to create a SOP.

The RMA has discussed the condition ofhiatus hernia, and this is just provided as anillustration of how hard it is sometimes toknow if an entity is a disease or not. Hiatushernia is where part of the stomach slides upinto the chest. We know from radiologystudies that at least 60% of the normalpopulation have a sliding hiatus herniawithout symptoms, and sometimes it’s anincidental finding. The problem for ussometimes is knowing whether a commoncondition can actually be considered adisease. What people come to the generalpractitioner with are symptoms, but if youhave a sliding hiatus hernia you may havereflux, and that reflux may cause symptoms.However, there are some people who have ahiatus hernia, no reflux and no symptoms,and other people can have symptomswithout reflux. Whether the “disease” ishiatus hernia, or whether the entity is“reflux”, occupies our time in discussion. Ina large number of cases the literature onparticular entities is unhelpful and


sometimes it’s hard to separate in thepublished literature reflux and symptomsfrom hiatus hernia itself.

The next step, after we have consideredwhether a condition is a disease under theAct, is to ask, “Is there published peerreviewed evidence?”. If there is, we thencollate a range of factors which can bepotentially implicated in the causation ofdisease X with potential for military serviceexposure. This is a huge task and ourSecretariat is involved heavily in this area.RMA members then view the evidence.Each RMA member brings with them theirown experience on whether a factoridentified can be considered “causative”.

We tend to use the option of clinical judgmentif we are considering a relatively rarecondition for which there is no peer reviewedpublished evidence on causation. One that Ihad personal experience with years ago isMerkel cell tumour of the skin. You willnever, ever, ever see a published paperlooking at causation in a strictly scientificfashion because the tumour is so rare.However, because I’ve had experience andseen patients with that condition, to me, it wasfairly obvious that this particularly rare skincondition is highly related to sun exposureeven though there is very little publishedevidence. You will see that in the skin cancerSOP, Merkel cell cancer therefore appears.

Once we have assessed causation, we thenapply the two standards of proof, either atthe reasonable hypothesis level or at thebalance of probabilities level. I will discussthose particular issues afterwards.

The second major achievement is thedevelopment of an RMA website whichcontains an introduction to the RMA, a shorthistory, profiles of the RMA members and alist of current investigations and reviews.The various RMA publications are publishedand there are useful links provided.

The third achievement is a more efficientreview process. Back in the year 2000 newpowers were introduced in amendments tothe VEA (section 196CA). The RMA is nowallowed to decline a request for formalreview if there is no new evidence provided.An investigation is a laborious process and ittends to tie up an enormous amount of RMAtime. This amendment means that we don’tneed to look at reviewing SOPs if there islittle in the way of new evidence. There isalso the power to collate requests so that theycan all be folded into the one investigation.In addition, we have had additional medicalresearch officers appointed to the RMAbecause of the volume of work.

The fourth achievement is independencefrom the Department of Veteran Affairs.That has been maintained, and we haveseparate legal advice and staff.

The fifth achievement is that the expertise ofthe RMA membership has been maintained.We have always had an epidemiologist andmost of us have epidemiology experience.The subject matter expertise of RMAmembers ranges across most diseases andwe all have a broad experience, whether thatbe in public health, public administration,research and/or clinical practice.

RMA members have varying periods ofappointment to avoid all of us retiring at thesame time and therefore a loss of corporatememory. It’s very important that we calibratewith each other so that the mind of the RMAin five years is the same mind of the RMA asit currently is, assuming no change in theVeterans’ Entitlements Act.

The sixth achievement is our pro-activeapproach to communication between theRMA and the community of SOP users.There’s never a sense that we have, increating SOPs, tried to slip one in undercover. Any proposed change is alwaysadvertised. We always have that transparency


so that people know what is happening withthe factors. In fact, ESO representatives arealways consulted if there is a potential tomake a factor harder to meet in a Statementof Principle or in deed, if a factor is to beremoved.

There are regular mail outs and publicationof annual reports and explanatory notes.Forums and conferences have been held andRMA members and the Secretariat attendrelevant conferences and meetings. Peoplealso understand that when they ring theRMA they will get a good informal hearingby either the Registrar or by other membersof the Secretariat.

The seventh and final achievement relates tothe RMA’s other roles. The RMA is notalways the RMA in that we haveoccasionally, as individuals, been appointedto various expert committees, and most ofthese expert committees or working partieshave been followed by publications andproceedings. We have looked at spina bifidaoccurrence in children of Vietnam veterans;the health effects of depleted uranium;radiation exposure; and the health of SASveterans. In the mid 90s there were bigconference, on smoking and malignantneoplasm of the prostate gland; and stressand health. There are some negatives in thatthe expert groups and conferences do takeup a lot of time for the RMA and theSecretariat.

The second half of my talk is an attempt toillustrate some of the legislative andscientific limitations which affect the RMAprocesses. I think most people understandthat the remit of the RMA is not to carry outresearch so I will not discuss that anyfurther.

We will consider these limitations underthree headings. The first is: “Whatconstitutes sound medical-scientificevidence (SMSE)?”. The second limitation

is, “How we define disease?”. The thirdlimitation is the standards of proof whichapply to particular types of service.

Section 5AB(2) of the VEA specifies whatconstitutes SMSE. SMSE is taken to beinformation that has been published in amedical or scientific publication and hasbeen subjected to a peer review process. Itcan also be information which, inaccordance with generally-accepted medicalpractice, would serve as a basis for diagnosisand management of a medical condition.Material on the Internet is often not peer-reviewed unless published by a reputableorganisation, so generally we don’t acceptInternet material as SMSE. Expert opinionsalso don’t usually count as SMSE, unlessthey are supported by published evidence.

I should expand on the word “hypothesis”,because there has always been someconfusion about that word. In the Veterans’Entitlements Act, “hypothesis” is used in alegal sense, whereas scientists think of“hypothesis” more in a scientific sense. Inthe legal sense, “hypothesis” means thatthere is substantial evidence that pointstowards a causal association between afactor or an exposure and a disease. Thelinkage must not be fanciful, and it hassometimes been quantified by peopleoutside the RMA as roughly a 1 in 20probability that factor X is linked withdisease Y.

However, in the scientific sense a hypothesisis really a hunch and when scientists conductexperiments they have to state theirhypothesis: “We think this is what happens.This is such and this is what we intend toprove”. In the legal sense, for the purposesof the RMA, the evidence needs to be therefirst.

I will now give some examples of theproblems with expert opinions. It would bepossible to pay an expert witness, I suspect,


to show that cannon fire causeshaemorrhoids. If you think about it, cannonfire is a shock that could increase strainingor increase pelvic pressure, and if repeatedover a time, could lead to haemorrhoids.That might have happened in the past, but ifyou look up the scientific literature, I doubtwhether there will be any evidence ofcannon fire as a causative agent inhaemorrhoids in the legal sense that we workwith.

Another example of expert opinion is a letterwhich came from an eminent heartspecialist. The issue at hand was whethersmoking was related to thrombosis, orclotting in the veins, and the cardiologistsaid:

“A specific cause for the thrombosis inthis person was not found, but there is awell-established connection betweencigarette smoking and venous thrombosisand it seems that cigarette smoking was afactor in its causation. In summary, then,this person has a history of femoral veinclotting and coronary artery disease.There is little doubt that cigarettesmoking is causally related to both ofthese conditions.”

That opinion was sent in as a letter withoutany supporting evidence. A review of theliterature between cigarette smoking andvenous thrombus, showed that:

“.. few controlled studies that dealt withvenous clotting alone were found thatsupported a positive association withsmoking, hence a smoking factor has notbeen suggested for the SOP on deepvenous thrombosis.”

The RMA takes seriously the opinion of amedical colleague. We would look up theliterature to see whether any literature hadchanged, but as people can see in thisinstance, the literature had not changed and

there was still insufficient peer-reviewedpublished evidence to support that link.

In the scientific literature it is part of themethodology of a study to clearly defineexposures and outcomes, whereas the RMAis often asked to investigate chemicals orgroups of exposures. We find out thatveterans are concerned about aircraftexhaust, or petroleum fumes or smoke asbeing causative factors for disease or illness.Our understanding, as the RMA, is that weneed to actually drill into what the chemicalcompounds are rather than just make a factorthat says “exposed to smoke”. It is ourunderstanding that we need to know thechemicals that cause harm, rather than just amixture of chemical substances.

Similarly, it is neither legally norscientifically correct to make factors forgroups of veterans. In the past the RMA hasmade a factor for prisoners of war. Thatfactor was made very early in the history ofthe RMA because it was almost impossibleto tease out what the noxious exposure reallywas. There were very few prisoners of warand, of course, they were subject to appallingconditions. Since then we have tried to stayaway from including groups of veterans inthe SOP factors.

We come now to the definition of “disease”,which is defined in section 5D(1) of thelegislation. I have already described theissue with hiatus hernia and tried tohighlight the problems that the RMA facesin dealing with conditions that are commonin apparently normal people. It is importantto understand that a disease does not includea temporary departure from normalphysiological state or the accepted ranges ofphysiological or biochemical measures thatresult from normal physiological stress.

Another limitation related to 5D is theinability to make a SOP for a condition thatis not recognised as a disease. For instance,


studies of Gulf War veterans in Australia andoverseas have found an excess of self-reported symptoms, but these are not able tobe classified as a disease under the Act.Constellations of symptoms do not alwayscluster in a consistent way to enable asyndrome or disease to be defined.

It is not clear sometimes when a risk factorbecomes a disease. For many years withinthe RMA, we have debated whetherconditions like obesity, hypertension or highserum cholesterol actually constitute adisease or whether they are simply acontinuum of normal life. The issue withhypertension has cropped up over manyyears and people will understand thathypertension, obesity andhypercholesterolaemia are certainly riskfactors for other diseases such as stroke.These are bad prognostic factors but whetherthe risk factors themselves actuallyconstitute a disease is open to some debate.

Of course, most of these risk factors aretreated, and so if you go to your GP and youhave high blood pressure or you are obese, oryou have hypercholesterolaemia, you willalmost certainly be treated. Takehypertension as an example. In the past, ablood pressure of 140 on 90 has beenconsidered to be “normal”, but if you have ablood pressure of 140 on 90, you can stillsuffer one of the side effects or one of thesequelae from hypertension. The lower yourblood pressure is, it is claimed, the longeryou will live, and blood pressure orhypertension is seen as a continuum ratherthan a fixed disease entity. The mind of theRMA in relation to this issue is still in a stateof flux. However, as uou would be aware, wedo in fact have SOPs for hypertension andmorbid obesity.

The next legislative issue relates to Sections196B(2) and 196B(3) which specify the twostandards of proof. The RMA must

determine the Statement of Principles inrespect of injuries, disease or death, settingout the factors that must, or must as aminimum, exist and which of those factorsmust be related to service. Our view is thatany factor, or any circumstance, can berelated to service, apart from genetic factors.

The RMA is not able to include a factorunless the body of SMSE points to orsupports a reasonable hypothesis of a causalassociation between the factor and theoutcome. This standard of proof applies toinjury, disease or death incurred onoperational, hazardous or peacekeepingservice. For the balance of probabilitiesstandard, which applies to eligible warservice and defence service, the SMSE hasto show that it is more probable than not thatthe factor in question is causally related tothe disease. For many diseases, despiteextensive review of the literature and a verygenerous standard of proof, few or no causescan be identified. Sometimes factors relatedto service cannot be included due to lack ofevidence.

In looking at causation, the RMA starts witha particular disease rather than starting witha particular symptom or exposure. Ingeneral, the RMA is not required to examinerisk factors or exposures and consider alltheir potential adverse health effects. TheRMA is required to focus on particular kindsof injury, disease, or death, as they relate toservice. The starting point is a disease ratherthan a risk factor or exposure. In addition,the RMA is not required to determinematters of fact concerning an individualveterans’ service record and the link betweenhis or her service and a factor. That is therole of DVA.

Synergistic effects are not usuallyconsidered. Let us take cancers of the tongueas an example. If you smoke, then you are atrisk of cancer of the tongue. If you drink


alcohol, you are at risk of developing cancerof the tongue. If you both smoke and drink,then you are at a much greater risk ofdeveloping cancer of the tongue. Theproblem here is that the literature often is notof sufficient quality that we can tease outexactly how much of a contribution smokingand alcohol consumption may contribute to agiven condition.

Likewise, the RMA does not considerpositive effects of exposures, only thenegative effects. That is in accord with thelegislation. Our remit is to considerexposures that are responsible for illness, notprotective effects. In looking through factorscausative for illness or disease wesometimes find protective factors, such asDDT in relation to breast cancer and alcoholin low doses in relation to heart attacks, butthey do not enter into our considerationwhen we create factors for SOPs. We mayhowever, consider inability to undertakeprotective actions, such as exercise orconsumption of dietary fibre.

The last issue is that of idiosyncratic factorsin disease causation. These are largelygenetic factors which we don’t take intoconsideration. One example is UV lightexposure and skin cancer. It is obvious that aperson’s skin has an impact on whether theyare predisposed to developing skin cancer.Some people have fairer skin than others,but this is again not a consideration that wetake into account.

In concluding my talk, I hope that I havebeen able to set the scene for speakers whowill follow. I have attempted to provide anoverview of the journey that the RMA hastaken over these 10 years, and also to pointout to you our achievements over that periodof time. As the RMA we will continue to actappropriately, carefully and, most of all,lawfully. We will continue to seek sound

medical-scientific evidence for causationand we will continue to be open to surprisesin our work. Thank you.


Stress and Stressors

Professor Beverley Raphael, RMA Member


March 2004



Even though my topic was psychiatricconditions and stress, the core issue whichcomes before the RMA repeatedly relates tothe whole issue of stressors, stress and therelationship to health. I think it’s very fittingthat this follows on from the discussion ofboth the strengths and the barriers to how theRMA functions, because in the area ofstress, the confusions that exist in thescientific literature and the advances thatcome progressively through a scientificexpansion in this field, have to inform whatwe do. Yet this has still left us with thingsthat may appear nebulous and difficult whenwe try to relate them to the real lifeexperience of veterans for matching againstSOPs.

Now, stress is a broad concept. In theEnglish language and probably in manyother languages there are similar or differentconcepts related to cultural perceptions.However, in the scientific literature it isoften very poorly defined and variablymeasured. That has to be on the table fromthe beginning because, as indicated in theearlier presentation, we have to base ourdecisions for the SOPs on the scientific andmedical literature.

We think it is useful to consider separatelythe stressors, the things that happen, eventsor ongoing circumstances, and stress, thereaction. The reaction may havepsychological, physiological, biological,social or cultural components.

When we think about stress and its effects,stress may range from something like beingin a motor vehicle accident to the fact thatstress is part of everyday life. From themoment a baby is born and indeed from themoment a baby is conceived, there arestressor effects which may impact ondevelopment either biologically,psychologically or socially. The adaptations

made, the coping strategies evolved and thepsycho-physiological mechanisms ofresponse may be linked to genetic pre-dispositions as well as to learning processes.These are all part of the fact that stress ispart of life.

In our earlier conference on stress we alsolooked at stress as challenge because it isquite clear that without adequate stimulus,which may be stressful at times, there is aninadequate process of development. Theresponse to the challenge component ofstressor exposures is often a critical aspect ofdevelopment and, as has been suggestedincreasingly by some research, it’s necessaryfor personal growth. There are nowquestionnaires that are looking at personalgrowth as a consequence of what might becalled psychologically traumatic stress.

Stress exposures, stress and individualreactivity can be influenced by geneticfactors and there’s a body of research whichhas looked at behavioural genetics and therelationships between reactivity to stress andpatterns of genetic connection. Learning canalso affect the response to stressors. If yougrow up in a very anxious environment; sayfor example you have a parent who isworried and anxious about how some threatmay hurt you, you may learn to be morereactive to external stressor exposures.

We know from a range of studies in theliterature that people are variously exposedto stressful exposures. When we look at theend of the spectrum of what might be seen aspotentially psychologically traumaticexposures, we know that while everybodygets some exposures in those circumstances,certain groups in the population may haveexcessive exposures. The question that thenarises is what part of that individual or thatgroup might seek or be vulnerable to greaterexposures to stressors for a variety ofpotential reasons.


There are social and cultural factors whichaffect the responses to stressors. We woulddefine something as bad for example ifsomeone attacked you savagely as youwalked out of a social venue. On the otherhand, if you were tackled the same way onthe football field everyone would be saying,“This is great.” Whether or not you ended upwith the same wounds, you might perceive itdifferently and its cultural and social contextwould be quite different. You would be seen,if your team won, as a victorious person andthe significance of the stressors and thestress effects would be seen as a greatcontribution to your team’s positiveachievements. In the other situation youmight be seen as a victim.

People’s perceptions of stressors come up inthe discussions about criterion A in post-traumatic stress disorder (PTSD). Perceptionis frequently part of what we try to take intoaccount if we can measure the impact itmight have as reported in the scientificliterature.

We also know that it is critical, if we arebeing fair to science, to bear in mindprotective influences, resilience andpersonal strengths. These factors might alsoinfluence how we respond to stress. Studiesof children and others in very adverse andhighly stressful environments have shownthat there are personal and othercharacteristics that favour resilience so thatthere is not a negative outcome from theexposure.

Positive support from the environment bothbefore, during and after an exposure, mayinfluence outcome as may learning andtraining. For example, in my work in thefield of disaster (which also relates to themilitary) we have found that people who areprepared and exercised in handling certainstressors or exposures are likely to havebetter outcomes in terms of their mental

health. Much of this may relate to the degreeto which we perceive or learn or developcontrol over the potential stressor, ourcapacity to have some control in our reactionto it, and the skills and a sense of masterywhich may come from past experience. Pastexperience may make you vulnerable but itmay also give you skills and strengths inhandling the next exposure.

Another important factor in stressorreactions is that when we look back we oftenattribute things to particular life experiences.How we separate real causes from thatattribution is a difficult thing in science, as itis clinically. We know, for example, thatsignificant events in our lives are oftenremembered very clearly. Whether thatmemory is on a spectrum with the traumaticmemories of a psychologically traumatisingexperience often remains to be identified inthe process of assessing its impact as well asthe relationship of the memories to it.

Recent research has suggested that even withwartime experiences, which may seem to bequite clear cut, there may be change overtime in the memory of what actuallyhappened. So, what actually happened, one’sperception of the event and the impact ofmemory progressively changing and dealingwith it in different ways, may mean that inretrospect the event takes on a greater or alesser significance as a potential causativefactor.

We know there are often quite significantdifferences between an acute exposure to aone-off, major, horrendous life event andmore chronic stressors which may seem tobe at a lower level and yet nevertheless mayseem to impact in a range of ways on healthand mental health. These differentiations areoften not clearly distinguished in theliterature and yet the science may point toone or other of these components as beingcritical in aetiology. How many, how


frequently and how often we are exposed aswell as how sudden, unexpected anduncontrollable stressor exposures are, maycontribute to their impact. However we knowthat we may actually attribute something to alife experience as a cause when it may ormay not be a cause.

Some years ago, some researchers looking atpeople who had a range of psychiatricillnesses asked them what they thought werethe causes. Most people attributed theirillnesses to stressor exposures. It is a strong,social attribution that we believe ourselves,if we’ve been stressed, that it may make ussick. It may certainly make uspsychiatrically unwell but we may believethat stress might have contributed to ourillness, very often without the science thatcan support our contention.

There is nobody in this room, includingmyself, who would not agree that war isstressful for all involved and all who watchand all who know anything about it. I startedmy life as a young general practitioner in aveterans’ practice. The principal in thepractice had survived Changi so I grew up asa doctor working with veterans and I endedup a psychiatrist, I am sure, because I keptsaying, “What happened to them in thewar?”. Nobody else seemed to be askingthose questions.

Dr Lars WeisÊth, who is a Professor ofDisaster and Military Psychiatry at theUniversity of Oslo and a consultant toNATO, suggested that there are three formsof war stress: shock traumas of briefduration, repetitive or serial trauma andprolonged exposure to danger characterisedby varying degrees of predictability andcontrol. Marshall, another researcher in thisfield, again talks about a range of lowintensity events which people might see ascoming into play in what is sometimes calledthe malevolent environment. This consists of

the more chronic type of events as opposedto high magnitude events and conditionswhere people’s lives are threatened. So, evenwhen people have tried to look at stress andwar, operational definitions of what stressmight be have varied between researcherswho are extensively experienced in lookingat this field.

Not all stressor exposures lead to problemsof illness and we know now from a largenumber of studies of soldiers and many otherexposed populations that even with severeexposures to horrendous traumas, noteverybody is likely to either develop PTSDor to have a psychiatric condition as aconsequence. For example, in a US study ofVietnam veterans, 15% were reported tohave PTSD.

In the general community in Australia thelevel of post-traumatic stress disorder foundby a national epidemiological study was3.5%, which is quite high. It is linked to arange of exposures to trauma in thecommunity. We know too that while combatexposure might be one of the high riskfactors for developing an illness like post-traumatic stress disorder, military servicealso has positive effects. There are a range ofstudies which suggest that these might be todo with learning, development of personalstrengths, being part of a team, coping skillsand a sense of independence and maturity.

Many young men and now young womencome in to the services at a time of personalmaturation- late adolescence and youngadult life. It is a time of development andgrowth, where there are particularvulnerabilities but opportunities for strength.Social cohesion, mastering the experience,personal characteristics, training,experience, and recognition and responsefrom others, are critical factors for copingwith an exposure.

If other people respond to you by saying, “It


was nothing” or “It’s your fault”, it tends tomake it more likely that this will be adifficult issue for you. When I was a generalpractitioner working with veterans I couldnever understand why their entitlement cardswere labelled, “Inadequate PersonalityDisorder,” because it seemed to me going towar meant you were adequate in the firstplace. There was a sense that people oftenfelt they were blamed for what had happenedto them, or that it was some reflection onthem as people. We know that betterrecognition is critical.

We are well aware that the RMA have arange of definitions of severe stressors.These have evolved in different waysbecause of the literature relevant to theparticular conditions and the science thatwas available at the time that the SOPs werebeing prepared. We understand this does notmake things easy, and indeed in someinstances we may not have adequatelyclarified some of these things because theliterature was uncertain. This will lead us tolook much more closely in a review of thesestressor exposures.

This links to the work that’s been done forDSM-IV, and before it DSM-III, in trying todefine better what a stressor exposure is.Encompassed in the DSM-IV definition ofan acute stressor is the reaction to thestressor: “Which event or events mightevoke intense fear, helplessness or horror”The need for a reaction in the definition hasbeen debated frequently. Some scientificliterature suggests that there can be adissociation from reactions in which it’s asthough you weren’t there. This dissociationmay be indicative of heightened risk ofdeveloping a condition later. So, sometimesthe definition includes the response andsometimes it doesn’t.

Where there has been exposure tocatastrophic stress a consequence may be

enduring personality change. This mayreflect part of a spectrum of PTSD. The term“psychosocial stressors” refers to a differentlevel of stress, and can be related to thebroader range of conditions than the veryspecific type of stressor identified for post-traumatic stress disorder. Here we look at therange of things that could occur that wouldbe extremely distressing for us.

In our own attempt to be more scientific,there is often a complexity betweendescribing the exposure and the reaction toit. In the scientific literature there is often noclear separation of the exposure and thereaction to it and there is a broad, ill defineduse of the word “stress”. The confusion inthe literature makes it difficult for the RMAto make conclusions which can be applied ina SOP.

Some of the questions that have arisen fromESOs, from veterans and from the courts areincluded here. I’m putting these up, notbecause there are sound scientific answers tothese, but because they are commonquestions that come up across the spectrum.

The first question is, does the stressorcomponent for PTSD have to be an actualthreat of death or serious injury? Does ithave to be a threat which anyone couldobjectively judge, or is it really to do withperceptions? What do we mean when we saythe person was confronted by an event? Ifyou’re rung up and told about an event, isthis stressful? What about your individualperceptions? If you thought it meant you ora loved one were going to die, is thatobjective or is that a perception? What if youfound out later that there wasn’t really abasis for the perception of threat?

Another question is what types of eventsmight constitute stressors, and how severemight they be? People have made manyattempts over the years to grade life eventstressors. In the early work of Holmes and


Rahe the death of a spouse rated the higheston the scale and we have to go from this totrying to identify the rating corresponding tobeing exposed to a terrorist threat, or indeedto any other horrific life threateningexperience. It’s interesting to note in theaftermath of September 11 there have been avery large number of studies in the USlooking at the rates of PTSD in thepopulation. New York hotels were bookedout as soon as September 11 happened withpeople coming in to do trauma counsellingfor everybody.

As it turns out, the better literature ishighlighting the fact that resilience wasstrong and early high levels have settleddown to quite low levels in the generalpopulation. We have to be very careful thatin our understanding of stressor exposure weare both appropriately recognising thepotential impacts as perceived by membersof the armed services and veterans, as wellas protecting people from developing adisabling view of the experience.

The RMA is committed to understandingand supporting the reality of the nature ofthe conditions described, and veterans’experiences, but also making sure that we dothis properly. To do it properly we have toprovide the scientific basis and in ourongoing review we have to answer some ofthe questions that have arisen and been putto us by veterans or the courts. We have totake into account what evidence there isabout subjective and objective realities.

The question of the malevolent environmentas a more profound and ongoing stressor hascome up but is still in the early stages ofstudy in the scientific literature. Otherquestions being studied are the nature of thedifferent stressor exposures and experiences,how memory may alter and change these inadaptive and non-adaptive ways, and howexposures may lead to different outcomes

and disorders. Our work, of course, then hasto be picked up in the clinical side for theassessment and diagnosis of individualveterans. Thank you.


“Normal” Population Abnormalities

Versus Risk Factors

Professor Andrew Wilson, RMA Member


March 2004


“Normal” Population AbnormalitiesVersus Risk Factors

When does something which is basicallyfairly normal for our society becomesomething which is abnormal? This issuegoes back quite a long way, and perhaps theeasiest way to look at it is to take theexample of blood pressure. When peoplefirst started to think about high bloodpressure as being a medical problem, back inthe pre-60s days, there was a thing called“malignant hypertension”, which was adisease in which the blood pressure was veryhigh.

These people had systolic blood pressureswhich were over the two hundreds, whichwas frequently rapidly fatal. People wouldhave strokes or renal failure within years ofthe development of this condition. Therewere medications that were available to treatit, but they were very limited in terms ofwhat they could do, and they had quiteserious side-effects.

Over time various studies started to point tothe fact that you didn’t have to havemalignant hypertension, but having highblood pressure which was something lessthan that malignant hypertension also placedyou at increased risk of heart disease, strokeand renal disease. The development of newmedications which didn’t have quite thesame side-effect profile meant that peoplewith high blood pressure could be treated ata lower level. The level at which we’ve beenprepared to treat high blood pressure, thelevel which is called hypertension, hasprogressively decreased, and we now accepta level of around 120 systolic as being alevel above which we think somebody hashypertension.

We now know from studies of over a millionpeople, studies from countries all around theworld, China, Australia, the UK, US and

Europe, that even as you go below thoselevels that we call hypertension, people whohave higher levels of blood pressure have ahigher risk of stroke and heart attack thanpeople who have lower levels of bloodpressure. However, there is still only onegroup that we’re calling hypertensive, that isthose people whose blood pressure is abovethis magic mark of 120, which is the pointwhen we start to treat it.

A similar problem exists in relation to bloodlipids, or cholesterol levels. I am part of acommittee which is being convened for thethird time in 10 years to look at what levelsof blood cholesterol we should treat withdrugs. Progressively over that 10 year periodwe have seen a lowering of the threshold aswe started to understand that lower levels ofcholesterol seem to be associated withincreased protection from heart disease andstroke. One in seven dollars of a budget ofabout $7 billion for the PharmaceuticalBenefits Scheme in Australia now goestowards supporting cholesterol loweringtherapy.

What is more, the most recent data suggeststhat you can treat people down to a level ofat least 3.5 and probably even get people’scholesterols down to a level of about 2.7, andthey’ll still gain some benefit for it. Now, ifwe look in the Australian population there’svirtually nobody who has a cholesterol levelof 3.5 or lower. It’s the exceptional person,perhaps elite athletes or SAS servingmembers, who would have cholesterol levelwhich might get down that low, and evenmany of them will have levels which arehigher than that.

So, we’re going to have to make a decision –do we give everybody in Australiacholesterol lowering therapy? Do weconsider that everybody has an abnormalcholesterol as the case may be? Clearly, as asociety, that’s a fairly significant issue that


we’re going to have to come to grips with.

That brings me on to the last item that I wantto talk about, and that’s weight. In theAustralian population weight is increasing.It increases with age across the population,and it’s unfortunate that we are now alsoseeing an epidemic of obesity in youngchildren.

What do we mean by obesity? There is oneset of definitions which have come out fromthe World Health Organisation which statethat a BMI of 25 to 29 will be calledoverweight and a BMI of 30 or over will becalled obese. In and of itself beingoverweight or obese is not unhealthy. I don’tconsider myself to be sick just because Icarry an extra 10 kilos of weight. I will startto think of myself as sick if I develop somecomplications as a result of that, and I’dprobably think of myself as sick if I was madenough to let somebody try and treat that insome way.

The difficulty that the RMA is facing intrying to deal with this is that by the basis ofour legislation we are required to makeSOPs for diseases. How do we handle thesethings which are almost normal patterns inour community and how do we try and thinkabout the association between that and someexposure that may or may not have occurredduring service?

That’s a challenge, and it’s one that we’regoing to have to continually keep underreview. We’ve tended to take a view that if adoctor diagnoses it as a treatable problem orif there is some complication, then we’ll callit a disease. Some examples arehypertension and sliding hiatus hernia.

Will it stop there? Is it going to stop withweight? Is that the last thing that we’re goingto have to look at? No, there are other thingsthat are already here that we’re going to haveto consider in the same way. Another issue

that we’re going to have to consider, forinstance, is low bone density orosteoporosis.

Osteoporosis is a disease defined byreaching a certain threshold of bone density,much in the same way as hypertension.There’s a lot of argument going on in themedical community about the cut-off pointat which we define osteoporosis. It is beingargued that perhaps it is too high. Again, weare dealing with a condition which willaffect virtually all women once they becomepost-menopausal, and which is increasinglybeing recognised as a common problem formen, particularly as men survive longer.

Although we don’t necessarily have anyanswers, we felt we needed to discuss theseissues, to try and give you an idea of the typeof thinking that we have to go through. As asociety we are trying to face this problemwhich is in almost plague-like proportions,but which is not easily understood in thetraditional way that we think about diseaseand abnormality physiologically. Thank youvery much.


Critical Appraisal and Causal Inference

Professor John Kaldor, RMA Member


March 2004


Critical Appraisal and CausalInference

Introduction

The primary purpose of this presentation isto convey the main scientific ideas behindcritical appraisal and causal inference, thetwo fundamental processes involved in thecreation of a Statement of Principles by theRMA.

The development of the SOP always beginswith a literature search. The RMASecretariat uses computer databases and arange of other mechanisms to identify allpublications that might conceivably qualifyas “sound medical-scientific evidence” ofrelevance to the SOP. Each publicationgenerally reports the results of a single studythat investigated the role of one or morefactors in causing the disease referred to bythe SOP.

Once these publications have been gatheredtogether, the process of critical appraisal isapplied to each one in turn. Critical appraisalis a systematic method developed in the fieldof health research for reviewing the resultsof published studies. Although it demands agood understanding of the principles ofepidemiology and statistics, criticalappraisal can be undertaken by people whoare not specialised in the content area of theresearch being reviewed. Indeed, the wholepremise of critical appraisal is that a generalreader can identify standard features ofstudies from their published reports, and thatconsideration of these features can thenallow a judgement to be made on the qualityof the study and the implications of itsfindings.

At the RMA, the literature search andcritical appraisal steps are undertakenbetween the regular monthly meetings, andresult in what is known as the submissionrelating to a proposed SOP. At the monthly

meetings, the RMA members jointlyexamine the submission, and use themethods of causal inference to determinewhich, if any, of the various factors that havebeen considered in relation to a particulardisease are causal. Causation can bedetermined at the reasonable hypothesis orbalance of probabilities level.

What is a cause?

Before discussing critical appraisal andcausal inference in more detail, it will beuseful to consider what we actually mean bythe cause of a disease or an injury. For someconditions, causes seem obvious: forexample, a vehicle crashes, a person whowas perfectly healthy before the crash isrescued and is found to have a fracture. Inthis situation, it seems indisputable that thecrash caused the fracture. There is no reasonto look for other causes, even though there isa remote probability of another factor havingbeen responsible. Nor do we feel compelledto do a literature search to find whether thereare any published studies that compare thenumbers of fractures in people who have justexperienced a road crash with the numbersin those who have not!

Similarly, if a person goes on a long marchon a hot day and at the end of the day hasblistered feet and a headache, there is littledoubt about the causal pathways. It isentirely reasonable to conclude that ill-fitting footwear caused the blisters anddehydration from sun exposure caused theheadache. Causation is unequivocallydemonstrated by the close proximity in timeof the causal factor and its effect, incombination with the obvious physicalpathways that provide the connectionbetween the two.

Defining a factor as a cause of a disease ismuch more complicated for diseases thatoccur a long time after exposure to the


factor. If a person is exposed to herbicide inthe context of Vietnam service, or someother situation, and then 20 years laterdevelops leukaemia, was exposure to theherbicide the cause?

We know little about the causes ofleukaemia, and the person could have beenexposed to many other factors both beforeand after the herbicide exposure. In thesesituations, a factor can only be determined tobe a cause of the disease in question if thereis corroborating evidence fromepidemiological studies. Specifically, weneed studies that determine whether therewas a higher rate of disease occurrence inthose exposed to the factor than in those whowere not. Although “clinical judgment” andanimal experiments may play some role, werely almost entirely on the humanepidemiological studies to tell us whetherthere was a difference in the disease ratesbetween those exposed and unexposed to thefactor.

These studies are essentially measuring therelative probability of the disease occurringin people who are exposed, compared tothose who are unexposed. If the probabilityof disease is higher in the exposed group,then we can say there may be an associationbetween the factor and the disease. Once anassociation has been established, we can thengo to the next step and consider whether itlikely to be causal. It is important to notethat a factor may be associated with adisease, without actually being a cause.

Through the science of epidemiology, anumber of different methods have beendevised for comparing people who areexposed to a factor with those who are not,and determining whether exposure isassociated with the development of disease.The various methods have various strengthsand weaknesses, and none is perfect, norapplicable in all situations. They do,

nevertheless, generally result in the estimateof a measure of association, usually oneknown as the relative risk.

If there is no association at all between thefactor and disease, the resulting relative riskis one (Figure 1). Simply stated, the amountof disease in the exposed group is the sameas in the unexposed group, and the ratioresulting from dividing one by the other isone. A relative risk substantially above oneprovides a strong suggestion that exposuremay be causally related to the disease. Forexample, a relative risk of three means thatpeople in the exposed group were found tohave the disease three times more often thanthose in the comparison group.

Figure 1. Relative risk

Sometimes studies show that the amount ofdisease in the exposed group is less than theamount in the unexposed group. Thisfinding suggests that exposure to the factormay actually be protective, in the sense thatit reduces the chance that the disease willoccur.

The RMA does consider protective factors,but more often is faced with studies thatreport factors that seem to have a weak ormoderate effect, in that the relative risk isincreased above 1.0, but not by very much.Most often, relative risks are reported in therange 1.25 up to about 1.75 (indicatingincreases in risk from 25 up to 75%),providing some suggestion of causality, but

0

0.5

1

1.5

2

2.5

3

3.5

4

protective none weak strong

interpretation

rela

tiv

e r

isk


still leaving room for doubt unless thestudies are of very high quality, or there area number or studies of reasonable quality allshowing a consistent result. Relative risksthat are well above two are hard to dispute asproviding strong evidence of causality. Therewould usually have to be some majorweaknesses or inconsistencies in availablestudies for the RMA not to draw aconclusion of causality in this circumstance.

Steps in critical appraisal

Evaluation of study quality is the role ofcritical appraisal. The published record ofeach study provides the raw material, and asystematic approach to reading the paper istaken to examine exactly what methods wereused to conduct the study, and thereby assessthe likelihood that the published relative riskis valid.

In carrying out its critical appraisal, theRMA depends crucially on the quality ofpublished reports. Even if a study was offundamentally high quality and published ina prestigious journal, it is not always clearlydescribed in the written report. Aconsiderable amount of experience in criticalappraisal is required before a practitionercan confidently decide what has been leftunsaid or partially stated in a publishedreport.

When conducting a critical appraisal of apublished report, the first step is todetermine what the author of the paper wasinvestigating. Usually there will be a statedresearch question, such as “does exposure toherbicides cause leukaemia?”, stated eitherin the title of the paper, the abstract, or theintroduction.

Next, it is essential to decide what particularstudy design is being used by the researchersto answer the question. Generally, the designwill be a randomised trial, a cohort study, acase control study, a cross-sectional study or

a correlational study. Many publications willstate the design, but a number do not, andspecialised knowledge is required to be ableto infer what design was used.

Each type of study has different qualities andweaknesses as an epidemiologicalinformation gathering device. Some of themhave strengths in being able to measure thelevels of exposure, while others are moresuited to accurate assessment of diseaserates. They all have fundamental limitations,and none is perfect.

Having identified the study design, weproceed to considering what factors arebeing investigated as possible causes of thedisease, and how they are being measured.Ideally, a study has been able to measure theexposure to the factors directly but, moreoften, epidemiological investigation relieson surrogate measures of exposure. Forexample, to determine whether a person wasexposed to a herbicide, the most directapproach would be direct measurement ofair or blood levels at the time of exposure,but such information is rarely available. Thestudy may therefore have relied on indirectmeasures, such as the self-report of the studyparticipants (often years after exposureoccurred).

Critical appraisal also requires that thereviewer of a paper determine how the studyassessed the presence or absence of thedisease of interest. Again, the ideal will bedirect, objective determination by theresearch team, but there are some diseasesthat can only be assessed indirectly. Forexample, a study of headache relies on self-report to make the diagnosis.

The main results of a study are generallypresented as relative risks, accompanied bysome measure of their statistical precision.For this purpose we use confidenceintervals, which give an indication of therange that there might be around the


estimate of the relative risk. Anotherstatistical adjunct is the P-value, whichassists in judging whether any observedassociation may be a chance finding.

Once the basic elements of the study havebeen established as clearly as possible fromthe published report, the process of criticalappraisal then involves assessing the extentto which chance, bias and confounding mayhave operated as alternative explanations ofthe reported findings.

If we see a relative risk that is clearlyincreased above one, there are generallythree broad alternatives that must beexcluded.

First, it could be a chance finding. Perhapsthe people exposed to the factor were at riskof getting the disease more than the peoplewho were not exposed, for reasonscompletely unrelated to the presence of thefactor, for example their genetic makeup.Although generally impossible to prove,when studies are very small, chance cannever be ruled out as the reason for theobserved difference.

A second explanation is that the study wassubject to some form of methodologicalbias. Ideally study measurements are carriedout in an objective way, but in practice, asubjective element is often introduced as aresult of the study design. Consider, forexample, studies that collect information onpast exposure from people who have aserious illness and a comparison group ofpeople without the illness. It is entirelyplausible that the recollections andperceptions of the two groups may differ,even if they in fact had identical exposurehistories.

The third circumstance that might give riseto apparent associations is the phenomenonof confounding, whereby people exposed tothe factor of interest differ systematically

from those unexposed, with regard to someother factor that is actually a cause of thedisease. Critical appraisal provides astandard framework for reviewing publishedreports of studies, and determining thepotential role of one or more of thesealternative explanations. With so manypapers published in so many journals, by somany different authors using so manydifferent approaches, critical appraisal hasbecome an essential tool for reducing theresulting information to a common,comparable form that allows it to besynthesised in an objective manner.

Evidence for causation

Finally, the RMA takes the informationresulting from the critical appraisals of allavailable studies, and forms a judgement asto whether there is a basis for determiningcausal relationships between specific factorsand the disease under consideration. At thispoint, there are a number of criteria that canbe considered, but there is no simple formulafor making the determination.

Most important seems to be the strength andconsistency of an association. If ten studiesof the relationship between a factor and adisease all show a relative risk of four orfive, it is very likely to be identified as acausal factor, at both the reasonablehypothesis and balance of probabilities level.

Big relative risks are hard to dismiss ashaving alternative explanations, andgenerally provide strong support forcausality, but they are not necessary for thedetermination of causality. A factor canincrease the probability of disease by twentyper cent or less, and still be a true cause. Inthis situation, other criteria are oftenassessed to support causality.

If a study has been able to quantify exposurein some way, rather than simply classifyingparticipants as exposed or unexposed, then it


may be possible to examine the relationshipbetween exposure levels and the relativerisk. A steady increase in risk withincreasing exposure, sometimes referred toas a dose-response, provides evidence infavour of a causal relationship (Figure 2).

Figure 2. Dose-response effect

It is also crucial to assess whether the studiesprovide clear evidence that the exposure waspresent before the disease. Although thiscriterion seems obvious, several studydesigns (case-control and cross-sectional)involve assessing exposure at the same timeas the disease is diagnosed. It is not alwayspossible to determine from the informationpresented that exposure indeed occurredbefore the onset of disease.

Another important aspect of judgingcausality is the plausibility of the linkage interms of the known biology of the disease,and its overall coherence with the reportedepidemiological findings. Although weclearly do not understand the biologicalbasis of all relationships that we observe, wewould be very cautious about attributingcausation in circumstances that directlycontradicted our understanding of biology.

Conclusion

The RMA draws on its ever-accumulatingexperience to try to establish someconsistency in the process of drawingtogether critical appraisals of publishedliterature and making causal inferencesabout the relationships between specificfactors and diseases. As you will be aware,we work under legislation that suggests thatwe should look for reasons to attributecausality if we can and that is indeed whatwe do.

As a result of this underlying philosophy, wetend to make judgements of causality,particularly at the reasonable hypothesislevel, on the basis of somewhat weakerevidence than might be accepted in othercontexts. Nevertheless, we aim to do so in aconsistent manner, using comprehensive andsystematic reviews of all the informationavailable. There is no compromise to theoverall scientific integrity of the process.

0

10

20

30

40

50

60

70

80

90

None Low Medium High

dose

incid

en

ce

dose-response

no dose-response


The Expectations from Health Studies

Summary of presentation by:

Professor Ken Donald, RMA Chairman


March 2004


The Expectations from Health Studies

This talk is about health studies and the wayin which they may or may not be of use tothe RMA and the SOP system. The firstthing to note is that the compensation systemin Australia depends upon causes of diseasesas defined by exposures. It does not dependupon classes of veterans. In fact, factors inSOPs made on classes of veterans are notcovered by the legislation.

The RMA has had a lot of requests over theyears to make factors for groups of veterans,everything from groups down to less than adozen men, through to whole theatres. If wedid go down that path, we would actuallyfinish up with hundreds of factors basedaround increasingly smaller and smallergroups, and that’s not what the parliament, tothis point in time, has decided.

One example is provided by the Korean Warveterans’ cancer study, which showed anexcess of smoking related cancers. Althoughthere is no factor for Korean War veterans,they can make claims under the SOP system.Korean Veterans who have lung cancer canclaim under a number of factors that alreadyexist in lung cancer – smoking, diesel fumes– just to name a couple. If they havecirrhosis of the liver, they can claim underthe alcohol factor.

It comes back to the point that not all peoplein a deployment have the same exposures.There is a tendency for a health study toassume that a particular group of soldiers,whether they be Korean veterans or whetherthey be a smaller group, have all had thesame exposures when in fact they haven’t.

Furthermore, it is sometimes quite difficultto sort out even what the exposures were 50years ago, or 20 years ago, let alone whichindividuals were exposed to which particularexposures. It has been our experience thatquite often health studies have raised

expectations that the RMA is in no position,legally, to meet. The current F-111deseal/reseal study brings up – from lookingat what we have received so far- it brings upanother issue about exposures. For asignificant proportion of the chemicals thatappear to have been part of the exposure,there is no sound medical-scientificevidence, no published literature, about whatthe effects of those chemicals might be.Again, for the RMA, that’s a blind alleybecause we can’t guess at what might be theeffects of a chemical. Occasionally, if achemical were very closely related to amolecule that has been extensivelyinvestigated, we might be able to draw someconclusions about potential effects.

Another issue that has come up in the GulfWar 1 study, is that the outcomes measuredby the study don’t fit the definition of“disease” in the legislation or, for thatmatter, don’t fit the definition of “disease”in the standard medical text books. The GulfWar study turned up a constellation ofsymptoms but no specific identifiabledisease. And looking at the F-111 study, it’sclear that people have a lot of symptoms andthey also sometimes appear to have signs,but unless there is a definable disease,legally we have no starting point.

Those are two of the problems withretrospective health studies, the problem ofmeasuring exposures and the problem of thesorts of outcomes that health studiessometimes produce. Retrospective studiesdepend to a significant extent on people’srecall and there is really very little opportunityto measure exposures. It is not that the healthstudies are not as well done as possible, it isthat the retrospectivity and the nature of thedata preclude some of the outcomes that youwould get from a hypothesis drivenprospective study. The general principle is thatretrospective studies often only really raiseproblems, they don’t solve them.


Another problem with health studies is thefinding of many different symptoms, whichcould just be a chance finding or a finding ofunknown significance. If you look at thegeneral practices around Australia on anyparticular day, there’s a percentage ofAustralians who go to see their GP, say 5%.Now, that 5% who go to see their GP on anyone day, bring with them about a dozendifferent symptoms and many of them neverget a diagnosis specifically attached to them.On that same day, 20% of the communityhad those same symptoms and did not go toa doctor. If you go to the next day, 5% go totheir GP again, but it’s a different 5%. Whenyou’re dealing with human beings and theirsymptoms, the facts of the matter are thatmany of the symptoms that human beingshave mean nothing, because the next daythey’re gone and another group of peoplehave got the same symptom. It raises a lot offalse alarms, particularly if people have gota heightened susceptibility to observing theirown symptoms.

There can even be problems withabnormalities detected by laboratory tests. Itis well-known that if you do blood tests oneverybody in the room here, you findabnormalities by chance, and if you do thatsame level of blood tests a week or a monthlater, you get false positive results on adifferent group of the population. Anenormous amount of health resources in thiscountry is wasted following false trails ofwhat are chance laboratory findings insomebody who has got absolutely nothingwrong with them. Laboratories in generalhave got a false positive rate, so bydefinition, some of the results you get out ofa laboratory are going to be wrong. Some ofthe ones that are wrong are going to be high,by chance.

These are the sorts of background issues thatimpact upon retrospective non-hypothesisdriven studies. They raise problems that

actually don’t exist, and you finish up withthese constellations of symptoms whichoften don’t add up to anything more thanwhat’s going on in the average community,day in and day out. It is hard to untanglewhen you’ve got something real, and whenyou’ve got something that is just one of thesefalse positives.

In retrospective studies exposure to risksfrom medical interventions are oftenstudied. In every situation in treating apatient, or deciding on a vaccination regimefor children or soldiers or anybody else, youare balancing the risks against the benefits.All immunisation has a complication rate.Aspirin has a complication rate- if you takeaspirin, you’re more likely to die ofhaemorrhagic stroke than you are if youdon’t take aspirin. There is no medicalintervention that does not have a risk. Thesafest and most cost effective medicalintervention ever invented is still aspirin, ineconomic terms, but it still has a down side.It is important to compare the risk ofintervention with what would have happenedif there had been no intervention. Becausethe risks occurred a long time ago they canbe difficult to assess.

Many retrospective health studies – not justretrospective studies in veterans, butretrospective studies in general – are oftenunder-powered. What happens is that even ifyou start with a large number, by the timeyou get down to specific disease entities, orspecific findings, or specific syndromes, thenumber of people left in that category do notgive you the statistical power to calculate astatistically significant outcome.

When studying small groups another dangeris that when the number of cases of aparticular disease is measured, instead offinding three cases of cancer X, they findseven. That is two and a half times what isexpected and it looks like something is really


going on. What you don’t know is that in thenext 500 people, there are going to be nocases because the statistics will catch up andthe rate will balance out over time. The lackof statistical power means that the results areunreliable, and could be purely due tochance.

Some ethics committees actually view astudy that has small sample size andtherefore has low power as being unethicalin its own right. They might produce a non-finding and create more questions than youactually answer. That is considered by someethics committees who do their homework toactually be an unethical thing to ask peopleto approve.

One of the other problems of health studiesis that they aren’t that, they are actuallysickness studies. They ignore the fact that,quite often the group that is being studiedhas a number of things that positively affecttheir health. One of the concerns of publichealth practitioners about these sorts ofstudies is they have a great propensity tomake a group of people who are wellbecome sick, because they overlook thepositive effects of whatever the exposuresmight have been.

Those are the main issues that impact uponthe RMA’s ability to use retrospectivestudies, whether they be retrospective healthstudies of veterans; or whether they beretrospective studies of bus drivers inDenmark; or occupational studies of leadworkers in factories. That is not to say thatretrospective health studies are not capableof providing key information. A lot of theRMA’s findings in health have come fromretrospective studies when there has been noother option. The limitations- small samplesizes, poor exposure documentation and lackof a predefined hypothesis- restrict theirusefulness.

Certainly, retrospective studies will be the

only way you’re going to get efficientcollection of information about rare forms ofcancer or motor neurone disease or otherdiseases that don’t occur very much. On theother hand, prospective studies give you theopportunity to study the symptomatologythat you see in a reasonably short amount oftime after deployment.

The RMA is very pleased that Defence ismoving towards being able to prospectivelylook at exposures of troops in the field.Some of our allies appear to be far ahead ofus in doing it, and are actually plotting whereeach individual soldier was and his or herparticular exposures. I think in the future,those sorts of prospective studies willactually be able to produce some outcomeswhich may well be more useful to the RMAthan the retrospective studies have tended tobe.

Soldiers will be prospectively tagged fortheir exposures, not necessarily theirdeployments. That will mean that a soldierover a lifetime, or over his/her period ofservice, will build up an exposure profilewhich may not necessarily reflect thedeployment arrangements. Scientificallythat is a more powerful tool than trying tomeasure diseases by deployments becausenot every soldier on the same deploymenthas the same exposures. Also, soldiers in thecurrent Defence Force will go on manydeployments. I think when you look atexposures rather than geographic parts of theworld or different particular conflicts, thereare going to be a whole different set ofquestions and a whole different set ofoutcomes that potentially might be moreuseful in the future.

Prospective studies can give us, in anevolving way, responsiveness to a wholedifferent set of needs and ways of doingmore. It is only through those types ofstudies that we will be able to discover the


evolution of problems, whether they start asconstellations of symptoms or whatever, andknow if there is going to be a progression toillness and have the opportunity to influencethat course in a more positive way.

We also shouldn’t forget that the mostpowerful use of this research is not lookingat health effects or claims for compensation,but the question of duty of care to make surethat people are not exposed unnecessarily toharmful things during deployments.


Protective Effects of Exposures

Professor John Kaldor, RMA Member


March 2004


Protective Effects of Exposures

In the course of reviewing the medicalscientific literature for a SOP, most factorsthat emerge as potentially causally related tothe disease that is the subject of the SOPshow a positive association; in other words,exposure to the factor is associated with anincreased likelihood that the disease willoccur. Nevertheless, we are increasinglyfaced with data on factors that show aninverse association with the disease, in thatexposure is associated with a lower risk.

As with positive associations, we need toexclude chance, bias and confounding asalternative explanations, before we cantranslate an inverse association into a claimthat the exposure is protective, in that it iscausally related to reducing the likelihood ofdisease occurrence.

At first glance, it might seem that protectivefactors would not need to be considered in aSOP on causal factors, given that thepurpose of the SOP is to describe causes ofdisease. In fact, if exposure to a factor canreduce the risk of disease, it is reasonable toconclude that a reduction in the exposurelevel may actually increase the likelihood ofthe disease occurring. The absence of anyfactor that is established as being protectivecan therefore be transformed into a causalfactor if a person is unable to maintainexposure to the normal or beneficial level ofthe factor.

For example, a number of studies have nowshown that exercise protects againstcolorectal cancer. Therefore, thecircumstance of being unable to exercisecould be interpreted as being a causal factorfor colorectal cancer.

For a causal factor, we would generallyexpect to see consistent estimates of therelative risk that were above about 1.25, suchthat exposure to the factor gives a 25 per

cent increase in the chance of getting thedisease, compared to people not exposed tothe factor. Similarly, it would be appropriateto define a factor based on inability tomaintain protective levels of exposure,provided the estimated relative riskassociated with the absence of exposure wasconsistently observed to be greater thanabout 1.25.

For factors in the RH instrument, the RMAhas accepted relative risks as low as 1.1, incircumstances where the combined evidencefrom epidemiological studies providesparticularly strong evidence for causality.While the RMA is encouraged to begenerous in its interpretation of the evidencefor causation, it clearly cannot go beyond thelimits of scientific credibility.

When a person develops a disease that iscovered by a SOP, the relevance of any factorincluded in the SOP depends on thethreshold level of exposure specified in theSOP. This level is determined from theavailable published literature, as beingassociated with a measurable increment inrisk.

For some factors, the increase in riskassociated with exposure is calculated inpublished studies by making comparisonswith people who were considered to havelittle or no exposure. For example, in studiesof exposure to herbicide, it is generallyassumed that comparison groups wereunexposed, and that any increment in riskassociated with exposure is with reference tothis background.

On the other hand, there are factors forwhich exposure is ubiquitous in humanpopulations, and the calculation of riskincrements in published studies must bemade with reference to people whoseexposure was below a specified level, ratherthan being zero. Consider fat consumption,or sunlight exposure as examples. In these


situations, the true relative risk associatedwith the threshold level of exposurespecified in the SOP will vary from personto person, depending on the individual levelsof exposure arising from sources other thanservice.

To illustrate this point, think of twoindividuals, one of whom had been exposedto 90% of the specified threshold level of afactor prior to the start of qualifying service,and another who had been exposed to 10%.If each of these individuals then has servicerelated exposure that takes the cumulativelevel exactly to the threshold specified in theSOP, they will both qualify for acceptance ofthe factor. However, the service-relatedexposure will clearly have been associatedwith a much smaller increment in risk for thefirst individual than the second.

To illustrate the point in another way, takethe hypothetical circumstance that our SOPsystem covered not just Australia but anumber of countries of our region, andconsider a SOP for malaria.

If a person lives predominantly in Australiaand develops malaria following deploymentto a malaria endemic area, the relative riskassociated with the deployment would bevery high. On the other hand, for a personwho lives in an endemic area, and is thendeployed in an endemic area, the incrementin risk associated with service may be quitesmall. Thus the background risk of anindividual can substantially influence aperson’s real relative risk.

The current system is effectively based onaverages, in that it makes acceptance of afactor more difficult for people who have avery low background, while it advantagespeople who actually have a very highbackground.

Formally, the legislation governing the SOPsdoes not make provision for taking account

of the levels of exposure experienced by anindividual outside service, although there isone major factor for which suchconsiderations have been built into theoperation of the corresponding SOPs.

Exposures that become SOP factors throughtheir absence or the inability to maintainprotective levels would be of this secondkind. Clearly if removal of the exposure is tobe judged to cause an increase in diseaserisk, there must be widespread exposure tothe factor under normal circumstances.Consider exercise levels or fruit andvegetable consumptions as illustrations.Thus individuals who qualify under factorsfor which it is the absence that increases riskare doing so at a variety of relative risks,because of the variation in their backgroundlevels of exposure.

The incorporation of protective factors (or atleast their absence or reduction) in SOPs isthus an evolving area of interest for theRMA. It has encouraged us to look moreclosely at some key methodological issuesunderpinning the calculation of thresholdlevels for exposure, and will certainlyprovide a wider range of factors forconsideration in SOPs.


Syndromes and Causes – Illness and Disease

Dr Justine Ward, RMA Secretariat

Paper presented at the RMA Forum

March 2004


Syndromes and Causes – Illness andDisease

Introduction

The health problems of veterans of the firstGulf War have been extensively studied byvarious governments. The difficulty ofdefining and finding the causes of the diversesymptoms reported by veterans of the GulfWar has once again highlighted the problemof war syndromes and their evaluation. TheRepatriation Medical Authority has astatutory responsibility to make Statementsof Principles for syndromes, provided thatthey meet the legal requirements fordefinition of a disease under Section 5D ofthe Veterans’ Entitlements Act (VEA). Theproblems with defining syndromes andevaluating their causes as they relate to thefunctions of the RMA will be outlined in thispaper. The broader question of understandinghealth and illness will also be considered, asthis question relates to how disease is definedand how symptoms are understood byindividuals.

Defining a Syndrome and PotentialConsequences

A syndrome is defined in Dorland’s Medicaldictionary as “a set of symptoms which occurtogether; the sum of signs of any morbidstate; a symptom complex”. To be able torecognise a syndrome there needs to be arepeatable grouping of particular symptoms.Often there is no known cause of a syndromeand sometimes a cause will subsequently bediscovered. For example, acquired immunedeficiency syndrome (AIDS) was initiallydescribed by a set of symptoms recognisedby an astute group of clinicians and study ofthe syndrome eventually led to the discoveryof a new virus.

Illnesses that are well defined as diseaseshave more status than symptoms and thisleads to a pressure to find a cause or at least

define a syndrome. The naming of asyndrome, while providing the comfort ofvalidation to sufferers, also has potentialconsequences which may be both adverse orbeneficial to individuals. It essentiallymedicalises a set of symptoms, with majorpotential medical, legal, political and socialconsequences because of flow on effects toapproaches to treatment, compensation andpatient’s lives.

Giving a set of symptoms a medical namecan place people in the sick role and actuallyimpede their recovery. This was realised asfar back as World Wars I and II, when it wasfound that soldiers with acute combat stressreaction were more likely to return to duty ifthey were treated quickly and near theircombat units than if they were treated aspatients in a hospital (Hyams, Wignall andRoswell 1996).

An individual’s expectations of his/herprognosis can have the effect of self-fulfilling prophecy. A poor self-rated healthstatus has been found to be a predictor ofmortality in longitudinal studies and thisphenomenon is not entirely explained byexisting illnesses and symptoms (Idler andBenyamini 1997). Self-rated health caninfluence behaviours that influence healthstatus, for example smoking, alcohol use,less engagement in preventive practices suchas physical activity and screening and poorcompliance with medications. Reducedexpectations may be reinforced by socialfactors, such as reduced employmentopportunities for those with disabilities,financial incentives and behaviour of peers(Lupton and Najman 1989). Concerns aboutpotential reproductive effects may makepeople worry about their children’s health oreven decide not to have children. Because ofsuch potential adverse consequences, it isimportant that syndromes are not definedunless there is a sound scientific basis fordoing so.


Exposures and Possible Causes

The problem of post combat syndromes hasprompted considerable research and debateas to their aetiology. One possibility is thatpost combat syndromes are physiologicaldiseases caused by unique environmentalexposures. In every conflict there are uniqueexposures related to the place, the time andthe conditions of war. In the Vietnam Warthere was Agent Orange/dioxin; in theBosnian conflict there was depleteduranium; and in the Gulf Wars uniqueexposures have included smoke and oilcloud, vaccinations against plague andanthrax and the anti-nerve agentpyridostigmine bromide.

Each conflict also has distinctivepsychological stressors, such as heavyshelling of trenches in World War 1; notknowing which villagers were Viet Cong inthe Vietnam conflict; and threat of beingexposed to chemical, biological orradiological weapons in the Gulf Warconflicts. Pervasive, unknown threats can bevery hard to cope with psychologically. Oneveteran wrote in a letter home: “I can dealwith getting shot at, because even if I got hit,I can be put back together- a missile, I caneven accept that. But gas scares the hell outof me...” (Berstein and Kelley, 1995)

There has been a great deal of effort madetowards attempting to discover possibleassociations with exposures and variousdiseases. Despite this large research effort,chronic somatic symptoms have notconsistently been linked to any particularexposure (Hyams, Wignell and Roswell1996, Sim et al 2003).

For more recent conflicts, there is apossibility that symptoms represent the earlystage of disease or diseases which have yetto manifest fully with demonstrable physicalsigns or changes in laboratory tests. Furtherfollow up is needed to ensure that diseases

do not develop. Sufferers of the so called“effort syndrome” in World War I werefollowed up but did not show an increasedmortality (Hyams, Wignell and Roswell1996, Jones et al 2002). Studies of mortalityin Gulf War veterans have not so far shownan increase in overall mortality relative tonon-deployed veterans and no diseasecategories were significantly elevated inveterans (Research Working Group ofMilitary and Veterans Health CoordinationBoard 2002). Although follow up willinevitably demonstrate the development ofdiseases over time, a link to exposures stillneeds to be made to establish causation.

Another possible explanation for postcombat symptoms is that, despite someunique war experiences, there is somethingabout the overall war experience thatproduces a common response. It isinteresting to examine the historical recordrelating to post-combat syndromes. Hyamset al (1996) has described war syndromescharacterised by similar symptoms afterevery conflict since the US civil war. Thesesymptoms include fatigue, shortness ofbreath, headache, sleep disturbance,forgetfulness and impaired concentration.These war syndromes were given differentnames after each conflict: “irritable heartsyndrome” in the US Civil War; “effortsyndrome” in World War I; “battle fatigue”in World War II; posttraumatic stressdisorder after the Vietnam; and “Gulf Warsyndrome” after the first Gulf War. US, UKand Australian Gulf veterans report sufferingfrom more symptoms than non-Gulfveterans. A range of neuropsychologicalsymptoms was most commonly reporteddespite markedly different exposures of eachveteran.

Jones et al (2002) attempted to characterisepost-combat syndromes by doing a historicalcluster analysis of symptoms using warpension files from the Boer War to the first


Gulf War. In this study the authors examinedthe veterans’ own attribution of symptoms.Boer war and First World War servicemenwith disordered action of the heart believedit to be due to physical illness or physicalexertion. First World War veterans withneurasthenia attributed their symptoms toboth physical exertion and the psychologicalstress of military service. Gulf Warservicemen with predominantly debilitysymptomatalogy tended to attribute illnessto physical illness, injury or environmentalconditions, whereas those withneuropsychiatric symptoms tended toattribute symptoms to the psychologicalstress of war.

This historical examination shows thatexperiences of symptoms, diagnostic labelsand beliefs about causation are linked butchange according the nature of combat,contemporary medical knowledge andprevailing health beliefs. These authorsconclude that what has changed is not thesymptoms themselves but the way in whichthey have been reported by veterans anddoctors. It highlights the potential danger ofallowing preconceptions to get in the way ofscientific hypothesis formation and testing.

Shared symptoms may represent a commonreaction to stressors or other exposureswhether they occur in military or civilianlife. Australian Gulf War I veterans reportedall fatigue related outcomes more commonlythan the comparison group and had elevatedamounts of symptoms in the groupings ofpsychophysiological, cognitive and athro-neuro-muscular (Sim et al 2003). There is anoverlap in these symptoms and thoseexperienced by civilians affected by multiplechemical sensitivities, fibromyalgia andchronic fatigue syndrome. Unexplainedsymptoms have also been reported bycivilians after the World Trade Centreattacks (Clauw et al 2003).

All these conditions or postulated conditionshave in common that they are based on self-reported symptoms, lack objectiveverification of exposures and provencausative exposures, lack consistentabnormal physical findings and cannot beconfirmed with any clinical test (AmericanCollege of Occupational and EnvironmentalMedicine 1999, Working group chronicfatigue syndrome 2002). It has beensuggested that, rather than apply specificlabels to groups of symptoms, particularlyones that imply a pathogenesis, a more clearand unbiased terminology should be used,for example “medically unexplainedsymptoms” (Clauw 2003).

To add to the confusion, there is also anoverlap of symptom based conditions withvarious psychiatric disorders which oftenmanifest with somatic complaints, includinganxiety, depression and somatoformdisorders (Hyams 1998). This is not tosuggest that the diagnosis of a psychiatriccondition is in doubt, but it can be difficultto establish whether the psychiatric disorderis the primary cause of symptoms or a resultof debilitating fatigue or pain.

Methodological Problems WithStudying Syndromes

A syndrome is sometimes defined primarilyfor the purposes of study. Having adefinition does not necessarily mean that acondition exists. Study designs require thatpeople with a condition (cases) arecompared to people without a condition(controls) for the prevalence of particularrisk factors, or that exposed and unexposedgroups of people are compared to see whatproportion develop into cases. It is thereforenecessary to develop at least a workingdefinition of a case. However, the lack ofspecificity of such definitions means thatthey may not be reliable in distinguishingcases from non-cases (Hyams 1998). This


reduces the power of the study to detect adifference in exposures between the twogroups. When cases cannot be reliablyidentified it is also difficult to recognize theinfluence of bias and confounding (Hyamsand Roswell 1998). It is hoped that progressin the study of symptom based conditionswill eventually identify some consistentdistinguishing laboratory features and makecase definitions much more specific, at leastin some instances.

Chronic fatigue syndrome has been able tobe characterised by a common set ofdiagnostic criteria (Fukuda 1994). Incontrast, while the level of symptomreporting among Gulf War veterans washigher than in the comparison group,statistical analysis showed that the pattern ofsymptom reporting in the two groups wassimilar. This has suggested that the Gulf Warveterans do not have a unique symptomcomplex or cluster (Sim et al 2003).Similarly, no consensus for proposeddefinitions for multiple chemicalsensitivities has been able to be reached inthe scientific community, in a large partbecause of the lack of ability to identify aspecific group of symptoms.

There were many methodological problemsassociated with studying Gulf War illnesses,especially problems with objectivelymeasuring exposures and recall bias. Thelatter is a potential problem with allretrospective studies. People who areconcerned about their symptoms or whohave a disease are more likely to ascribe to aparticular exposure or experience thanpeople who feel well. Sufferers of symptom-based conditions have ascribed theirsymptoms to various temporally associatedenvironmental exposures, including modernoffices, chemicals, food allergies andsilicone implants (Hyams 1998). Extensivequestioning of veterans for health studies orhealth assessments has the potential to

remind them of traumatic events andprovoke symptoms.

The authors of the Australian Gulf Warhealth study state that some, but not all,symptom reporting could be explained byrecall bias (Sim et al 2003). Their analysisalso suggested a possible problem withrecall bias in reporting of exposures. A studyof recall of military hazards showed thatreporting of military exposures can changeover time, although consistency of reportingis better for some exposures than others(Wessely et al 2003). Furthermore, reportingnew exposures was associated withworsening health perception while forgettingprevious exposures was associated withimproved perception.

Ill-defined symptoms are commonpresenting complaints to primary carephysicians.

As many general practitioners can attest, it isnot always easy to diagnose a diseasebecause many diseases present with similarsymptoms and signs. Without the ability todistinguish between them on the basis ofsigns or pathological tests, some diseaseswould be impossible to diagnose. Even if aphysiological correlate of disease is found, itdoes not necessarily establish that it iscausally related, as the abnormality couldalso be an effect of symptoms rather than thecause of them.

Illness and Disease

Part of the difficulty with dealing withmedically unexplained symptoms may bedue to an artificial dichotomy betweendisease and health. Dorland’s medicaldictionary defines disease as “any deviationfrom or interruption of the normal structureor function of a part, organ, or system of thebody as manifested by characteristicsymptoms and signs; the etiology, pathology,and prognosis may be known or unknown.”


The VEA has a slightly broader definition:

(a) any physical or mental ailment,disorder, defect or morbid condition(whether of sudden onset or gradualdevelopment); or

(b) the recurrence of such an ailment,disorder, defect or morbid condition.

Disease definitions are useful for variouspurposes: for researchers to establishcausation, for governments to base policieson compensation and for the medicalprofession to make decisions abouttreatment, especially if symptom clustersconsistently respond to specific clinicalstrategies. However, disease definitions donot necessarily give an accurate reflection ofhow people experience their health on a dayto day basis. Health is defined as a state ofoptimal physical, mental, and social well-being, and not merely the absence of diseaseand infirmity.

In a review of the scientific literaturepertaining to stressors and the Gulf War,Marshall (1999)pointed out that:

“Illness and disease are overlapping, butdistinct, constructs. Whereas diseaserefers to constellations of symptoms thatdefine a diagnosable physical orpsychiatric disorder, illness refers to thesubjective experience of poor health.Illness manifests itself as somatic (bodily)or psychological symptoms, but may stemfrom multiple sources—includingcognitive and social processes—and mayor may not reflect the presence of anunderlying disease (Kleinman, 1988). Therelationship of illness to disease iscomplex. A person may experience illhealth with no underlying disease.Conversely, he or she may suffer from anunderlying disease without perceivinghimself or herself as ill (Weiner, 1992).”

Sociologists argue that it is not possible to

separate the experience of health or ill healthfrom its social context. Much poor health isa function of social circumstances andevents (eg broken marriages, boring andrepetitive work) and may have nopathophysiological basis (Lupton andNajman 1989). The traditional diseasemodel looks simplistically for single orcomponent causes with a biological basis,whereas an endpoint of complete healthallows for a model of causation whichaccommodates a complex interaction ofpsychosocial and physical factors.Psychosocial factors include context,knowledge, attitudes, beliefs, personality,past experiences, peer values, and the social,legal and economic environment. All thesefactors affect the meaning or significancethat is attached to symptoms. Causal modelsthat rely on a more integrative approach,with stress and central nervous systemresponses as a final common pathway, mayprove useful in explaining the effects of amultiplicity of environmental factors (Kipenand Fiedler 2002).

Hyams et al (1996) suggest that it will not bepossible to explain war-related syndromesuntil there is a better understanding of healthand illness in the general population.Population based studies suggest that morethan one third of symptoms may bemedically unexplained (Clauw et al 2003).Hyams proposes two basic questions:

1. What is the relation between chronic,non-specific symptoms andphysiologic and psychologicalillness?

2. What factors – medical,environmental, psychological, orsocial- create a personal sense of illhealth?

Many disease processes can be imagined asa continuum from complete health to illhealth to disease. The division between what


is called health and what is called disease isoften determined by a cut off point. Cut offpoints are usually set where symptomsbecome very functionally disabling and/orwhere treatment has been shown to bebeneficial to the majority of people from thispoint onwards. This cut off point is notfixed- it changes according to the sensitivityof diagnostic tests, scientific understandingabout the benefits of treatment and theavailability of treatment options.

Common examples of this include treatmentlevels for hypertension, diabetes andhypercholesterolaemia. If the cut-off point isset too high, some people who would benefitfrom treatment will miss out. If the cut-offpoint is set too low, some people’s healthmay actually be harmed because they will belabelled as sick and will receive treatmentunnecessarily. Decisions about wheresymptoms, signs or tests are labelled“abnormal” need to keep this risk/benefitbalance in mind and should be based ongood evidence. At the clinical and sociallevel, though, it is important to keep in mindthat there is often a continuity betweenpersons whose symptoms have been givendisease status and those with unexplainedsymptoms.

Just because medically unexplainedsymptoms do not appear to fit the traditionalmedical disease paradigm, are hard to studyand are not easily accommodated within thelegislative restrictions of the RMA, it doesnot mean that they are not real andlegitimate. Whether symptoms may or maynot be due to a disease process, they can bevery disabling functionally and greatlyimpair quality of life. Despite an absence ofproven causes, treatment is still veryimportant, although care must be taken toavoid well intentioned but harmfulinterventions.

There is little systematic evidence on which

to base strong recommendations forinterventions to prevent or mitigate postcombat syndromes (Clauw et al 2003).There is some evidence that cognitive-behavioural interventions that preparepersonnel for the realities of war offer somebenefit (Clauw et al 2003). Social supportappears to buffer the effect of stressfulevents. Critical incident debriefing does notappear to improve health outcomes and maydo more harm than good. Constructivetreatment after symptoms have developedinvolves management of symptoms andefforts to restore functioning, rather thanfocussing on exhaustive diagnostic testing.Outcomes are better if treatment is givenearly, before symptoms become chronic.

Conclusion

In summary, the objective strategies used todetermine when symptoms without anyidentifiable pathological basis become adisease or syndrome include statisticalclustering, response to treatment, expertconsensus and symptom severity. The bestapproach relies on more than one methodand should make clinical andpathophysiological sense. However,recognition of symptom clustering is alsoinfluenced by contemporary beliefs.

Studies to date do not rule out the possibilityof distinct clinical diseases being eventuallyfound to be responsible for at least somechronic unexplained symptoms. In relationto Gulf War related symptoms and otherchronic symptom-defined conditions, it isdifficult to know if they reflect physiologicalor psychological diseases with single ormultiple causes, or a normal response to thephysical and psychological stresses of war.The evidence we have to date is not veryhelpful in distinguishing between thesepossibilities.

A basic understanding of the prevalence of


symptoms in the general population isneeded. Studies of military populations alsohave much potential to help answer thequestions of health and illness, particularly ifthey are prospective in nature and collectaccurate data on exposures. It is important toknow the psychological and physical state ofsoldiers prior to deployment, in order toestablish that there has been a change.Follow up immediately after the conflict andperiodically thereafter will clearly establishthe timing of any adverse health effects.Adequate preparation may help prevent postcombat symptoms and early treatment ofsymptoms regardless of cause will minimisechronic effects.


Summary of Issues Raised by ESOs

Issues and responses collated from edited transcripts of the RMA Forum

March 2004


RMA Processes

Basis for Removing a Factor

ISSUE

Please clarify, when removing a SOP factor,the basis of sound medical-scientificevidence (SMSE) and how the new SMSEsupports the removal of a particular factor?

RESPONSE

Professor Donald explained that the issue isthat new studies emerge and those studiesoften contradict previous studies. They aresometimes better studies or bigger studieswith more power than the ones that werepreviously in the literature. They sometimescause the RMA to revise its opinion becauseof the quality of the studies or the size of theresult, or all of those things. The whole bodyof evidence is taken into account whenremoving a factor.

Pearce Report Recommendations

ISSUE

What is the progress of the Pearce Reportrecommendations? In relation torecommendation 2, has the RMA consideredthe question of military experience as adesirable selection criterion for futureappointments to the RMA?

RESPONSE

Professor Donald replied that the Minister ofthe day endorsed 18 of the 20recommendations and the RMA hasaddressed all of those within its remit. Theissue of how the RMA is constructed wasoutside the remit of the RMA. Mr BillMaxwell (DVA) said that the governmentresponse had been to factor in the matter ofmilitary experience as a criterion but not tomake it a prerequisite. However, there maybe other means by which Defenceexperience could be brought into the

process. There is already an establishedprocedure where Departmental staff attendthe informal part of the RMA meeting andcomment upon the operationalisation of thedraft SOPs. Some Defence Health peoplemight be able to do something similar.Professor Donald agreed that this might bean appropriate approach, especially in lightof the MRC Act and its applicability tocurrent serving members. He undertook totake the matter up with the Commission.

ESO Response Time

ISSUE

Time allowed for ESO response time whenasked to comment on RMA papers.

RESPONSE

Professor Donald said that the usual timegiven to response is about six to eight weeksbut it depends on the RMA meetingschedule. The RMA is aware that there aresometimes delays by peak organisations ingetting the papers out to individual ESOs.Viv Quinn (RSL) explained that theconsultation process within theirorganisation can take longer than six weeks.Mrs Carole Friedrichs (RMA secretariat)responded that it was also necessary torecognise that a longer response time woulddelay the making of the instrument, whichmight also include changes that positivelyaffect veterans. It was agreed to undertakefurther consultation with ESOs in relation toresponse time and electronic distribution.

RMA Protocols

ISSUE

What are the RMA Protocols foradministration and management?

RESPONSE

Professor Donald answered that the RMAoutlined its processes in this forum and the


previous Forum in 1988. These processes aredocumented in the proceedings of the 1998forum.

Expediting Reviews of SOPs

ISSUE

What (early) options are available when agenuine claim is restricted by a SOP in itscurrent format?

RESPONSE

Professor Donald said that at the moment theanswer is that there are no quick options,although there are sometimes other factorsin a SOP which the veteran could look at. MrJohn O’Connor Whyte (Specialist MedicalReview Council) made the point that there isonly a three month window to appeal adecision in respect of a SOP to the SMRC,although he warned that the Council’sprocesses also take some time. ProfessorDonald explained that even though there areover 50 investigations listed, but the RMAcan prioritise some investigations if theybear on a number of claims. There was asuggestion from Dr Keith Horsley (DVA)that an amendment to the legislation suchthat a single factor could be reviewed mightmake the RMA’s processes more efficientand responsive. Professor Donald agreed butsaid that there would need to be somecaveats around that and there would need tobe Government support for amending thelegislation.

The Specialist Medical Review Council

ISSUE

If the SMRC decides that there is sufficientevidence to include a factor that the RMAhas rejected, does the RMA take that at facevalue or do you have to review it again?

RESPONSE

Professor Donald replied that the SMRC canask the RMA to do a review or they can

instruct it to put a factor in. They haveinstructed the RMA to put a factor in onlyonce, in relation to chronic lymphaticleukemia and electromagnetic fields. TheRMA did as instructed. In relation to jobstrain and hypertension they asked the RMAto do a review. The RMA looked at it anddecided that on the whole of the evidenceavailable the proposition wasn’t sustained.However, the SOP factors put in by theSMRC are no more permanent than the SOPfactors put in by the RMA. All of the SOPsare subject to review and as the scientificliterature changes, they may change.

MRC Act

ISSUE

The passing of the MRC Act means that theSOPs will be used to relate service tocompensation claims for serving DefenceForce members. How will this affect therelationship of the RMA with the DefenceForce? How will the RMA consult with theDefence community?

RESPONSE

Professor Donald noted that the legislation isnew and will require careful reading, but thatthe process of making SOPs will not change.The RMA has foreseen a need to make somenew SOPs in relation to overuse injuries andhas advertised investigations into a selectionof overuse syndromes. There will also be aneed to consider diseases that affect youngerpeople because of extension of coverage toschool cadets.

Members of the RMA are observers on theMedical Advisory Panel and so are keptinformed of current discussions betweenDefence and DVA in relation to occupationalhealth and safety issues. In the past the RMAhas visited troops on deployment in Timorand Bouganville to gain an understanding ofconditions of service and may do so again.The RMA is looking at developing other


mechanisms for consultation with the ADF.

Professor Andrew Wilson added that inaddition to looking at specific conditions,there may also be specific exposuresrelevant to serving personnel that need to bethought about. The Centre for Military andVeterans’ Health which is being launched inMay is going to be an important resource forlooking at these issues.

Professor Donald agreed but pointed out thathealth and mortality studies of groups ofveterans often document exposures poorly.There are plans within ADF and DVA toovercome this problem by developing aprospective health surveillance programwhich will monitor exposures duringdeployment of small groups of soldiers.

Supporting documentation for requests

ISSUE

A request for a review of a SOP requiressupporting information which may bedifficult for a small organisation orindividual to collect, especially whenmaterial on the internet is not peer reviewed.Section 196CA allows the Authority torefuse a request for review when there isinsufficient new evidence. What is thequality and quantity of information requiredby the RMA in support of a request?

RESPONSE

Professor Donald explained that when theRMA first started it was important to getenough SOPs out there working so that thesystem would function. To that end, theAuthority would sometimes just look at aparticular factor. Subsequently it had legaladvice that if a SOP is to be opened forinvestigation, it is necessary to look at thewhole SOP, not just one factor. That meantthat reviewing a SOP is a major task,sometimes taking many months. Amechanism was needed to be able to

prioritise SOPs that had to be reviewed asagainst the ones where there was not asignificant case for them to be reviewed.

When the RMA receives a request for reviewa medical officer will check the literature tosee if there is anything new that would makea difference since the SOP was last made.Even if a minimal amount of information issupplied, staff of the secretariat will attemptto pursue it and identify relevant articles.The length of time since the SOP was lastreveiwed is taken into consideration.Members of the RMA sometimes also knowof new activities in the area underconsideration. There is a process fordeciding whether there is anything whichwould suggest to the Authority that it isreasonable to review the SOP, we don’t justsay “no” simply because of the lack ofinformation provided. Section 196CA is asafety valve so that the RMA can priortisewhat really needs to be done.

Andrew Leiboff (medical research officer)added that the request must nominate adisease and a risk factor so that a search ofthe published evidence can be made. Internetsearches are not dismissed as a source ofinformation, but if the information isnothing but opinions then the quality of theevidence may be questionable.

Professor Andrew Wilson reinforced thepoint that it is not the internet per se that isthe problem but the quality of theinformation that comes from it. The web is amajor source for published peer reviewedliterature and there are some fully peerreviewed journals that are only published onthe web.

Documentation of Previous Studies

ISSUE

Why is it so time consuming to reviewearlier scientific studies when these would


presumably have been documented fromprevious investigations?

RESPONSE

Professor Donald agreed that there is adocument available, that being a submissionto the Authority of the critical appraisal ofthe literature, at the time the SOP was madepreviously. There are two issues with suchdocumentation. One is that if a SOP isreviewed it is important to make sure thatthere were no mistakes. That requires havinga significant look at that previous criticalappraisal and maybe even re-reading someof the papers. The second thing is, if newpapers have been published, then thequestion is what do they do to the balance ofthe evidence. In other words, how do they fitinto the context of the balance of evidence inthe previous study. Sometimes one goodstudy will critically alter the balance ofevidence and lead to a different conclusion.

Professor John Kaldor further explained thatthe previous submission does summarise theevidence, but it is a judgement, not a onedimensional summary. You don’t come upwith some score to which you can just add afurther computation if a new paper comesalong. You might, for example, have thesituation of twenty studies that were done inthe past but they all had majormethodological flaws of different kinds andthen one terrific study comes along that hasgot over that flaw. You have to have a look atthe overall body of evidence in order to putthe new study in context.

Guidelines for Critical Appraisal

ISSUE

Is there an international convention thatdefines the factors that are important in theanalysis and appraisal of reports? Does thatmake it much easier to draw out theinformation?

RESPONSE

Professor Kaldor replied that there are anumber of different sets of guidelines thathave been put in place for doing a criticalappraisal. The Cochrane collaboration, forexample, is one very well known structurethat recommends that reviews be done in acertain way, primarily in the therapeuticsarea. Various learned institutions orindividuals have promulgated guidelines.You could not say there is one standard butthere’s a great deal in common among thedifferent types of standards that are used.The sort of categories that the RMA uses arethe ones in general use. Professor Kaldorexplained that internally at the RMA there isan attempt to summarise information, notnecessarily paper by paper, but in a way thatcombines the information with an awarenessof those key categories.

Professor Donald added that most of thestudies are done on “free-ranging humanbeings” who are pretty hard to round up.This limits the quality of evidence fromepidemiological studies. You just cannot geta dose response sometimes because people,in doing the studies, have been unable tostructure it in such a way that you could evencalculate a dose response.

ESO Consultation to Inform DecisionMaking

ISSUE

Epidemiological studies vary in quality andit is sometimes difficult to find evidence.Previously the RMA had a lot ofconsultation and communication betweenex-Service organisations and theDepartment of Veterans’ Affairs so that theRMA could take into account a far widerview of the clinical evidence in relation tocausality. Now there appears to be a greateremphasis on documentary evidence.


RESPONSE

Professor Donald responded that the RMAalways has the option of exercising clinicaljudgement. He explained that probablynearly all of the straightforward decisionswere made in those first three years and whatis left is the matters which are less clear,matters which are in contention and mattersin which the literature is sometimes veryextensive. Professor Donald acknowledgedthat the issues have become more difficultand that it is time once again to have moreface to face communication. The RMAintends to recommence holding meetingsinterstate.

Professor Kaldor further clarified themeaning of clinical judgement. Clinicaljudgment or “common sense” is distinctfrom or complementary to epidemiologicaldata. He emphasised that those processescome in after the RMA has gone through avery comprehensive review of what isavailable. Clinical judgements are amechanism of interpretation of the data andresolving ambiguities that cannot beresolved with the available data. They are nota substitute or an alternative pathway for thatanalysis.

Issues Relating To Particular Factors

Class of Veterans

ISSUE

Although you have stated that the legislationis not based on classes of veterans, there is afactor for 30 days service in Vietnam, sowhat is the difference?

RESPONSE

Professor Donald acknowledged that theVietnam factor was a class of veteran factorbut it came in about 1994/95 and the RMAhas had advice that it cannot repeat that sort

of factor. It was based on the fact thatalthough there was some publishedliterature, it was very difficult to untanglewhat the real exposures were. Not everysoldier in Vietnam, for example, had thesame exposure to Agent Orange. The 30 dayswas chosen because that was what wasrequired to get the Vietnam Medal for activeservice. Similarly, for radiation the RMA hasgone away from class of veteran factors todoses of radiation.

Professor Donald added that it is alsoimportant that the factors retain scientificcredibility. The SOP system was introducedbecause the Auditor General advisedparliament that the old system wasinconsistent and lacking in credibility. TheRMA must be able to point to a convincinglink to the science. Furthermore, inclusionof a factor must be based on the entire bodyof evidence, not just on one published paperout of many.

DDT

ISSUE

DDT is not recognised in the SOPs yet thereare studies which indicate damage to birdeggs.

RESPONSE

Dr Keith Horsley (DVA) agreed that DDTwas banned because it was making eggshellstoo thin and the American Bald Eagle wasbecoming extinct. He noted however thatDDT is in fact recognised in the SOPsbecause exposure to DDT is a factor in theSOP for pancreatic cancer. Professor Donaldexplained that the RMA has studied theeffects of DDT extensively but there haveonly been two studies of significance, theone which allowed the RMA to include it inthe SOP for cancer of the pancreas andanother which showed that it protectedAfrican-American women from breastcancer.


Definition of METs

ISSUE

It is difficult for widows to substantiate thefactor for “inability to undertake more than amildly strenuous level of physical activity”because the definition uses METs as the unitof measurement.

RESPONSE

Professor Donald said that the RMAunderstands that it is often difficult forwidows to remember details but this is aproblem for the Department as it relates tothe evidence. There is some guidance fromthe Department at two levels. The GARPdocuments provide some guidance on how totranslate activity levels into METS. TheDecision Support Unit also distributesexplanatory bulletins.

Brigadier Bill Rolfe (VRB) explained thatthere is an evidentiary onus to put forwardsome material that relates in a positive wayto the particular contention. The VRB takesinto account a lack of records due to thepassage of time but without some material itis difficult to make a case.

Lifting and Carrying Factors

ISSUE

Can you clarify the application of lifting andcarrying factors in SOPs for cervical,thoracic and lumbar spondylosis?

RESPONSE

Professor Donald said that the load and thefrequency of carrying or lifting are used tocalculate the cumulative total. If heavy loadsare lifted or carried more frequently, the totalwill be reached more quickly. Both pickingup and putting down count as a singleoperation. The instrument for cervicalspondylosis only has a carrying factorbecause the SMSE relates to carrying heavy

loads on the head. Carrying implies an initiallift hence the same formula as for lifting canbe used.

Dr Kym Hickey (RMA medical officer)added that it is just a matter of multiplyingthe weight of the load by how many timesyou lift it to get the cumulative loads.

Solar UV Formula

ISSUE

What are the results of recent RMAmeetings concerning the Solar UV Formula?

RESPONSE

Professor Donald stated the background tothis issue, that is, a potential problem withrunning the UV formula program on theDepartment’s new computer system and awish expressed by ex-Service organisationsnot to make any changes to the formulabased factor. After much discussion overseveral meetings the RMA has decided to tryto develop a new factor but retain theformula. The new factor will probably be asort of screening factor based on the amountof time and the latitude where the veteranserved. If the veteran does not succeed underthe new factor then he or she can go to theformula for a more detailed examination. MrNorm Clarke (Legacy) supported thisproposal because it would take theguesswork out of trying to obtain a detailedlifetime history from war widows about theirhusband’s clothing.


Stress of Perceived Failure

ISSUE

Quite often a member of the defence force orveteran may experience stress from failureor perceived failure.


RESPONSE

Professor Beverley Raphael agreed with thisstatement and added that in combatsituations your time sense is often distorted,and you think you had time to do thingswhen in fact you didn’t. In the Granvilledisaster people’s sense of frustration aboutnot being able to fulfil their functions andthe stress of that became apparent. Thoseissues come up in the consideration ofstressor exposures. The issue for the RMA iswhat does the scientific literature say andhow does it delineate those stressors.

Response to Judges’ Interpretations

ISSUE

When cases come to court, judges oftenprovide their interpretation on the words inthe SOP, for example in relation to intensefear, helplessness and horror. Does the RMAreview those psychiatric conditions whenthat happens?

RESPONSE

Professor Raphael responded that the RMAdoes take note of issues that are brought upfrom individual cases. The RMA will lookback and see what it based the wording on,and how that’s reflected in the literature. Oneof the limitations is what the literatureshows, even though the RMA tries to dissectout these things, sometimes the literaturedoesn’t dissect them out at all.

Severe Stressor Definition

ISSUE

Some SOPs specify that a stressor beextreme or severe and there are manysituations in which a veteran does notremember a severe stressor despite being ina combat situation for quite some time.

RESPONSE

Professor Raphael agreed that in the

circumstances of combat there are a range ofstressors, not all of which would meet thedefinition of a catastrophic stressor. Thequestion that has come up for the RMA isdoes there have to be an actual or objectivethreat or is it sufficient for there to be aperception of it. That is a complex boundary.Nobody would argue that in combat there isactual threat of death. Less severe stressorsmay result in other conditions which arecovered by the SOPs.

Lack of Evidence of Exposure to aStressor

ISSUE

We see cases of veterans with a psychiatricdisability but no incident or evidence of realstress, apart from service itself. The SOPsdon’t seem to apply to these cases. Are youlooking at that area? Shouldn’t the principleof beneficiality apply?

RESPONSE

Professor Raphael explained that it certainlycould come up in what the RMA is lookingat because there’s an evolving literature inthe whole aetiology of a range of psychiatricdisorders. She pointed out that at least one infive of the Australian community has adiagnosable psychiatric disorder. The yearsof being a soldier or being in the services arethe years in which these disorders come on,so there is also potentially a coincidentaleffect. There are a great many people in theages of 18 to 40 who develop significant anddisabling psychiatric illnesses who havenever had exposure to combat. So that’s whatcomplicates understanding whether theservice has contributed in this particularcircumstance. While the RMA is open tolooking at the issues, there has to be anidentified causal chain.

Professor Donald added that the principle ofbeneficiality applies at a number of levels in


the system. In the evidentiary area wherethere is reverse standard of proof, thencertainly you would expect that it wouldapply. At the reasonable hypothesis level, theevidence that is needed to put in a causalfactor is in fact a very generousinterpretation of the literature. It’s not reallya question of whether the generous nature ofthe legislation can be applied here. Theproblem for the RMA is that the issue is socomplex that simple solutions that can besustained are very difficult to find. Findingthe evidence in the literature to sustain orunderpin those simple solutions is verydifficult because the literature is verycomplicated indeed.

Professor Raphael further added that theliterature is becoming more sophisticated soit may become more obvious in future howto distinguish between service and otheraspects of life. Also, it may be that a systemof management might be more clearly linkedto care than compensation.

Prolonged or Cumulative Stress

ISSUE

In respect of the severe stressor, thestatement of principle for PTSD specifies anactual event, one single event, as opposed tosomeone who has been involved withprolonged stress or cumulative stress in acombat zone. For example, there is theveteran who has had the fear of beingattacked or overrun for his whole period ofservice in Vietnam. There doesn’t seem to bea window of opportunity for that veteranwho may be suffering from PTSD.

RESPONSE

Professor Raphael replied that there is notthe intent to exclude a veteran who has hadexposure to that sort of stress but the fact isthat in the literature a lower level of stressthat is chronic may be more associated with

anxiety and depressive disorders than withPTSD per se. So, sometimes the differentdefinitions are reflective of that. Generally,if a person has been exposed to the sort oflife threat continuously in combat, he wouldhave had at least one event like that, so itshouldn’t exclude him from being eligiblefor PTSD if that’s the condition that he has.This is a matter for ongoing consideration bythe RMA.

Stress from Natural Disasters

ISSUE

What is the literature in relation to stressfrom natural disasters such as Cyclone Tracyand the Brisbane floods and how can it beapplied to military settings?

RESPONSE

Professor Raphael responded that the RMAdoes take the general disaster literature intoaccount as well as military studies. Studiesof disasters world-wide have shown impactsfrom the acute stressor exposure as well asfrom the chronic stressors of the aftermath.

Health Studies

Separating Out Multiple Deployments

ISSUE

There is an overlap between Korean Warveterans and veterans of other conflicts;some Korean War veterans went on to servein Vietnam and some had already served inthe British atomic test program or World WarII. How is this taken into account?

RESPONSE

Dr Keith Horsely replied that work is beingdone to separate these groups. There wereabout 30% of Korean War veterans whowere World War II veterans and19% whowere also Vietnam War veterans. An analysis


of the veterans who only went to Korea isplanned. It will be difficult to construct acomparison group because nearly all theactive army went to Korea. A comparisonwith older Australian men is beingundertaken. Mr Bill Maxwell (DVA) addedthat the phenomenon of soldiers going onmultiple deployments is also a feature ofmore recent deployments.

Dealing With Findings of ExcessSymptoms

ISSUE

There are some diseases that may not havebeen identified from previous health studies.How can the system deal with the findingsof excess symptoms being identified bythese studies? Is it within the authority of theRMA to make recommendations to theDepartment?

RESPONSE

Professor Donald responded that there wasno easy answer to that problem, which is alsolikely to unfold when the F111 health studyis completed. Further investigation may benecessary to discover whether a new diseaseis being uncovered. The RMA has to workwith the legislation. With health studiesthere is sometimes an unrealistic expectationthat the RMA will be able to act upon thefindings when in fact it will not produce anoutcome that fits into the legislation.

Professor Donald added that the RMA doesmake recommendations to the Commission.The RMA has been advocating forprospective surveillance and ongoingsystematic record keeping for nearly tenyears. The RMA also has the opportunity tomake suggestions or comments through theMedical Advisory Panel.

Mr Bill Maxwell (DVA) explained that theVEA does not require a person to lodge aclaim for a diagnosed medical condition. It

requires them to lodge a claim for theacceptance of incapacity, which can bedescribed in terms of a range of symptoms.The department has to investigate the claim,establish the medical conditions present andthen apply the relevant SOPs. If no label forthose symptoms can be identified, the claimcan be dealt with as a non-SOP condition. DrKeith Horsley (DVA) added that it is DVApolicy to provide treatment until such timeas a label is applied to those symptoms.

Professor Wilson pointed out thatretrospective health studies can help identifyconstellations of symptoms which might goto form a new syndrome. The advantage ofprospective studies is that they provide theopportunity to study the exposures thatactually contribute in some way to thosesymptoms or syndromes.

RMA Response to F111 Health Study

ISSUE

What action has the RMA taken in regards tothe findings of the F111 Health Study andany other DVA/Defence Study?

RESPONSE

Professor Donald said that the RMA hasbeen watching the F-111 study carefully. ThePrinciple Medical Officer attends themeetings of the Steering Committee and theRMA gets the minutes of that meeting. TheRMA has advertised an investigation intotoxic encephalopathy which is specificallyaimed at determining whether there is adisease under which some of the F111deseal/reseal people would fit. Some ofthem may be covered by the existingpsychiatric SOPs. However, there willpotentially be people who were involved inthe reseal/deseal process who do not fit intothe categories that a SOP would cover.

The RMA did obtain a copy of the chemicalcomposition of the materials that were used


and were staggered by the list of chemicals.Dr Ian Gardner (Defence) pointed out thatthe study is not specifically about the use ofchemicals, but also about deseal/resealprocesses. Professor Donald furtherexplained that it was recognised right upfront that it would be impossible to identifythe contribution of this huge range ofchemicals, most of which had very poorlydescribed toxicological principles. Thereforethe study is really basically looking at theprocesses and asking the question, “Is therean association between working on theseprocesses and adverse health outcomes?”

Mr Andrew Leiboff (medical researchofficer) added that the members of the RMAand Secretariat have been made aware of thechemicals which were used by this groupand will routinely look for evidence of acausal link between those chemicals and thedisease under investigation.


Appendix 1


Determination of SOPs Flowchart – 2004 Forum

Request from eligible person ororganisation for an investigation relatedto Disease X

Is the condition under consideration adisease, under VEA 5D(1)?

NoSOP

Is there published peer reviewed evidence onDisease X, under VEA 5AB(2)(a)(i)?

SMSE indicates or points to a

causal association between a

potential factor and Disease X,

under VEA 196B(2)

RMA members’ assessment of causation ofDisease X in relation to each potential factor

No SOP factorSOP factor withdose for RH

SOP factor withdose for BoP

SMSE shows that it is more

probable than not that there is

a causal association between a

potential factor and Disease X,

under VEA 196B(3)

No SOP factorfor BoP

Clinical judgement,VEA Section5AB(2)(a)(ii)


Appendix 2


Background Terminology:

Introduction To Some Epidemiological Terms

Dr Alex Bordujenko, RMA Secretariat


Epidemiology

This is the study of variations in diseasefrequency among population groups, and thefactors that influence these variations. Theprinciple objective of epidemiology has beento determine factors which may cause orcontribute to disease processes in humans,so that preventive measures may be applied.

Epidemiologic observations have a longhistory, with much work developed throughthe study of acute epidemic diseases such ascholera and typhoid. The discipline hasburgeoned over the latter half of theTwentieth century, with interest in the studyof the cause, treatment and prevention ofcancer, cardiovascular and other chronicdisease, and of course the advent ofcomputer storage and analysis systems.

Approaches for Epidemiological Study

Hennekens and Buring (1987) defineepidemiology as “the study of thedistribution and determinants of diseasefrequency in human populations.”“Humans” distinguishes the approach fromresearch using animal or other systems inexperiments. “Populations” contrasts thepractise of individual investigation as inclinical research. “Frequency” indicates thequantification of disease occurrence and therisk attributable to various potential causes.The term “distribution and determinants”points to the two major approaches ofepidemiology:

1. examination of the distribution ofdisease frequency in populations (thiscan produce hypotheses about thecauses of disease) known asdescriptive studies; and

2. analytical studies which test thesehypotheses by reviewing personalcharacteristics or exposures amongindividuals within the populations.

Descriptive studies use population basedstatistics on mortality, disease incidence, andsurvival. Other registries for examplehospital based disease registries, may also beuseful. Obviously the studies concernpopulations and not individuals andmeasures of any exposures are usually broadand may be subject to confounding orinterfering factors. Selection of free livingpopulations may introduce biases andconfounding into the calculations.Examination of national and internationaltrends, migrant studies and time trends hasprovided valuable insights into the causationof a number of chronic diseases for examplebreast, prostate and lung cancers.

Analytical studies have provided muchuseful information concerning the discoveryand/or confirmation of a number of lifestyleand other environmental exposures as causesof chronic disease, including cancer.Examples of these include cigarettesmoking, where, for smokers of 40 or morecigarettes per day there is a risk of lungcancer of more than twenty times that of anon-smoker. Another well documentedexample is occupational exposure toasbestos and the development ofmesothelioma, where the relative risk is wellover 100 fold that of the unexposedpopulation. Analytical studies from severalinternational sources in the last decade havealso demonstrated that both the incidenceand recurrence of neural tube defects can begreatly reduced by maternal folatesupplementation in early pregnancy, even inthe absence of maternal folate deficiency.

In chronic disease epidemiology, the types ofanalytical studies encountered are:

A. Cohort studies identify groups ofindividuals with and without aparticular exposure, and follow themover time to examine diseaseincidence and/or mortality rates.


These may be current or pastexposures. An association issuggested when rates of disease ordeath differ between the groups.These are able to directly measureincidence and mortality rates relatedto a particular exposure (especiallywith prospective design) but theyrequire large numbers of exposedindividuals particularly whenconsidering uncommon diseases,before significant differences may benoted.

B. Case-control studies or case-referentstudies identify people with aparticular disease (case), and a groupof people without the disease(controls), and then collectinformation about past exposures, forexample by interview orquestionnaire. They provide a methodof studying rare diseases but may besubject to recall and other biases, anddifficulty in measuring pastexposures.

Data Presentation and Interpretation

The odds ratio (OR) is a measure ofassociation used in case control studies toestimate the odds of exposure in cases to theodds of exposure in controls. Thisapproximates, but is not synonymous with,the “relative risk” (RR) the measure ofassociation used in cohort studies. The termrelative risk (RR) is used to describe thecomparison of the risk of a known exposedgroup versus a known unexposed groupdeveloping a specific condition. Thus if therelative risk is one the risk is the same forboth groups and exposure is not seen to beassociated with the development of theparticular condition that is, there is noincrease in the risk of a studied outcomewith the exposure of interest. If the RR (orOR) is 1.5 then the risk for the studied

outcome in the exposed versus theunexposed group is increased by 50%. AnRR (or OR) of 2 implies a doubling of risk,and an RR (or OR) of less than one impliesa reduction of risk. Problems in decisionmaking occur when the described increase inrisk is weak (under a two to three foldincrease) and particularly when the relativerisk is close to one, for example 1.1 (10%increase) or 1.3 (30% increase) rather thanthe 20 fold increases for heavy cigaretteconsumption and the incidence of lungcancer and the much grater increases seenwith occupational asbestos exposure and theincidence of mesothelioma. Manyepidemiologists are reluctant to accept asreal, increases in risk of less than 100%(RR<=2) as likely to be causative unless the“Bradford Hill” types of criteria arestringently applied to the body of evidencepertinent to the putative association, andoverall, a considered case can then be madeto support causality.

Another term, the “confidence interval”(CI), is used to describe the range of relativerisk (or odds ratio) rates within which theactual result lies, to within, for example, a95% probability. Thus, if the confidenceinterval includes one then the result couldhave occurred due to chance and no trueeffect may exist. If the 95% confidencelimits exclude one it does not exclude thepossibility of a chance result, rather itindicates that chance would explain theobserved (or a greater) risk estimate onlyone out of 20 times.

Selected Measures of DiseaseFrequency

As well as the relative risk and odds ratio anumber of other measures of diseasefrequency need to be considered. Aconsideration of the basic concepts of thesemeasures includes the formulae used tocalculate such measures. In its simplest form


data from a two-by-two table from a case-control or cohort study with countdenominators would appear as:

DiseaseYes No Total

Exposure Yes a b a+bExposure No c d c+dTotal a+c b+d a+b+c+d

a = the number of individuals who areexposed and have the disease

b = the number who are exposed and donot have the disease

c = the number who are not exposed andhave the disease

d = the number who are not exposed andwho do not have the disease

As stated above for cohort studies the termrelative risk (RR) is used to describe thecomparison of the risk of a known exposedgroup versus a known unexposed groupdeveloping a specific condition, that is theincidence of the disease in the exposeddivided by the incidence in the unexposedIe/Io or the cumulative incidence of thedisease in the exposed divided by thecumulative incidence in the unexposedCIe/CIo.

The formula for calculating relative risk forcohort studies with count denominators isthus:

Ie = CIe = a/(a+b)Io Cio c/(c+d)

(where a,b,c,d are derived from the 2x2 tableoutlined above).

For case control studies with countdenominators the odds ratio is expressed as:

a/c = adb/d bc

(where a,b,c,d are derived from the 2x2 tableoutlined above)

The odds ratio is said to provide a validestimate of the relative risk for case-controlstudies where the cases are newly diagnosed,where prevalent cases are not included in thecontrol group and where the selection ofcases and controls is not based on exposurestatus.

Attributable risk is the measure whichprovides information about the absoluteeffect of the exposure and is the excess riskof disease in those exposed compared withthose who are unexposed to a specific factor.This measure is defined as the differencebetween the incidence rates in the exposedand unexposed groups and may becalculated in cohort studies as

AR = CIe – CIo = a/(a+b) – c/(c+d)

(where a,b,c,d are derived from the 2x2 tableoutlined above)

The attributable risk percent (AR%),attributable rate percent attributableproportion or etiologic fraction is calculatedas the attributable risk divided by the rate ofdisease among the exposed and is said torepresent the proportion of disease in thatgroup that could be prevented by absence ofthe exposure.

AR% = AR/Ie x 100 = (Ie – Io) x 100 =(1- Io/Ie) x100 = (1-1/RR) x 100 =

(RR-1/RR) x 100

Population Attributable Risk (PAR) is themeasure used to estimate the excess rate ofdisease in the total study population ofexposed and unexposed individuals that isattributable to the exposure. The PAR iscalculated as the rate of disease in thepopulation (incidence rate in totalpopulation = It) minus the rate in theunexposed group (Io):

PAR = It – Io

or by multiplying the AR by the proportion


of exposed individuals in the population(Pe):

PAR = (AR) x (Pe)

Population Attributable Risk Percent(PAR%) is represented by:

PAR% = 100 x (Pe) x (RR-1)/1+(Pe) x (RR-1)

Epidemiologic Studies Need CarefulExamination

(Refer to:- Darzins PJ, Smith BJ and HellerRF (1992). How to read a journal article.The Medical Journal of Australia, Vol 157 pp389-394.)

The size of the population studied isimportant – the larger the sample size thegreater the power (or ability) to detect aspecified risk, the smaller the sample sizethe weaker the power. Negative results fromsmall studies may not be conclusive as onlylarge studies may confidently exclude orinclude low to moderate levels of risk.

When examining any study results,consideration of the possibility of a non-causal association is necessary. Theobserved association between exposure anddisease may result from bias, confounding,chance, or cause-and-effect.

Bias is the term used for any systematicerror in a study and may occur during studyselection, information gathering or inreporting of the assessment of the exposureor outcome under investigation.Confounding bias is the possibility of theobserved effect being due to other variablesnot adequately considered in study design oranalysis of the results. Many types of studybias have been described includingselection, information, recall, andinterviewer bias. Confounding bias orconfounding is due to variables which maythemselves account for all or part of an

apparent association between an exposureand a disease. They may also obscure anassociation. Chance is considered previouslyin the discussion of study power andConfidence Intervals.

Study Types

Study design has an effect on the quality ofevidence which may be gained and arecognised ‘hierarchy’ of study types exist.In developing the following the “USPreventative Services Task Force:Guidelines for Quality of Evidence” (Fisher,1989) have been considered. Given thespecific needs of the RMA somemodification has been undertaken. In thisinstance the level of evidence available is atbest observational (cohort or case controlstudies). The following is broadly thedivision of available study designs and howthese may be considered in the informationgathering.

Analytic Studies:

1 Intervention Studies1a Randomised Controlled Trial1b Controlled Trial

2 Observational Studies2a Cohort-Prospective2b Cohort-Retrospective

3 Case Control Studies

Descriptive Studies:

4 Population (Correlational)

5 Individual5a Cross Sectional Surveys5b Case Series5c Case Reports

Where the numbering 1-5 refers to the gradeassigned to the quality of the evidence.Quality refers here to study design ratherthan individual study merit that is that theevidence from cohorts is graded as higher


than that from case control studies – this isthe method used by the US PreventativeServices Task Force.

While the RMA places emphasis on primaryresearch published in the leading peerreviewed journals of either broad ordiscipline specific type; published, peerreviewed, reports on the epidemiology ofdisease such as those produced from time totime by the International Agency forResearch into Cancer, the National Academyof Science, or the Surgeon Generals’ Reportsconcerning to smoking related disease; areconsidered appropriate sources forexamination. Published reports from sourcessuch as the National Health and MedicalResearch Council and other expertcommittees are also be considered wherecontemporary, applicable material isavailable.

Consideration of Individual Studies

(Refer to:- Darzins PJ, Smith BJ and HellerRF (1992). How to read a journal article.The Medical Journal of Australia, Vol 157 pp389-394.)

In the absence of interventional studies suchas randomised controlled trials most relianceis placed on well designed and reportedcohort and case control studies andProfessor Heller provides a method ofconsidering these. This forms a mentalcheck list in consideration of materials.

The following questions may be specificallyaddressed.

1. What is the research question?2. What is the study type?3. What are the outcome factors and how

are they measured?4. What are the study factors and how are

they measured?5. What important confounders are

considered?6. What are the sampling frame and

sampling method?7. How many subjects reached follow-up?8. Are statistical tests considered?9. Are the results clinically/socially

significant?10. What conclusions did the authors reach

about the study question?

After determining these features a decisionon adequacy of methods and clarity ofresults is made considering:

bias – are the results biased in one direction.If so, what is the direction and magnitude ofbias

confounding – are there any seriousconfounding or distorting influences? Hasan attempt been made to deal with these andhas this been adequate?

chance – is it likely the results occurred bychance? Consideration of the statisticalcontent of the study.

It is recognized that for many putativefactors evidence may only be available indescriptive studies. This is often the case forcase reports or case series of diseaseassociations or drug reactions.

Association and Causation

Association is the term used to describe thestatistical dependence between twovariables. In epidemiology it is the degree towhich the rate of disease in persons with anexposure of interest is either higher or lowerthan the rate of disease among those withoutthat exposure. Such an association does notmean, or even imply, that the observedrelationship is one of cause and effect(Hennekens and Buring, 1987).

Making judgements about causality fromepidemiologic data involves a logical


process which addresses two major areas:

1. Whether for any individual study, theobserved association between anexposure and disease is valid. Anassessment of validity requires aconsideration of the likelihood ofalternative explanations for the resultsand chance (the luck of the draw), bias(any systematic error in the study forexample in subject selection,information gathering or reporting), orconfounding (the observed effect beingdue to other variables not adequatelyconsidered in study design or analysisof the results); and

2. Whether the body of the evidenceconsidered supports a judgement ofcausality. In this process standardepidemiological criteria are used(Hennekens and Buring, 1987).

Epidemiologic criteria used to assist inthe assessment of causality

The RMA considers the individual studieswith respect to the above and then, inconsidering the available evidence usesstandard epidemiological criteria to make ajudgement regarding causality with regard tothe reasonable hypothesis and balance ofprobabilities standards of proof. TheBradford Hill criteria (Bradford-Hill, 1965),and more contemporary versions, are widelyaccepted in the interpretation ofepidemiological studies for the purpose ofassessing the possibility of a causalassociation.

Consideration of the body of evidenceavailable for each contention against thecurrent epidemiologic criteria will result in ajudgement regarding causality. As ProfessorHolman (1997) notes, more than 30 differentsystems of causal verification have beendescribed. In his technical appendix to thePearce Report he outlines ten criteria for

classification of evidence of causality, basedon work by Mervyn Susser. The RMA hasconsidered a number of such systemsincluding those of Bradford Hill, Susser andthose co-authored by Professor Holman in“The Quantification of Drug CausedMorbidity and Mortality in Australia, 1995”(English and Holman, 1995). The RMArecognises the underlying similarities whichunderpin these systems.

The exact description of these epidemiologiccriteria varies between authors and the RMArecognises the need to consider both internalstudy validity (for individual studies) andfactors important in the body of evidence(the applicable evidence available fromepidemiological, clinical, toxicological andother research) in these criteria.

Sir Austin Bradford Hill, as well as otherprominent statisticians and epidemiologists,including Mervyn Susser and KennethRothman, have described how the subjectivelikelihood (or the correct judgement) of acausal relationship is increased whenevidence relating to an association meetscriteria devised to consider the availableevidence. The Bradford Hill (Bradford-Hill,1965) criteria are as follows:

1. Strength of Association2. Consistency3. Specificity4. Temporality5. Biological Gradient6. Plausibility7. Coherence8. Experimental evidence9. Analogy

The criteria used by the Expert Committeeon Herbicide Exposure and Spina Bifida(1996) further refined the criteria toexplicitly include consideration of bias andconfounding in the criteria:

1. Statistical significance (that is the


possibility of chance being responsiblefor an apparent association; and studypower)

2. Strength of association3. Consistency of association between

studies4. Possibility of bias in measurement of

exposure or outcome5. Possibility of selection or confounding

bias6. Time sequence7. Dose response8. Biological plausibility (including

aspects of theoretical coherence,biological coherence and factualcoherence)

1. Statistical significance and power

If the criterion of statistical significance issatisfied then the evidence is supportive ofan association. The failure of a test to reachstatistical significance in the presence ofadequate statistical power provides evidenceagainst the association, however in theabsence of adequate statistical power it maynot necessarily detract from the association.

2. Strength of association

The greater the strength of association themore likely it is to be causal. Confounding isless likely to explain a strong associationbecause the strength of the associationbetween the confounding variable and theoutcome must also be strong. While a strongassociation is supportive of causality, a weakassociation may not necessarily detract fromthe evidence of causality however adequateconsideration of potential confounding orbias is essential.

3. Consistency of replication

Consistency of the evidence or the lack ofevidence in the face of study diversity intime, place, circumstances and population,as well as research design, strongly supports

or detracts from a causal hypothesis.

4. Possibility of bias in measurement ofexposure or outcome

Consideration of any systematic error in thestudy in information gathering or inreporting of the assessment of the exposureor outcome under investigation. Absence ofbias in the studies considered to show apositive association supports the existence ofa putative association. The presence of biasdetracts from the conclusions which may bedrawn from the information.

5. Possibility of selection orconfounding bias

Consideration of any systematic error in thestudy in subject selection; or the possibilityof the observed effect being due to othervariables not adequately considered in studydesign or analysis of the results. Absence ofbias or confounding in the studiesconsidered to show a positive associationsupports the existence of a putativeassociation. The presence of bias oruncontrolled confounding detracts from theconclusions which may be drawn from theinformation.

6. Time sequence

The exposure must precede the disease orinjury. This criterion is compatible with, butdoes not necessarily support causality.Reversal of the order of exposure anddisease or injury is the most persuasive basisavailable for rejection of causality.

7. Dose response

A response which is in proportion to thelevel of exposure is strongly persuasive of acausal relation. However, its absence doesnot necessarily detract from the association.

8. Biological plausibility

(aspects of theoretical coherence, biological


coherence and factual coherence)

Theoretical coherence: Findings plausible interms of pre-existing theory are supportiveof the association. Conversely, findings thatare implausible in terms of pre-existingtheory detract from the evidence.

Factual coherence: Compatibility of anew result with pre-existing facts issupportive of the association.Incompatible pre-existing facts stronglydetract from evidence of causality.

Biological coherence: Pre-existingknowledge which identifies a mechanismby which the chemical exposure mayproduce the disease or injury issupportive of case for the associationbeing causal. Observations from speciesother than humans may also be used tosupport the potential mechanism ofaction. Incoherence between biologicalknowledge and study observationsdetracts from the case for a causalassociation.

As Rothman and Greenland (1997)eloquently acknowledge inductively orientedcausal criteria are not sufficient withinthemselves and require sound scientificjudgement to traverse the path for which thecriteria are “the road map throughcomplicated territory”.


Bibliography

1. Bradford-Hill A: The Environment and Disease: Association or Causation. Proceedingsof the Royal Society of Medicine, 58:295-300, 1965.

2. Darzins PJ, Smith BJ and Heller RF (1992). How to read a journal article. The MedicalJournal of Australia, Vol 157 pp 389-394.

3. English D and Holman D (1995). “The Quantification of Drug Caused Morbidity andMortality in Australia, 1995.” National Drug Strategy; Commonwealth Department ofHuman Services and Health.

4. Expert Committee (1996). “Report of the Expert Committee Into the PossibleConnections Between Exposure to Herbicides in Vietnam and Spina Bifida in Children ofVietnam Veterans”. Report prepared for the Repatriation Commission, Australia.

5. Hennekens CH and Buring JE (1987) Epidemiology in Medicine First Edition LittleBrown and Co Boston

6. Holman D (1997). Technical Appendix: Criteria for assessing Causation in Pearce D,Review of the Repatriation Medical Authority and the Specialist Medical Review Council,Commonwealth of Australia, p120.

7. Rothman KJ and Greenland S Causation and Causal Inference. In: Detels R, HollandWW, McEwan J, Omenn GS eds, Oxford Textbook of Public Health, Volume 2: TheMethods of Public Health, pp 616-629. New York: Oxford University Press, 1997.

8. US Preventative Services Task Force: Guidelines for Quality of Evidence Ed M Fisher.William and Wilkins. Baltimore Maryland, 1989. Methodology pp 27-37 and Appendix App 387-397.


Appendix 3


CRITICAL APPRAISAL CHECKLIST

What is the study about?• What is the study hypothesis (research

question)?• What is the study type? • What is the outcome factor (disease of

interest) and how is it measured?• What are the risk factors and how are

they measured?• Where did the study subjects come

from?

What are the main results?• What is the size of the effect? • Are the results statistically significant?• What are the confidence intervals?

Can the results be explained byanything else apart from the riskfactor under consideration?• What are the important potential

confounders in this association andwere they taken into account?

• Is there any bias in the selection ofstudy subjects or in the measurementof exposures or outcomes?

• Could the results be explained bychance?

What is the evidence for causation?• What is the strength of the

relationship?• Is there a dose response effect?• Is it clear that the exposure precedes

the outcome?• Are the results consistent with other

evidence?• Do the results make sense from our

understanding of biology?

Conclusions• What conclusions do the authors

reach? Do you think that they arereasonable?

Workshop SummaryA workshop was conducted at the forum forthe purpose of giving delegates some directexperience of the processes by which theRMA works in deciding on which factorsshould be included in a Statement ofPrinciples. The workshop was divided intotwo parts. For the first part, participantswere divided into work groups of aroundeight people and provided with two studieseach on which to go through a criticalappraisal checklist (see box). Each groupwas allocated one of two topics relating toissues recently considered by the RMA:hepatitis C as a risk factor for diabetes, andinadequate intake of dietary fibre as a riskfactor for colorectal cancer.

For the second part of the workshop,participants came back together for adiscussion about whether the body ofevidence was strong enough to support theinclusion of these risk factors in either theRH or BoP instruments. Participants wereprovided with a table summarising all theavailable peer-reviewed studies concerningthose particular factors.

Professor Andrew Wilson led the discussionon hepatitis C and diabetes after Dr IanSmith (RMA Secretariat) had gone throughthe answers to the critical appraisal questionsfor the two hepatitis C studies. All the groupsfelt that the material was sufficient only toinclude the factor in the RH instrument. Thiswas in accordance with the decision of theRMA. Professor Wilson pointed out that inthis case the exposure was easy to define, asit was simply a positive blood test. The issueof latency (time interval between exposureand outcome) was harder to resolve becauseof the lack of available data, particularly alack certainty about a clear biologicalmechanism. Different mechanisms wouldinvolve different time frames. The mostgenerous interpretation of the data in thiscase was to have no latency period.

the challenges faced by the RMA whenmaking decisions about risk factors. Inrelation to hepatitis C and diabetes, thesituation was one of having to come up witha decision based on very limited informationof varying quality. In relation to dietary fibreand colorectal cancer, there was a lot moreevidence available but the results wereconflicting and there was an additional needto decide on dose. In either case, thedecisions could change if new informationfrom well conducted studies becomesavailable.


The final SOP factor (RH only) was:

“having hepatitis C virus infection beforethe clinical onset/clinical worsening ofdiabetes mellitus”

Professor John Kaldor led the discussion ondietary fibre and colorectal cancer afterAndrew Leiboff (RMA Secretariat) hadgone through the answers to the criticalappraisal questions for the two dietary fibrestudies. Professor Kaldor discussed twographs showing a summary of the resultsand confidence intervals of all availablecase-control and cohort studies. Cohortstudies are considered to be a stronger formof evidence. He pointed out that in the graphsummarising the case-control studies mostof the results were below the level of noeffect, giving the overall impression thatmost studies suggest that dietary fibreprotects against colorectal cancer. However,the graph summarising cohort studiesshowed that most of the results indicated noeffect, apart from one recent large study.

In conclusion, there was enough indicationto put a factor in the RH instrument, butinsufficient indication to put a factor in theBoP instrument. Professor Kaldor explainedthat the RMA then had to decide upon thedose of fibre and the length of time in whichfibre intake has to be reduced before cancerrisk starts to increase. These decisions werebased on the information in the availablestudies.

The final SOP factor (RH only) was:

“an inability to consume an average dailyintake of 20 grams of fibre in food (or atotal of 36 500 grams of fibre in food)over a continuous period of five yearswithin the ten years immediately beforethe clinical onset of malignant neoplasmof the colorectum”

Professor Kaldor concluded that these twoexamples were good illustrations of some of


Appendix 4


Bibliography/ReferencesAmerican College of Occupational and Environmental Medicine (1999). Position Statement:multiple chemical sensitivities: idiopathic environmental intolerance. JOEM 41: 940-942.

Baume P, Bomball R, Layton R. (1994) A fair go: report on compensation for veterans andwar widows, Australian Government Publishing Service, Canberra,.

Bernstein D and Kelley T. (1995) The Gulf War comes home: sickness spreads, but thePentagon denies all, The Progressive, Madison, USA.

Bingham S et al (2003). Dietary fibre in food and protection against colorectal cancer in theEuropean Prospective Investigation into Cancer and Nutrition (EPIC): an observational study.The Lancet, Vol 361, pp. 1496-1501.

Clauw DJ, Engel CC Jr, Aronowitz R, Jones E, Kipen HM, Kroenke K, Ratzan S, Sharpe M,Wessely S. (2003). Unexplained symptoms after terrorism and war: an expert consensusstatement. J Occup Environ Med, Vol 45(10), pp 1040-8.

Darzins PJ, Smith BJ and Heller RF (1992). How to read a journal article. MJA, Vol 157, pp389-394.

Donald K, Bordujenko A. (1999) Is it time to limit the role of the judiciary in compensationmatters? Aust NZ J Pub Health, Vol 23, pp 328-30.

Hill AB. (1965) The environment and disease: association or causation? Proc Roy Soc Med Vol58, pp 295-300.

Fukuda K, Straus SE et. al (1994). The chronic fatigue syndrome: A comprehensive approachto its definition and study. Annals of Internal Med, 121(12) pp 953-959.

Hyams KC, & Roswell RH. (1998) Resolving the Gulf War Syndrome question. AmericanJournal of Epidemiology, Vol 148(40 pp 339-42.

Hyams KC, Wignall FS, Roswell R. (1996). War syndromes and their evaluation: from the UScivil war to the Persian Gulf war. Annals of Internal Medicine, Vol 125 (5) pp 398-405.

Hyams KC. (1998). Developing case definitions for symptom-based conditions: the problemof specificity. Epidemiologic Reviews, Vol 20(2) pp 148-156.

Idler EL, Benyamini Y (1997) Self-rated health and mortality: a review of twenty-sevencommunity studies. Journal of Health & Social Behavior Vol 38(1) pp 21-37

Jones E, Hodgins-Vermaas R, McCartney H, Everitt B, Beech C, Poynter D, Palmer I, HyamsK, Wessely S. (2002). Post-combat syndromes from the Boer War to the Gulf War: a clusteranalysis of their nature and attriubtion. eBMJ, Vol 324, pp 1-7.

Kipen HM, Fiedler N (2002). The role of environmental factors in medically unexplainedsymptoms and related syndromes: conference summary and recommendations. EnvironmentalHealth Perspectives, 110(Suppl 4) pp 591-5.


Lupton G and Najman J (eds) (1989). Sociology of Health and Illness, Australian readings.Macmillan, Melbourne.

Mai V et al (2003). Dietary fibre and risk of colorectal cancer in the Breast Cancer DetectionDemonstration Project (BCDDP) follow up cohort. International Journal of Epidemiology, Vol32, pp. 234-239.

Mehta S, Brancati F, Strathdee S et al (2003). Hepatitis C Virus Infection and Incident Type 2Diabetes Mellitus. Hepatology, Vol 38, pp. 50-56.

Marshall GN, Davis LM, Sherbourne CD. (1999) A review of the scientific Literature as itpertains to Gulf War Illnesses, Vol 4: Stress, RAND, Available at:www.rand.org/publications/MR1018.4.1/MR1018.4.1.html

Okan V et al (2002). Increased frequency of HCV but not HBV infection in Type 2 diabeticpatients in Turkey. IJCP, Vol 56, pp. 175-177.

Pearce D, Holman D. (1997) Review of the Repatriation Medical Authority and the SpecialistMedical Review Council, Department of Veterans’ Affairs, Commonwealth of Australia,Canberra, pp 21-28.

Sim M, Abramson M, Forbes A, Glass D, Ikin J, Ittak P, Kelsall H, Leder K, McKenzie D, McNeil J, Creamer M, Fritschi L. (2003) Occupational and Environmental Health Unit MonashUniversity. Australian Gulf War Veterans’ Study. 2003. Canberra, Commonwealth Departmentof Veterans’ Affairs

Susser M. (1977) Judgment and causal inference: criteria in epidemiologic studies. AmericanJournal of Epidemiology.Vol 105: 1-15.

Research Working Group of Military and Veterans Health Coordinating Board (2002). AnnualReport to Congress: Federally sponsored research on Gulf War veterans’ illnesses for 2001.Department of Veterans Affairs USA.

Wessely S, Unwin C, Hotopf M, Hull L, Ismail K, Nicolaou V & David A. (2003) Stability ofrecall of military hazards over time. Evidence from the Persian Gulf War of 1991. BMJ, Vol183 pp 314-22

Working group. (2002). Chronic fatigue syndrome. Clinical practice guidelines – 2002. TheMedical Journal of Australia, Vol 176 Supplement pp s19-s56.

Worthy D, Greenslade A, Roberts P. (1992). Audit Report No. 8 1992-3. Efficiency auditDepartment of Veterans’ Affairs: compensation pensions to veterans and war widows,Australian National Audit Office, Australian Government Publishing Service, Canberra.


Appendix 5


Organisations Represented at the ForumAPPVA Australian Peacekeepers & Peacemakers Veterans Association

ASASA Australian Special Air Service Association

AVADSC Australian Veterans & Defence Services Council

DoD Department of Defence

DVA Department of Veterans’ Affairs

RDFWA Regular Defence Force Welfare Association

RMA Repatriation Medical Authority (the Authority)

RSL Returned & Services League of Australia Limited

SMRC Specialist Medical Review Council

TPI Australian Federation of Totally & Permanently Incapacitated Ex-Servicemenand Women

VRB Veterans’ Review Board

VVAA Vietnam Veterans Association of Australia

VVFA Vietnam Veterans’ Federation of Australia


Documents

Proceedings of 2004 CANBERRA FORUM - Introduction to the RMA