SHI2010 - Aalborg Universitet · We hope this conference in 2010 in Copenhagen will be as successful as the previous confe-rences. It has been a pleasure to receive a larg e number

SHI2010

Proceedings 8th Scandinavian Conference

on Health Informatics Copenhagen, August 23 - 24, 2010

Editors Ann Bygholm

Pia Elberg Ole Hejlesen

ISBN 978-82-519-2606-5

SHI2010



Editors Ann Bygholm


ISBN 978-82-519-2606-5

SHI2010



Editors Ann Bygholm


ISBN 978-82-519-2606-5

SHI2010



Editors Ann Bygholm


ISBN 978-82-519-2606-5

SHI2010 Proceedings Editors Ann Bygholm, MA, PhD Department of Communication and Psychology Aalborg University Kroghstræde 1 DK-9220 Aalborg Denmark Pia Elberg, MSc Department of Health Science and Technology Aalborg University Fredrik Bajersvej 7 D1 DK-9220 Aalborg Denmark Ole Hejlesen, MSc, PhD Department of Health Science and Technology Aalborg University Fredrik Bajersvej 7 D1 DK-9220 Aalborg Denmark © SHI 2010 & Tapir Academic Press, Trondheim 2010 ISBN 978-82-519-2606-5 This publication may not be reproduced, stored in a retrieval system or transmitted in any form or by any means; elec-tronic, electrostatic, magnetic tape, mechanical, photocopying, recording or otherwise, without permission. Copyright remains with the authors. All papers contained in these proceedings may be printed for use and republished by non-profit organizations on condition that a copy of the publication is sent to the publisher at the address below, and SHI2010 Proceedings are listed as the source. Layout: The editors Printed and binded by: Tapir Uttrykk, Trondheim, NORWAY Tapir Academic Press publishes textbooks and academic literature for universities and university colleges, as well as for vocational and professional education. We also publish high quality literature of a more general nature. Our main product lines are: � Textbooks for higher education � Research and reference literature � Non-fiction We only use environmentally certified suppliers. Tapir Academic Press NO–7005 Trondheim, Norway, Tel.: + 47 73 59 32 10, Email: [email protected], www.tapirforlag.no, Publishing Editor: [email protected]




EDITORIAL ......................................................................................................................................................................... 5

KEYNOTE SPEAKERS ........................................................................................................................................................... 6

eHÄLSOSTRATEGIERNAS REALISERANDE –EN UTVECKLIN UTAN ÄTERVÄNDO! Kristina Bränd Persson, Sverige EHELSE/IKT FOR PASIENTER- HVORDAN BØR POLITIKERNE STYRE? Åse Kari Haugeto, Norge UDBREDT ANVENDELSE AF TELEMEDICIN I DANMARK -FRA VISION TIL HANDLING! Lars Hulbæk, Danmark

SCIENTIFIC PAPERS: SISSEL BØCKMANN, BJØRN CHRISTIAN HAUGE .................................................................................................................. 9

Implementation of electronic patient records (EPR) in the curriculum for health and social worker students: ”OLD HABITS DIE HARD”

MORTEN HERTZUM, JESPER SIMONSEN ............................................................................................................................ 14

Clinical Overview and Emergency-Department Whiteboards: A Survey of Expectations toward Electronic Whiteboards METTE JENSEN, HEIDI KLITGAARD PEDERSEN AND OLE HEJLESEN ...................................................................................... 19

Decision Support for Management of Anticoagulation Patients in Community Nursing METTE DENCKER JOHANSEN, JENS SANDAHL CHRISTIANSEN, OLE K. HEJLESEN .................................................................. 24

A large dataset analysis of the long-term glucose counter-regulation to hypoglycemia in continuous glucose data to facilitate decision support in diabetes

HENRIETTE MABECK ......................................................................................................................................................... 29

Can computerized physician order entry systems increase security? LOUISE PAPE-HAUGAARD, ANNE R RASMUSSEN, KIRSTINE H ROSENBECK, STIG KJÆR ANDERSEN ..................................... 34

Pitfalls when integrating terminology systems and EHRs KIRSTINE HJÆRE ROSENBECK, ANNE RANDORFF RASMUSSEN ........................................................................................... 39

Does mapping to SNOMED CT improve precision of subjective clinical evaluations? MIE HVIID SIMONSEN, DANIEL SIMONSEN, LARS PETER MADSEN, OLE HEJLESEN .............................................................. 44

Suggestions for a reference method for quality assessment of sound and vibration levels in neonatal ambulances TERJE SOLVOLL, STEFANO FASANI, ASHOK BABU RAVURI, ANNELIES TIEMERSMA, GUNNAR HARTVIGSEN ........................ 49

Evaluation of an Ascom/trixbox system for context sensitive communication in hospitals MORTEN VILLUMSEN, SØREN THORGAARD SKOU, SIDSEL MARIA MONRAD RØNVED, OLE HEJLESEN ................................ 54

Health Technology Assessment of ultrasonography as a guidance tool for peritendinous steroid injections in patients with Achilles tendinopathy

SCIENTIFIC ABSTRACTS: CHARLOTTE D. BJØRNES, BIRGITTE S. LAURSEN, CHARLOTTE DELMAR, CHRISTIAN NØHR .................................................. 59

How can an Online Patient Book improve quality in the contact between male cancer patients and healthcare professionals?

CLAUS EHLERS, MORTEN THOMSEN, JENS ULRIK NIELSEN, BIRGIR SIGURDSSON, HERBERT JESSEN .................................... 60

Correlation between Electronic Medical Records and treatment outcomes KATHRINE HAUMANN, SOLVEIG ØKLAND, ELI SOFIE BERG. ANNE KRISTIN PAULSEN .......................................................... 61

E-learning for patients with Diabetes type 2: A four-stage model ANN MERETE DUEDAL JENSEN, MERETE MARTLEV JENSEN, ANNE SOFIE KORSAGER, ANNE RANDORFF RASMUSSEN, PIA BRITT ELBERG ............................................................................................................................................................. 62

Using the SNOMED CT-model for standardization of clinical information in labour documentation

EDITORIAL ......................................................................................................................................................................... 5






















EDITORIAL ......................................................................................................................................................................... 5






















EDITORIAL ......................................................................................................................................................................... 5






















NIS JOHANNSEN............................................................................................................................................................... 63

Multiplicity and Negotiations of Interests in Health IT Projects TRINE JØRGENSEN ........................................................................................................................................................... 64

Defining Clinical Content – Lessons Learned from implementing an EHR in Regions-hospital Randers ANNA-BRITT KROG ........................................................................................................................................................... 65

The Electronic Patient Record as a Media of Communication KITTA LAWTONA, STINE LOFT RASMUSSEN ....................................................................................................................... 66

Experiences with In Situ Simulation used in Evaluation of Telemedicine ANNE MOEN, ASTRID BREVIK SVARLIEN , MARI SYNNØVE BERGE, GRETE NETTELAND ...................................................... 67

Informatics Competencies – necessary conditions for sustainable eHealth LARS K MUNCK, SUZANNE KONGSGREN, BIRTHE GUNLUND, PETER M TORP, KARINA R HANSEN, ANNEGRETHE MØLBAK, LENE STENBEK, HELLE BALLE, CARSTEN VRANG, JENS D LOMHOLDT, MOGENS L FRIIS ....................................... 68

Effects of Shared Medication Record at admission- a randomised clinical trial JENS ULRIK NIELSEN, MORTEN THOMSEN, CLAUS EHLERS, BIRGIR SIGURDSSON, HERBERT JESSEN ................................... 69

Simplification of IT in the Health Care sector LARS CHR. RAGUS ............................................................................................................................................................ 70

Formation of one complete and total integrated nationwide eHealth structure in Greenland in 2015 – the Greenlandic Healthcare System’s ICT Strategy

INGRID STORRUSTE SVAGÅRD, RUNE FENSLI .................................................................................................................... 71

Decentralizing the Holter service through improved primary-secondary care collaboration KEN WACKER, LISE MARIE OLSEN, LOUISE PAPE-HAUGAARD ............................................................................................ 73

Application of an Electronic Patient Medication Management System in ambulatory units HELLE WENTZER, ANN BYGHOLM ..................................................................................................................................... 74

Minutes on the role of an EPR in outpatient consultation

WORKSHOPS: RONG CHEN, DANIEL KARLSSON, MARIE SANDSTRÖM, ERIK SUNDVALL, HANS ÅHLFELDT, TOR ANDRE SKJELBAKKEN, JOHAN GUSTAV BELLIKA, PER HASVOLD, STEIN ROALD BOLLE .......................................................................................... 75

Scandinavian openEHR Implementation – from Vision to Action

EIRIK ØVERNES & VIBEKE FLYTKJÆR ................................................................................................................................. 79

Kvalitetssikring gjennom kreativ bruk av e-læring i helsetjenesten

JENS SCHIERBECK, PATRICK HULSEN ................................................................................................................................. 80

Medicin-modulet bliver til tværfaglig kommunikations-platform

SØREN E. TVEDE ............................................................................................................................................................... 80

Decision Support for drug selection and dosing LYKKE BURMØLLE ANDERSEN .......................................................................................................................................... 81

Klinisk overblik fra det præhospitale skadested til stabilisering og behandling i sygehuset

NIS JOHANNSEN............................................................................................................................................................... 63



















SØREN E. TVEDE ............................................................................................................................................................... 80



NIS JOHANNSEN............................................................................................................................................................... 63



















SØREN E. TVEDE ............................................................................................................................................................... 80



NIS JOHANNSEN............................................................................................................................................................... 63



















SØREN E. TVEDE ............................................................................................................................................................... 80



- 5 -

Editorial This 8th Scandinavian Conference on Health Informatics (SHI) is held in Copenhagen with the University of Aalborg as the responsible organization. The first two conferences were held in 2003 and 2004 in Arendal, Norway, the third and fourth in 2005 and 2006 in Aalborg, Denmark, the fifth and sixth at the eHealth Institute in Kalmar, Sweden, and the se-venth in Arendal, Norway. We hope this conference in 2010 in Copenhagen will be as successful as the previous confe-rences. It has been a pleasure to receive a large number of papers-, abstract-, and workshop- submissions for the conference. Un-fortunately, this also means that we have had to reject several submissions The purpose of the conference is to stimulate exchange of information, co-operation and networking in development, implementation, use and assessment of informatics in the health sector in Scandinavia with special focus on clinical use of informatics and the role of the citizen in healthcare. The theme for this year is: Health Informatics from vision to action. The theme reflects that health informatics has devel-oped from ideas, pilots and sparse applications to widespread use in all parts of the health care sector. We hope that the conference will support the synergy naturally found in the linguistically and culturally relatively closely related countries by offering contributions on both local activities and research projects, on national activities in the Scan-dinavian countries, and on themes on the international research agenda for health informatics. As with the previous conferences, SHI2010 is arranged in co-operation between the following organizations: � Norway: Agder University, KITH – Norwegian Centre for Informatics in Health and Social Care, Norwegian Centre

for Telemedicine (NST), Norwegian Society for Medical Informatics � Sweden: eHälsoinstitutet, Högskolan in Kalmar, Swedish Federation for Medical Informatics � Denmark: Aalborg University, Danish Society for Medical Informatics (DSMI), and Virtual Centre for Health Infor-

matics (V-CHI) All the submitted contributions have been reviewed and commented on by at least two independent members of the scien-tific committee, and we want to thank all the reviewers for their excellent work in providing constructive feedback to the authors. This valuable input has improved the quality of many of the submissions. We also want to thank Louise B. Pape-Haugaard, Aalborg University for keeping track of all the contributions and ensuring they found their way to the final pro-ceedings. Enjoy the conference!

Forord Denne 8. Skandinaviske Konference i Sundhedsinformatik (SHI) afholdes i København med Aalborg Universitet som an-svarlig organisation. De første to konferencer blev afholdt i 2003 og 2004 i Arendal, Norge, den tredje og fjerde i 2005 og 2006 in Ålborg, Danmark, den femte og sjette i Kalmar, Sverige og den syvende i Arendal, Norge. Vi håber at denne ot-tende konference i København bliver et lige så vellykket arrangement som de tidligere konferencer. Det har været en stor fornøjelse at modtage de mange indsendte papers, abstracts og forslag til workshops. Dette betyder også at vi har været nødt til at afvise adskillige bidrag. Konferencens formål er at stimulere til erfaringsudveksling, samar-bejde og netværksdannelse inden for udvikling, implementering, anvendelse og vurdering af informatik i sundhedssektoren i Skandinavien med særligt fokus på den kliniske anvendelse af informatik og på borgerens rolle i sundhedsarbejdet. Dette års tema er: Sundhedsinformatik fra Vision til handling. Temaet reflekterer at sundhedsinformatikken har udviklet sig fra ideer, pilotforsøg og spredte anvendelser til udbredt brug overalt i sundhedsvæsnet. Vi håber, at konferencen vil understøtte den synergi, som naturligt findes i de relativt nært sprogligt og kulturelt beslægte-de lande ved at tilbyde indlæg om både lokale aktiviteter og forskningsprojekter, om nationale aktiviteter i de skandinavi-ske lande og om temaer på den internationale forskningsdagsorden i sundhedsinformatik I lighed med tidligere konferencer er SHI2010 arrangeret i et samarbejde mellem følgende organisationer: � Norge: Universitetet i Agder, Kompetencesenter for IT i helse- og socialsektoren (KITH), Nasjonalt senter for tele-

medicin (NST), Forum for databehandling i Helsesektoren (FDH) � Sverige: eHälsoinstitutet, Högskolan i Kalmar, Svensk Förening för Medicinsk Informatik (SFMI) � Danmark: Aalborg Universitet, Dansk Selskab for Medicinsk Informatik (DSMI), Virtuelt Center for Sundhedsin-

formatik (V-CHI) Alle de indsendte bidrag er blevet bedømt og kommenteret af mindst to uafhængige medlemmer af den videnskabelige komité, og vi ønsker at takke alle reviewere for godt arbejde og for at give konstruktiv feedback til forfatterne. Dette har forbedret kvaliteten i mange af bidragene. Endelig vil vi gerne takke Louise B. Pape-Haugaard for at holde styr på alle bi-drag, kommunikere med forfattere og reviewere og opsætning af proceedings. Vi ønsker jer alle en god konference!

Editorial board: Ann Bygholm


- 5 -









- 5 -









- 5 -









- 39 -

Does mapping to SNOMED CT improve precision of subjective clinical evalua-tions?

Kirstine Hjære Rosenbecka, Anne Randorff Rasmussena aTaMiCS, Department of Health Science and Technology, Aalborg University, Aalborg, Denmark

Abstract

Generally clinical documentation should be precise and unambiguous, but often this is not the case. This is the case for clinical subjective evaluations like “decrease in appetite”. This paper examines if the precision of the evaluation can be improved by mapping to SNOMED CT. Two examples of subjective clinical evaluations from real clinical documentation are mapped to SNOMED CT using three different modelling approaches in order fit pre-defined levels of precision. The findings suggest that mapping to SNOMED CT alone cannot improve the pre-cision, but by mapping the context of the subjective evaluation or the underlying objective observations, the precision might be improved. To determine the appro-priate level of precision fitting a clinical situation, Health Care Professionals (HCPs) should preferably validate the different possible mappings.

Keywords: SNOMED CT, Computerized Medical Re-cords Systems, Terminology as topic, Documenta-tion/standards

Introduction

When studying what should characterise clinical docu-mentation proposed quality indicators are among others accurate, correct, comprehensible and consistent[1]. However, health records often include imprecise clinical evaluations; consider common expressions like:

� “Decrease in appetite”

� “Impairment of vision”

The problem is that in itself, the evaluation cannot be interpreted since they are. Therefore there is a risk of misinterpreting the information, as the precision of the expressions are rather low. What does a decrease in ap-petite mean? For patients having surgery for obesity, a decrease in appetite might indicate that the surgery was successful. For an anorexic patient a decrease in appetite might be a serious worsening of the disease. On the other hand, in health records the evaluations are often nested with other clinical statements. It is for in-stance common to document the rationale behind deci-sions or evaluations. This nesting in the record is an important context for the interpretation of the statements within.[2] When introducing Electronic Health Records, reusing the information for both primary and secondary might

end up fragmenting the information, and hereby with-draw important context information. The importance of precise clinical documentation will dependent on the use:

� In shared care the importance of precision is greater than when treating a patient within the same department. This is highlighted by Eccher et al. They state that “semantic accuracy is essential in order to avoid possible misunderstandings among the different actors involved in the process of care.”[3]

� The needed ability of the retrieved information to be accumulated for secondary purposes as quality assurance and research. For statistics to be ex-pressive high precision of data is a prerequisite.

The consequences of imprecise documenting will change drastically as more interoperable EHR systems are intro-duced because one department might know exactly what they mean by “Decrease in appetite”, but when sharing this information with another department or using the information for secondary purposes the implicit defini-tion might be lost. Therefore it is no surprise that linking to standardised terminology is regarded an important task by researchers i.e. [4] and standardisation organisa-tions1

It has been proposed, that when introducing Electronic Health Records (EHRs) there is a possibility of avoiding ambiguity and inconsistency of the clinical documenta-tion by introducing a standardised terminology. Both Chute et al. and Cimino[5][6] state that ideally clinical terminologies among other should be characterised by:

developing, using and refining information models for clinical IT-systems.

� Concepts does not change with respect to time, perception or use

� It should be possible to identify duplications, am-biguities and synonyms

Also Elkin et al. describe general quality metrics of clinical terminology as “non-vagueness, non-ambiguity and non-redundancy.”[7] These statements hint that unambiguity and consistency are goals of introducing standardised clinical terminol-ogies. The aim of this paper is to study, the precision of subjec-tive evaluations in EHRs. What nesting of information is necessary for the subjective evaluation to be useful both

1http://www.openehr.org/wiki/display/term/Terminology+and+openEHR

- 39 -



Abstract



Introduction















- 39 -



Abstract



Introduction















- 39 -



Abstract



Introduction















- 40 -

primarily and secondarily? Three levels of precision for two subjective clinical evaluations are studied in order to add to the knowledge of how mapping to SNOMED CT can improve the precision of clinical statements.

Materials and Methods In this section the clinical cases including the clinical evaluations will be described. Afterwards the three dif-ferent modelling approaches applied to obtain different levels of precision will be presented. The purpose of all three approaches is documentation in clinical informa-tion systems and they all use SNOMED CT. For each precision level, the two clinical cases was interpreted and mapped to SNOMED CT.

The clinical cases

Both cases are Danish and are identified in Master’s the-ses written at Department of Health Science and Tech-nology at Aalborg University. As both are written in Danish the evaluations are translated to English, which might modify the semantics. Therefore both Danish and English terms are included in the descriptions. Evaluation 1: Tolerable bleeding

The first clinical case is related to maternal blood loss during labour. The case originates from an interview regarding clinical documentation in EHR systems in an obstetrics and gynecology department. The case is do-cumented in detail in [8]. When documenting maternal blood loss during labour, the clinicians used to evaluate if the bleeding was toler-able (“tilladelig” in Danish). When working with elec-tronic clinical documentation they realised that they did not have the same perception of what the expression meant and therefore saw the need of consensus. They specified that by tolerable bleeding they meant that the mother was bleeding less than 100ml and that the con-sciousness was unaffected. Evaluation 2: Eating too little The second clinical case is related to nutrition screening and treatment. The case originates from a project study-ing how to support electronic nutrition documentation using SNOMED CT. The case is based on both inter-views with nurses and analysis of guidelines and sche-mas regarding nutrition screening and treatment. The case is documented in detail in [9]. Whenever a patient is admitted to the hospital, a nutri-tion screening is preformed. Here it is determined if the patient is in nutritional risk. If the patient is at risk, the first thing the nurse will do is to document “characteris-tic of patient” or “nursing diagnose”. There are four primary diagnoses: eating too little, weight loss, BMI or other. These can be qualified by explaining the diagno-ses i.e. decrease in appetite, nausea or pain. Document-ing the diagnose forms the basis of the treatment regime. (Eating too little is “småtspisende” in Danish)

Levels of precision

The levels of precision take their point of departure in:

� Existing pre-coordinated SNOMED CT terms consisting of a finding and a qualifier. In the in-troducing example of decrease in appetite, this would mean identifying findings related to appe-tite in the SNOMED CT hierarchy and see if “findings related to appetite” have any children which pre-coordinates the finding with the quali-fier “decrease”. This level try to map the evalua-tion as is, without interpretation.

� The clinical situation in which the evaluation is obtained. In the introducing example the decrease in appetite meant something different because of the context; the patient being either overweight or anorexic. Hereby this level interprets what clini-cal situation would distinguish the present evalua-tion from similar evaluations, hereby improving the precision.

� Nesting the subjective evaluation with the ration-ale behind the evaluation. The rationale behind an evaluation is often patient observations, to im-prove precision these should preferably be struc-tured. If you want to evaluate that a patient is hy-pertensive, measurements of the blood pressure would be convincing observations to use as ra-tionale. Hereby this level interprets the observ-able rationale on which the evaluation can be based.

The mapping to SNOMED CT is performed by the first author and reviewed by the second. The SNOMED CT® version 0907 was used in the study. To illustrate the three different levels, nesting of infor-mation is necessary. For the first two precision levels, this is rather simple using only SNOMED CT expres-sions. For the last precision level a more complex nest-ing of clinical statements is needed. Therefore Mock-Ups of user-interfaces are used to illustrate this.

Results

The result section consists of the interpretation and map-ping of the two clinical evaluations using the three dif-ferent modelling approaches.

Pre-coordinated finings and qualifiers

Evaluation 1: Tolerable bleeding When identifying bleeding (finding) in SNOMED CT both the associated morphology haemorrhage and the child finding of vaginal bleeding has pre-coordinated expressions consisting of bleeding and a qualifier. Ex-amples of different qualifiers are illustrated in figure 1 and figure 2. The qualifier “tolerable” cannot be found, but as it is tolerable you could presume that it is a mod-erate vaginal bleeding, you could also argue that since it generally bleed a lot when giving birth, is must be a

- 40 -



The clinical cases



Levels of precision






Results




- 40 -



The clinical cases



Levels of precision






Results




- 40 -



The clinical cases



Levels of precision






Results




- 41 -

massive haemorrhage. Generally it is difficult to dis-criminate the SNOMED CT qualifiers as they are just as imprecise as “tolerable”.

Hemorrhage

Actute Massive Subacute ChronicDiffuse

Bleeding(finding)

Associated morphology

Figure 1- Examples of clinical statements in SNOMED CT identified when trying to find “tolerable bleeding”

Finding of vaginal bleeding

Normal menstual Fresh Scanty Moderate Profuse Abnormal

Bleeding(finding)


Evaluation 2: Eating too little

Equivalent results are showed for the example of “eating to little”. SNOMED CTs hierarchies were searched in order to find qualified expression regarding eating. The finding of eating pattern had the terms picky eater and greediness. For the child finding of quantity of eating examples of qualifiers are binging, excessive, fasting and overeating as can be seen in figure 3. The qualifier “too little” could not be found. Even though “too little” might mean picky eater or fasting, this alters the meaning of the expression, it does not make it more precise.

Finding of quantity of eating

Binging Excessive Fasting Overeating

Finding of eating pattern

Picky eater Greediness

Figure 3- Examples of clinical statements in SNOMED

CT identified when trying to find “Eating too little” Context dependent expressions in SNOMED CT Evaluation 1: Tolerable bleeding The context that would separate this finding from similar findings is that the bleeding occurs during labour. There-fore the children of labour finding in SNOMED CT were searched. We found the expression maternal blood loss that had the child:

Maternal blood loos within normal limits (289244001) Here the context “maternal” add to the precision, how-ever “within normal limits” points towards an implicit observation of how much blood is lost and is therefore still imprecise.

Evaluation 2: Eating too little A basic question to ask is what comparison is made when evaluating that the patient eats too little? Looking at the other nursing diagnoses, the second specifies weight loss and the third BMI. Therefore the purpose of specifying “eating too little” is probably to say that we cannot see an alteration in the weight, and the BMI is not critical, but the patient does not meet the required energy intake and therefore there is a risk of weight loss. Also the time might be a decisive factor, if the patient did not eat enough yesterday, you would probably not diagnose that the patient “eat too little”. So a suggestion would be: Inadequate dietary caloric intake (88202002) CLINICAL

COURSE Continual (263730007)

The first term specify that the energy intake is inade-quate. The CLINICAL COURSE post-coordination specify that it is a continual problem, not just one day. Here the precision is better, but we do not know how much less than the required energy intake the patient eats as well as its unknown for how long time it has been a problem.

Registering both observations and evaluation

When linking together different observations, evalua-tions and treatments in a clinical IT system, there would typically be a need for choosing an information model to handle this or to link them by long post-coordinated ex-pressions. But since the aim of this paper is to study how to improve the precision of clinical statements, there is no need for choosing one model over another. Instead small mock-ups of user-interfaces are presented to show how the observations and evaluations could be interre-lated in order to avoid ambiguity. SNOMED CT is used to find labels and content of drop down menus, but there are no underlying post-coordinated expressions. Evaluation 1: Tolerable bleeding The observations, on which it can be based if the bleed-ing is tolerable, are, according to the example, the quan-tity of maternal blood loss and the consciousness of the patient. Semantically it does not make sense that you can evalu-ate something about the bleeding by looking at con-sciousness and when looking at the expression maternal blood loos within normal limits, it is clear that this only refers to the quantity of blood measured. But since the consciousness is also important, it might be that we were evaluating something broader than the bleeding i.e. the maternal condition during labour and in this perspective the consciousness is important. These are the presumptions leading to the design of the Mock-Up in figure 4. All terms in the interface are drawn from SNOMED CT except for the unit “ml”. When documenting consciousness, the precision of the observation would be improved, if using the Glasgow Coma Scale, but since the case is about bleeding it was decided not to specify consciousness in detail, and just

- 41 -


Hemorrhage


Bleeding(finding)





Bleeding(finding)
















- 41 -


Hemorrhage


Bleeding(finding)





Bleeding(finding)
















- 41 -


Hemorrhage


Bleeding(finding)





Bleeding(finding)
















- 42 -

draw expressions from SNOMED CT that was directly related to consciousness findings.

Figure 4- Illustration of nested clinical statements that could explain the rationale behind the evaluation “toler-

able bleeding”

Evaluation 2: Eating too little What should be objectively measured in order to deter-mine if the patient eat too little? For this example we presumed that the patient’s energy requirement com-pared to the daily intake of energy was needed. This is illustrated in figure 5, where the expressions Calorie requirement and dietary calorie intake are found in SNOMED CT. Continual inadequate dietary caloric intake refers to the post-coordination: Inadequate dietary caloric intake CLINICAL COURSE Continual.

Figure 5- Illustration of nested clinical statements that could explain the rationale behind the evaluation “eat-

ing too little” Since calories/day is not easy to measure, some help for the calculations would probably be needed. This explains the blue links in figure 5.

Discussion

In this section it is discussed how to choose between the levels of precision in different clinical situations. After-wards the methodological approach of this paper is dis-cussed. In the end importance of this type research is discussed, since it differs from most SNOMED CT re-lated research.

Method discussion

In this study only two examples of subjective evaluations are evaluated, and a limited number of persons are per-forming the SNOMED CT mappings. Therefore it is not possible to generalise based on the results from this pa-per. The aim of this paper is merely to introduce another way of perceiving SNOMED CT mapping. Based on our results, we hypnotise, that we cannot map directly from existing documentation to SNOMED CT and expect more precise documentation. Therefore it is interesting to perform research, where the focus is to discuss single mappings in detail.

In 2008 a review concluded, that most SNOMED CT related research is concerned with the comparison of SNOMED CT with other terminology systems - typically providing coverage percentages. [10] Here there is no or little information in the papers regarding the details of how each mapping is performed and what arguments the terminologists, HCPs or researchers performing the mapping have for one mapping over another. In [11] 864 expressions are mapped from a General Medical Evalua-tion Template to SNOMED CT and MEDICIN®. Even though four examples of mapping are presented, the al-ternatives are not discussed. Detail oriented SNOMED CT research would allow re-search communities and other SNOMED CT interest groups to share experience regarding representation of clinical documentation using SNOMED CT.

Choosing between levels of precision clinically

The findings suggest, that merely mapping the subjective evaluation itself to SNOMED CT will not improve the precision, the consequence could on the other hand be, that the clinical meaning of the expression could be al-tered, hereby adding to the imprecision. Therefore, using a standardised terminology as SNOMED CT is not a guarantee that subjective evaluations automatically be-come more precise. This also suggests that whenever possible the first level of precision should be avoided. For the two clinical cases presented, which of the two remaining precision level would be clinically useful? We acknowledge that this decision should be qualified by clinical experts, but we will never the less put forward a solution based on the clinical situation in which the in-formation is documented. For “tolerable bleeding” it would be adequate to choose the third level, where the observations and evaluations are linked together. It seems to be important, that similar maternal blood loss is not evaluated differently. This could lead to the case where two patients having similar blood losses, would be treated differently. I.e. Blood transfusion would be considered and after the delivery the fluid balance would be monitored closely if the bleeding was evaluated as intolerable. Furthermore, the patient record is written after the delivery, so there will not be a tight time-limit of documentation related to a busy clinical situation.

- 42 -



able bleeding”




Discussion


Method discussion





- 42 -



able bleeding”




Discussion


Method discussion





- 42 -



able bleeding”




Discussion


Method discussion





- 43 -

In the case of “eating too little” it would be adequate to choose the second level, where the context of the evalua-tion is included. This is mainly due to the fact, that in the clinical situation the relevant observations are not avail-able. The question of characteristics of the patients is the first to be answered after the nutrition screening, which is performed just after patient admission to a hospital. Therefore the information of how much the patient eats can only be answered by the patient or a relative. Going through all meals i.e. the preceding week in order to de-termine the precise calorie intake would be nonsense and presume a non-existing data-quality. Instead the nurse would probably question the patient about nutrition and based on this give an evaluation of the adequacy or in-adequacy of the energy intake. Different modelling approaches to specify subjective evaluations In this paper it was attempted to model subjective evaluations according to three pre-specified levels of precision. It made sense for this study to predefine the levels, but to categorise detail-levels is of course an ab-straction as there are levels in between and even more precise ways to document. For example the term inade-quate dietary caloric intake could be post-coordinated with a term specifying that it was “one week since the onset” instead of the more imprecise “continual”. An-other example of improved precision is using the Glas-gow Coma Scale to represent findings of consciousness. In order to perform this type of research, it is however interesting to discriminate between different modelling approaches of different types of clinical expressions.

Conclusion The findings of this paper are only based on two exam-ples of subjective evaluations. However preliminary findings suggest that:

� Using SNOMED CT is not a guarantee that sub-jective evaluations become more precise. When trying to map the evaluation directly, the conse-quence could on the other hand be that the clini-cal meaning of the expression could be altered.

� Two modelling approaches are suggested taking into account the context of the finding or specify-ing the underlying observations related to the evaluations. This could possibly improve the pre-cision of clinical evaluations.

� Through the process of making the information precise using SNOMED CT presumptions and choices were made. Choosing the appropriate level of precision requires clinical validation.

In the future it would be interesting to study the map-pings of other types of clinical expressions this could i.e. be treatment onsets and underlying findings. More clini-cal expressions and HCP validation of the different mod-

elling approaches would add to the significance of the findings. Acknowledgments Authors are partly granted by CSC Scandihealth, Region Nordjylland and Trifork Software Solutions References [1] Stetson PD, Morrison FP, Bakken S, Johnson SB. Pre-

liminary development of the physician documentation quality instrument. J.Am.Med.Inform.Assoc. 2008;15:534-41.

[2] Kalra D, Ingram D. Electronic health records. Informa-tion Technology Solutions for Healthcare 2006:135-181.

[3] Eccher C, Purin B, Pisanelli DM, Battaglia M, Apolloni I, Forti S. Ontologies supporting continuity of care: The case of heart failure. Comput.Biol.Med. 2006;36(7-8):789-801.

[4] Sundvall E, Qamar R, Nyström M, Forss M, Petersson H, Karlsson D, et al. Integration of tools for binding ar-chetypes to SNOMED CT. BMC Medical Informatics and Decision Making 2008;8(Suppl 1):S7.

[5] Cimino JJ. Desiderata for controlled medical vocabular-ies in the twenty-first century. Methods Inf.Med. 1998;37:394-403.

[6] Chute CG, Cohn SP, Campbell JR. A framework for comprehensive health terminology systems in the United States: development guidelines, criteria for selection, and public policy implications. ANSI Healthcare Infor-matics Standards Board Vocabulary Working Group and the Computer-Based Patient Records Institute Working Group on Codes and Structures. J.Am.Med.Inform.Assoc. 1998;5:503-10.

[7] Elkin PL, Brown SH, Carter J, Bauer BA, Wahner-Roedler D, Bergstrom L, et al. Guideline and quality in-dicators for development, purchase and use of controlled health vocabularies. Int J Med Inform 2002;68:175-86.

[8] Rosenbeck KH. Deling, genbrug og standardisering af Sundhedsfagligt Indhold - Brobygning i det danske EPJ-landskab. Specialeprojekt fra Institut for Sundheds-videnskab og Teknologi 2009.

[9] Andersen P,Lindgaard A. Strukturering og standardise-ring af sygeplejefaglig dokumentation indenfor ernæring vha. SNOMED CT. Specialeprojekt fra Institut for Sundhedsvidenskab og Teknologi 2009.

[10]Cornet R,de Keizer N. Forty years of SNOMED: a lit-erature review. BMC Med.Inform.Decis.Mak. 2008;8 Suppl 1:S2.

[11]Brown,SH.Rosenbloom,ST.;Bauer,BA.;Wahner-Roedler,D.;Froehling,DA.;Bailey,KR.;Lincoln,MJ; Montella,D.;Fielstein,EM.;Elkin,PL.Direct Comparison of MEDCIN® and SNOMED CT® for Representation of a General Medical Evaluation Template. : AMIA; 2007.

Address for correspondence Kirstine Hjære Rosenbeck, Email: [email protected]

- 43 -




















- 43 -




















- 43 -




















Documents

SHI2010 - Aalborg Universitet · We hope this conference in 2010 in Copenhagen will be as successful as the previous confe-rences. It has been a pleasure to receive a larg e number