Postgrad MedJ 1995; 71: 217-221 C) The Fellowship of Postgraduate Medicine, 1995

High dependency units in the UK: variable size,variable character, few in number

Fiona-Jane Thompson, Mervyn Singer

Bloomsbury InstituteofIntensive CareMedicine, DepartmentofMedicine,University CollegeLondon MedicalSchool, London, UKF-J ThompsonM Singer

Correspondence toDr Mervyn Singer,Bloomsbury Institute ofIntensive Care Medicine,Department of Medicine,UCL Medical School, RayneInstitute Building,University Street, LondonWClE 6JJ, UK

Accepted 30 November 1994

SummaryAn exploratory descriptive survey wasconducted to determine the size and char-acter ofhigh dependency units (HDUs) inthe UK. A telephone survey and subse-quent postal questionnaire was sent to the39 general HDUs in theUK determined bya recent survey from the Royal College ofAnaesthetists; replies were received from28. Most HDUs (82%, n = 23) were geo-graphically distinct from the intensivecare unit and varied in size from three to13 beds, although only 64% (n = 18)reported that all beds were currentlyopen. Nurse: patient ratios were at least1:3. Fifty per cent ofunits had one or moredesignated consultants in charge,although only 11% (n = 3) had specificallydesignated consultant sessions. Juniormedical cover was provided mainly bythe on-call speciality team. Twenty unitsacted as a step-down facility for dis-charged intensive care unit patients and21 offered a step-up facility for patientsfrom general wards. Provision offacilities and levels of monitoring variedbetween these units. Few HDUs exist inthe UK and they are variable in size and inthe facilities and monitoring procedureswhich they provide. Future studies areurgently required to determine cost-effectiveness and outcome benefit of thisintermediate care facility.

Keywords: dependency, intensive care, highdependency care

Introduction

Britain differs from mainland Europe in that ithas fewer intensive care beds per 100 000population.' As a consequence, a higher pro-portion of intensive care patients in the UKrequire mechanical ventilation and, by implica-tion, more acutely ill, though non-ventilatedpatients are managed in non-intensive careward areas. An intermediate care area, such as ahigh dependency unit (HDU), would thereforebe beneficial for many of these patients.An HDU is defined by the Association of

Anaesthetists of Great Britain and Ireland2 as'an area for patients who require more intensiveobservation, treatment and nursing care thancan be provided on a general ward. It would notnormally accept patients requiring mechanical

ventilation, but could manage those receivinginvasive monitoring'.A 1988 survey' from the Association of

Anaesthetists identified only 55 hospitals in theUK with HDU facilities. However, a morerecent audit4 identified that only 39 of 256hospitals now have a designated HDU. Werecently published a study' costing intensivecare and high dependency care in our hospital.Our case mix and levels of monitoring andstaffing may significantly differ from generalHDUs in other hospitals, thus making mean-ingful comparison difficult. We thereforeundertook a survey to determine the size andcharacter of HDUs identified by the recentreport.4

Methods

The 39 units identified by the Royal College ofAnaesthetists Audit4 were provided by theauthor (J Stoddart, personal communication).A telephone survey was conducted to confirmtheir current existence, identify any otherHDUs within the hospitals and to ascertain thenamed consultant in charge of each unit towhom a questionnaire would be sent.A postal questionnaire comprising 32 closed

questions was sent. Two weeks were providedfor questionnaire completion and all non-respondents were contacted with a follow-uppostal questionnaire. The total duration of datacollection was six weeks. A second telephonesurvey was conducted after the initial dataanalysis to clarify any points of ambiguity.

Results

From the initial telephone survey six of theprevious 39 HDUs identified were now closeddue to budgetary constraints. The total numberof HDUs was 37 from a total of 33 hospitals,since two hospitals were subsequentlyidentified as having more than one HDUfacility. The response rate for completed ques-tionnaires was 76% providing information on28 HDUs in 26 hospitals. The geographicallocations of these hospitals were as follows:London (n = 6), South England (n = 2), NorthEngland (n = 11), Scotland (n = 3), Wales(n = 3) and Northern Ireland (n = 1). The typeand size ofeach responding hospital is shown infigure 1 and other specialist units is identifiedwithin the hospitals are displayed in table 1.Most HDUs (82%, n = 23) were geog-

raphically distinct from the intensive care unit

on April 19, 2020 by guest. P

rotected by copyright.http://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.71.834.217 on 1 April 1995. D

ownloaded from

Thompson, Singer

No. of hospitals No. of HDUs

:.:

0UU

District General

Post Graduate

Teaching

University-affiliated

Total hospital beds

Figure 1 Type and size of responding hospitals

and 39%O (n = I 1) of these units were part of anacute general ward. Figure 2 shows the numberof beds designated for high dependency care.Sixty four per cent (n = 18) ofHDUs reportedthat their total number of designated beds werecurrently fully staffed for high dependencycare. Fourteen (78%) of the 22 fully open unitshad four beds or less while the six (21 %) unitswith closed beds ranged in size from 4- 12 beds(median 8); 2-4 beds (median 2.5) were cur-rently closed due to budgetary constraints.

STAFFINGNurse:patient ratios were at least 1:3 althoughthe smaller units (four beds or less) had propor-tionally more nurses. Only 50%0 (n = 14) ofthese units reported having one or more desig-nated consultants in charge, namely: surgeon210% (n = 6), anaesthetist 180% (n = 5), inten-sivist 110% (n = 3), and physician 7 O (n = 2).However, only 110% (n = 3) of these units hadspecifically designated consultant sessions.

Junior medical cover during weekday work-ing hours (09.00-17.00), was mainly providedby the speciality under which the patient wasadmitted (85%, n = 24) or by the intensive careunit or designated HDU team (15%, n = 4).Similarly, junior doctor cover at nights andweekends was provided by the on-callspeciality team in 75%O (n =20) of HDUs.Treatment strategies were executedpredominantly by the primary or on-call teamof the speciality under which the patient wasadmitted.

ADMISSIONSTwenty two (78%) HDUs collected statisticaldata on their patient admissions. Figure 3depicts the number of patients admitted tothese HDUs in the last year; this includes

Table 1 Specialist units identified within the responding hospitals

Number of hospitalsUnit having unit (oo)

Intensive care unit 26 (100)Accident and emergency department 23 (88)Coronary care unit 23 (88)Post-operative recovery ward 20 (77)Acute admission ward 13 (50)Renal unit 8 (31)

1 2 3 4 5 6 7 8 9 10 11 12 13 14

Number of beds in HDUFigure 2 Number of designated beds per HDU

No. of HIDUs

El0UU

District General

Post Graduate

Teaching

University-afliliated

Annual patient admissions

Figure 3 Annual patient throughput per HDU

estimations where statistics were not available.Very few units (l8o%, n = 5) utilised scoringsystems for determining severity of illness,such as the Acute Physiology and ChronicHealth Evaluation (APACHE) III and theTherapeutic Injury Severity Scoring (TISS)6systems.The majority of patient categories admitted

to HDUs were classified as shown in table 2.The length of patient stay varied from 1.5 to

7 days (median 2.5). Twenty HDUs (710o),reported that their HDU acted as both a'step-down' facility for patients dischargedfrom an intensive care unit and as a 'step-up'unit (750o, n = 21), for patients from generalward areas.

Fifty seven per cent (n = 16) of HDUsreported having an admission criteria policy.Although 5000 (n = 14) of HDUs stated thatthey had refused patient admissions in the last12 months due to nursing staff and HDU bedshortages, these problems were sporadic.

Provision of facilities and monitoring proce-dures provided by these units and on generalwards within their hospitals are shown in table 3.

218 on A

pril 19, 2020 by guest. Protected by copyright.

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.71.834.217 on 1 A

pril 1995. Dow

nloaded from

High dependency units in the UK

Table 2 Categories of patients admitted to

HDUs

Category Number (o%)

Post-operative elective 20 (74)Post-operative emergency 16 (59)Non-operative trauma 16 (59)Medical (non-cardiac) 14 (52)Medical (cardiac) 10 (37)Burn injury 5 (19)Paediatrics 3 (11)

Table 3 Provision of facilities/treatments in responding hospitals

Provision in HDUs Provision in generalFacility/treatment (n = 28) (O%) wards (n = 26) (O%)

MonitoringIntra-arterial pressure 20 (71) 0 (0)Pulmonary artery catheter 12 (43) 0 (0)Central venous pressure 27 (96) 13 (50)Pulse oximetry 26 (93) 11 (42)Continuous ECG 28 (100) unknown

Respiratory supportMechanical ventilation 3 (11) 0 (0)Nasal positive pressure ventilation 3 (32) 0 (0)Facial continuous positive 16 (57) 1 (4)

airways pressureTracheostomy 22 (78) 10 (28)

Cardiovascular supportVasodilator infusion(s) 23 (82) 6 (23)Inotropes 23 (82) 6 (23)Cardioversion 11 (39) 1 (4)Pacing 11 (39) 3 (12)Intra aortic balloon pump 1 (4) 0 (0)

Renal supportHaemofiltration 6 (21) 0 (0)Haemodialysis 4 (14) 0 (0)Peritoneal dialysis 6 (21) 3 (12)Plasma exchange 1 (4) 0 (0)

Analgesia/sedationEpidural analgesia 20 (71) 7 (27)Patient-controlled analgesia 24 (86) 15 (58)Continuous intravenous sedation 20 (71) 7 (27)

Discussion

Despite the relative paucity of intensive carebeds compared to mainland Europe, fewHDUresources currently exist in the UK. There isalso a wide variation in facilities, proceduresand medical cover which the HDUs provide. Inspite of recommendations for the provision ofHDUs,2 there has, in fact, been a decrease inthe availability of these facilities. This mayprove to be a false economy since theHDU may

'. . . provide the appropriate care for many

acutely ill patients who do not require the fullfacilities of an intensive care unit'.2 Thus, theHDU may offer a step-up facility for a patientwhose condition gives cause for concern, wheregreater monitoring and supervision may pre-

vent the need for admission to intensive care

which may prove considerably more lengthyand expensive. Conversely, the HDU could actas a step-down unit for patients who no longerrequire intensive care, but who are not ideallysuited to return to a general ward.

Less than 10% of patients in a study atUniversity College Hospital' required a step-up from HDU to intensive care. The highproportion of intensive care patients who aremechanically ventilated (80%) andhaemofiltered (21%) suggests more effectiveutilisation of intensive care facilities. Thus,intensive care units with less than 5000 of theirpatients being ventilated may benefit fromhaving more HDU beds and fewer intensivecare beds. The total cost of an intensive carepatient day in 1991 on the University CollegeHospital unit was £1148.72 in contrast to£437.68 for an HDU patient day.

Readmission rates to intensive care units areshown in table 4. It is noteworthy that theUniversity College Hospital patient readmis-sion rate was half that of the others quoted7' 8;whether or not the presence of an HDU at thishospital was a significant factor requires fur-ther study though there can be no denying thehigher subsequent morbidity and mortality ofpatients requiring readmission (box 1).9Rowan et allo also identified that 17.9% of

UK intensive care unit admissions died in theunit while a further 9.8% died in hospital afterdischarge from the unit. Bion" questioned whythese patients were dying on the wards andsuggested one or more of the following pos-sibilities: premature discharge from the inten-sive care unit due to pressure on beds, lack ofhigh dependency facilities and inappropriateadmission to intensive care. However, no datayet exist to answer this question. The NationalConfidential Enquiry into PerioperativeDeaths (NCEPOD: 1991/1992)12 in the UKprovided some national data on postoperativedeaths (box 2). How many ofthese deaths couldhave been prevented by anHDU facility is alsoopen to question.A 'snapshot' study'4 of the subjective

dependency of postoperative surgical patientsin eight UK hospitals suggested that up to13.5% of general ward patients would have

Table 4 Readmission rates to intensive careunits (ICUs)

Total ICU ReadmissionsStudy and site admissions (%)

UCH' 114 3 (2.6%)26 UK ICUs7 8786 460 (5.20°)13 US ICUs8 17440 818 (4.7%/)

Box 1

Characteristics ofpatientsrequiring readmission to intensivecare9

* mortality 41.500 (vs 7.30o in control ICUpatients matched for age, sex, primary disease)

* mean length of stay 47.8 ± 42 days (vs20.8 ± 14.2 days in control patients)

* 54°0 of readmissions had respiratory failure* 30 o0 of readmissions had recurrence orworsening of original problem

219 on A

pril 19, 2020 by guest. Protected by copyright.

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.71.834.217 on 1 A

pril 1995. Dow

nloaded from

220 Thompson, Singer

Data from UK enquiry intoperioperative deaths12

* 1616 deaths studied* 985 died on a general ward, 414 in ICU and 44

in HDU* precise cause ofdeath and the proportion of

unexpected, potentially avoidable orpermitted deaths were not determined

* less than 20% ofdeaths occurred on the day ofthe operation or on the following day

* 70% of the 1616 patients were classified ashaving moderate to severe chronic systemicdisease (see'3)

* despite this high risk grading, 1001 patientswere sent post-operatively to the general wardrather than to a higher dependency area.

Box 2

been appropriate for HDU care, while approx-imately 50%O had inappropriately placed inten-sive care unit patients. However no data onoutcome was provided. Kilpatrick et all5 alsofound that 40% of their intensive care unitadmissions had a low calculated and actual riskofhospital mortality. Many were admitted withmetabolic disorders, post trauma and for res-piratory observation and monitoring. Theauthors suggested that these patients may bemore appropriately managed on an HDU.The American College of Physicians con-

ducted a Clinical Efficacy Assessment Project(CEAP)'6 to ascertain from published datawhether it was possible to identify patientsunlikely to benefit from intensive care beforeadmission to an intensive care unit and tosubsequently treat them in a step-down(intermediate care) area. Patients consideredunlikely to benefit from intensive care in thestudy were either low risk monitored patientsor high risk patients who die despite aggressivetherapy. However, they were unable to find anydata as studies identified these patients afteradmission to the intensive care unit rather thanbefore. Thus, they shifted the emphasis of theirstudy to find, within the first 24 hours ofintensive care unit admission, those patientswho could be managed in a less resource-intensive environment without having anegative effect on outcome. Again, no studieswere found so they recommended a multi-centre study should be performed to determineoutcomes on patients identified as low risk whoare randomly assigned to alternative hospitallocations for treatment compared with thoseassigned to continued intensive care unit treat-ment until routine discharge.A further justification for HDU facilities

arises from the recent initiative to reduce juniordoctors' working hours. This has resulted indecreased cover at nights and weekends soconcentration of sicker patients in a betterstaffed and monitored area may be advan-tageous. Nurse practitioners'7 with higherlevels of training may possibly have a usefulrole in this area.No data exist on the cost-effectiveness of

high dependency care. Obviously, meaningful

Learning/summary points

* the HDU is an intermediate care facilitybetween the general ward and intensive careunit, offering both step-up and step-downfunctions

* few HDUs exist in theUK* those that do vary in size (3-13 beds), patient

throughput, and monitoring and treatmentcapability

* the majority ofHDU beds are used forpost-operative care

* though 50% ofHDUs have designatedconsultants in charge, few have dedicatedconsultant sessions

* studies have yet to be performed to determinewhether the HDU is both cost-effective andbeneficial in terms of patient outcome

Box 3

comparison between different hospitals wouldbe difficult in view ofthe wide variation in theircharacter and the facilities provided. However,such a study is urgently needed because of everincreasing financial restraints and limitation ofresources. In a study conducted in two BritishDistrict General Hospitals in 1988, Shiell et al'8estimated costs of /525 and £465 per intensivecare unit patient day. At University CollegeHospital in 1988' our equivalent costs were£779 for an intensive care unit patient day and£163 for an HDU patient day, with an averagecost of £327 for a combined intensive careunit/HDU patient day. This is considerablycheaper than for the two hospitals above whichhad a median APACHE II score on admissionof 11 in contrast to a 24-hour APACHE IIscore of 18 in our intensive care unit.

In an attempt to limit costs and relievepressure on intensive care beds in the US, thedevelopment of intermediate care facilitieshave been advanced. A 1985 national sample ofUS intensive care units'9 revealed that 10-50%ofpatients were not considered to have receivedactive intensive care. Oye and Bellamy20claimed that 41% of admissions to theirmedical intensive care unit did not receiveacute treatments and that these 'monitored'patients had a 99% intensive care unit survivaland 90% hospital survival. In 1988 Wargovichet aPn' found that a third of all designatedintensive care unit beds in the state of NewJersey were now staffed and equipped asintermediate care areas. Many of the patientscurrently admitted to intensive care units in theUK appear to fall into this category and thuscould be cared for in a high dependency setting.

In conclusion, HDU facilities are generallylacking in the UK. Proper utilisation of thisresource could have significant implications onoutcome and benefit. Studies of cost-effectiveness and outcome benefit are urgentlyneeded to validate this.

The authors would like to thank Dr JC Stoddart, RoyalVictoria Infirmary, Newcastle upon Tyne for supplyingthe list of 39 HDUs, and the 28 HDUs who completedthe questionnaire.

on April 19, 2020 by guest. P

rotected by copyright.http://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.71.834.217 on 1 April 1995. D

ownloaded from

High dependency units in the UK 221

1 Singer M, Myers S, Hall G, Cohen SL, Armstrong RF. Thecost of intensive care: a comparison on one unit between1988 and 1991. Intensive Care Med 1994; 20: 542-9.

2 The Association of Anaesthetists of Great Britain andIreland. The High Dependency Unit-acute care in the future.London: Association of Anaesthetists, 1991.

3 The Association of Anaesthetists of Great Britain andIreland. Intensive Care Services-provision for the future.London: Association of Anaesthetists of Great Britain andIreland, 1988.

4 The Royal College of Anaesthetists. National ITU Audit.London: Royal College of Anaesthetists, 1992/1993.

5 Knaus WA, Draper EA, Wagner DP, Zimmerman JE.APACHE II: a severity of disease classification system. CritCare Med 1985; 13: 818-29.

6 Keene AR, Cullen DJ. Therapeutic Intervention ScoringSystem: update 1983. Crit Care Med 1983; 11: 1-3.

7 Rowan KM. Outcome comparisons of intensive care units inGreat Britain and Ireland using the APACHE II method.PhD thesis, University of Oxford, 1992.

8 Knaus WA, Wagner DP, Zimmerman JE, Draper EA.Variations in mortality and length of stay in intensive careunits. Ann Intern Med 1993; 118: 753-61.

9 Durbin CG, Kopel RF. A case-control study of patientsreadmitted to the intensive care unit. Crit Care Med 1993;21: 1547-53.

10 Rowan KM, Kerr JH, McPherson K, Short A, Vessey MP.Intensive Care Society's APACHE II study in Britain andIreland-II: outcomes comparisons of intensive care unitsafter adjustment for case mix by the American APACHE IImethod. BMJ 1993; 307: 977-81.

11 Bion J. Outcomes in intensive care - are related to skill mix,but we still need much better measures. BMJ 1993; 307:953-4.

12 Campling EA, Devlin HB, Hoile RW, Lunn JN. The reportof the national confidential enquiry into perioperative deaths1991/1992. London, 1993.

13 American Society of Anesthesiologists: New classification ofphysical status. Anesthesiology 1963; 24: 111-3.

14 Crosby DL, Rees GA. Provision ofpostoperative care in UKhospitals. Ann R Coll Surg Engl 1994; 76: 14-8.

15 Kilpatrick A, Ridley S, Plenderleith L. A changing role forintensive therapy: is there a case for high dependency care?Anaesthesia 1994; 49: 666-70.

16 Bone RC, McElwee NE, Eubanks DH, Gluck EH. Analysisof indications for early discharge from the intensive careunit. Clinical Efficacy Assessment Project: American Col-lege of Physicians. Chest 1993; 104: 1812-7.

17 Mirr MP. Advanced clinical practice: a reconceptualizedrole. Clin Iss Crit Care 1993; 4: 599-602.

18 Shiell AM, Grifliths RD, Short AIK, Spiby J. An evaluationof the costs and outcome of adult intensive care in two unitsin the UK. Clin Intensive Care 1990; 1: 256-62.

19 Wagner DP, Knaus WA, Draper EA. Identification oflow-risk monitor admissions to medical-surgical intensivecare units. Chest 1987; 92: 523-30.

20 Oye RK, Bellamy PE. Patterns of resource consumption inmedical intensive care. Chest 1991; 99: 685-9.

21 Wargovich RM, Bekes CE, Cernaianu A. Intermediateintensive care unit in New Jersey. Crit Care Med 1988; 16:1167-8.

Medical AnniversaryWILLIAM HARVEY, 1 APRIL 1578

William Harvey (1578-1657) wasborn at Folkestone, where his fatherhad a thriving business carrying mailand other items, initially throughoutKent and later across the EnglishChannel. Young Harvey was a smallman, sharp-eyed and of olive com-plexion. His broad head narrowed to apoint at the chin, which bore a smallbeard beneath a thick moustache. Araised left eyebrow gave rise to aquizzical expression. His numerousidiosyncrasies included keeping sugarin his salt cellar, combing his hairwhile walking out of doors, fingeringthe dagger he always wore, andtreating his own gout by putting hislegs in a bucket of cold water and thenroasting in front of a fire. He wasaddicted to that new beverage, coffee.He was educated at King's School,

Canterbury, Gonville and Caius Col-lege, Cambridge, Padua Universityand then became vhvsician at St Bar-tholomew's Hospital. He wrote his immortal De motu cordis (1628) when he was aged 50.Until that time, the Galenic concept of the blood flow was ofsome slow drift of blood fromthe right to the left heart. Harvey argued that valves in veins must suggest an onward flow ofa venous and arterial circulation and not ebb and flow like the tide.

William Harvey's discovery revolutionised the concept ofthe circulation ofthe blood withits constant irrigation ofthe tissues by a vast network ofblood vessels, ranging from the mainartery with a calibre of two thumbs down to the finest capillaries with a diameter ofone-tenth of a human hair. The aggregate length of this complex system of vessels isthousands of miles, and yet it is filled with less than six litres of blood.Harvey died on June 3, 1657 in Roehampton and is buried at the Church of St Andrew,

Hempstead, Essex.

Courtesy of National Portrait Gallery

on April 19, 2020 by guest. P

rotected by copyright.http://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.71.834.217 on 1 April 1995. D

ownloaded from