Trauma services requirements in a district general hospitalserving a rural area

S J Kinny, D H A Jones

Accident Unit, GwyneddHospital, Bangor,Gwynedd LL57 2PWS J Kinny, FRCSED,orthopaedic registrarD H A Jones, FRCS,consultant orthopaedic surgeon

Correspondence to: MrJones.

Br Alfcd 1990;300:504-8

AbstractObjective-To assess the demands made on a

regional trauma centre by a district trauma unit.Design-Two part study. (1) Prospective analysis

of one month's workload. (2) Retrospective analysisof one year's workload by using a computer basedrecords system. Comparison of two sets of results.Setting-Accident unit in Gwynedd Hospital,

Bangor.Patients-(1) All patients who attended the

accident unit in August 1988. (2) All patients whoattended the accident unit in the calendar year April1988 - April 1989.Main outcome measure-Workload of a district

trauma unit.Results-In August 1988 there were 2325 atten-

dances; 2302 of these were analysed. In all, 1904attendances were for trauma; 213 patients wereadmitted to the trauma ward and 103 required anoperation that entailed incision. Patients whoattended the unit had a mean (range) injury severityscore of 2-13 (0-25). Only two patients had injuriesthat a district general hospital would not be expectedto cope with (injury severity score >20). In the year

Alap ofGwynedd

April 1988 - April 1989, 21007 patients attended theunit. In all, 17 958 attendances were for orthopaedicinjuries or injuries caused by an accident; 1966patients were admitted to the unit.Conclusions-Most trauma is musculoskeletal

and relatively minor according to the injury severityscore. All but a few injuries can be managed indistrict general hospitals. In their recent report theRoyal College of Surgeons has overestimated therequirements that a British district general hospitalwould have of a regional trauma centre.

IntroductionAll staff providing trauma services throughout the

United Kingdom are facing major changes in organisa-tion and practice. There is overwhelming evidence insupport of centralisation of trauma care and reductionin the number of hospitals receiving injured patientswithin a district. ''- In particular the recent RoyalCollege of Surgeon's report on the management ofpatients with major injuries' recommends that eachdistrict should have only one major accident andemergency department.A separate study by the British Orthopaedic Associ-

ation endorses this view and recommends that eachdistrict should have only one hospital receiving patientswith major trauma.4 The management of multipleinjuries and complex musculoskeletal trauma isgenerally not possible in small units, and patientsshould be taken directly to hospitals that have theequipment and staff to deal with them.24 Non-designated hospitals might incur an increased workloadofpatients admitted for elective orthopaedic procedureswho may be relocated from the district trauma centre.

Set against these arguments for centralising traumaservices are factors particularly relevant to rural areas,such as transport difficulties and the traditional bondsand loyalties of patients and staff to local hospitals.5Changes in accident and emergency departments have"knock-on" effects on other departments-forexample, reduction or closure of an accident andemergency unit may deprive another specialty ofadmission facilities for patients with severe problems.The reasons for altering the services must be explainedto staff in other departments and the community.Community health councils should be involved early,and they in turn should help educate the public whosewelfare is their brief. With regard to the establishmentof regional trauma units the British OrthopaedicAssociation states that it is not possible to assess howpressing the need is for such units until the results ofaccurate audit are available from district traumacentres.Gwynedd is a large county comprising Snowdonia,

Anglesey, the Lleyn Peninsula, and the coastal stripincluding the population centres of Bangor, Llan-dudno, and Caernarfon (figure). The resident popula-tion was approaching 240 000 in 1988 and is subject to aconstant and large seasonal increase caused by holiday-makers.The major accident centre is in our hospital, which is

BMJ VOLUME 300 24 FEBRUARY 1990504

TABLE I InJury severn'ty scoresin patients with trauma whoattended in August 1988

Scorc No of patients

0l234s6

9101314161725

71385

31l

284212s

12662

l

l

2

sited well geographicallyN, being in the heart of thecountv and close to the A55-the North Wales coastroad. The accident unit is part of the trauma andorthopaedic service for Gwynedd, which is based atBangor. The total medical staffing consists of threeorthopaedic consultants, an associate specialist, twoorthopaedic registrars, and seven senior house officers.All staff work at various times in different parts of theservice. At the time of writing the unit comprised anaccident and emergency department, a 23 bed severetrauma ward, and two trauma theatres. Accident unitstaff deal with all emergencies within their capabilities,but they normally offer continuing care only for thosepatients with musculoskeletal, head, thermal, chest,and spinal injuries. Patients within the province ofother specialties are referred directly to the relevantdepartments; in cases of multisystem injury thedestination of the patient is decided according toclinical priorities. First line cover in the accident unit isprovided by one senior house officer in the accidentand emergency department (two at busy times) and onein the trauma ward, each working a one in three rota.More senior staff back up the senior house officers,conduct daily trauma operating lists, and undertakefracture and hand clinics in premises adjoining theunit.There are lesser accident and emergency depart-

ments in hospitals at Llandudno, Holyhead, Porth-madog, Pwllheli, and Blaenau Ffestiniog. Patientswith serious injuries presenting to these hospitals arereferred to our hospital. There are also communityhospitals at Dolgellau and Tywyn at the edges of thecounty, and patients presenting there may be referredto nearer hospitals in adjoining counties. Ambulancestaff usually take patients with major trauma directly toour hospital.We present the workload of our hospital in an

attempt to quantify the demands made on it and itsregional supporting services.

MethodsOur study comprised two parts. Firstly, we per-

formed a prospective detailed analysis of all patientswho attended the unit in August 1988, a seasonallybusy month. Accident and emergency and necropsyrecords were reviewed by one of us (SJK), and thefollowing were assessed: diagnosis, injury severityscore,6 and outcome. The demands created by thesepatients on the trauma ward, theatres, and outpatientclinics were also assessed.

Secondly, for the calendar year April 1988 to April1989 we performed a retrospective analysis of thegeneral workload of the unit by using the computerbased accident and emergency records system togetherwith a manual study of the necropsy records for thesame period. These figures were correlated with thosefor August 1988. We also estimated the error factor ofsuch a computer system.

ResultsAUGUST 1988

In August 1988 there were 2296 patients who created2325 attendances (there were 29 revisits); 37 patientsdid not wait to be seen. The accident and emergencyrecords of 13 patients were missing, and two patientscould not be classified owing to inadequate informationhaving been recorded. Howev'er, the presentingcomplaint as recorded by the clerical staff at the time ofarrival of the patient in the incomplete records and therecords of those who did not wait allowed many ofthese cases to be classified at least as trauma or non-trauma, and in some cases the injury severity scorecould be calculated.

We analysed 2302 patient attendances. Revisitingpatients or any presenting late after trauma (forexample, with an infected laceration) were given thesame injury severity score as would have applied to theinitial injury. We thought that this would mostaccurately reflect the time demand such patients hadmade on the unit.Of the 2302 attendances, 1904 (82.7%) were for

trauma. Of these, 478 (25 1%) were for contusions andabrasions, 413 (21-7%) for fractures or dislocations,366 (19.200) for lacerations, 242 (12-7%) for ligamentsprains, and 120 (6 3%1o) for the presence of foreignbodies (including those affecting the eye); all othertypes of injury each accounted for less than 4% of atten-dances. Seventy one patients (3 1 %) had more than oneinjury.

In all, 213 (9.2%) patients were admitted to thetrauma ward. Of these, 103 required an operation thatentailed incision. The operations included fixation orhemiarthroplasty for fractured neck of the femur (33cases), fixation of ankle fractures (14 cases), and repairof hand injuries (21 cases). Eighty patients requiredmanipulations that did not entail incision and wereperformed under general, regional, or local anaesthesia,either as overnight or day cases. These included 56patients with displaced forearm or wrist fractures,seven with displaced midshaft tibial fractures, andeight with dislocated shoulders.Some patients who required an operation or manipu-

lation were admitted directly to the trauma ward fromother hospitals and were included in the operation ormanipulation figures. Other admissions to the unit(those not requiring an operation) included mainlyelderly patients with minor injuries who were unable tomanage on their own and patients with head injurieswho were admitted for observation.Two patients died in the unit, one after failed

resuscitation after drowning and another after sufferingtrauma (an elderly person with a fractured neck of thefemur). Two patients required transfer to the regionalneurosurgical unit at Liverpool. A patient with acomplex acetabular fracture was transferred forconsideration ofpelvic reconstruction. Only one patientwas admitted to general surgery after trauma (thepatient had perirenal haematoma). Two patients diedoutside hospital: one was a pedestrian who had been hitby a car and received instantly fatal head injuries, andthe other had a C2/3 fracture dislocation after fallinginto shallow water.Those who attended the hospital had a mean (range)

injury severity score of 2-13 (0-25) (table I). Allpatients with an injury severity score >11 requiredadmission. All those with a score of > 8 either requiredadmission or had a displaced fracture requiringmanipulation under anaesthesia as a day case. Alldisplaced radial fractures such as Colles' fracturesscored nine (32), as did all fractured necks of the femur.Only two patients had a score of >20. The patients whodied outside hospital both had a score of 75.

APRIL 1988-APRIL 1989In the calendar year April 1988 to April 1989, 21 007

patients attended the unit. Tables II and III show thenature of injury and outcome (by the method ofdeparture) in the patients. Notable features are that88-2% (17 958) of the attendances were for orthopaedicinjuries or those caused by an accident, or both,including deliberate self harm and assault, and 8-6%(1746) were for medical emergencies; this compareswith similar figures of 86-2% (1980) and 9.5% (218),obtained for Augtust. A total of 2175 (10-7%) werereferred to one of the outpatient clinics directly, and1966 (9.7%) were admitted to the unit. Most of thesepatients were also followed up in the outpatient clinics;some, however, were referred back to their own

BMJ VOLUME 300 24 FEBRUARY 1990 505

regions for follow up or to their general practitioner.Tables IV and V give a breakdown of the patients'

age and type of injury. The annual figures for the fourlargest age groups are very similar to those obtainedabove for August alone. About three fifths (10601) ofour patients were aged 30 or under, and over threequarters (13 658) were aged 45 or under. Table VIshows workload in terms of place of residence. In all,84 4% (17 603) of the patients were resident inGwynedd, 15 6% (3255) coming from elsewhere in theUnited Kingdom or overseas. In August, proportions

I'ABLE VI -I'lace of residence oJ patientts with trauima who attended inAugust 1988 and those who attended in April 1988 - April 1989.Figzures are numbers (percentages ) ofpatients

Place of residence

Gwynedd (permanently)Gwynedd (temporarily)Othcr parts of \X'alesOther parts in rest of UKOserseas

'totalCoding crrorActual total

August 1988 April 1988 -April 1989

1460 (63-6)3 (0 1)

58 2-5)743 (32 4)32 (1-41)

1728s 82 9)320 (19O331 (16)

2804)13 4)120 (0-6)

2296 20898149

21007

TABLE II-Category, of injur-N, or circumstance in patietits who attenidedin April 1988 - April 1989

No 'N,

Orthopaedic/caused bs accidentSelf harmAssaultMedicalSurgicalPsychiatricNo abnormalitVDid not waitBrought in deadReferred from other hospital

Total by computerCoding errorActual total

17577265116

1746346451271011231

20366641

21007

86-31 30 68 6170-20 60 50 10 1

100

TABI III-Outcome in patients who attended in April 1988 - April1989

t)ischargedRef'erred to general practitionerAdvised to revisitRef'erred to outpatient department:With fractureWith orthopaedic injuryWith hand injurs

Admitted to accid nt unitAdmitted to other specialtiesReferred to other specialtyLeft or refused treatmentDied

Total by computer( oding errorActual total

No

105253436148

200933143

1966126367613730

20366641

21007

16 90-7

rABIE IV-Age ofpatients with trauma who attended in April 1988 -

April 1989

Age (sears) No 'l

0-15 4646 26-416-30 5955 34-131 -45 3057 17 446-60 1773 10 161-75 1274 7-376-90 748 4-391-110 68 0-4

l'otal by computer 17521 100Coding error 3486Actual total 21007

[ABLE V-Tvpes of injury in patients with trauma who attended inApril 1988 - April 1989

No v

Contusion/abrasionFracture/dislocationLacerationSprainPresencc of forcign bodySoft tissue infectionBites/stingsThermal injuryPoisoning/optical defectOther*

Total by computerCoding errorActual total

*Includes injuries that wcrc not classificd.

4734392930332424104945733828023S1042

175213486

21007

IABLE VII-Cautses of death il patients with trauma according toGwvnedd county necropiv records for April 1988 -April 1989

Cause of dcath No

Head injury (including two self inflicted gunshot wounds) 11Cervical spine injury (tso acute) 3Crushed chest 2Traumatic aortic rupturc 2Ruptured liver one acute) 2Multiple injuries (road traI'fic accident and falls) 9Explosion 2Exanguination (self-inflicted neck laceration) IHanging 3Mlyocardial infarct seconidary to burmis IPulmonarv embolus aftcr leg injury 3Medical complications after fractured neck of femur 4Drowninig 7

Total 50

,, '[ABLE VIII-Place atnd cause of death in patients with traumaaccording to Gwvnedd county necropsy records for April 1988 - April

51-7 1989

Place and cause of death Total

Scene of inicidentResuscitation room:

After drowning, failed resuscitationOf multiple injuries, including lacerated liverOf ruptured thoracic aortaOf ruptured aorta and thoracic spine fracturc/dislocation

Another hospital before transfer:Of multiple injuries, including fatal head injuryOf laceratcd liver and disrupted pelvisOf massive cervical spine disruption

At home or in another hospital after discharge or transfer:Of pneumonia after cervical spine injuryOf pneumonia after fractured neck of femur and lymphomaOfpulmonarv embolus at home after fractured neck offemur

Gswynedd Hospital:Of pulmonary embolus after lower leg fracturcOf pulmonarv embolus after leg haematomaOf pneumonia after fracturcd neck of femurOf pneumonia after lacerated liverOf myocardial infarct after 25"% burinsOf fatal head injursOf mvocardial infarct after fractured neck of fe'mur

31

11ll

2jlj

l j

I j21J21

I

II

3

were 63-7% (1463) and 36-3% (833) respectively,reflecting the large seasonal increase due to holiday-makers.

Necropsy records showed 50 deaths from traumaduring the same year. Tables VII and VIII list thecauses and places of death.

COMPUTER ERROR

There was an input omission of 2-0% regarding ageand a similar error regarding place of residence; 14.0n15of patients had not had their type of injury codedcorrectly. We calculated that reliable data were avail-able on 83-4% of patients.

DiscussionThe Royal College of Surgeon's report makes the

following recommendations-':* Regional trauma centres should be established fortreating patients with an injury sexrerity score of >20.Patients who would benefit from transfer to a traumacentre would be those with major damage to thoracic

BMJ VOLUME 300 24 FEBRUARY 1990506

and abdominal viscera, severe head injury, or complexorthopaedic trauma such as major pelvic fractures* The centres should serve a population catchmentarea of about two million people* The centres should be staffed by people workingthe equivalent of three eight hour shifts per day, withstaff on standby exclusively for treating victims ofmajor injury. This would entail more than 1000admissions per year (1095 = 1 per shift) to give suchstaff a justifiable workload and adequate experience.We thought that some evidence of workload was

necessary to establish the accuracy of such proposedfigures. There was good correlation between themanual study and the one year computer analysis,which, though subject to certain user induced errors,still accurately reflects a unit's workload and can beused for study purposes.The analysis for August 1988 found only two

patients with injuries that a district general hospitalmight not "be expected to provide a full service for"according to the Royal College of Surgeon's report(injury severity score >20), although other studieshave used slightly lower scores to define seriousinjury.' One of the two patients had a head injury withan intracranial collection and required transfer forneurosurgical drainage (injury severity score = 5'225), the other as a result of a road traffic accident had aflail chest with haemopneumothorax plus a basal skullfracture (injury severity score = 4 + 32 = 25) and wassuccessfully managed in our intensive care unit. Wethink that our unit would encounter about 24 suchcases per year-that is, only one for every 10 000 of thepopulation served. On this basis a regional traumacentre would need to serve a population not of twomillion, but 10 million, to justify its independentexistence with "set aside" staffing. Using the injuryseverity score definition of major trauma (a 16) we stillfound only four patients fitting the definition, meaningthat a regional trauma centre would still require apopulation of about five million. Clearly Britain doesnot generate as many patients with major trauma as anequivalent American population, and thus a modifiedapproach to centralisation of trauma facilities that ismore suitable to British conditions is necessary.This does not mean that we are reluctant to acceptcentralisation oftrauma services. We strongly advocateit and already practise it within our district to allowquality care of traumatised patients, the benefits ofwhich are clearly shown in previous studies. 009Though the injury severity score has proved to be of

benefit in predicting prognosis in terms of mortality,morbidity, and length of hospital stay,'" in cases ofblunt trauma and most penetrating injuries other thangunshot wounds" we found that it had many draw-backs. For instance, it cannot predict the need forpatients' transfer; it is difficult to ascribe a score topatients with head injuries on admission; it was notable to predict whether a patient would require anoperation, with its associated risks; and in patientswith multiple limb fractures more than one fracturedoes not merit an increased score.Thus we think that the score alone should not

become the benchmark for the need for transfer ofpatients when organising more centralised traumaservices such as those currently contemplated.The Royal College of Surgeon's report also suggests

the following breakdown of the workload ofan accidentand emergency department in a typical district generalhospital: 60% patients with trauma, 20% patients withacute non-traumatic conditions requiring surgery, and20% patients with medical, paediatric, and psychiatricconditions. Our figures, which are at variance withthese, are 88-2% patients with trauma, only 1-7%patients with acute surgical problems, and 8o8%

patients with acute medical, paediatric, and psychiatricproblems.Our figures were achieved by the following policies:

* Patients with acute surgical and medical problemswho require admission and are seen firstly by generalpractitioners should be referred directly to the relevantspecialties and not to the accident and emergencydepartment unless the problems are life threatening.This requires the cooperation and goodwill of thegeneral practitioner and hospital colleagues plus regularexplanation of policies to new doctors

* Patients must not be allowed to use an accident andemergency unit as an after hours general practice. Apatient attending with a condition that is not urgentshould be examined but then normally redirected toseek treatment from a general practitioner* Rates of reattendance should be kept to a minimumand patients directed to separate outpatient clinics orback to their general practitioners for follow up. Inparticular, removal of suture from simple lacerationsshould be done by general practitioners in almost allcases. As 17-3% of our annual attendances were forlacerations, most of which would have requiredsuturing, our workload could have been increased byabout 15% if these patients all came back to theaccident and emergency department for removal ofsuture.

Only three of the 2296 patients attending in Augustwere directed to reattend the unit. All three wereholidaymakers who had sustained superficial burnsand were asked to attend for dressings. Otherwise,dressings can be done by district nurses or familydoctors. If all patients with wounds attended forsecond dressings the workload of the unit might beincreased by up to 50%.From our results a district general hospital can

expect to generate the following minimum workload:8 75% of the local population served can be expected toattend annually; almost 10% of those presenting arelikely to require admission for trauma, and halfof theseare likely to require an open operation; an additional3-5% will require closed manipulations; a further 10%will need referral for follow up in the outpatient clinic.As nearly all manipulations and most admissions to

the trauma ward also require local follow up in clinicsabout 20% of all patients attending the accident andemergency department will eventually need outpatientfollow up. Many will require more than one visit. Acorollary ofthese figures is that an inefficiently managedaccident unit, which permits or even encouragespatients to reattend for aftercare of wounds or followup of fractures rather than redirecting them to generalpractitioners and separate clinics, would be able toreduce its workload by up to 25% by narrowing theservice to the quality care of acute trauma.Our study supports the view that major trauma

should be centralised at district level, wherein a singledistrict general hospital builds up expertise in themanagement of all patients with trauma of mediumseverity4 and at least competent initial managementand stabilisation of patients with major multipleinjuries. Such expertise can be gained only whendistrict health authorities, with the advice of seniormedical staff, designate one hospital within theirdistrict as the receiving centre for all but trivial cases oftrauma. The medical staffresponsible for that hospital'saccident and emergency department must then have anactive role in consultation with colleagues in otherspecialties to make sure that it becomes an efficientunit.With regard to the planning of regional trauma

centres we believe that further studies from otherdistrict general hospitals are needed before the criteria

BMJ VOLUME 300 24 FEBRUARY 1990 507

for their establishment can be decided. In general,musculoskeletal as opposed to multisvstem trauma iswell managed at district general hospital level inBritain, and the Royal College of Surgeon's report maywell have overestimated district general hospitalrequirements for regional trauma centres.

We thank Mrs B Davies for her help in preparing themanuscript.

I Department ot Healhti and Social Securitv. Ilospitl medical staffing-achievinga balance: a plan for action. London: D)HSS, 1987.

2 Roval College ot Surgeonis. Report of the working party on the management ofpatietits nith major in urtes. ILonidon: RCS, 1988.

3 Buck N, D)evlin HIS, Lunn JN. Report of the confidential enzquir%y intoperioperative deaths. London: Nuftield Provincial Hospitals Trust and theKings Fund, 1987.

4 Trauma Subconamittce ot the British Orthopaedic Association. Thtemanagement of trat(ma in Great Brntain. London: BOA, 1989.

S Ramsav A. N\'VoS1tor Journal of Ncw Souith Wales branch of the AustralianMedical Associationi March 1989:32.

6 Civil I), Schlwab CWV. 'I'he abbreviated injury scale, 1985 revision: acoiidciiscd chart for clinical use. 7 Irumnta 19X6;28:87-90.

7 Christian IMIS. Morbiditv and mortality of car occupants: comparative surveyover 24 months. Br ledj 1984;289: 1525-6.

8 NW'est J(i, Calcs RH, Grazzaniga AB. Impact of regionalisatioi: the OrangeCountv expcrience. Arch Surg 1983;188:740-4.

9 Eastman AB, Shacklord SR, Hollingworth-Fridlund P, Cooper LN. 'Thecffe'ct of' regionalisation upon the quality of trauma care as assessed byconcurrent audit bcfore and al'ter the institution of a trauma system; apreliminary report. 7 Trurna 1986;26:X12-20.

10 Bull J1. 'I'he injury severity score of road traffic casualties in rclation tomortality, time ol' death, hospital treatiiecnt time and disability. Accid .AnalPrey 1975;7:249-55.

11 Seinmlow J l, Conc R. Utility of the injury severity score: a confirmation.Hi'alth S'r Res 1976;11:45-52.

12 Bcverland D)E, Ruithicrford WH. Ani asscssmcnt of the validity of' the injurNseverity scorc swhen applied to gunshot woutids. Inj;urv 1983;15:19-22.

13 M,acKenzie EJ, Shapiro S, Eastham JN. Rating AIS severity using emergeicydepartmeint shects vs inpatient charts. 37 rauma 1985;25:984-8.

.Accepted1-5lDcemnber 1989

Hearing disability in people aged 50-65: effectiveness and acceptabilityof rehabilitative intervention

S D G Stephens, D E Callaghan, S Hogan, R Meredith, A Rayment, A C Davis

Welsh Hearing Institute,University Hospital ofWales, Cardiff CF4 4XWS D G Stephens, MRCP,audiological physicianD E Callaghan, audiologicaltechniczianS Hogan, MSC, audiologicalscientistR Meredith, audiologicaltechnicianA Rayment, hearing therapist

Medical Research CouncilInstitute of HearingResearch, University ofNottinghamA C Davis, I'HD,epidemiologist

Correspondence to: DrStephens.

BrAled]7 1990;300:508-11

AbstractObjective-To determine the best means of

detecting hearing disability in subjects aged 50-65and whether rehabilitative intervention is acceptablein this age group.Design-Questionnaire survey of patients on

general practice age-sex registers. Two types ofquestionnaire were used, one being based on theclosed set approach of the Institute of HearingResearch questionnaire, which had been used in apilot study, and the other being a simplified versionofthis questionnaire developed by the Welsh HearingInstitute and based on open set questions. Question-naires were sent up to three times, and any patientswho had not responded two months after the lastposting were personally contacted.Setting-Two general practices in Glyncorrwg

and Blaengwynfi in the Afan valley, West Glamorgan.Patients-271 Patients in Glyncorrwg (136 men,

135 women) and 333 patients in Blaengwynfi (173men, 160 women) aged 50-65.

Interventions-All patients indicating hearingdisability in answering the questionnaires wereinvited to attend for a evaluative session in theirvillage. After audiometric testing advice andarrangements for fitting a hearing aid were offered asappropriate.Main outcome measures- Response rates and

prevalence of hearing disability before interventionand of possession of hearing aids before and afterintervention.

Results -After three postings and personal contactthe response rate was 98% (266/271) in Glyncorrwg,where the complex questionnaire was used, and 97%(322/333) in Blaengwynfi. The prevalence of hearingdisability was respectively 53% (141/266) and 46%(148/322) and the prevalence of owning a hearing aid7% (19/266) and 8% (24/322). After intervention thepossession of hearing aids rose to 24% (64/266) inGlyncorrwg and 22% (71/322) in Blaengwynfi; sixmonths later the aids were being used regularly. Adirect comparison of the two questionnaires in 69subjects from Blaengwynfi showed no significantdifferences in the amount of disability detected byeach one. The first posting ofquestionnaires detected65% (189/289) of the hearing disability in the twovillages or 78% (72/92) of those prepared to accepthearing aids for the first time; 96% (88/92) of those

who accepted hearing aids were detected by twopostings.Conclusions-Simple questionnaires are effective

in detecting hearing disabilities in people aged 50-65,and intervention was acceptable in many of thosewho reported having difficulties in hearing. Theresponse rates from successive postings suggest thattwo postings are sufficient in terms of the return indetecting those who will accept intervention.

IntroductionThe average patient presenting at a hearing aid or

rehabilitation clinic for the first time is aged about 70years and has had hearing problems for about 15years.' 7 By then the patient and his or her family haveexperienced considerable frustration because of thisdisability. Difficulties in adapting to new listeningconditions and to handling hearing aids are more likelyin such patients than in those presenting earlier.One way ofreducing this delay would be to introduce

a secondary prevention programme by screening forhearing disability and impairment. The most effectivepeople to target with such a programme are those aged50-65, in whom the prevalence of hearing impairmentand disability begins to increase noticeably.4 At anearlier age screening would not be cost effectivebecause of the low prevalence. At a later age it would betoo late.We consequently performed a pilot study in subur-

ban Cardiff in which we contacted patients aged 50-65from a group practice age-sex register by either sendinga disability questionnaire or performing domiciliaryaudiometric screening. Rates for the acceptance ofhearing aids were similar with the two approaches, andafter intervention the rate of using hearing aids rosefrom 3 5% to 8 9% in this population.'

Several substantive and clinical problems wereassociated with the pilot study, and we extended it to adifferent population with a higher proportion ofsubjects from the manual social classes. We werealso interested in comparing two types of screeningquestionnaire, a closed set approach as used previouslysand an open set question as used in the Cardiff healthsurvey.4 We therefore approached all patients aged 50-65 in two practices (Glyncorrwg and Blaengwynfi) inthe upper Afan valley in West Glamorgan to determinethe best techniques for detecting disability and to

508 BMJ VOLUME 300 24 FEBRUARY 1990