Joumal of Advanced Nursing, 1995, 22,694-699

The health helief model and use of accident andemergency services hy the general puhlic

Mike Walsh PhD RGN BA(Hons) DipN(Lond) PGCE A&E JBCNS CertHead of Department of Nursing Studies, University College ofSt Martin, LancasterEngland

Accepted for pubhcation 6 December 1994

WALSH M (1995) Journal of Advanced Nursing 22, 694-699The health belief model and use of accident and emergency services by thegeneral publicThere has heen much dehate ahout the use made hy the general puhlic ofaccident and emergency services A strong element of professional dissaprovalhas heen present, as shown hy phrases such as 'inappropriate attender' Thispaper examines the reasons why people attend accident and emergency and thefactors that delay or accelerate attendance, utilizing a framework espoused inthe medical sociology literature, l e the Health Belief Model This predicts thatindividuals carry out a treatment cost-henefit analysis when making decisionsahout seeking medical assistance A sample of 200 adult, amhulatory accidentand emergency patients was lnterview^ed whilst waiting to see the casualtyofficer for this study The data demonstrated that much ofthe medical,sociological hterature concerning patient consultation with doctors isapplicahle to the accident and emergency situation, m particular the HealthBelief Model A range of factors was shown to make statistically significantdifferences to the delay times mvolved in deciding to attend accident andemergency and the time it took to then suhsequently attend and register as apatient These factors also fit the cost-henefit analysis which the Health BenefitModel predicts takes place Accident and emergency attendance therefore needsto he seen as a logical decision-making process that requires hospitals toprovide appropriate services, rather than merely lahelling the patients asinappropriate

^ 'inappropriate attenders' by many She rightly pointsout the need to consider the patient s percepbons before

There has been much discussion concermng the way the arriving at such judgemental labelsgeneral public utilize accident and emergency (A&E) ser- Typical of many medical research papers into A&E usagevices The medical literature is replete with papers that is the work of Davison et ai (1983), who claimed that 39%claim that a high proportion of attenders are at best 'map- of a sample of 587 A&E pabents were neither accident norpropnate' or at worst 'misusers' or 'abusers' of the A&E emergency cases, according to the clinical judgement ofservice (Fouroughi & Chadwick 1989) the casualty officers, and therefore should not be seen m

This feeling that many pabents who attend A&E should an A&E deparbnent Myers (1982) claimed that 54% of anot be there has been explored by authors such as Sbaih sample of 1000 A&E pabents could have been dealt with(1993), who showed that they are seen as 'rule breakers' by a general pracbboner (CP) according to his clmiccd

judgement Many other similar studies have been camedCorrespondence DrM Walsh Branthwaite House Crosby Ravensworth. OUt, SUch as that of Hobday (1986) who allocated 5 1 % ofnearPennth Cumbna CAio 3IP England a sample of A&E pabents to a grouping that could have

694 ® 1995 Blackwell Science Ltd

Accident and emergency services

been dealt with by the GP, Davies (1986) had a figure of63% of inappropriate attendances, while DnscoU et al(1987) stated that 35% of A&E attenders m their sampleshould have been seen by their GP One common featureof these and other studies is the lack of stated, objectivecntena by which to judge what constitutes 'appropriate'attendance at A&E Such subjectivity makes a questionablebase for sweeping claims about the appropriateness of A&Eattenders and makes meaningful replication of these stud-ies impossible

Interestingly, one of the few pieces of work to set outobjective cntena for appropnate attendance (Worth &Hurst 1989) found a much lower rate of 'inappropriate'attendance, l e 14% Perhaps this mdicates that subjectivefactors have a great deal to do with labellmg patients as'mappropnate' rather than considenng the possibility thatit IS the service that is inappropriate for the needs ofmany patients

ATTRIBUTION THEORY

The literature m the fields of social psychology and medi-cal sociology assist m the understanding of this labellingprocess Attnbution theory offers a partial explanation inthat staff may be making attnbutions to patients in connec-tion with their attendance that are intemal in nature ratherthan extemal (Heider 1958, Hewstone & Antaki 1988) Thistendency to ignore outside factors and see only the personas responsible for their behaviour has been tenned the'fundamental attribution error' by Ross (1977), lndicatmgthe importance of this error in perception Attributiontheory would therefore suggest that A&E staff may ignorethe possibility that there are very good reasons affectingthe person's everyday life which make it logical for themto present to A&E rather than their GP Instead, the patientIS held solely responsible for their attendance and anintemal attribution is made Thus hes the way to victimblaming

Helman (1991) has pointed out tbat the perceptions ofillness held by the public and health professionals are usu-ally very different, particularly with regard to the signifi-cance of symptoms This msight suggests that the generalpubhc may well behave m ways that seem inappropriateto health professionals when confronted by illness andinjury Evidence to support this diffenng perception aboutA&E comes from Wood & Cliff (1986), who found that acommon patient perception was that A&E was better orquicker than the GP

Health Belief Model

A further key theory that seeks to explain attendance pat-tems IS the Health Behef Model (HBM) (Rosenstock 1974)This proposes that individuals assess their susceptibilityto ill health, the nsks mvolved, the costs and likely benefits

of treatment before deciding to seek medical help Thissuggests that the patient makes a cost-benefit analysis andwill decide to attend A&E if the perceived benefits out-weight the perceived costs Jones et al (1991) support theapplication of the HBM to the A&E field, with their workshowing that the model's use achieved sigmficantimprovements in comphance amongst patients treated inthe emergency department This approach has its cntics,however Safer et al (1979) consider it too abstract, thereal world is much more complicated than this simplecost-benefit model m their view, whilst Rogers (1991) con-siders that the HBM depersonalizes the individual andreduces the person to the status of a calculating machine

There are, however, strong echoes of other theoreticalexplanations of patient attendance patterns in the HBMZola (1973), for example, talks of tnggermg factors thatlead to consideration of attendance, and cites the followingkey ingredients in the decision-making process

1 availability of care,2 cost of care,3 failure or success of alternative home remedies,4 patient perception of the problem,5 perceptions of significant others m the patient's life

This analysis contains strong elements of patient percep-tion and cost-benefit analysis m response to sjonptoms[triggering factors)

The other important body of work in this field stemsfrom the work of Mechanic (1978,1992), who lists variouskey factors involved in the decision to seek medical helpThese include the salience, persistance and perceivedsenousness of s}Tnptoms, disruption of everyday life andcompeting needs, accessability of treatment and the pamand anxiety generated by the condition Again, there arestrong elements of cost and threat being weighed againstthe benefits of treatment in this approach

The theoretical framework outlined briefiy here suggeststhat there may be altemative ways of looking at the waypatients utilize the A&E service rather than labelling themas inappropnate and indulging m victim blaming It waswith this framework in mind that the study descnbedtook place

THE STUDY

The study consisted of interviewing a sample of 200 ambu-latory A&E patients who presented at the minor injunessection of a major city A&E department in Englandbetween February and Apnl 1991 The sample consistedof 100 male and 100 female patients, aged 16-60, whosecondition did not m any way mfiuence cognition (suh-stance-use related attendances, head lnjunes, etc , weretherefore excluded) Patients were interviewed by theresearcher before seeing the casualty officer and wereselected on a sample of convenience basis on the days the

© 1995 Blackwell Science Ltd, Joumal of Advanced Nursing, 22, 694-699 695

M Walsh

researcher was in the department All days of the weekwere sampled, interviews occumng between 09 00 and21 00 hours

The structured interview lasted approximately 20minutes and had a mix of closed and open-ended ques-bons The schedule was piloted on a group of sue pabentsand minor amendments made m response The interviewsyielded a large amount of data and content analysis wasused to analyse responses to the open questions Moreformal statistical analysis was also camed out

Reliability and validity testmg

Denvabon of categones from open-ended quesbons is asubjective exercise Reliability was cbecked by having acolleague work through the data to denve categones andassign responses to those categones independently Thetwo sets of data were compared for levels of agreementutilizing the kappa coefficient technique (Eventt & Hay1992) to eliminate agreement by chance The results indi-cated strong agreement and hence the content analysiscould be assigned a high degree of reliability The abilityto clarify unclear responses witb subjects during the inter-view also enhanced the reliability of the data

Interview scbedule

The interview schedule was drafted after careful readingof the literature and also after the first phase of the study,which had involved analysis of a sample of 2000 A&E casenotes This helped ensure the content validity, althoughconsultabon with a panel of experts could have furtherenhanced content validity The data were felt to havestrong face validity as the interview was structured aroundthe patient giving their percepbons of their anxiety, painand understanding Finally, the researcher dressed m sucha way as not to be identified with the hospital A&E clinicalstafi' in order to minimize his impact upon the situation,thus reducing threats to the ecological validity of theresearch (Hammersley & Atkinson 1983)

RESULTS

The interviews were wide ranging The data presentedhere are that which was most relevant to understandingwhy the patient came to A&E when he nor she did Pabentswere asked about the bme it took them to decide to go toA&E (illness delay time, T,] and the time it took from thatdecision to registration (ublizabon delay bme, T^], as wellas their reasons for attending The marked skew that waspresent m the distnbubon of the delay bme data led tothe use of the median as a more meanmgful measure ofcentral tendency

Content analysis of the open-ended quesbon 'Why didyou come to A&E today rather than your GP'' produced

the responses identified m Table 1 There was a strongsense of pabents attendmg because they perceived it asmore beneficial to them

The diffenng distnbutions of ublizabon delay bmes mthe groups in Table 1 frequently achieved stabstical sig-nificance (Table 2) This suggests that there are relabon-ships between tmung of attendance and reason forattendance

Influence of others

When asked why the pabent had come to A&E at thisparbcular time, analysis of the answer again revealed astrong element of perceived advantage and also showedthe infiuence of others m makmg the decision (Table 3)

The marked vanations m utilizabon and lllness tuneswere striking, and were consistent with the reasons givenfor the timing of the attendemce This suggests that thereIS an internally consistent set of factors at work mfiuencmgdecisions about the timing of A&E attendances for manypabents When the data m Table 3 were tested for statisti-cal sigmficance, significant results were acbieved(Table 4) This supports the view that pabents makerational choices about their A&E attendance

When pabents were asked if they had to make any special

Table 1 Reasons for preferring A&E to GP

Reason NoMedianTu hours

1 A&E more appropnate or better than GP2 GP would send me here anyway3 Quicker/wait too long for GP appointment4 Sent by GP after lnibally going to GP5 Advised to go to A&E by others than GP6 More convenient than GP7 GP surgery closed/GP not available8 No GP or GP > 25 miles awayOthers

Total

40353529272321184

1 352 602 306 002 100 801 001 35

233

Some patients gave more than one answer

Table 2 Significant differences in utilizabon delay bmes

Pairs of reasons

6-46-26-34-76-all other reasons2-all other reasons7-all other reasons

f

10 911916 01546 01548 36517 8955

16 01546 3879

P

<0 01<0 001<0 05<0 02<0 02<0 001<0 05

696 © 1995 Blackwell Science Ltd, Journal of Advanced Nursing, 22, 694-699

Accident and emergency services

Table 3 Reasons for attendmg A&E at tbis particular tune

Median bours

Reason No

1 Convenience 54 8 40 2 802 Told to come by another person 46 4 25 1253 Pam and discomfort 33 1 50 l 204 Concem condition needs seemg now 31 0 10 0 805 Left to see if It got better, It did not 21 2100 1506 Tbougbt waiting times would be less 17 8 25 3 50

now7 Others 17 0 45 0 65

Total 219

Total exceeds 200 as some patients gave two answers

Table 4 Significant differences in components of delay times

Pairs of reasons ^ value

Illness delay times (TJ1-42-44-55-25-3

All otber reasons-4All otber reasons-5Utilization delay times |

1-4All otber reasons-4All otber reasons-1

11 7071

8 3191

22 4112

12 7231

15 1650

14 5018

17 5348

16 1250

9 1970

8 5114

<0 001<0 02<0 001<0 01<0 001<0 001<0 001

<0 001<0 02<0 02

arrangements to attend A&E at this time, 118 replied no, 44stated that they had to make arrangements to have time offworlc/college and 16 stated that they had to make arrange-ments for the care of dependant family members Of thislatter group of 16 patients, 15 were women

The lnfiuence of advice from others in attending A&Ewas demonstrated by the fact that 128 patients stated theyhad been advised to attend by vanous people who couldbe summanzed as fnends/relations (61), work/college col-leagues (31) and health professionals (36) Tbere was amarked gender difference in that men were twice as likelyto attend without advice than women, who m their tumwere almost three times as likely as men to have consulteda health professional before attendance

Anxiety and fear

Anxiety and fear were pronounced m the sample Whenasked to self-rate their anxiety on a scale of 0-5 where 0represented no womes and 5 extremely anxious, only11 5% of patients scored 0 whilst 30% scored 4 or 5

Table 5 Median delay times and anxiety levels

Low anxiety (0-2)(x" = 17 0641P<0Low anxiety (0-2)

7,0 90 bours Higb anxiety r ,(3-5) bours

Tu 0 85 hours Higb anxiety r»2 40(3-5) bours

Table 6 Most worrying aspect of ASbE attendance

Patient responses No of patients

Nothing 39Present condition 37Disruption work/study 31Disruption everyday bfe/leisure 29Possible complications of complaint 20Fear of tbe unknown 18Treatment required 8Waitmg time in A&E 8Others 10

Total 200

Whether patients were experiencing high or low levels ofanxiety produced a statistically significant difference inthe distnbution of the illness and utilization delay times,sucb tbat high anxiety levels were associated with longerdelay times (Table 5)

It was also possible to demonstrate a modest but signifi-cant positive correlation between pam and anxiety levelsfor the sample, giving a value of Spearman's correlationcoefficient of 0 411 (P<0 01)

When patients were asked what they found most worry-ing about their A&E attendance there was a number ofcategones of response (Table 6)

Again, significant differences were found m utilizationdelay times such that, for example, those concemed aboutdisruption to their everyday life had significantly longerdelays than those concemed about the present condition,future complications or those who had no womes at all

Those patients who talked of a fear of the unknown alsohad significantly longer utilization delays than those withno womes or those who were most womed about thepresent condition

DISCUSSION

The picture that emerges from these patient responses isnot one of people wilfully abusmg the A&E service, ratherIt IS one that is consistent with a rational decision-makingprocess The factors that emerge m patient decisionmaking are congruent with the HBM and the wntings ofMechanic (1978, 1992) and Zola (1973)

In Table 1 is listed a set of reasons that show that thepatient considers it to be to his or her benefit to have come

© 1995 Blackwell Science Ltd, Joumal of Advanced Nursing, 22, 694-699 697

M Walsh

to the AScE rather than the GP because it is seen as quicker,will produce better treatment, will save time as the GPwould only send them to AgcE anyway, or is more con-venient Patients without a GP or with the perception thatthey cannot access their GP felt that A&E is the only routeinto to medical care Both Zola (1973) and Mechanic (1978)stress the importance of access and availability m decisionmaking with regard to seeking medical assistance Patientsalso descnbed the importance of having been persuadedto attend by significant others (Zola 1973} Table 1 andTable 3 show a range of external factors and patient percep-tions that are important m influencing attendance, whichunderlines the importance of not making the fundamentalerror of intemal attribution

Perceived reason for attending

The validity of the convenience factor is supported by thesignificantly shorter utilization delay times associatedwith patients who give this as their reason for coming toA&E Once the person has decided to go to A&E becauseIt IS most convenient, it would be expected that she or hewould have significantly shorter delays in actuallyattending A&E than other patients, and this was mdeedfound The data show that the perceived reason for attend-ance can significantly affect the time it actually takes thepatient to attend

The reasons patients gave for attending when they didare again consistent with a cost-benefit analysis Reasonssuch as convenience and thinking the queue would be lessnow are about benefits to the patient, whilst factors suchas pam and concem suggest the costs of not attending havebecome too high Again the key role of significant othersm influencing attendance was apparent, whilst variationsm reasons were associated with significantly differentdelay times The small but significant number of almostexclusively female patients (8% of sample) who had tomake special arrangements for dependeint family membersbefore attending indicates the role of extemal factors minfluencing attendemce An extemal attnbution to behav-iour would be correct here, rather than the consistent errorof intemal attribution that we tend to make

Anxiety

The data showed a great deal of anxiety and worry amongstthe patients that was linked to pam and also delay timesPatients with higher anxiety levels seemed to delayattending A&E, even once they had decided to go, morethan less anxious patients Are A&E nurses aware of thisdegree of anxiety m what are usually referred to as 'minor'injuries' It might be a minor injury to the nurse, but to thepatient, whose perception of the importance of S5rmptomsmay be very different (Helman 1991), it clearly is not aminor condition

Health belief modelEvidence to support the HBM emerges again with factorssuch as disruption of daily activities (a cost) causing themost anxiety to 30% of patients This reason again indi-cates the importance of extemal rather than intemal factorsm influencing attendance The notion that this factor is acost IS supported by the findmg that it is associated withlengthened delay times in attendmg for treatment com-pared to other groups of patients

Anxieties relating to the present complamt or futurecomplications were uppermost for a further 28 5%, forwhom treatment could be seen as of benefit The otherfactors m Table 6 are also consistent with a cost-benefitanalysis, as fear of the unknown can be seen as somethingto overcome (cost) and is something that is associated withsignificant delays m seeking treatment compared to someother groups of patients

CONCLUSIONS

The patients in this sample gave a variety of reasons fortheir attendance at A&E that were consistent with the HBMand the work of Zola (1973) and Mechanic (1978) A rangeof factors showing evidence of cost-benefit analysis wasfound that also influenced the very timmg of attendance,as well as the basic decision to attend A&E itself

The important role of significant others m influencingattendance is supported by this study There is extensiveevidence that a range of extemal factors affect the decisionto attend A&E The labelling of patients as 'inappropriateattenders' is derived from intemal attributions aboutbehaviour, which this work suggests are usually inaccur-ate A predominantly extemal attribution would be morecorrect and less judgemental

The evidence in this study suggests that the term 'inap-propriate' should not be applied to patients in A&EAttendance is at the end of a logical decision-makingsequence, influenced by extemal factors emd consistentwith the HBM and the work of writers m the field of medi-cal sociology such as Zola (1973) and Mechanic (1978)The challenge to the A&E service is therefore to make theservice appropriate for the patient's needs rather thanblame the patient for being there

References

Davies T [1986] Accident Department or GP' BMJ 292, 241-243Davison A , Hildrey A & Floyer M (1983) Use and misuse of an

A&E department in the East End of London Journal of the RoyalSociety of Medicine 76, 37-40

DnscoU P , Vincent C & Wilkinson M (1987J The use of theaccident and emergency department Archives of EmergencyMedicine 4, 77-82

Everitt B & Hay D (1992) Talking About Statistics EdwardArnold, London

698 © 1995 Blackwell Science Ltd, Joumal of Advanced Nursing, 22, 694-699

Accident and emergency services

Foroughi D & Chadwick L (1989) Accident and emergencyahusers Tiie Practitioner 233, 8 May, 657-659

Hammersley M & Atkinson P (1983) Ethnography, Principles inPractice Tavistock, London

Heider F (1958) The Psychology of Interpersonal RelationsWiley, New York

Helman C (1991) Cuifure Heaifh and Illness Butterworth-Hememann, Cbdord

Hewstone M & Antaki C (1988) Attribution theory and socialexplanations In Introduction to Social Psychology(Hewstone M , Stroebe W , Codol J & Stephenson G eds)Blackwell Scientific, Oxford, pp 111-141

Hobday P (1986) Patients attendmg casualty from one group prac-tice The Practitioner 232, 9 May, 552-555

Jones S , Jones P & Katz J (1991) Comphance m acute and chronicpatients receivmg a health hehef model intervention in theemergency dept Social Science of Medicine 32(10), 1183-1189

Mechanic D (1978) Medical Sociology Collier Macmillan,London

Mechanic D (1992) Health and illness hehaviour and patient-practitioner relationships Social Science of Medicine 34, 12,1345-1350

Myers P (1982) Management of nunor medical prohlems andtrauma, general practice or hospital? foumal of tiie RoyalSociety of Medicine 75, 879-883

Rogers W (1991) Explaining Health and Illness Wheatdisaf,London

Rosenstock I (1974) Histoncal ongms of the health behef modelHealth Education Monographs 2, 328-335

Ross L (1977) The intuitive psychologist and his shortcomingsdistortions in the attribution process In Advances m Exper-imental Social Psychology [BerkovrtizL ed),Vol 10 AcademicPress, New York

Safer M ef al (1979) Determinants in three stages of delay inseeking care at a medical clinic Medical Care 17(1), 11-30

Shaih L (1993) Accident and emergency work, a review of someof the literature fournal of Advanced Nursing 18, 957-962

Wood T & Chff K (1986) A&E departments, why people attendwith minor injuries and ailments Public Health 100, 15-20

Worth C & Hurst K (1989) False alarm Nursing Times 85(15),24-27

Zola I (1973) Pathways to the doctor, from person to patientSocial Science and Medicine 7, 677-689

© 1995 Blackwell Science Ltd, foumal of Advanced Nursing, 22, 694-699 699