Fmily Prctice Vol. 20, No. 1 Oxford University Press 2003, ll rights reserved. Printed in Gret Britin The provision of sme-dy cre in generl prctice: n observtionl study Helen Stoddrt, Mggie Evns, Tim J Peters nd Chris Slisbury Stoddrt H, Evns M, Peters TJ nd Slisbury C. The provision of sme-dy cre in generl prctice: n observtionl study. Fmily Prctice 2003; 20: 41 47. Bckground. Surveys suggest tht generl prctice is perceived to be inccessible, with long delys before it is possible to be seen by doctor. Although there is demnd for rpid ccess to cre, this hs the potentil to decrese continuity of cre, which reserch consistently shows to be vlued by ptients but stressful for doctors. Objective. Our im ws to investigte the provision of sme-dy cre nd the chrcteristics, tretment expecttions nd priorities of ptients ttending for routine nd sme-dy ppointments. Methods. The uptke of sme-dy nd routine ppointments ws determined from ppointment system records. A self-completed questionnire survey of people ttending 362 sme-dy nd 362 routine ppointments t 15 generl prctices in Avon nd Gloucestershire enbled comprisons with respect of ptient chrcteristics nd consulttion experiences, ptients preferences for rpid ccess, seeing known prctitioner or seeing doctor rther thn nurse. Results. The uptke per 1000 prctice popultion per dy ws 7.6 [95% confidence intervl (CI) 7.1 8.1] for routine ppointments nd 3.8 (95% CI 3.5 4.2) for sme-dy ppointments. In univrible nlyses, use of sme-dy cre ws greter by people who were younger, non-white, in work nd with eductionl qulifictions. They hd ttended the prctice less often, wited less time to see doctor previously, hd their current problem for shorter time nd were more likely to expect prescription. From multivrible nlyses, only ge, speed of previously obtining ppointments, expecting prescription nd hving the problem for short time were independently ssocited with receiving sme-dy cre. Rpid ccess to cre ws rted s more importnt thn seeing known doctor for two clinicl scenrios. Conclusions. Much cre is provided currently on the dy it is requested in generl prctice. The use of sme-dy cre ws relted to cute illness nd expecttions but ws not ssocited with deprivtion or ptient chrcteristics prt from ge. Keywords. Generl prctice, sme-dy cre. Introduction There currently is much emphsis on improving ptient ccess to primry cre. The NHS Pln proposes tht ll ptients should be ble to see primry cre professionl within 24 h nd GP within 48 h. These proposls re prtly response to the perceived inccessibility of GPs, supported by ntionl survey which highlighted delys tht ptients experienced in obtining n ppointment. 1 Received 13 Mrch 2002; Revised 31 July 2002; Accepted 9 September 2002. Division of Primry Helth Cre, University of Bristol, Cothm House, Cothm Hill, Bristol BS6 6JL, UK. Correspondence to Dr Chris Slisbury; E-mil: c.slisbury@bristol. c.uk All prctices hve systems for mnging ptients who wish to be seen on the sme dy, such s reserving some ppointments for urgent cses, seeing ptients fter routine surgeries or providing open surgeries. However, GPs find tht providing cre for ptients on the sme dy is stressful. 2 The use of terms such s extrs 3,4 to describe these ptients implies tht this demnd is beyond the norml service provided. Previous reserch hs suggested tht demnd for sme-dy cre is more relted to socil fctors thn medicl need, being ssocited with younger ge, lower socil clss, single mothers, minor self-limiting illnesses nd when mking n ppointment ws considered inconvenient. 4 6 Possible wys to improve ccess include involving nurses in ssessment nd tretment, nd orgniztionl chnges such s the recently promoted Advnced Access model. 7 However, more rpid ccess potentilly 41
42 Fmily Prctice n interntionl journl could reduce continuity of cre from one helth professionl, which is relted to ptient stisfction nd possibly improved helth outcomes. 8 It is uncler whether ptients give higher priority to speed of ccess or continuity of cre. 9 This study investigted the proportion of cre currently provided on sme-dy bsis within generl prctice in the UK, nd compred the chrcteristics of ptients ttending for routine nd sme-dy ppointments. We lso exmined ptients expecttions nd the priority they give to rpid ccess versus continuity of cre, nd to seeing doctor rther thn nurse. Methods Smpling nd dt collection Fifteen GP prctices in Avon nd Gloucestershire were selected purposively to provide rnge of settings in terms of geogrphicl re, popultion chrcteristics nd prctice size. A questionnire ws distributed to smple of sme-dy nd routine ptients ttending these prctices. Sme-dy ptients were defined s those who sw the doctor on the dy they mde the ppointment. Some prctices used nurses to mnge ptients with minor illnesses, nd these consulttions were included. Routine ptients were defined s those who hd booked routine surgery ppointment before the dy of ttendnce. Sme-dy ptients did not include ptients who requested routine ppointment nd were given one on the sme dy, for exmple becuse of cncelltion. Tretment room ppointments nd specil clinics, such s dibetes or ntentl clinics, were excluded. For ech prctice, on two rndomly smpled consecutive weekdys, the trget ws to pproch the first 15 routine nd 15 sme-dy ptients per dy nd request tht they self-complete questionnire whilst t the surgery. Ech weekdy ws smpled n equl number of times. The ge nd gender of ll those pproched ws recorded. Prctice ppointment records were exmined to scertin the uptke of sme-dy nd routine ppointments in the study period. Questionnire Following n initil qulittive phse of 20 interviews with ptients on the dy of consulttion nd piloting, questionnire ws finlized covering socio-demogrphic sttus, pst experience of seeking ppointments nd detils of the current ppointment. Sme-dy ptients were sked their resons for requesting sme-dy rther thn pre-booked ppointment. Tretment expecttions nd preferences were explored by sking ptients to respond in three wys to two illness scenrios, one describing n cute illness (sore throt) nd one chronic illness (bck pin) (Box 1). First, ptients were sked wht sources of help they would consider in response to ech scenrio. Secondly, they were sked the extent to which they greed or disgreed with sttements relting to seeing doctor quickly, seeing doctor they knew nd seeing doctor rther thn nurse, using Likerttype scles. Thirdly, they rnked these three issues in order of importnce. With ptients consent, their medicl records were exmined t lter dte to scertin whether prescription ws issued, if secondry cre referrl ws mde nd the number of consulttions in the previous 12 months. Dignoses were obtined from the medicl records nd clssified independently by ME nd HS ccording to the body system ffected. Sttisticl issues Ignoring clustering effects, 183 ptients in ech group would yield 80% power to detect s significnt t the 5% level difference of 15 percentge points in the proportion of ptients expecting prescription (50 versus 35%). Allowing for clustering nd non-response (ssuming intr-prctice correltion of 0.02, 60 ptients per prctice nd 10% non-response) incresed this to 450 routine nd 450 sme-dy cre ptients, which ment 15 prctices were required. Univrible nd multivrible reltionships between type of ppointment nd ptient chrcteristics were investigted using descriptive sttistics nd logistic regression with robust stndrd errors to djust for clustering effects, using tests for heterogeneity or liner trend s pproprite. To void overlooking potentilly differentiting chrcteristics, 20% threshold for sttisticl significnce ws used to select vribles for multivrible nlyses; otherwise, 5% ws used. Ptients preferences for rpid ccess, seeing known prctitioner or seeing doctor rther thn nurse were investigted using logistic regression compring cross ppointment type within scenrios, nd the Sturt Mxwell test to compre between scenrios. Scenrio 1: Imgine you hve hd cold for couple of dys. During the night you noticed tht your throt ws very sore nd you hd temperture. Box 1 Scenrio 2: Imgine you hve hd bck pin for the lst couple of yers. It comes nd goes nd the doctor hs sid it s probbly rthritis. For the lst week it s been pinful gin.
Results Rte of sme-dy nd routine ppointments The verge list size of prticipting prctices ws 9401 ptients (rnge 4893 15 300). Appointment rtes were clculted from ll routine nd sme-dy ppointments over the 2-dy period. This ws only possible for 13 prctices since in two it ws not possible to distinguish when the ppointment hd been mde (one rurl, one urbn, both with list sizes of ~7500). A totl of 1419 ptients hd been seen per dy in the 13 prctices (941 routine nd 478 sme-dy). The rte per 1000 prctice popultion per dy of routine ppointments ws 7.6 [95% confidence intervl (CI) 7.1 8.1] nd of sme-dy ppointments 3.8 (95% CI 3.5 4.2). Questionnire survey Of the 783 ptients pproched in ll 15 prctices, 724 (92%) returned questionnire. Responders did not differ from non-responders by ge, sex or type of ppointment. Consent to ccess medicl records ws given by 638 responders (88%), with 23 (3.6%) notes unvilble. Chrcteristics of responders were similr to those in ntionl study of ptients consulting in UK generl prctice in terms of gender nd employment sttus, but the study smple contined more people ged 17 64 yers (68 versus 62%) nd more mrried people (70 versus 60%). 10 Resons for requesting sme-dy ppointment re shown in Tble 1. Needing prescription, hving the problem for long enough or hving to wit too long for n ppointment were cited frequently. Most TABLE 1 The resons given for why it ws importnt to see doctor/nurse on the sme-dy Reson n % (332 responses b ) My helth problem hs gone on for long enough 100 30 Need prescription or some medicine 97 29 I would hve to wit too long for n ppointment 83 25 Feel too ill to wit 78 23 Worried I might hve serious illness 49 15 Worried I might be infectious to others 38 11 It ws more convenient to come tody 34 10 Need sick note for work 18 5 The doctor I like to see is on duty tody 17 5 Wnt results of some tests 5 2 Other c 19 6 Response by sme-dy ptients only. b Totls exceed 100% s respondents could choose ny number of nswer options. c Included dvised to ttend by someone else. The provision of sme-dy cre in generl prctice 43 consulttions were with doctors rther thn nurses; 90% of 347 sme-dy nd 99% of 350 routine ppointments. As shown in Tble 2, sme-dy cre ws used more frequently by people who were younger, non-white, not retired nd those with eductionl qulifictions. Smedy ptients were more likely to be working or in the other ctegory, which included those studying or cring for children. Unemployment, not working due to illness or disbility nd receiving benefits (excluding stte pension) were not relted to ppointment type. Smedy ptients were more likely to be living with nonrelted dults. They hd problems of more recent onset, but third hd hd their problem for few weeks or more. Sme-dy ptients hd fewer ppointments in the previous yer nd were more likely to hve seen doctor previously on the sme dy. They were more likely to expect prescription nd prescriptions were issued more frequently to sme-dy ptients. Although only smll numbers of hospitl referrls were mde, this ws more common for routine ptients. Ptients seen on the sme dy with known dignoses (n = 291) were more likely thn routine ptients (n = 296) to hve ENT (24 versus 8%) or skin (14 versus 9%) problems, nd less likely to hve crdiovsculr (2 versus 13%), endocrine (1 versus 7%), urinry/reproductive/ obstetric (13 versus 18%) or mentl helth (4% versus 10%) problems. Only few of the chrcteristics studied were ssocited independently with the type of ppointment in the multivrible model (Tble 3). Younger ge, previously obtining n ppointment with ny doctor quickly, expecting prescription nd hving the problem for less thn week were ssocited with greter likelihood of receiving sme-dy cre. Tble 4 describes the speed with which prticipnts would seek help for ech of the illness scenrios. These decisions were similr by ppointment type. For the sore throt scenrio, 78% of 354 routine nd 72% of 349 sme-dy ptients would del with it themselves, with 9 nd 12%, respectively, visiting phrmcy nd only four people using wlk-in centre. For the bck pin scenrio, sme-dy ptients were slightly more likely thn routine ptients to choose to see doctor tody, but this ws of borderline significnce. In the ttitudinl questions, routine ptients were more likely to stress the importnce of seeing doctor they knew for both scenrios (trend P = 0.004). For the bck pin scenrio only, routine ptients lso hd stronger preference for seeing doctor rther thn nurse (trend P = 0.004). There were no significnt differences between sme-dy nd routine ptients in the priority they gve to seeing doctor tody (trend P 0.6 for the two scenrios). In the rnking exercise, Tble 5 shows tht being seen quickly ws rted s the priority for both scenrios. Seeing known doctor ws more importnt for bck pin, being seen quickly ws more importnt for sore
44 Fmily Prctice n interntionl journl TABLE 2 Descriptive chrcteristics of responders nd univrible nlyses of ssocitions with the use of sme-dy cre Chrcteristic Routine Sme-dy P-vlue b n % n % 1 Socio-demogrphic Age (yers) 0.005 0 4 7 2 34 10 5 14 15 4 44 12 15 24 22 6 51 14 25 44 105 29 119 33 45 64 135 37 70 20 65+ 77 21 41 11 Gender 0.60 Mle 136 38 129 36 Femle 225 62 231 64 Ethnicity 0.051 White Cucsin 344 96 327 92 Other 15 4 30 8 Eduction 0.038 None 107 32 86 25 CSE, GCSE, O levels 98 29 115 34 A levels, degree, higher degree 73 21 63 22 Other 62 18 12 19 Employment 0.052 Full-time 119 33 133 38 Prt-time 69 19 74 21 Retired 90 25 46 13 Other 80 22 100 28 Benefits received 0.69 Yes 66 19 71 20 No 276 81 277 80 Living sitution 0.0032 Husbnd/wife 258 72 242 68 Other dults 40 11 66 18 Alone/residentil unit 59 16 46 13 Children under 16 living with them 0.0006 Yes 107 30 168 48 No 245 70 179 52 Telephone vilble 0.98 Yes 342 97 338 97 No 12 3 12 3 2 Pst experience of GP ppointments No. of ppointments in lst yer c 0.018 d 1 2 44 15 60 21 3 4 72 24 80 38 5 6 55 18 53 18 6 7 38 13 36 12 9 93 31 61 21 Time to get n ppointment with ny doctor 0.014 Sme dy 131 39 187 55 Next dy 94 28 62 18 2 3 dys 81 24 63 18 4+ dys 33 10 34 10 Time to get n ppointment with prticulr doctor 0.020 Sme dy 20 6 42 12 Next dy 29 8 43 13 2 3 dys 123 35 94 28 4+ dys 178 51 169 47 3 Experience with current ppointment How long with problem 0.0001 Tody, few dys 29 8 163 47 About week 18 5 72 21 Few weeks 72 20 52 15 More thn month 233 66 63 18
The provision of sme-dy cre in generl prctice 45 TABLE 2 Continued Chrcteristic Routine Sme-dy P-vlue b n % n % Expected dvice 0.89 Yes 227 64 225 64 No 126 36 127 36 Expected informtion 0.81 Yes 107 30 103 29 No 246 70 249 71 Expected prescription 0.0002 Yes 159 45 227 64 No 194 55 125 36 Expected tretment other thn mediction 0.062 Yes 58 16 39 11 No 295 84 313 89 Expected to be sent to hospitl 0.56 Yes 13 4 10 3 No 340 96 342 682 4 Consequences of current ppointment Prescription given 0.0091 Yes 158 50 185 62 No 157 50 115 38 Referred to hospitl 0.018 Yes 25 8 12 4 No 289 92 288 96 Slight vritions in totls for ech chrcteristic reflect missing dt. b P-vlues were derived from univrible logistic regression models of the ssocition between the chrcteristic nd the type of cre, djusted for clustering by prctice. c Ptients who were registered for 1 yer re not included in this vrible. d P-vlue from test for trend cross ctegories. throt, nd there ws no difference between scenrios for seeing doctor rther thn nurse. Discussion This lrge nd brodly representtive study is the first to exmine the uptke nd chrcteristics of sme-dy nd routine ptients in generl prctice in Englnd. Although ccessing n ppointment in primry cre is perceived to be difficult, 1 bout third of ll consulttions were ctully delivered on the dy cre ws requested, lthough this did vry by prctice (14 65% of ll ppointments). Our findings do not support the previous reserch which suggested tht ptients requesting sme-dy cre might be seen becuse of socil rther thn medicl resons. 4 6 Ptients who ttended on the sme dy hd been seen less frequently in the lst yer thn routine ptients, nd the decision to consult quickly ws relted to the ptient s ge, recent onset of illness, expecttions nd previous experience. The study hs severl limittions. It is likely tht the chrcteristics of sme-dy ptients re prtly dependent on the prctices rrngements s well s ptients choices. A further limittion is tht consecutive ptients were pproched strting in the morning, rther thn rndom smple. This might hve resulted in n overrepresenttion of certin types of ptients. Finlly, the reserch ws bsed on smll smple of dys from reltively smll number of purposively selected prctices. It is interesting tht lthough mny of those seen on the sme dy hd their problem for very short period of time, third hd experienced it for over week. Why these individuls decided to be seen on the sme dy is uncler, nd non-medicl fctors might hve contributed. 11 Mny reported they felt tht they hd the problem for long enough, felt too ill to wit ny longer, would hve to wit too long for n ppointment or expected prescription. The multivrible model suggests tht the longer they wited in the pst the less likely they were to receive sme-dy cre on this occsion. Since those expecting prescription were more likely to receive sme-dy cre (djusted odds rtio of 2 in Tble 3), there might be some potentil to modify the demnd for such cre by ptient nd professionl eduction bout the benefits of mediction for minor illnesses nd the possibility of lterntively using overthe-counter tretments nd dvice from phrmcists. The use of sme-dy cre declined with ge, which my be relted to greter preference for continuity or to more chronic illness.
46 Fmily Prctice n interntionl journl TABLE 3 Vribles ssocited with the use of sme-dy cre: multivrible model Chrcteristic Odds rtios 95% CI 1 Socio-demogrphic Age group 0 4 4.3 1.9 9.7 5 14 2.3 0.8 6.8 15 24 3.2 1.3 8.2 25 44 2.4 1.2 5.1 45 64 0.9 0.5 1.6 65+ 1.0 P-vlue 0.028 2 Pst experience of GP ppointments Time to get ppointment with ny doctor Sme dy 3.1 2.1 4.6 Next dy 1.0 2 3 dys 1.1 0.6 2.1 4+ dys 2.4 1.0 5.3 P-vlue 0.0001 3 Experience with current ppointment Expected prescription No 1.0 Yes 2.0 1.4 2.7 P-vlue 0.0006 4 Time hd current problem Tody, few dys 1.3 0.5 3.1 About week 1.0 Few weeks 0.2 0.1 0.4 More thn month 0.1 0.03 0.1 P-vlue 0.0001 Ech vrible djusted for ll others in the finl model. It is uncler whether rpid ccess is comptible with continuity of cre, 8 which hs been rted by ptients nd GPs s being more importnt thn convenient ppointment for serious clinicl or psychologicl problems. 12 The findings support the notion tht continuity of cre is more importnt for chronic problems, lthough for both scenrios the most common preference ws to be seen quickly. Whether this would lso pply to other chronic conditions is unknown. A recent study hs lso found tht only few ptients nd GPs vlued personl GP more thn convenient ppointment for somewht similr scenrios (pinful knee nd severe cough nd cold). 12 There is considerble ptient nd politicl pressure to provide more rpid ccess to primry cre in the UK. However, this study shows tht much cre is lredy being provided on the dy it is requested. In the current debte bout ccess to primry cre, decisions will hve to be mde bout the trde-offs between speed nd continuity, between ptient demnd nd clinicl need, nd bout the most pproprite roles for doctors nd nurses. The rdicl re-orgniztion of ppointment systems is being hevily promoted currently, 7 bsed on some positive experience but little forml evlution. The costs nd benefits of these chnges for ll ptients, including the elderly nd those with chronic illnesses, re unknown. Acknowledgements We would like to thnk ll the prctices for their involvement nd the ptients for completing the questionnire. This study ws funded by the South West GP Trust. References 1 Airey C, Errens B. Ntionl Surveys of NHS Ptients: Generl Prctice 1998. London: NHS Executive, 1999. 2 Luthr M, Mrshll M. How do generl prctices mnge requests from ptients for sme-dy ppointments? A questionnire survey. Br J Gen Prct 2001; 51: 39 41. 3 Kendrick T, Kerry S. How mny surgery ppointments should be offered to void undesirble numbers of extrs? Br J Gen Prct 1999; 49: 273 276. 4 Virji A. A study of ptients ttending without ppointments in n urbn generl prctice. Br Med J 1990; 301: 22 26. TABLE 4 Illness scenrios: ptient s choice of ction Wht ptient would do Routine Sme-dy P-vlues n % n % Sore throt scenrio See the doctor tody b 23 7 38 11 See the doctor in few dys 19 5 14 4 Other 312 88 297 85 Totl 354 100 349 100 Bck pin scenrio See the doctor tody b 61 17 81 23 See the doctor in few dys 204 58 163 47 Other 88 25 103 30 Totl 353 100 347 100 0.15 0.058 Adjusted for ge nd clustering. b Includes seeing GP, hving home visit or ttending csulty.
The provision of sme-dy cre in generl prctice 47 TABLE 5 First priority of cre cross both scenrios Bck pin Know Quickly Dr not nurse Totl (%) Sore throt Know 42 17 5 64 (10%) Quickly 91 311 57 459 (72%) Dr not nurse 25 27 62 114 (18%) Totl (%) 158 (25%) 355 (56%) 124 (19%) 637 P 0.0001 Sturt Mxwell test. 5 Tylor B. Ptient use of mixed ppointment system in n urbn prctice. Br Med J 1984; 289: 1277 1278. 6 Field J. Problems of urgent consulttion within n ppointment system. J R Coll Gen Prct 1980; 30: 173 177. 7 Murry M. Ptient cre: ccess. Br Med J 2000; 320: 1594 1596. 8 Guthrie BWS. Does continuity in generl prctice relly mtter? Br Med J 2000; 321: 734 736. 9 Freemn G, Richrds S. Is personl continuity of cre comptible with free choice of doctor? Ptients views on seeing the sme doctor. Br J Gen Prct 1993; 43: 493 497. 10 McCormick A, Fleming D, Chrlton J. Morbidity Sttistics from Generl Prctice: Fourth Ntionl Study 1991 1992. London: OPCS, 1995. 11 Zol I. Pthwys to the doctor: from person to ptient. Soc Sci Med 1973; 7: 677 689. 12 Kerley KE, Freemn GK, Heth A. An explortion of the vlue of the personl doctor ptient reltionship in generl prctice. Br J Gen Prct 2001; 51: 712 718.