Fmily Prctice, 2016, Vol. 33, No. 3, 302 308 doi:10.1093/fmpr/cmw014 Advnce Access publiction 18 Mrch 2016 Helth Service Reserch Antibiotic prescribing for sore throt: cross-sectionl nlysis of the ReCEnT study exploring the hbits of erly-creer doctors in fmily prctice Anthe Dlls, *, Mieke vn Driel, Simon Morgn b, Amnd Tpley b, Kim Henderson b, Jen Bll c, Chris Oldmedow c, Andrew Dvey b, Kte Mulquiney b, Joshu Dvis d,e, Neil Spike f,g, Lwrie McArthur h nd Prker Mgin b,i Discipline of Generl Prctice, University of Queenslnd, Brisbne, b Generl Prctice Trining Vlley to Cost, Newcstle, c HMRI/CReDITTS, Newcstle, d Deprtment of Infectious Diseses, John Hunter Hospitl, Newcstle, e Globl nd Tropicl Helth Division, Menzies School of Helth Reserch, Drwin, f Victorin Metropolitn Allince Generl Prctice Trining, Melbourne, g Deprtment of Generl Prctice, University of Melbourne, Melbourne, h Adelide to Outbck GP Trining Progrm, Adelide, Austrli nd i Discipline of Generl Prctice, University of Newcstle, Newcstle, Austrli. *Correspondence to Anthe Dlls, Discipline of Generl Prctice, University of Queenslnd, Level 8, Helth Sciences Building, Building 16/910, Royl Brisbne nd Women s Hospitl, Brisbne, Queenslnd 4029, Austrli; E-mil: nthe.dlls@nd.edu.u Abstrct Bckground. Acute sore throt is common condition presenting to fmily prctitioners. It is usully self-limiting, with ntibiotic tretment recommended only for high-risk presenttions. Overprescribing of ntibiotics contributes to individul nd community resistnce. Lerning to prescribe in the context of dignostic uncertinty nd ptient pressures is chllenge for erlycreer doctors. Prescribing hbits develop erly nd tend not to chnge with time. Objective. To estblish the prevlence nd ssocitions of ntibiotic prescribing for cute sore throt by Austrlin voctionl trinees in fmily prctice. Method. A cross-sectionl nlysis of dt from the Registrr Clinicl Encounters in Trining (ReCEnT) study. This ongoing, multicentre prospective cohort study documents the nture of trinees consulttion-bsed clinicl experiences. Univrite nd logistic regression nlyses were conducted on dt recorded in consulttions for sore throt in nine collection periods during 2010 14. Results. Dt from 856 individul trinees (response rte 95.2%) were nlysed. Sore throt ws mnged in 2.3% encounters. Antibiotics were prescribed for 71.5% of sore throt dignoses. The vribles ssocited with prescribing were inner-regionl loction nd higher socio-economic re. There ws no significnt ssocition with younger ge of ptient or greter trinee experience. If n ntibiotic ws prescribed, the trinee ws more likely to seek informtion from guidelines or supervisor. The Author 2016. Published by Oxford University Press. All rights reserved. For permissions, plese e-mil: journls.permissions@oup.com. 302
Antibiotic prescribing for sore throt by erly-creer doctors in fmily prctice 303 Conclusions. The high frequency of ntibiotic prescribing nd the lck of ttenution in prescribing with incresed experience suggest current eductionl interventions nd the pprenticeship model of trining is not fostering pproprite prctice in this importnt clinicl re. Trgeted eductionl interventions, for supervisors s well s trinees, re indicted. Key words. Antibcteril gents, drug resistnce, evidence-bsed medicine, generl prctice, grdute medicl eduction, microbil, physicin prescribing ptterns. Introduction Acute sore throt presents prticulr chllenge for fmily prctitioners. It is common condition in Austrli 2.5 of every 100 presenttions to generl prctice re for sore throt, mking it the 10th most common reson for ttendnce in the period 2013 14 (1). Complints of sore throt in generl prctice settings interntionlly rnge between pproximtely one nd four per cent of presenttions (2,3). Group A streptococcl phryngitis is responsible in the USA for 5 15% of presenttions of sore throt in dults, nd 20 30% in children (3), with comprble popultion sttistics being demonstrted in Austrlin studies (4). The pek incidence is in school-ged children (4). Antibiotic therpy reduces the smll risk of complictions such s rheumtic fever, otitis medi, cute sinusitis nd quinsy if the etiology is bcteril (3). However, not only re these complictions very rre, but determining bcteril etiology of sore throt is difficult (5) nd there re significnt differences in guideline recommendtions cross the world. Both UK (NICE guidelines) nd Austrlin (6) guidelines suggest voiding ntibiotic use in sore throt, prt from the smll minority of ptients who re t risk of rheumtic fever (for exmple in Austrlin Aboriginl communities) or who hve severe clinicl fetures (such s high fever, lymphdenopthy nd difficulty swllowing). Overuse of ntibiotics contributes to resistnce in communities nd individuls (7) nd to unnecessry dverse side effects nd cost. Rtes of compliction ssocited with sore throt cused by Group A streptococcus re reduced by the use of ntibiotics. These include the forementioned suppurtive nd non-suppurtive complictions, but these re uncommon nd usully confined to prticulr high-risk popultions. The use of ntibiotics for sore throt in generl is ssocited with only modest benefits, shortening illness durtion by n verge of 16 hours nd with high number needed to tret in high-income countries to prevent severe complictions (8). Thus, the prescription of ntibiotics for sore throt remins controversil. Prescribing in the context of dignostic uncertinty nd ptient pressures is chllenge for doctors in fmily prctice (9), especilly erly-creer or trinee GPs, s they dpt to the different disese spectrum of community s opposed to hospitl prctice (10). Antibiotic prescribing prctices in these erly-creer stges re of prticulr importnce s prescribing ptterns, once estblished, tend not to chnge over time (11). In Austrli, voctionl trinees prctice s independent clinicins (including for prescribing purposes) while under the oversight of experienced clinicin supervisors in n pprenticeship model. They lso receive forml eduction from their regionl trining providers (RTPs). Apprenticeship models of Generl Prctice eduction re common worldwide. This study ims to estblish the prevlence nd ssocitions of ntibiotic prescribing for cute sore throt by Austrlin voctionl trinees in fmily prctice. Method This ws cross-sectionl nlysis of dt from the Registrr Clinicl Encounters in Trining (ReCEnT) study. ReCEnT is n ongoing, multicentre prospective cohort study of GP trinees (registrrs) in 5 of Austrli s 17 RTPs in 5 of the 6 Austrlin sttes. These RTPs hve opted to prticipte in the eductionl, reflective feedbck nd reserch components of the ReCEnT project. They provide brod nd representtive smple of the different trining conditions (urbn, regionl, remote nd very remote) tht re present in the Austrlin Generl Prctice environment. ReCEnT documents the nture nd ssocitions of the in-prctice consulttion-bsed clinicl nd eductionl experiences of GP trinees. The study protocol is described in detil elsewhere (12). Briefly, trinees complete pper-bsed forms recording detils of 60 consecutive consulttions in ech of their three 6-month generl prctice trining terms (12 monthly for prt-time trinees). Trinees t one of the five RTPs lso collected dt during n optionl fourth trining term. Trinees complete this process s compulsory prt of their trining, nd my consent to their dt to be used for reserch purposes. Trinee demogrphics nd prctice dt re documented t the strt of ech collection period, nd ptient demogrphics nd clinicl detils re recorded for ech ptient encounter. Trinee vribles recorded re ge, gender, trining term, whether in full-time or prttime (<8 hlf-dy clinicl sessions per week) trining, plce of primry medicl qulifiction (Austrli or interntionl), nd whether the trinee hd worked t the prctice in previous term. Ptient vribles recorded re ge, gender, Aboriginl or Torres Strit Islnder sttus, non-english speking bckground (NESB) sttus, the ptient being new to the prctice nd the ptient being new to the trinee. Prctice vribles recorded re size (smll prctice considered <6 doctors) nd billing policy (whether the prctice routinely bulk bills tht is, government subsidy is ccepted s full pyment nd there is no cost to the ptient). Prctice postcode ws used to determine the Austrlin Stndrd Geogrphicl Clssifiction- Remoteness Are (ASGC-RA) clssifiction to define the prctice loctions degree of rurlity (very remote, remote, outer regionl, inner-regionl or mjor city loction) nd Socioeconomic Index for Are (SEIFA) Index of Disdvntge. Consulttion vribles recorded re durtion, the nture of problems/dignoses mnged, if the problem/dignosis ws new or preexisting one, nd whether pthology or imging tests were ordered, or referrls or follow-up rrnged. Eductionl fctors included whether the trinee sought dvice or informtion during the consulttion (from their supervisor or other resources, such s specilists, books or electronic resources), or generted lerning gols. Problems/dignoses re coded ccording to the Interntionl Primry Cre Clssifiction (ICPC-2) nd medictions ccording to the Antomic Therpeutic Chemicl (ATC) Clssifiction.
304 Fmily Prctice, 2016, Vol. 33, No. 3 The nlyses in this study used dt from nine collection periods during 2010 14. Individul RTPs contributed from one to nine rounds of dt depending on the RTPs dte of commencement in the study. We defined cute sore throt s those problems/dignoses coded s ICPC-2 codes R72 (strep throt), R76 (tonsillitis, cute), R74008 (phryngitis, cute), R74006 (infection, throt), R74017 (phryngitis) nd R21005 (sore throt). Codes R72001 (scrlet fever), R72003 (scrltin) nd R72004 (scrlet fever) were excluded from the nlysis, s ntibiotics re recommended for ll cses of these conditions in the uthorittive Austrlin guidelines (6). Sttisticl nlysis The unit of nlysis ws the individul problem/dignosis rther thn the trinee consulttion. The proportion of presenttions for R72, 76, 74008, 74006, 74017 nd 21005 coded problems/dignoses were clculted with 95% confidence intervls (CIs). Proportions of R72, 76, 74008, 74006, 74017 nd 21005 coded problems/dignoses for which ntibiotics were prescribed were clculted with 95% CIs. Proportions of prticulr ntibiotics prescribed were lso clculted. For our primry nlysis, the outcome fctor ws whether n ntibiotic hd been prescribed. To test ssocitions of n cute sore throt being treted with ntibiotics, simple nd multiple logistic regression models were used within generlized estimting equtions (GEEs) frmework to ccount for the repeted mesures on trinees. Exct methods were used for covrites with low expected vlues in 25% or more of cells in the cross tbultion of the covrite nd the outcome. These re mrked s Exct in the univrite tbles. All vribles with P-vlue <0.2 nd relevnt effect size in the univrite nlysis were included in the multiple regression models. Vribles which hd smll effect size nd were no longer significnt in the multivrite model were removed from the finl model s long s removl of the vrible did not chnge the resultnt model. For those sore throt problems/dignoses where in-consulttion informtion or dvice ws sought, proportions of prticulr sources consulted were clculted with 95% CIs. Sttisticl nlyses were completed using STATA 13.1 nd SAS v9.4. Predictors were considered sttisticlly significnt if the P-vlue ws <0.05. Results A totl of 856 individul trinees (response rte 95.2%) contributed 1832 trinee rounds of dt (including detils of 108 759 individul consulttions nd 169 303 problems/dignoses). The demogrphics of the prticipting trinees nd prctices re presented in Tble 1. Sore throt ws mnged in 2495 (2.3%) (95% CI: 2.2 2.4) encounters nd comprised 1.5% (95% CI: 1.4 1.5) of problems/ dignoses mnged. Antibiotics were prescribed for 1783 (71.5%) (95% CI: 69.7 73.2) of sore throt (R72, 76, 74008, 74006, 74017 nd 21005 coded problems/dignoses). The individul ntibiotics prescribed re presented in Tble 2. The most commonly prescribed ws phenoxymethylpenicillin or benzthine phenoxymethylpenicillin (62.5% of totl ntibiotics). Associtions of ntibiotic prescribing for sore throt Univrite ssocitions of prescribing n ntibiotic re presented in Tble 3. There were significnt univrite ssocitions of ntibiotics hving been prescribed with ptient ge, with the ptient not being new to the prctice, with the problem being new (i.e. the initil presenttion of sore throt), with in-consulttion dvice or informtion being sought, with follow-up of the ptient being rrnged nd with greter number of problems being mnged in the index consulttion. The multiple logistic regression models for prescribing of n ntibiotic re presented in Tble 4. An ntibiotic ws more likely to be prescribed by trinees working in inner-regionl res compred to mjor cities (Odds rtio 1.37, 95% CI: 1.05 1.79), nd ptients t Tble 1. Chrcteristics of prticipting trinees, prctices nd consulttions (ReCEnT dt collection rounds 2010 14) Vrible Clss n % (95% CIs) or men (SD) Trinee vribles (n = 856) Trinee gender Femle 562 65.7% (62.4 68.8) Pthwy trinee enrolled Generl 641 75.2% (72.2 78.0) Rurl 211 24.8% (22.0 27.8) Qulified s doctor in Austrli Yes 664 78.5% (75.6 81.1) Trinee ge (yers) Men (SD) 32.5 (6.3) Trinee-term or prctice-term vribles (n = 1832) Trinee trining term Term 1 765 42.8% (39.5 44.0) Term 2 538 29.4% (27.3 31.5) Term 3 454 24.8% (22.9 26.8) Term 4 75 4.1% (3.3 5.1) Trinee works full-time Yes 1395 77.8% (75.8 79.6) Trinee worked t the prctice previously Yes 486 26.9% (24.9 29.0) Prctice routinely bulk bills b Yes 317 17.4% (15.8 19.2) Number of GPs working t the prctice 1 4 604 33.7% (31.6 35.9) 5 10+ 1187 66.3% (64.1 68.4) Rurlity of prctice Mjor city 1060 57.9% (55.6 60.1) Inner-regionl 521 28.4% (26.4 30.6) Outer regionl/remote/very remote 251 13.7% (12.2 15.4) SEIFA (decile) of prctice c Men (SD) 5.4 (2.9) My not dd to 865 or 1832 due to missing dt. b Consulttion t no cost to the ptient. c Socioeconomic Index for Are (SEIFA) Reltive Index of Disdvntge.
Antibiotic prescribing for sore throt by erly-creer doctors in fmily prctice 305 Tble 2. Individul ntibiotics prescribed by GP trinees for presenttions of sore throt Antibiotic nme Frequency % Phenoxymethylpenicillin 1 079 60.3 Amoxicillin 297 16.6 Roxithromycin 136 7.6 Ceflexin 103 5.8 Amoxicillin + enzyme inhibitor 41 2.3 Benzthine phenoxymethylpenicillin 39 2.2 Erythromycin 38 2.1 Cefclor 21 1.2 Other ntibiotics 37 2.1 Totl 1 791 100 prctices locted in higher SEIFA deciles (tht is, in res with less socio-economic disdvntge) were more likely to be prescribed n ntibiotic (OR 1.06, 95% CI: 1.01 1.10). Informtion or dvice ws sought by trinees in 19.9% of consulttions for sore throt. Trinees who prescribed n ntibiotic were more likely to ccess informtion or dvice during the consulttion (OR 2.37, 95% CI: 1.65 3.40). Exmining this further, this included sking supervisor [on 17.8% (95% CI: 14.1 22.2) of occsions when dvice ws sought] or using online [71.5% (95% CI: 66.5 76.0)] or hrdcopy [10.5% (95% CI: 7.7 14.1)] resources. There ws no reduction in ntibiotic prescribing with greter level of experience (lter trining term). In fct, there ws non-significnt trend to more prescribing in Term 2 compred to Term 1 (OR 1.19, 95% CI: 0.90 1.58) nd significntly higher prescribing in Term 4. There were no sttisticlly significnt differences between prescribing for ptients in younger ge groups (0 2 nd 3 14 yers compred with 15 50 yer ge group). There ws significntly less prescribing for ptients ged over 51 yers. Aboriginl nd Torres Strit Islnder sttus ws not significntly ssocited with ntibiotic prescriptions, however, the totl number of ptients from this popultion ws smll. Discussion Ptients presenting with cute sore throt were prescribed n ntibiotic in 71.5% of instnces. This is very lrge proportion given tht uthorittive Austrlin guidelines stte tht most ptients with sore throt do not require ntibiotic tretment. It is recommended tht ntibiotics be reserved for ptients t high risk of non-suppurtive complictions or ptients with prticulrly severe clinicl fetures suggestive of streptococcl infection (6). The rte of prescribing is well in excess of qulity indictors for ntibiotic prescribing in cute tonsillitis (13), nd suggests tht trinees re not prescribing ntibiotics in n evidence-bsed mnner. Estblished Austrlin GPs ntibiotic prescribing rtes re comprble t 88.1% of presenttions for sore throt (14). To compre this with other high-income countries, one study of Dutch GPs reported ntibiotics prescribed in 33% of consulttions for sore throt (15), nd US physicins prescribed ntibiotics in 60% sore throt presenttions (2). This report builds on previous work describing trinees prescribing of ntibiotics for upper respirtory trct infection nd cute bronchitis (16), which similrly showed non-evidence-bsed prescribing ptterns. GP trinees see more cute presenttions thn their more estblished collegues, including high rte of respirtory infections (17). Inpproprite ntibiotic prescribing by trinees in these conditions is therefore concern in its own right s contribution to high rtes of community ntibiotic use, s well s for the longer term effects on trinee prescribing ptterns. Penicillin is the recommended first-line ntibiotic for tretment of bcteril phryngitis (6) trinees re dhering to this choice in somewht more thn hlf the cses, nd fll short of qulity indictors for this condition (13). Our dt does not include individul ptients contextul informtion including penicillin llergy so we cnnot ssess the ppropriteness of prescribing decisions for other ntibiotics, for exmple ceflexin nd roxithromycin. However, the frequency of use of moxicillin, nd moxicillin plus enzyme inhibitor (together ccounting for 18.9% of ntibiotics prescribed) suggests tht penicillin llergy my not be mjor fctor in the inpproprite prescribing rte. The lck of significnt ssocition of prescribing with younger ptient ge group is notble. Though there were significnt differences on univrite nlyses, these were of smll effect size nd on multivrible nlyses the differences were no longer significnt. Streptococcl phryngitis is most prevlent in the ge group 5 15 yers (3). Tht those in this ge group did not receive significntly more prescriptions for ntibiotics thn those in younger (0 2 yers) nd older (15 50 yers) ge groups suggests tht trinees my not be considering the epidemiology of streptococcl s opposed to virl infections in their prescribing decisions. Another explntion for these high prescribing rtes could be ptient pressure. This is often cited in the literture s mjor driver for ntibiotic prescribing (9), in spite of evidence from multiple sources tht ptient demnd for ntibiotics might be relted to lck of informtion or concerns bout the course of their disese (18). Ptients from higher socio-economic res were more likely to be prescribed n ntibiotic, potentilly due to pressure plced on the doctor by more ffluent demogrphic. We found significnt ssocition of ntibiotic prescribing with prctice loction in n inner-regionl re nd non-significnt ssocition with rurl/remote loction (both compred with mjor city prctices). This my relte to the greter ccess to cre in mjor cities. Lesser ccess to follow-up consulttion my led trinee to prescribe more liberlly t the index consulttion. Concerns regrding ccess to medictions nd further medicl ttention re resons noted by trinees (10) nd experienced prctitioners for incresed likelihood to prescribe n ntibiotic. The ssocition of in-consulttion informtion nd dvice seeking with ntibiotic prescribing is seemingly nomlous. Authorittive Austrlin evidence-bsed guidelines, if ccessed, recommend restricted prescribing (6). In consulttions where n ntibiotic ws prescribed, 19.9% of trinees sought dvice. Of these, mny used electronic (71.5%) or hrdcopy (10.5%) resources [including Austrlin Therpeutic Guidelines (6)], nd 17.8% consulted supervisor. Trinees my seek dvice or informtion for more severely ill ptients (in whom ntibiotic prescription my be pproprite) nd this my contribute to the ssocition of informtion seeking with ntibiotic prescribing. Given the very high prescribing rtes nd use of electronic resources, however, we hypothesize tht trinees re consulting resources for dose checking without encting the cler recommendtions ginst ntibiotic prescribing for most cses of sore throt. In previous qulittive study, we found tht GP trinees cknowledge the trnsition to generl prctice from their initil trining in the hospitl to primry cre s shift from n environment of low-prevlence/high-morbidity infections nd (ppropritely) ggressive ntibiotic tretment to community environment of high-prevlence/low-morbidity infections tht often do not require ntibiotics (10). We would therefore expect tht trinees
306 Fmily Prctice, 2016, Vol. 33, No. 3 Tble 3. Associtions of prescribing of ntibiotics by GP trinees for presenttions of sore throt: univrite nlyses Vrible Clss Antibiotics prescribed No (n = 712) Yes (n = 1783) P Ptient fctors Ptient ge group 0 2 60 (8.5%) 149 (8.4%) 0.002 3 14 193 (27.5%) 566 (32.0%) 15 50 365 (51.9%) 923 (52.2%) 51+ 85 (12.1%) 130 (7.4%) Ptient gender Mle 256 (37.0%) 701 (40.2%) 0.103 Femle 435 (63.0%) 1042 (59.8%) Aboriginl or Torres Strit Islnder No 672 (99.1%) 1671 (98.9%) 0.650 Yes 6 (0.9%) 18 (1.1%) Non-English speking bckground No 645 (94.9%) 1623 (95.5%) 0.653 Yes 35 (5.1%) 76 (4.5%) Ptient/prctice sttus Existing ptient 230 (32.8%) 456 (26.1%) 0.0001 New ptient to trinee 436 (62.1%) 1132 (64.9%) New ptient to prctice 36 (5.1%) 156 (8.9%) Trinee fctors Trinee gender Mle 271 (38.1%) 655 (36.7%) 0.507 Femle 441 (61.9%) 1128 (63.3%) Trinee ge men (SD) 33.5 (7.3) 33.4 (7.1) 0.744 Employed full-time or prt-time Prt-time 160 (22.9%) 350 (20.0%) 0.215 Full-time 538 (77.1%) 1399 (80.0%) Trining term Term 1 310 (43.5%) 726 (40.7%) 0.092 Term 2 210 (29.5%) 547 (30.7%) Term 3 173 (24.3%) 421 (23.6%) Term 4 19 (2.7%) 89 (5.0%) Worked t the prctice previously No 521 (73.9%) 1289 (72.8%) 0.734 Yes 184 (26.1%) 481 (27.2%) Qulified s doctor in Austrli No 194 (27.8%) 478 (27.1%) 0.928 Yes 505 (72.2%) 1286 (72.9%) Prctice fctors Prctice size Smll 244 (34.9%) 554 (31.5%) 0.089 Lrge 456 (65.1%) 1206 (68.5%) Prctice routinely bulk bills No 586 (82.5%) 1447 (81.8%) 0.931 Yes 124 (17.5%) 321 (18.2%) Rurlity Mjor city 453 (63.6%) 1058 (59.3%) 0.078 Inner-regionl 172 (24.2%) 518 (29.1%) Outer regionl/remote 87 (12.2%) 207 (11.6%) SEIFA Index (decile) Men (SD) 5.7 (2.9) 6.0 (2.8) 0.052 Consulttion fctors Sought in-consulttion dvice No 649 (91.2%) 1429 (80.1%) <0.0001 Yes 63 (8.8%) 354 (19.9%) Pthology ordered No 642 (90.2%) 1586 (89.0%) 0.352 Yes 70 (9.8%) 197 (11.0%) Imging ordered No 711 (99.9%) 1783 (100%) 0.285 (exct) Yes 1 (0.1%) 0 (0%) Follow-up ordered No 501 (70.4%) 1174 (65.8%) 0.024 Yes 211 (29.6%) 609 (34.2%) Lerning gols No 637 (91.9%) 1601 (92.3%) 0.848 Yes 56 (8.1%) 133 (7.7%) Referrl ordered No 685 (96.2%) 1737 (97.4%) 0.103 Yes 27 (3.8%) 46 (2.6%) New problem No 132 (19.8%) 133 (8.2%) <0.0001 Yes 533 (80.2%) 1489 (91.8%) Number of problems Men (SD) 1.5 (0.8) 1.3 (0.6) <0.0001 Consulttion durtion (min) Men (SD) 14.5 (7.2) 14.0 (6.0) 0.137 SEIFA Socioeconomic Index for Are (SEIFA) Reltive Index of Disdvntge. in the lter stges of their GP trining would prescribe ntibiotics less frequently. However, our findings re contrry to this expecttion, which suggests tht the current pprenticeship model of GP trining (in plce in Austrli nd mny other countries) my not be (in the re of ntibiotic prescribing) dequtely supporting trinees through the trnsition to community prctice (19). Our finding, however, should be treted with some cution. Our nlysis is crosssectionl nd the lck of difference my reflect cohort effects rther thn lck of temporl chnges in trinees prescribing. Strengths nd limittions of the study To our knowledge, ReCEnT is the lrgest study of trinee GPs. The lrge smple size includes trinees from five trining regions cross five Austrlin sttes, locted in mjor city to very remote res. The profile of trinees smpled by ReCEnT is similr to tht of the popultion of GP trinees in Austrli. The high response rte 95.2% is n importnt strength. Our dt, however, do not llow judgement of the clinicl ppropriteness of individul prescribing decisions (20). We re lso
Antibiotic prescribing for sore throt by erly-creer doctors in fmily prctice 307 Tble 4. Associtions of prescribing of ntibiotics by trinees for presenttions of sore throt: multiple logistic regression nlysis Vrible (referent) OR (95% CI) P-vlue Ptient fctors Ptient ge (15 50) 0 2 0.95 (0.68, 1.34) 0.778 3 14 1.12 (0.89, 1.43) 0.336 51+ 0.66 (0.48, 0.90) 0.009 New ptient to trinee 1.00 (0.79, 1.25) 0.981 New ptient to prctice 1.57 (1.03, 2.39) 0.036 Trinee fctors Femle 0.92 (0.76, 1.13) 0.425 Trining term (1) 2 1.19 (0.90, 1.58) 0.232 3 1.05 (0.80, 1.39) 0.704 4 2.10 (1.08, 4.08) 0.029 Prctice fctors Rurlity (urbn) Inner-regionl 1.37 (1.05, 1.79) 0.019 Outer regionl/remote/very remote 1.32 (0.88, 1.99) 0.181 Prctice size (lrge) 1.13 (0.89, 1.44) 0.316 SEIFA Index decile 1.06 (1.01, 1.10) 0.018 Consulttion fctors Follow-up ordered 1.26 (1.00, 1.58) 0.053 Sought in-consulttion dvice/informtion 2.37 (1.65, 3.40) <0.0001 New problem 2.66 (1.99, 3.55) <0.0001 Number of problems 0.75 (0.64, 0.87) 0.0002 Consulttion durtion 1.00 (0.97, 1.02) 0.643 SEIFA Socioeconomic Index for Are (SEIFA) Reltive Index of Disdvntge. unble to determine whether prescriptions written were filled or consumed by ptients however, prescribing decisions mde by trinees precede subsequent ptient behviour. A further limittion is tht the lrgest burden of complictions of Group A streptococcl phryngitis is in the Austrlin Aboriginl nd Torres Strit Islnder popultion (6). Our nlysis is not powered to exmine the effect of Aboriginl nd Torres Strit Islnder sttus on trinees ntibiotic prescribing. Conclusions The very high frequency of ntibiotic prescribing nd the lck of ttenution in prescribing frequency in more senior terms suggest (despite the cvets outlined bove) tht both current eductionl interventions nd the pprenticeship model of trining re filing to foster pproprite clinicl prctice during the trnsition from hospitl to community prctice in this importnt clinicl re. Trgeted eductionl interventions, which my need to include supervisors s well s trinees, re indicted. Our results suggest tht eduction encourging ppliction of evidence, epidemiology nd guidelines to everydy clinicl prctice might be of benefit to erly-creer doctors in generl prctice. Declrtion Funding: Generl Prctice Eduction nd Trining Registrr Reserch Fund (grnt number 024/12 to AD) nd the University of Queenslnd. Ethicl pprovl: Humn Reserch Ethics Committee, University of Newcstle, Reference H-2009-0323. Conflict of interest: none. References 1. Britt H, Miller GC, Henderson J et l. A Decde of Austrlin Generl Prctice Activity 2004 05 to 2013 14. Sydney, New South Wles, Austrli: Sydney University Press, 2014. 2. Brnett ML, Linder JA. Antibiotic prescribing to dults with sore throt in the United Sttes, 1997 2010. JAMA Intern Med 2014; 174: 138 40. 3. Wessels MR. Streptococcl phryngitis. N Engl J Med 2011; 364: 648 55. 4. Dnchin MH, Rogers S, Kelpie L et l. Burden of cute sore throt nd group A streptococcl phryngitis in school-ged children nd their fmilies in Austrli. Peditrics 2007; 120: 950 7. 5. Del Mr C. Mnging sore throt: literture review. I. Mking the dignosis. Med J Aust 1992; 156: 572 5. 6. Acute Phryngitis nd/or Tonsillitis [Revised 2014 Mrch], in etg Complete. [Internet]. Melbourne, Austrli: Therpeutic Guidelines Limited, 2015. www.etg.tg.com.u (ccessed on August 2014). 7. Goossens H, Ferech M, Vnder Stichele R, Elseviers M, Group EP. Outptient ntibiotic use in Europe nd ssocition with resistnce: crossntionl dtbse study. Lncet 2005; 365: 579 87. 8. Spinks A, Glsziou PP, Del Mr CB. Antibiotics for sore throt. Cochrne Dtbse Syst Rev 2013; 11: CD000023. 9. Teixeir Rodrigues A, Roque F, Flco A, Figueirs A, Herdeiro MT. Understnding physicin ntibiotic prescribing behviour: systemtic review of qulittive studies. Int J Antimicrob Agents 2013; 41: 203 12. 10. Dlls A, vn Driel M, vn de Mortel T, Mgin P. Antibiotic prescribing for the future: exploring the ttitudes of trinees in generl prctice. Br J Gen Prct 2014; 64: e561 7. 11. Bjornsdottir I, Kristinsson KG, Hnsen EH. Dignosing infections: qulittive view on prescription decisions in generl prctice over time. Phrm World Sci 2010; 32: 805 14. 12. Morgn S, Mgin PJ, Henderson KM et l. Study protocol: the Registrr Clinicl Encounters in Trining (ReCEnT) study. BMC Fm Prct 2012; 13: 50. 13. Adrienssens N, Coenen S, Tonkin-Crine S et l. Europen Surveillnce of Antimicrobil Consumption (ESAC): disese-specific qulity indictors for outptient ntibiotic prescribing. BMJ Qul Sf 2011; 20: e772. 14. Ntionl Prescribing Service. Prescribing Prctice Review 9: Use of Antibiotics in Respirtory Trct Infection. Surry Hills, New South Wles, Austrli: Ntionl Prescribing Service, 2000. 15. Akkermn AE, Vn der Wouden JC, Kuyvenhoven M, Dielemn JP, Verheij TJM. Antibiotic prescribing for respirtory trct infections in Dutch primry cre in reltion to ptient ge nd clinicl entities. J Antimicrob Chemother 2004; 54: 1116 21.
308 Fmily Prctice, 2016, Vol. 33, No. 3 16. Dlls A, Mgin P, Morgn S et l. Antibiotic prescribing for respirtory infections: cross-sectionl nlysis of the ReCEnT study exploring the hbits of erly-creer doctors in primry cre. Fm Prct 2015; 32: 49 55. 17. de Jong J, Visser MRM. Opening the blck box: the ptient mix of GP trinees. Br J Gen Prct 2011; 61: e650 7. 18. Butler CC, Rollnick S, Pill R, Mggs-Rpport F, Stott N. Understnding the culture of prescribing: qulittive study of generl prctitioners nd ptients perceptions of ntibiotics for sore throts. Br Med J 1998; 317: 637 42. 19. Ajjwi R, Thistlethwite JE, Aslni P, Cooling NB. Wht re the perceived lerning needs of Austrlin generl prctice registrrs for qulity prescribing? BMC Med Educ 2010; 10: 92. 20. Dekker AR, Verheij TJ, vn der Velden AW. Inpproprite ntibiotic prescription for respirtory trct indictions: most prominent in dult ptients. Fm Prct 2015; 32: 401 7.