Family Practice, 2015, Vol. 32, No. 5, 525–532 doi:10.1093/fampra/cmv047 Advance Access publication 18 June 2015 Health Service Research GP trainees’ in-consultation informationseeking: associations with human, paper and electronic sources Downloaded from https://academic.oup.com/fampra/article-abstract/32/5/525/689617 by guest on 26 May 2020 Parker Magina,b,*, Simon Morganb, Susan Wearnec,d, Amanda Tapleyb, Kim Hendersonb, Chris Oldmeadowe,f, Jean Balle, John Scottb, Neil Spikeg,h, Lawrie McArthuri and Mieke van Drielj a Discipline of General Practice, University of Newcastle, Newbolds Building, University Drive, Callaghan, NSW 2308, General Practice Training Valley to Coast, Corner of Gavey and Frith Streets, Mayfield NSW 2304, cAcademic Unit of General Practice, Australian National University Medical School, Canberra, ACT 2606, dDepartment of Health, Health Workforce Division, Woden Town Centre, Canberra, ACT 2606, eClinical Research Design IT and Statistical Support Unit (CReDITSS), Hunter Medical Research Institute, Newcastle, fSchool of Medicine and Public Health, University of Newcastle University Drive, Callaghan, NSW 2308, gVictorian Metropolitan Alliance General Practice Training, 15 Cato Street, Hawthorn, VIC 3122, hDepartment of General Practice, University of Melbourne, 200 Berkeley Street, Carlton, VIC 3053, iAdelaide to Outback General Practice Training, Lower Level, 183 Melbourne Street, North Adelaide, SA 5006 and jDiscipline of General Practice, School of Medicine, The University of Queensland, L8 Health Sciences Building, Royal Brisbane and Women’s Hospital, Brisbane, Australia. b *Correspondence to Parker Magin, Discipline of General Practice, University of Newcastle, Newbolds Building, University Drive, Callaghan, NSW 2308, Australia; E-mail: parker.magin@newcastle.edu.au Abstract Background. Answering clinical questions arising from patient care can improve that care and offers an opportunity for adult learning. It is also a vital component in practising evidence-based medicine. GPs’ sources of in-consultation information can be human or non-human (either hard copy or electronic). Objectives. To establish the prevalence and associations of GP trainees’ in-consultation information-seeking, and to establish the prevalence of use of different sources of information (human, hard copy and electronic) and the associations of choosing particular sources. Methods. A cross-sectional analysis of data (2010–13) from an ongoing cohort study of Australian GP trainees’ consultations. Once each 6-month training term, trainees record detailed data of 60 consecutive consultations. The primary outcome was whether the trainee sought in-consultation information for a problem/diagnosis. Secondary outcomes were whether information-seeking was from a human (GP, other specialist or other health professional) or from a non-human source (electronic or hard copy), and whether a non-human source was electronic or hard copy. Results. Six hundred forty-five trainees (response rate 94.3%) contributed data for 84 723 consultations including 131 583 problems/diagnoses. In-consultation information was sought for 15.4% (95% confidence interval = 15.3–15.6) of problems/diagnoses. Sources were: GP in 6.9% of problems/diagnoses, other specialists 0.9%, other health professionals 0.6%, electronic sources 6.5% and hard-copy sources 1.5%. Associations of information-seeking included younger patient age, trainee full-time status and earlier training stage, longer consultation duration, referring the patient, organizing follow-up and generating learning goals. Associations of choosing human information sources (over non-human sources) were similar, but also included the trainee’s training © The Author 2015. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. 525 Family Practice, 2015, Vol. 32, No. 5 526 organization. Associations of electronic rather than hard-copy information-seeking included the trainee being younger, the training organization and information-seeking for management rather than diagnosis. Conclusion. Trainee information-seeking is mainly from GP colleagues and electronic sources. Human information-sources are preferentially sought for more complex problems, even by these early-career GPs who have trained in the ‘internet era’. Key words: Australia, evidence-based medicine, family practice, general practitioners, information seeking behaviour, internet, internship and residency, primary health care. Methods Answering clinical questions arising from patient care can improve that care and also offers an opportunity for learning (1,2). With clinical information overload, generating and answering clinical questions is increasingly important (2). The number of physicians’ clinical questions generated increased markedly during 2009–13 in a US study (3). Answering clinical questions is particularly important for generalist clinicians, and GPs generate more patient-specific questions than do other specialists (4). In-consultation ‘real-time’ answering of questions during patient encounters is practicable (5) (and acceptable to patients (6)). But clinical questions generated in-consultation by GPs and GP trainees may be more likely answered post-consultation than in-consultation (7,8). One study of GP trainees, however, found most questions were answered in-consultation (5). The frequency of in-consultation information-seeking of GPs has varied markedly between studies (7,9–11), but has been less than that of GP trainees (5,8). GPs’ sources of information in answering in-consultation questions are of particular importance. Human resources (other specialists, GP peers and GP supervisors/trainers) have been preferred over written formats by GPs (12), and the initial uptake of electronic sources was poor (9,13). But during 2009–13 the use of technological formats to answer physicians’ clinical questions increased more rapidly than traditional sources in the USA (3). Evidence for the associations of GPs’ information-seeking is sparse. A recent systematic review has identified a need to research such associations to inform development of information-retrieval systems personalized to the clinician’s characteristics (1). It also identified evidence gaps regarding information-seeking behaviours pertaining to patient subpopulations of interest such as complex and aging patients (1). The aim of this analysis was to establish the prevalence of information-seeking in general practice consultations. We also sought to establish the sources of information accessed and the purpose of that information-seeking (to inform diagnosis or management). In response to identified evidence gaps, we aimed to establish clinician, patient and practice associations of choice of information-source. We have previously specifically examined GP trainees’ informationseeking from their supervisor (trainer/preceptor) (14). In the current analysis we sought to establish the prevalence and associations of trainees’ information-seeking from all sources. We also sought to establish the prevalence and associations of use of different information-sources (human, electronic and hard copy). Our analysis focusses on GP trainees, for whom ‘acquiring information’ is a key component of their vocational training in evidencebased-medicine (2). Furthermore, familiarity of use from training and early clinical practice are dominant influences on later resource use by established GPs (12). This analysis took place within the Registrar Clinical Encounters in Training (ReCEnT) project. ReCEnT is an ongoing multi-site cohort study of GP registrars (trainees). Participants included in this analysis are GP trainees of four of Australia’s 17 GP Regional Training Providers (RTPs) across four Australian states. RTPs are government-funded, not-for-profit, geographically-defined GP vocational training organizations. The methodology has been described elsewhere (15). GP trainees collect data once during each 6-month general practice training term (12 months for part-time trainees). This results in data collection on three or four occasions during training. Trainee demographics and characteristics of practices in which they are working are recorded by each trainee, each term. Trainees then record, each term, the details of 60 consecutive clinical consultations contemporaneously on a paper-based encounter form. Data collection is designed to reflect a ‘normal’ week of general practice. Thus, consultations in a specialized clinic, e.g. vaccination clinic are excluded. Only office-based (not home visits or nursing home visits) consultations are included. Outcome factors The primary outcome factor was whether the trainee sought ‘realtime’ in-consultation information during the consultation. This information was recorded as being sought from another GP (including the trainee’s supervisor), a non-GP specialist, another health professional, electronic sources or hard-copy sources. Secondary outcomes were: 1. Whether information sought was from a human source (GP, other specialist or other health professional) or from a nonhuman source (electronic or hard copy) 2. Whether a non-human source was electronic or hard copy Whether in-consultation information was sought for diagnosis, for management, or for both was also recorded as was the problem/ diagnosis for which information was sought. Problems/diagnoses were coded according to the International Classification of Primary Care, second edition (ICPC-2 plus) (16). ICPC-2plus classifies problems/diagnoses into 17 Chapters—16 body system/discipline-specific Chapters and a ‘General and Unspecified’ Chapter. Independent variables Independent variables related to trainee, patient, practice and consultation. Trainee factors were age, gender, training term, place of medical qualification (Australia/international), full-time/part-time status, Downloaded from https://academic.oup.com/fampra/article-abstract/32/5/525/689617 by guest on 26 May 2020 Introduction 527 GPs’ in-consultation information-seeking Statistical analysis This was a cross-sectional analysis of consultations from eight rounds of the longitudinal ReCEnT study, 2010–13. The unit of analysis was the individual problem/diagnosis rather than the consultation, as it is for individual problems that information is sought. The proportion of problems for which trainees sought in-consultation information was calculated with 95% confidence intervals (CIs). The proportions of problems for which trainees sought in-consultation information from a supervisor, a non-GP specialist, another health professional, an electronic source, or a hard-copy source were calculated with 95% CIs. Of those problems for which the trainee sought information, the proportions for which information was for diagnosis, management, or both diagnosis and management, were calculated with 95% CIs. To test associations of trainees seeking in-consultation information, simple and multiple logistic regression were used within a generalized estimating equations (GEE) framework to account for clustering of patients within trainees. All variables with a P-value less than 0.2 and relevant effect size in the univariate analysis were included in the multiple regression models. Further logistic regression models within GEEs were constructed with the outcome factor ‘human or non-human source of information’ (for problems/diagnoses for which information was sought) and with the outcome factor ‘electronic or hard-copy source of information’ (for problems/diagnoses for which information was sought from non-human sources). Data for problems/diagnoses for which information was sought from more than one category (human and non-human, electronic and hard copy, respectively) were excluded from these further analyses. Statistical analyses used STATA 13.1 and SAS v9.4. P-values < 0.05 were considered statistically significant. Results A total of 645 trainees (response rate 94.3%) contributed data for 1426 trainee-rounds, and 84 723 consultations including 131 583 problems/diagnoses. Table 1 displays trainee and practice characteristics. Prevalence and types of information-seeking In-consultation information was sought for 20 328 [15.4% (95% CI = 15.3–15.6)] problems/diagnoses [in 17 440 (20.6% (95% Table 1. Characteristics of GP trainees and practices of four Regional Training Providers participating in ReCEnT study data collection, 2010–13 Variable Trainee variables (n = 645) Trainee gender Qualified as a doctor in Australia Trainee age (years) Trainee-term or practice-term variables (n = 1426) Trainee training term Trainee works fulltime Trainee worked at the practice previously Does the practice routinely bulk bill Number of GPs working at the practice Rurality of practice SEIFAa index (decile) of practice a Socio-economic Index for Area Relative Index of Disadvantage. Class n, % (95% CIs) or mean (SD) Male Female No Yes Mean (SD) 220, 34.1% (30.4–37.8) 425, 65.9% (62.2–69.6) 155, 24.4% (21.1–27.8) 480, 75.6% (72.2–78.9) 32.8 (6.6) Term 1 Term 2 Term 3 Term 4 No Yes No Yes No Yes 1–5 ≥6 Major city Inner regional Outer regional or remote Mean (SD) 557, 39.1% (36.5–41.6) 488, 34.2% (31.8–36.7) 306, 21.5% (19.3–23.6) 75, 5.3% (4.1–6.4) 302, 21.7% (19.5–23.8) 1091, 78.3% (16.2–80.5) 994, 70.6% (68.3–73.0) 413, 29.4% (27.0–31.7) 1179, 83.4% (81.5–85.4) 234, 16.6% (14.6–18.5) 454, 32.5% (30.1–35.0) 941, 67.5% (65.0–69.9) 827, 58.0% (55.4–60.6) 424, 29.7% (27.4–32.1) 175, 12.3% (10.6–14.0) 5.4 (2.8) Downloaded from https://academic.oup.com/fampra/article-abstract/32/5/525/689617 by guest on 26 May 2020 whether the trainee had previously worked at the practice or had prior health qualifications or post-graduate qualifications, and the trainee’s RTP. Patient factors were age, gender, Aboriginal or Torres Strait Islander status, non-English speaking background, whether the patient was new to the practice, and new to the trainee. Practice factors were rurality/urbanicity, practice size (fulltime equivalent GPs), if the practice had internet access at the consultation desk, and if the practice routinely bulk-bills (that is, there is no financial cost to the patient for the consultation). Practice postcode was used to define the Australian Standard Geographical Classification-Remoteness Area classification (the degree of rurality) of the practice location (17) and to define the practice location’s Socioeconomic Index for Area Relative Index of Disadvantage (18). Consultation factors were whether the problem/diagnosis was a chronic disease (classified according to the methodology of O’Halloran et al. (19)), whether the problem/diagnosis was new, and the number of problems addressed in the consultation. Other consultation variables were consultation duration, whether imaging was ordered, pathology was ordered, referrals were made, follow-up was organized and learning goals were generated. 528 CIs = 20.3–20.8)) consultations]. The sources of information were: GP in 6.9% (95% CI = 6.8–7.1) of problems/diagnoses, other specialists 0.9% (95% CI = 0.9–1.0), other health professionals 0.6% (95% CI = 0.5–0.6), electronic sources 6.5% (95% CI = 6.4–6.6), hard-copy sources 1.5% (95% CI = 1.4–1.5) and other sources 0.7% (95% CI = 0.7–0.8). It should be noted that trainees could seek information from more than one source for a problem. Information was sought in 9.9% (95% CI = 9.5–10.4) of instances for diagnosis, in 61.3% (95% CI = 60.6–62.0) for management, and in 28.8% (95% CI = 28.1–29.5) for both diagnosis and management. The most common ICPC-2 chapters for which information was sought were Skin (16.0%), General and Unspecified (12.7%), Musculoskeletal (10.8%) and Respiratory (10.7%). The relationships of information-seeking to independent variables are presented in Table 2. The logistic regression model with outcome factor ‘Sought any in-consultation information’ is presented in Table 3. Significant patient associations of information-seeking were younger age and male gender, and not being new to the practice or new to the trainee. Trainee associations were younger age, full-time status, and earlier training stage. The practice-level association was not having internet at the desk. Consultation-level associations were the problem/ diagnosis being new, longer consultation duration (though a lesser number of problems were dealt with), referring the patient, organizing follow-up and generating learning goals. Associations of information-seeking: human versus non-human sources The logistic regression with outcome factor ‘Sought in-consultation information from a human source’ (when information was sought) is presented in Table 4. Significant patient associations of information-seeking from a human (rather than non-human) source were age (younger or older age range than 15–34), male gender and the patient not being new to the practice or new to the trainee. Trainee associations were earlier stage of training and the RTP trained with. Consultation associations included the problem/diagnosis not being new, longer consultation duration (though a lesser number of problems were dealt with in the consultation), referring the patient, organizing follow-up, and generating learning goals. The information or advice being for diagnosis rather than management was also associated with a human source of information. Associations of information-seeking: electronic versus hard-copy sources The logistic regression with outcome factor ‘Sought in-consultation electronic information’ (when advice was sought from a non-human source) is presented in Table 5. Significant associations of seeking electronic information included the trainee being younger, the RTP trained with, and information being sought for management rather than diagnosis. Discussion Main findings and comparison with previous literature Information-seeking was common in our study: for 15.4% of all problems/diagnoses and in 20.6% of consultations. Comparisons can be made with other studies that have elicited in-consultation information-seeking contemporaneously. Our methodology is not directly comparable, but it can be established from previous papers that US and Australian GPs sought answers to questions at 6.8 to 16.2 answers per 100 consultations (9–11) but that Spanish GPs sought only 1.7 answers per 100 consultations (7). In studies of US family medicine trainees, 20.7 (8) and 113.7 answers (5) were sought per 100 consultations. The variability in prevalence may reflect differences in practice context (for example, the comparatively short Spanish consultation duration) (7) as well as differences in experience between trainees and established GPs. The prevalence of information-seeking in our study was greater than in studies of established GPs. This is consistent with the relative clinical inexperience of our trainee participants and the proposed central importance of information-seeking skills in GP training (2). In our study, information was sought in 9.9% of instances for diagnosis, in 61.3% for management, and in 28.8% for both diagnosis and management. In a Spanish study (with short consultation-time and few in-consultation questions), of GPs’ clinical questions generated, 53% were for diagnosis and 33% were for treatment or management (7). But in a more comparable study of US GPs, 73% of questions were for treatment and 27% for diagnosis (9). The emphasis on management compared to diagnosis in our study may reflect information-seeking on management often being relatively straightforward (and thus easily addressed in-consultation) once a diagnosis is established. Information regarding diagnosis, often a more complex cognitive task than selection of treatment, may be less satisfactorily addressed in ‘real-time’, inconsultation. Out-of-consultation information-seeking (not measured in this study) may often be appropriate. It should also be noted that we found generation of learning goals to be strongly associated with in-consultation information-seeking—suggesting a more complex learning process as well as utilitarian dealing with a problem in the consultation. Dermatology and musculoskeletal were the most common ICPC-2 chapters for information-seeking. This may reflect the mismatch of limited hospital-based training exposure but high general practice prevalence of dermatological and musculoskeletal problems in Australia. In Canadian GP training, these were only the fifth and sixth most common disciplines prompting information-seeking (20). Regarding source of information, for 8.4% of problems/diagnoses it was a human source, for 6.5% an electronic source, and for 1.5% a hard-copy source. A study of Canadian GP trainees (20) found that in 65.5% of instances information-seeking was from human sources (mostly the trainee’s supervisor), 20.7% from electronic sources and 14.0% from hard-copy sources. In a US study of family medicine trainees 47.4% of questions were answered by human sources, 25.9% by electronic sources and 22.9% by hardcopy sources (8). Human sources were also deemed to be of more clinical utility than electronic or hard-copy sources in a Canadian trainee study (20). There are no comparable studies of sources of information of established GPs to provide context for these seemingly high proportions of human sources. It should be noted that trainees are inexperienced in in-consultation informationsearching as well as in clinical practice and will usually have ready recourse to advice and information from a supervisor/trainer/preceptor. But in an Australian qualitative study, experienced GPs as well as trainees highly valued human resources—though for GPs it was non-GP specialist colleagues and for trainees it was their supervisor (12). Downloaded from https://academic.oup.com/fampra/article-abstract/32/5/525/689617 by guest on 26 May 2020 Associations of information-seeking: all sources Family Practice, 2015, Vol. 32, No. 5 529 GPs’ in-consultation information-seeking Table 2. Associations of seeking any in-consultation information with all independent variables Sought assistance Variable Class No (n = 111 255) Yes (n = 20 328) P Patient age group 0–14 15–34 35–64 65+ Male Female No Yes No Yes Male Female Part time Full time Term1 Term2 Term3 Term4 No Yes No Yes No Yes No Yes Small (<6 GPs) Large (>5 GPs) No Yes Major city Inner regional Outer regional, remote and very remote 1 2 3 4 No Yes Patient known to trainee New to trainee New to Practice No Yes No Yes No Yes No Yes No Yes No Yes No Yes Mean (SD) Mean (SD) Mean (SD) Mean (SD) 14 254 (81%) 27 857 (84%) 45 607 (86%) 21 789 (86%) 40 240 (84%) 68 182 (85%) 104 390 (85%) 1205 (82%) 99 734 (85%) 6547 (82%) 37 630 (86%) 73 625 (84%) 24 058 (85%) 84 713 (84%) 41 759 (79%) 38 389 (87%) 25 000 (89%) 6107 (91%) 76 564 (83%) 33 236 (88%) 25 434 (81%) 84 092 (86%) 97 698 (85%) 12 320 (85%) 83 029 (85%) 27 381 (83%) 35 949 (84%) 72 896 (85%) 92 093 (85%) 18 234 (84%) 65 060 (85%) 32 201 (83%) 13 994 (86%) 40 420 (85%) 13 416 (87%) 10 930 (81%) 46 489 (84%) 2680 (87%) 108 575 (84%) 48 700 (85%) 52 200 (84%) 7200 (82%) 56 383 (86%) 54 872 (83%) 85 992 (85%) 24 900 (84%) 92 845 (85%) 18 410 (82%) 103 767 (85%) 7488 (77%) 64 313 (88%) 46 942 (80%) 101 309 (90%) 9946 (53%) 99 422 (86%) 11 833 (75%) 32.9 (6.7) 5.4 (2.8) 17.9 (9.4) 2.0 (1.0) 3372 (19%) 5248 (16%) 7685 (14%) 3625 (14%) 7675 (16%) 12 113 (15%) 19 145 (15%) 265 (18%) 18 081 (15%) 1423 (18%) 6264 (14%) 14 064 (16%) 4299 (15%) 15 552 (16%) 10 845 (21%) 5850 (13%) 3060 (11%) 573 (8.6%) 15 362 (17%) 4681 (12%) 6004 (19%) 13 968 (14%) 17 880 (15%) 2245 (15%) 14 570 (15%) 5491 (17%) 6733 (16%) 13 279 (15%) 16 650 (15%) 3537 (16%) 11 436 (15%) 6523 (17%) 2369 (14%) 7130 (15%) 2031 (13%) 2504 (19%) 8663 (16%) 410 (13%) 19 918 (16%) 8652 (15%) 9582 (16%) 1558 (18%) 8864 (14%) 11 464 (17%) 15 611 (15%) 4682 (16%) 16 308 (15%) 4020 (18%) 18 129 (15%) 2199 (23%) 8697 (12%) 11 631 (20%) 11 625 (10%) 8703 (47%) 16 310 (14%) 4018 (25%) 33.5 (6.9) 5.2 (2.8) 21.7 (10.9) 1.8 (0.9) <0.0001 Patient gender ATSI NESB Trainee full-time or part-time Training term/post Worked at practice previously Qualified as doctor in Australia Previous health qualifications Post grad medical qualifications Practice size Practice routinely bulk bills Rurality RTP Internet at desk Patient/practice status New problem Chronic condition Pathology ordered Imaging ordered Follow-up ordered Learning goals generated Referral ordered Trainee age SEIFA index Consultation duration Number of problems 0.2740 0.3332 0.0948 0.0198 <0.0001 <0.0001 0.0003 0.8453 0.3352 0.4464 0.2679 0.0063 0.0194 0.0918 0.0100 <0.0001 0.0306 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 0.0012 0.8322 <0.0001 <0.0001 SEIFA index, socio-economic index for area relative disadvantage; ATSI, Aboriginal and Torres Strait Islander; NESB, Non-English speaking background. Downloaded from https://academic.oup.com/fampra/article-abstract/32/5/525/689617 by guest on 26 May 2020 Trainee gender <0.0001 Family Practice, 2015, Vol. 32, No. 5 530 Table 3. Results of logistic regression with outcome ‘any in-consultation information sought’ Univariate Adjusted Class OR (95% CI) P OR (95% CI) P Patient age group Referent 15–34 0–14 35–64 65+ Female Female Part time Term2 Term3 Term4 Yes Yes Inner regional Outer regional/ remote/ very remote 2 3 4 Yes New to practice New to trainee Yes Yes Yes Yes Yes 1.22 (1.16–1.29) 0.91 (0.87–0.94) 0.88 (0.83–0.94) 0.93 (0.90–0.96) 1.16 (0.97–1.38) 0.84 (0.72–0.97) 0.63 (0.58–0.69) 0.52 (0.46–0.59) 0.39 (0.30–0.51) 0.70 (0.65–0.77) 0.68 (0.55–0.84) 1.22 (1.07–1.39) 1.26 (1.02–1.55) 0.80 (0.62–1.03) 1.30 (1.00–1.70) 1.04 (0.87–1.25) 1.59 (0.93–2.72) 1.10 (1.02–1.19) 0.98 (0.94–1.03) 1.30 (1.25–1.35) 1.19 (1.14–1.24) 1.65 (1.57–1.72) 6.68 (6.11–7.31) 1.91 (1.81–2.02) 0.97 (0.95–0.99) 1.03 (1.03–1.04) 0.77 (0.75–0.80) <0.0001 <0.0001 <0.0001 <0.0001 0.0948 0.0198 <0.0001 <0.0001 <0.0001 <0.0001 0.0003 0.0032 0.0344 0.0869 0.0519 0.6406 0.0918 0.0165 0.4374 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 0.0012 <0.0001 <0.0001 1.21 (1.14–1.29) 0.94 (0.90–0.99) 1.02 (0.95–1.08) 0.95 (0.92–0.99) 1.06 (0.89–1.26) 0.86 (0.75–0.98) 0.67 (0.60–0.75) 0.61 (0.53–0.70) 0.44 (0.35–0.56) 0.97 (0.87–1.09) 0.95 (0.76–1.19) 1.03 (0.92–1.17) 0.89 (0.70–1.14) 0.94 (0.69–1.28) 1.26 (0.95–1.68) 1.21 (1.00–1.47) 0.76 (0.62–0.94) 0.86 (0.79–0.94) 0.95 (0.91–1.00) 1.21 (1.17–1.26) 0.95 (0.90–1.00) 1.27 (1.20–1.34) 5.54 (5.07–6.05) 1.33 (1.25–1.41) 0.98 (0.97–1.00) 1.04 (1.04–1.04) 0.72 (0.69–0.75) <0.0001 0.0146 0.6127 0.0112 0.4850 0.0223 <0.0001 <0.0001 <0.0001 0.6419 0.6646 0.5839 0.3578 0.6775 0.1071 0.0505 0.0093 0.0005 0.0296 <0.0001 0.0351 <0.0001 <0.0001 <0.0001 0.0179 <0.0001 <0.0001 Patient gender Trainee gender Trainee full time or part time Training term/post Referent term 1 Worked at practice previously Qualified as doctor in Australia Rurality RTP Referent RTP 1 Internet at desk Patient/practice status Referent patient known to trainee New problem Pathology ordered Follow-up ordered Learning goals Referral ordered Trainee age (years) Consultation duration Number of problems OR, odds ratio. Associations of information-seeking Regarding associations of information-seeking, there is little literature with which to compare our results. Ely et al., however, contrary to our results, found patient female gender and older age to be associated with information-seeking. Theirs’ was a study of established US GPs’ and these associations were with question-asking rather than information-seeking (10). While we found no association with rurality, urban GPs sought more information than rural GPs in another US study (9). Our finding that full-time status was associated with more in-consultation information-seeking may possibly indicate more out-of-consultation information-seeking by part-time trainees (who would possibly have more opportunity for this). The association of information-seeking with not having internet access at the consultation desk is surprising, but we also found that having internet access isn’t significantly associated with use of electronic versus hard-copy resources. Most trainees had internet access, so the clinical significance is limited. But these results may reflect many electronic resources used by trainees being incorporated in their general practice-specific computer software rather than accessed online. Associations with consultation duration, organizing follow-up, making referrals and fewer other problems being dealt with in the consultation likely reflect the intrinsic difficulty or complexity of the problem (or the difficulty the particular trainee has with it). As also does the generation of learning goals for further information-seeking out-of-consultation. We found that seeking information from a human rather than nonhuman source was associated with both younger and older patient agegroups (than the referent: 15–34 years). It may be that young adult patients in general present less complex clinical challenges than paediatric or older adults and that these challenges can be simply addressed via books or electronic sources rather than calling on the expertise of supervisors or other specialists. Similarly, associations with consultation duration, organizing follow-up, making referrals, generating learning goals and fewer other problems being dealt with in the consultation likely reflect the intrinsic difficulty or complexity of the problem. The association of information-seeking from a person with the problem involving diagnosis rather than treatment would, again, be consistent with information-seeking for more complex clinical challenges being directed to human rather than non-human sources. But as the trainee becomes more senior, they are better able to cope with complex clinical scenarios. Consistent with this proposition, we found decreasing reliance on human sources and increasing ‘self-directed’ informationseeking from books or electronic sources in the association of information-seeking from human sources with later stages of training. Information-seeking from a human source being associated with the problem being old and the patient known to the practice may reflect preferential information-seeking from another GP in the practice who had been involved in the patient’s previous care and, so, was familiar with the patient and their problems. The difference in information-seeking from a human source that we found between RTPs suggests structural differences in supervisory arrangements, or differences in supervisory ‘culture’ between RTPs. Regarding information-seeking from an electronic rather than hard-copy source, the association with younger trainee age is expected. The association with treatment rather than diagnosis may reflect treatment being more likely to change with time (thus requiring more up-to-date electronic sources), with the principles of diagnosis being more stable and less rapidly changing than treatment. The reason for the difference between RTP is not obvious, but may reflect resourcing or training in use of electronic resources. Downloaded from https://academic.oup.com/fampra/article-abstract/32/5/525/689617 by guest on 26 May 2020 Variable 531 GPs’ in-consultation information-seeking Table 4. Results of logistic regression with outcome ‘information from a human source’ (when in-consultation information sought) Univariate Adjusted Class OR (95% CI) P OR (95% CI) P Patient age group Referent 15–34 0–14 35–64 65+ Female Yes 1.08 (0.99–1.19) 1.21 (1.13–1.30) 1.52 (1.37–1.69) 0.89 (0.84–0.95) 0.92 (0.81–1.04) 0.0901 <0.0001 <0.0001 0.0004 0.1946 1.24 (1.11–1.40) 1.15 (1.04–1.27) 1.57 (1.38–1.79) 0.88 (0.82–0.96) 0.91 (0.77–1.06) 0.0003 0.0046 <0.0001 0.0020 0.2300 Term2 Term3 Term4 Yes 0.73 (0.65–0.81) 0.51 (0.44–0.60) 0.47 (0.36–0.61) 0.77 (0.67–0.87) <0.0001 <0.0001 <0.0001 <0.0001 0.69 (0.59–0.81) 0.55 (0.45–0.67) 0.38 (0.25–0.57) 0.98 (0.81–1.17) <0.0001 <0.0001 <0.0001 0.7898 Yes 1.16 (0.95–1.42) 0.1462 1.19 (0.91–1.55) 0.2011 2 4 5 New to practice New to trainee 1.50 (1.15–1.95) 1.18 (0.89–1.56) 1.60 (1.33–1.93) 0.87 (0.77–0.98) 0.78 (0.73–0.84) 0.0025 0.2419 <0.0001 0.0257 <0.0001 1.62 (1.14–2.30) 1.02 (0.74–1.39) 1.48 (1.16–1.88) 0.75 (0.65–0.88) 0.78 (0.72–0.85) 0.0070 0.9240 0.0013 0.0004 <0.0001 Yes Yes Yes Yes Yes Diagnosis 0.84 (0.79–0.90) 1.80 (1.62–2.00) 1.15 (1.08–1.23) 1.55 (1.44–1.68) 2.39 (2.17–2.63) 1.75 (1.52–2.01) <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 0.81 (0.75–0.87) 1.46 (1.29–1.66) 1.05 (0.97–1.13) 1.28 (1.16–1.41) 2.02 (1.78–2.28) 1.99 (1.70–2.32) <0.0001 <0.0001 0.2462 <0.0001 <0.0001 <0.0001 Diagnosis & management 2.11 (1.86–2.39) 0.97 (0.96–0.99) 1.04 (1.03–1.04) 0.96 (0.92–1.00) <0.0001 0.0002 <0.0001 0.0538 2.10 (1.84–2.40) 1.00 (0.98–1.01) 1.04 (1.03–1.04) 0.85 (0.81–0.90) <0.0001 0.6786 <0.0001 <0.0001 Patient gender Non-English speaking background Training term/post Referent term 1 Worked at practice previously Qualified as doctor in Australia RTP Referent RTP 1 Patient/practice status Referent patient known to trainee New problem Imaging ordered Follow-up ordered Learning goals generated Referral ordered Sought help for diagnosis/management Referent management Trainee age (years) Consultation duration Number of problems OR, odds ratio. Table 5. Results of logistic regression with outcome ‘information from an electronic source’ (when in-consultation information sought from a non-human source) Univariate Adjusted Variable Class OR (95% CI) P OR (95% CI) P Patient age group Referent 15–34 0–14 35–64 65+ Yes Yes Inner regional Outer regional/ remote/ very remote 2 3 4 Yes Yes Yes Diagnosis 0.87 (0.73–1.04) 1.01 (0.88–1.16) 1.27 (1.05–1.53) 1.65 (1.19–2.30) 0.54 (0.35–0.84) 0.56 (0.37–0.84) 1.23 (0.26–5.78) 0.1369 0.9160 0.0150 0.0027 0.0062 0.0048 0.7965 0.83 (0.69–1.01) 0.98 (0.84–1.15) 1.15 (0.95–1.40) 1.17 (0.81–1.71) 0.55 (0.35–0.87) 0.85 (0.59–1.22) 2.02 (0.56–7.34) 0.0644 0.8404 0.1459 0.4052 0.0098 0.3744 0.2862 1.27 (0.80–2.04) 1.36 (0.86–2.15) 4.26 (3.02–6.00) 1.24 (1.07–1.44) 0.75 (0.61–0.92) 0.77 (0.59–1.01) 0.66 (0.52–0.84) 0.3131 0.1923 <.0001 0.0034 0.0069 0.0545 0.0006 0.84 (0.40–1.76) 1.05 (0.54–2.03) 2.70 (1.82–3.99) 1.14 (0.97–1.35) 0.73 (0.57–0.93) 0.84 (0.67–1.06) 0.65 (0.50–0.84) 0.6517 0.8956 <.0001 0.1126 0.0126 0.1367 0.0012 Diagnosis & management 0.78 (0.63–0.97) 0.0223 0.80 (0.64–1.00) 0.0477 0.94 (0.92–0.96) 1.10 (1.01–1.20) <.0001 0.0297 0.96 (0.94–0.99) 1.06 (0.96–1.17) 0.0026 0.2720 Qualified as doctor in Australia Previous health qualifications Rurality Referent major city RTP Referent RTP1 Chronic condition Imaging ordered Learning goals Sought help for diagnosis and/ or management Referent management Trainee age (years) Number of problems OR, odds ratio. Downloaded from https://academic.oup.com/fampra/article-abstract/32/5/525/689617 by guest on 26 May 2020 Variable Family Practice, 2015, Vol. 32, No. 5 532 References The sample size and response rate are strengths of the study as is the contemporaneous recording of information-seeking. The large number of independent variables recorded allows a detailed exploration of associations of information-seeking. Limitations are our lack of data on whether the information sought adequately answered the in-consultation question or contributed to patient care. Our large sample size also raises the issue of clinical as opposed to statistical significance, and the effect sizes of particular associations we have found should be considered when assessing their impact. 1. Del Fiol G, Workman TE, Gorman PN. Clinical questions raised by clinicians at the point of care: a systematic review. JAMA Intern Med 2014; 174: 710–8. 2. Phillips R, Glasziou P. Evidence based practice: the practicalities of keeping abreast of clinical evidence while in training. Postgrad Med J 2008; 84: 450–3. 3. Salinas GD. Trends in physician preferences for and use of sources of medical information in response to questions arising at the point of care: 2009–2013. J Contin Educ Health Prof 2014; 34 (suppl 1):S11–6. 4. Bennett NL, Casebeer LL, Kristofco R, Collins BC. Family physicians’ information seeking behaviors: a survey comparison with other specialties. BMC Med Inform Decis Mak 2005; 5: 9. 5. Ramos K, Linscheid R, Schafer S. Real-time information-seeking behavior of residency physicians. Fam Med 2003; 35: 257–60. 6. Kahane S, Stutz E, Aliarzadeh B. Must we appear to be all-knowing?: patients’ and family physicians’ perspectives on information seeking during consultations. Can Fam Physician 2011; 57: e228–36. 7. González-González AI, Dawes M, Sánchez-Mateos J et al. Information needs and information-seeking behavior of primary care physicians. Ann Fam Med 2007; 5: 345–52. 8. McCord G, Smucker WD, Selius BA et al. Answering questions at the point of care: do residents practice EBM or manage information sources? Acad Med 2007; 82: 298–303. 9. Ely JW, Burch RJ, Vinson DC. The information needs of family physicians: case-specific clinical questions. BMJ 1992; 35: 265–9. 10. Ely JW, Osheroff JA, Ebell MH et al. Analysis of questions asked by family doctors regarding patient care. BMJ 1999; 319: 358–61. 11. Barrie AR, Ward AM. Questioning behaviour in general practice: a pragmatic study. BMJ 1997; 315: 1512–5. 12. Robertson J, Moxey AJ, Newby DA, Gillies MB, Williamson M, Pearson SA. Electronic information and clinical decision support for prescribing: state of play in Australian general practice. Fam Pract 2011; 28: 93–101. 13. Coumou HC, Meijman FJ. How do primary care physicians seek answers to clinical questions? A literature review. J Med Libr Assoc 2006; 94: 55– 60. 14. Morgan S, Wearne S, Tapley A, et al. In-consultation information and advice seeking by Australian GP trainees from GP trainers—a cross sectional analysis. Educ Prim Care (in press) 2014. 15. Morgan S, Magin PJ, Henderson KM et al. Study protocol: the Registrar Clinical Encounters in Training (ReCEnT) study. BMC Fam Pract 2012; 13: 50. 16. Britt H. A new coding tool for computerised clinical systems in primary care—ICPC plus. Aust Fam Physician 1997; 26 (suppl 2): S79–82. 17. Australian Bureau of Statistics. Australian Standard Geographical Classification (ASGC)—2006 http://www.abs.gov.au/AUSSTATS/abs@.nsf/Latest products/1AE106C101420508CA2571A900170741 (accessed on 3 January 2015). 18. Australian Bureau of Statistics. 2039.0—Information Paper: An Introduction to Socio-economic Indexes of Areas (SEIFA), 2006. http://www.abs. gov.au/ausstats/abs@.nsf/mf/2039.0/ (accessed on 3 January 2015). 19. O’Halloran J, Miller GC, Britt H. Defining chronic conditions for primary care with ICPC-2. Fam Pract 2004; 21: 381–6. 20. Allan GM, Ma V, Aaron S, Vandermeer B, Manca D, Korownyk C. Residents’ clinical questions: how are they answered and are the answers helpful? Can Fam Physician 2012; 58: e344–51. Conclusion We found that trainee GPs access information within more than onefifth of consultations. They mainly seek information from GP colleagues and electronic sources. Human sources of information seem to be preferentially sought for more complex problems, even by these early-career GPs who have trained in the ‘internet era’. Implications for practice structure and workload are in the time-demands on the colleagues consulted and in provision of adequate and appropriate resources to meet information-seeking needs. More broadly, the patterns and associations of information-seeking of these earlycareer GPs may indicate future resource needs in the evolving area of GP information-seeking for patient care. For example, electronic resources, especially those providing information on up-to-date therapy and information on particular topics (in Australia, for example, perhaps especially dermatology and musculoskeletal medicine). But even as the future composition of GP practitioners increasingly reflects the facility with information technology of our trainee study population, the complexity of practice is likely to increase (with, for example, greater complex multimorbidity). Our findings suggest access to expert human information sources must still be available for these complex problems, irrespective of technological advances in electronic information-sources. As our study was confined to GP trainees, future research including established GPs may explore the applicability of our findings to a clinically more experienced (but possibly less information technology-literate) practitioner population. Acknowledgements SW’s views are not necessarily those of the Department of Health. Declarations Funding: 2014 General Practice Education and Training (GPET) research grant (D14/11354). The participating educational organizations: General Practice Training Valley to Coast, the Victorian Metropolitan Alliance, General Practice Training Tasmania and Adelaide to Outback GP Training Program are funded by the Australian Government. Ethical approval: University of Newcastle Human Research Ethics Committee, Reference H-2009-0323. Conflict of interest: none. Downloaded from https://academic.oup.com/fampra/article-abstract/32/5/525/689617 by guest on 26 May 2020 Strengths and limitations