UBC Faculty Research and Publications

The implications of the feminization of the primary care physician workforce on service supply: a systematic… Hedden, Lindsay; Barer, Morris L; Cardiff, Karen; McGrail, Kimberlyn M; Law, Michael R; Bourgeault, Ivy L Jun 4, 2014

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-12960_2013_Article_427.pdf [ 306.92kB ]
JSON: 52383-1.0224016.json
JSON-LD: 52383-1.0224016-ld.json
RDF/XML (Pretty): 52383-1.0224016-rdf.xml
RDF/JSON: 52383-1.0224016-rdf.json
Turtle: 52383-1.0224016-turtle.txt
N-Triples: 52383-1.0224016-rdf-ntriples.txt
Original Record: 52383-1.0224016-source.json
Full Text

Full Text

REVIEW Open AccessThe implications of the feminization of the primarycare physician workforce on service supply: asystematic reviewLindsay Hedden1,2*, Morris L Barer1,2, Karen Cardiff2, Kimberlyn M McGrail1,2, Michael R Law1,2 and Ivy L Bourgeault3AbstractThere is a widespread perception that the increasing proportion of female physicians in most developed countries iscontributing to a primary care service shortage because females work less and provide less patient care compared withtheir male counterparts. There has, however, been no comprehensive investigation of the effects of primary carephysician (PCP) workforce feminization on service supply. We undertook a systematic review to examine the currentevidence that quantifies the effect of feminization on time spent working, intensity and scope of work, and practicecharacteristics. We searched Medline, Embase, and Web of Science from 1991 to 2013 using variations of the terms‘primary care’, ‘women’, ‘manpower’, and ‘supply and distribution’; screened the abstracts of all articles; and enteredthose meeting our inclusion criteria into a data abstraction tool. Original research comparing male to female PCPs onmeasures of years of practice, time spent working, intensity of work, scope of work, or practice characteristics wasincluded. We screened 1,271 unique abstracts and selected 74 studies for full-text review. Of these, 34 met the inclusioncriteria. Years of practice, hours of work, intensity of work, scope of work, and practice characteristics featured in 12%,53%, 42%, 50%, and 21% of studies respectively. Female PCPs self-report fewer hours of work than male PCPs, havefewer patient encounters, and deliver fewer services, but spend longer with their patients during a contact and dealwith more separate presenting problems in one visit. They write fewer prescriptions but refer to diagnostic servicesand specialist physicians more often. The studies included in this review suggest that the feminization of the workforceis likely to have a small negative impact on the availability of primary health care services, and that the drivers ofobserved differences between male and female PCPs are complex and nuanced. The true scale of the impact ofthese findings on future effective physician supply is difficult to determine with currently available evidence,given that few studies looked at trends over time, and results from those that did are inconsistent. Additionalresearch examining gender differences in practice patterns and scope of work is warranted.Keywords: Primary health care, Physicians, Gender differences, Health human resources planning, Workforceplanning, Practice patterns, Female, MaleBackgroundThe primary care physician (PCP) workforce in many in-dustrialized nations is increasingly female. In several indus-trialized countries, the proportion of PCPs who are womenhas doubled or nearly doubled over the last 30 years [1,2].Globally, 32% of all physician graduates worldwide arefemale, and that percentage is higher, on average, in familymedicine [3]. Thirty-four percent of family medicine/general practice physicians and 55% of family medicineresidents in the United States (US) are women [4]. InCanada, women now make up 58% of medical schoolenrollees (up from 14% in 1968) [5] and more women thanmen are choosing to specialize in primary care [6].Amidst often highly-charged claims of physician short-ages from the public and medical leadership alike, futurephysician workforce planning has been identified as apriority for both research and policy action in many in-dustrialized countries, and is essential for the rationalmanagement of health care systems [7]. If they are to be* Correspondence: lindsay.hedden@ubc.ca1Centre for Health Services and Policy Research, University of BritishColumbia, 201-2206 East Mall, V6T 1Z3 Vancouver, BC, Canada2School of Population and Public Health, University of British Columbia, 2206East Mall, V6T 1Z3 Vancouver, BC, CanadaFull list of author information is available at the end of the article© 2014 Hedden et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the CreativeCommons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, andreproduction in any medium, provided the original work is properly credited. The Creative Commons Public DomainDedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article,unless otherwise stated.Hedden et al. Human Resources for Health 2014, 12:32http://www.human-resources-health.com/content/12/1/32an effective policy tool, physician workforce planning willneed to go beyond simply projecting the traditional factorsof population growth and ageing, and physician head-counts [8-10], to include variables that affect both servicerequirements (population need) and availability [11-13].The rapid feminization of the PCP workforce over thepast half-century is a significant demographic change thathas the potential to influence service availability. For ex-ample, claims that changes in the gender balance of thePCP workforce will change the effective overall supply ofprimary health services (for example, because female phy-sicians work fewer hours than their male counterparts)and/or the mix of available services (for example, becauseof differences in styles of practice) have a certain intuitivevalidity. Unfortunately, debate in this arena has, for themost part, not advanced much beyond these simplisticclaims.Thus far, even where workforce planning models ac-count for changes in physician workforce demographics(such as feminization), they commonly apply a simplisticcalculus, using simple service or headcounts, or assum-ing the work of a female physician as a fixed proportionof a male physician (typically using full-time equivalentmeasures) [14,15]. It is very difficult to find supply pro-jection models that embody evidence about the differ-ences between male and female physicians in life-courseproductivity, changes over time in trends in retirement,or recent changes reflecting shifting work-life prioritiesamongst younger cohorts of physicians. The focus of thispaper is to synthesize the evidence relating to the first ofthese factors - male-female differences in physician serviceprovision over a life-cycle. Our specific population ofinterest is general practice and family medicine (which wewill henceforth refer to as PCPs); other primary care spe-cialties such as internal medicine and pediatrics will bediscussed in a subsequent manuscript.This systematic review examines evidence related tothe effect of the PCP workforce, defined here asfeminization on the supply of physician services. Spe-cifically, we reviewed studies that compared male andfemale PCPs in terms of the amount of time they spentworking, how intensely they worked (that is the num-ber of services or patient encounters per unit time),and whether their practice and service characteristicsdiffered.MethodsSearch strategy and inclusion criteriaIn an effort to ensure comprehensiveness, we usedmultiple search strategies to locate both peer-reviewedand grey literature sources. Peer reviewed literaturewas selected from Medline (OVID), Embase, and Webof Science. We limited our search to English languagearticles published between January 1990 and January2013. Our database-specific search terms included varia-tions on ‘physician’, ‘women’, and ‘workforce’ (see Additionalfile 1 for the full search strategies). We identified rele-vant grey literature using the Canadian Health ResearchLibrary, ProQuest Dissertations and Theses, and theCanadian Health Human Resource Network Library (http://www.hhr-rhs.ca/index.php?option=com_content&view=article&id=168&Itemid=78&lang=en). We also conducted searchesof the websites of organizations, groups, governments, as-sociations, and professional bodies identified using theCanadian Agency for Drugs and Technologies in Health’s‘Grey Matters’ guide to grey literature [16]. Additionally,we completed forward and reverse citation searches (snow-balling) of included peer-reviewed articles using GoogleScholar.We imported search results into a reference managerand removed any duplicates. We screened all abstractsfor relevance to the research topic and pulled relevantarticles. Two reviewers independently reviewed all full-text articles using the inclusion and exclusion criteria inTable 1 and thematic typology in Table 2, and disagree-ments were resolved by discussion. We computed aKappa statistic for inter-rater reliability. Studies were notexcluded due to quality issues; however, methodologicalconcerns are presented as part of both the Results andDiscussion sections.Data extraction and article typologyWe abstracted and summarized the following data fromall included articles: citation; country; objectives; studysample, response and drop-out rates (where applicable);study design (cross-sectional or longitudinal); data col-lection (administrative, survey, or other primary data);analytic methodology; outcome measure(s); and results.We coded articles using a typology designed with theintention of capturing any practice differences betweenmale and female physicians that could, either directly orindirectly, affect the availability of primary health careservices. It includes variations in what care is delivered,to whom, and how much. The typology consists of fivethemes and eleven subthemes (Table 2). Table 2 includesexamples of how each thematic area may be linked tochanges in service availability.We conducted a qualitative examination of study qualityby assessing the following items: clarity of researchquestions and objectives; appropriateness of study design;sample size and representativeness; validity of measures;addressing possible confounders; and generalizability.ResultsSearch resultsThe initial search of Medline, Embase and Web of Sciencelocated 1,476 citations, of which 205 were duplicates. Theabstracts from the remaining 1,271 were screened forHedden et al. Human Resources for Health 2014, 12:32 Page 2 of 11http://www.human-resources-health.com/content/12/1/32relevance to the topic, and 1,224 were excluded, leaving47 peer-reviewed articles. An additional 27 studies wereidentified from grey sources and through snowballingof references in selected articles. These 74 sources wereretained for full-text review (Figure 1). Of these, 34studies met the inclusion criteria; they are summarizedin Additional file 2. The K-coefficient for inter-rateragreement beyond change was 0.84.Thirty of the 34 included studies (88%) had been pub-lished in peer- reviewed fora. Fifteen of the 34 (44%)were conducted in Canada, four (12%) in the US, andfive (15%) in the United Kingdom. Twenty-seven studies(79%) used a cross-sectional methodology. Of these, 21(78%) used retrospective survey data, five (19%) used ad-ministrative data, and one employed prospective primarydata collection. Of the seven (21%) studies that used lon-gitudinal methods, four (57%) used administrative data,one combined administrative and survey data, and two(29%) used surveys alone.Thematic resultsHours of work, intensity of work (defined here as num-ber of services or patient encounters per unit time), andscope of work featured in 18 (53%), 14 (42%) and, 17(50%) studies respectively (Figure 2). Practice character-istics were examined in seven (21%) studies, and years ofpractice was a focus in only four (12%). Themes with adirect impact on service availability (years of practice,hours and intensity of work) were more commonly fea-tured (26 articles, 76%) than those that affect supply orTable 1 Inclusion and exclusion criteriaInclusion criteria Exclusion criteriaPublication DetailsPublished between January 1990 and January 2013; published in English Published before January 1990 or after January 2013;published in a language other than EnglishParticipants/PopulationPCPs (studies focusing on all physicians were included only if resultspertaining to PCPs were presented separately)Other physician specialties; all physicians, where separateanalysis for PCPs is not presentedComparisonMale to female PCPs1 Does not compare male and female physiciansOutcome MeasuresA measure of one or more of the following: time spent working, intensityof work, scope of work, or practice characteristics2None of time spent working, intensity of work, scope of work,or practice characteristicsDesignOriginal research Editorials, comments or commentaries, letters; reviews articles;reports with no primary data analysis1Specialist physicians (such as pediatricians, or general internists) who may practice like PCPs on occasion (that is acting as a point of entry to the health caresystem, providing person-focused care over time, and acting as a coordinator for care provided elsewhere) were not included.2Raw or adjusted results for one or more of these measures must be presented. If these measures were included as covariates in a multivariate modeling exercise(for example, for income), the study was excluded unless raw comparisons on one of these outcomes are also presented.PCP, primary care physician.Table 2 Article typologyTheme Subtheme Potential effect on supply - Direct/IndirectYears of practice • Retirement Direct - for example, shortening of career or more lengthy absences from practice• Leaves of absenceHours of work • Full- versus part-time work Direct - for example, less time spent working overall, or less time spent on directpatient care in favour of other responsibilities• Time spent on patient care• Time spent on administrativeresponsibilities, professional developmentIntensity of work • Number of services/time Direct - (lower service or patient volumes)• Number of patients/timeScope of work • Patient characteristics Indirect - (restrictions in scope of practice, or basket of services delivered; restrictedpatient population; reduced availability of out-of-office or off-hours care)• Service provisionPracticecharacteristics• Location Indirect - (imbalance between urban- versus rural-based practices leading to shortagesin some areas, oversupply in others)• Group practice versus solo practiceHedden et al. Human Resources for Health 2014, 12:32 Page 3 of 11http://www.human-resources-health.com/content/12/1/32availability of services indirectly (practice characteristics,scope of practice) (18 articles, 53%). Slightly more thansixty percent of the included studies focused on a singlethematic area.Hours of workAll 18 studies that examined hours of work found thatfemale PCPs tended to self-report working fewer hoursthan their male counterparts. Few of these studies, how-ever, presented results that adjusted for physician age,practice characteristics or other factors that may con-found the relationship between physician sex and workhours (for example, [17,18]). In their survey of Englishgeneral practitioners, Gravelle and Hole found that theaverage difference in hours per week worked betweenmales and females was 11.8 hours [17]. Forty-five per-cent (5.3 hours) of this difference was due to the greaterproportion of male PCPs at each age working full-time,and 46% (5.4 hours) was due to female PCPs reducingtheir hours more than male PCPs who have the same fam-ily circumstances. The final 9% (1.1 hours) of the differ-ence was due to differences in physician demographics(for example, age) and practice characteristics (for ex-ample, size of practice) [17].In their European study, Boerma and van den Brink-Muinen found that, on average, male PCPs worked morehours per week, excluding on-call time (45.1 versus36.2) [18]. In countries where the difference in hourswas statistically significant (12 of 32 study countries),male PCPs worked more in ten, and female PCPsworked more in two [18]. Results from North Americaare similar, with female PCPs working between four and14.5 fewer patient-care hours per week [8,19-23].Female PCPs were more likely to report working part-time (31.6% versus 11.1%) [17,24], and billed Canadianprovincial health insurance plans for fewer months ofthe year [25]. Having children under the age of 18 in-creased the probability that female PCPs worked part-time, but had no effect on male PCPs [17].Despite consistent differences found in hours workedoverall, and specifically in hours spent on patient care,male and female PCPs tended to spend a similar amountof time on-call [21,23,26].Three of the included studies examined longitudinaltrends in work hours for male and female physicians[8,27,28]. In their study on PCP labour supply inCanada, Crossley et al. found a secular decline in hoursof patient care between 1982 and 2003 [28]. Althoughfemale physicians were found to have worked fewerhours than male physicians, a change in the behavior ofmale PCPs accounted for a greater proportion of the de-cline in hours of patient care than did the growing pro-portion of females in the workforce. The gap in hoursworked between male and female PCPs diminished overthe study period [28]. They also reported that, for femalephysicians only, there was a significant age effect onhours of patient care: hours declined up to approxi-mately age 38, and then gradually increased with agePotentially relevant citations identified through Medline, Embase, and Web of Science: N = 1476Abstract screening: N = 1271Duplicates removed: N = 205Abstracts excluded for lack of relevance: N = 1224Records included in full-text review: N = 47Citations identified through snowballing: N = 5Citations identified in grey literature: N = 22Ful-text articles excluded: N = 40Not original research: N = 8Not primary care: N = 5No male/female comparison: N = 11Didn’t include outcome of interest: N = 15Out of date range: N = 1Full-text articles included: N = 34Figure 1 Search results.Hedden et al. Human Resources for Health 2014, 12:32 Page 4 of 11http://www.human-resources-health.com/content/12/1/32[28]. This would be consistent with a ‘childbearing years’effect. Aasland and Rosta found that the gap betweenmale and female PCPs’ hours of work is also narrowingin Norway, with female PCPs having worked significantlyfewer hours than male PCPs between 2000 and 2006,but not in 2008 [27]. In that country, however, physi-cians’ hours have, on the whole, increased rather thandeclined, with the increase in hours obviously beingmore marked amongst female physicians [27].Intensity of workEleven studies compared the number of services per unitof time delivered or number of patients seen for maleand female PCPs. Of these, five presented multivariateresults, controlling for the effect of physician and patientcharacteristics, or other confounders.Cohen et al., Woodward and Hurley, and the Canad-ian Institute for Health information all found that Can-adian male PCPs bill for more services compared withtheir female colleagues, and that physician gender contrib-uted significantly to explaining variation in service activity[25,29,30]. Boerma and van den Brink-Muinen similarlyfound that European female PCPs have on average 4.1 (or14%) fewer office contacts per day. This difference in of-fice contacts was only significant in 12 of the 32 studycountries, and in half of these, female physicians had sig-nificantly more daily contacts than male physicians [18].Additionally, when results were restricted to only includephysicians who worked full-time, the sex-related differ-ence in contacts dropped to 2.3 fewer contacts per day forfemale physicians, and a significant difference was foundin only six of 32 countries. Of these, women had signifi-cantly more contacts per day in three [18].Consistent with the age-stratified results repented forhours worked, Constant and Legere reported that thedifference between male and female PCPs peaks betweenthe ages of 36 and 40, and declines thereafter [14].Unadjusted results from the remaining studies wererelatively consistent: male PCPs were reported to delivermore services than female PCPs (700 versus 399/month)[24], and to have more patient encounters (between 32and 72/week) (for example: [19,21,26]). Female PCPs,however, were found to manage more problems per pa-tient encounter (157.8 versus 145.4 per 100 encounters)and spend 40% more time with each patient (20.5 versus14.4 minutes) [31,32].In their longitudinal examination of intergenerationaldifferences in workloads of physicians from six Canadianprovinces, Watson et al. found that between 1992 and2001, female PCPs reduced their workloads (defined asnumber of visits per year) by 6.1%, while male workloadsremained stable. The result was an accentuated differ-ence in workload over time: female physicians’ work-loads were, on average, 74% of the workloads of theirmale counterparts in 1992, and 68% in 2001 [8].These results run somewhat counter to those reportedby Crossley et al. who found that the gap in self-reportedhours worked between male and female physicians wasnarrowing [28]. It is possible that these conflicting resultscould be caused by some combination of differences intime periods used for analysis (1982 to 2003 versus 1992Chambers et al., 1996Chan et al., 2003Chaytors, 2001Cohen et al., 1991Constant et al., 2008Cree, 2001Dumontet et al. 2012French et al., 2006Harrison et al., 2011Kazanjian et al., 2000Keane et al., 1991Mayorova et al., 2005McKinstry et al., 2006 McMurray et al., 2002Norton et al., 1994Raymont al., 2005Slade et al., 2002Watson et al., 2006Weeks et al., 2006Weyrauch et al., 1995Woodward et al., 1995Ubokudom, 1997Leese et al., 2002Gravelle et al. 2007Crossley et al., 2009Aasland et al., 2011Atkin, 2000Hours of workYears ofpracticeBensing et al., 1993Boerma et al., 2000Bergeron et al., 1999Brett et al., 2009Britt et al., 1996CIHI., 2001Carek et al., 2003RetirementLeaves of absenceFull vs. part timePatient care hoursOther activitiesNumber of patients/timeNumber of services/timePatient characteristicsService provisionPractice locationSolo vs. group practiceScope of workPractice characteristicsIntensity of workFigure 2 Frequency of thematic categories.Hedden et al. Human Resources for Health 2014, 12:32 Page 5 of 11http://www.human-resources-health.com/content/12/1/32to 2001), outcome measure (hours versus billed consulta-tions) or other differences in methodology [28]. If onetakes both sets of results at face value and attempts to rec-oncile them, a possible conclusion would be that malePCPs are reducing their hours while maintaining visitcounts, while female PCPs are maintaining their hours,but are decreasing their visits. Taking account of other re-sults cited here, it may be that female PCPs are simplychanging their style of practice, taking more time witheach patient and dealing with more problems per visit.The other conclusion that can be drawn from these resultsis that measuring physician productivity is difficult, andthat the numerator (outputs or outcomes per unit of activ-ity) matters [33].Scope of workPatient characteristics Compared with male PCPs, fe-male PCPs saw a higher proportion of female patients[24,25,31,34] in all age groups [27], but especially in the15 to 49 age category [24,25]. They also saw fewer older-aged patients than their male counterparts [1,23]. Theseresults survived multivariate analyses that accounted forthe age of physician, practice location, and graduationperiod [25].Care delivered Controlling for patient and physiciandemographics, female PCPs were significantly morelikely to manage issues related to the reproductive or fe-male genital system [1,31,34], as well as psychologicaland social problems [1,31,34]. Female physicians wereless likely to manage issues of the musculoskeletal, ormale genitourinary systems [1,31].With respect to obstetrical and prenatal care, resultsfrom US-based literature were inconsistent with thosefrom Canada. In the US, male and female PCPs wereequally likely to provide prenatal care, with or withoutdelivery [23]. In contrast, in Canada, female physicianswere more likely than their male counterparts to provideprenatal care, but were less likely to provide intrapartumcare [24].After adjusting for problems per encounter, as well asphysician, practice and patient characteristics, Australianmale PCPs had a higher rate of prescribing (4.3% moremedications per 100 patients) [1]. Female PCPs recorded19.5% more clinical treatments (for example, educationand counselling), 18.5% more referrals, 8.1% more im-aging ordered and 9.6% more pathology tests ordered[1]. In their 1993 study on service delivery trends formale and female PCPs in the Netherlands, Bensing andcolleagues found that female physicians wrote fewer pre-scriptions and performed fewer technical interventionscompared with male physicians; however, they orderedmore laboratory tests [34]. They found no difference inthe rate of referrals to specialists [34].Chan and colleagues examined the referral rates forCanadian male and female PCPs. Like Harrison et al. [1]they found that female physicians referred to specialistsabout 10% more frequently than their male colleaguesafter making adjustments for patient age and gender [35].Boerma and van den Brink-Muinen found that maleEuropean PCPs were more involved in technical proce-dures; however the difference was smaller in countrieswith a gatekeeping system [18].Out-of-office and Off-hours care Five studies exam-ined the provision of out-of-office and/or off-hours care[18,23,24,26,36]. In 1991, Keane et al. reported that asmaller proportion of Canadian female than male PCPsbilled for home visits (1.5 versus 3.7 per 100 patients)and after hours care (7.0 versus 9.6 per 100 patients,after controlling for the effects of place and date of MDgraduation, practice location, certification status, andwork status [24].Adjusted for patient, physician and practice character-istics, male PCPs also more routinely made long-termcare facility visits (50.6% versus 35.5% for females), andhome visits (49.0% versus 33.8% for females) [18]. MalePCPs were also more likely than their female counter-parts to bill for time in the hospital (14.8% versus 13.1%,emergency room (37.0% versus 14.2%), or for surgicalassists (64.8% versus 47.2%) [18].Consistent with the multivariate results from Keane et al.and Boerma and van den Brink-Muinen, the two stud-ies that report only bivariate results found that femalePCPs were less likely to provide after-hours services[23,26], make house calls (for example, 12.7% versus15.2% for men), and spend significantly more of theirwork time in office or clinic practice (87.9% versus80.9% for men [23]. This is in contrast to findings re-ported by Bergeron et al. who report that althoughmale physicians make more home visits compared withfemale physicians, they spend an almost equal amountof time on this activity (5.7 versus 5.2 hours/week) [36].Years of practicePatterns of retirement (or practice leave) were examinedin four of the included studies [37-40], and results aremixed. French et al. found that a similar proportion ofmale and female PCPs in Scotland intend to retire at age59 [40]. In their study of Australian physicians, Brett et al.report that male PCPs were more likely to intend to retirebefore age 65: 75% of women compared with 59% ofmen reported that they intended to work to normal re-tirement age (rather than retiring early) [37]. In theirsurvey of physicians who had recently left practice,however, Leese et al. found that female leavers tendedto be younger, and to have children under the age of 18[38]. This suggests that childrearing responsibilitiesHedden et al. Human Resources for Health 2014, 12:32 Page 6 of 11http://www.human-resources-health.com/content/12/1/32play a key role in decisions to leave practice, and thatfemale PCPs are more likely to leave practice for rea-sons other than full retirement, compared with theirmale counterparts.Leaves of absence, for reasons of childbearing orotherwise, were not a focus in any of the articles in-cluded in this review.Practice characteristicsFemale PCPs practicing across Europe and in Australiawere less likely than men to work in solo practice (ratherthan in small or large groups (Europe: 27% of womenfound to work in solo practice versus 45.2% of men [18];Australia: 4.6% of women work in solo practice, versus13.2% of men)) [1]. In the US, male and female PCPs areabout equally likely to practice within a small group(32.7% versus 38.3%) [23].Female PCPs practicing in Europe were significantlyless likely to practice in rural areas compared with theirmale counterparts (14.9% versus 27.2% rural). In con-trast, in the US and Australia, women and men wereequally likely to choose rural practice [1,23]. FemalePCPs in Europe were more likely to work in inner citylocations (33.7% versus 18.0%) [18].DiscussionThe intent of this systematic review was to examine theimpact of the increasing proportion of women in thePCP workforce on service delivery in five areas thatcould affect such projections of service supply: years ofpractice, hours of work, intensity of work, scope of work,and practice characteristics. Compared with their malecolleagues, female PCPs: Self-report fewer hours of work (excluding on-calltime) Have fewer patient encounters, and deliver fewerservices (perhaps as an artifact of working fewerhours), but spend longer with their patients during acontact and deal with more separate presentingproblems during each visit Write fewer prescriptions, but order more laboratorytests, and refer patients on to specialists morefrequently See more female patients and fewer geriatric patients Provide less out-of-office (including home, nursinghome and hospital visits) and off-hours careThe scale of the impact of these findings on future ef-fective physician supply is difficult to determine withcurrently available evidence, given that very few studieslooked at time trends or years of practice, and resultsfrom those that did are inconsistent. Also, the full im-pact will depend critically on future trends in thefeminization of the workforce. In Canada, and in the UKand other parts of Europe, the proportion of medicalstudents who are female ensures that the overall supplyof physicians will continue to become increasinglyfemale in the near term.Given that fact, the differences in practice patterns be-tween male and female PCPs could result in increasedderived demand for specialist physician services, labora-tory technicians, imaging technicians or other healthprofessionals, outside of primary health care. The factthat female PCPs spend less time in off-hours care, andare less likely to serve patients at home and in nursinghomes, could increase the reliance on already-stretchedemergency departments and walk-in clinics as a sourceof primary health care, and force a rethinking of howmedical care is delivered to patients outside standard of-fice hours and locations.It is important to consider the effects of childbearingand childrearing, which were mentioned in several stud-ies, but were seldom explicitly investigated, and werenot the primary focus of any of the research documentsreviewed here. Female PCPs who had children under age18 worked fewer hours per week and were more likelyto have self-reported part-time status compared withwomen who did not. The dampening effect of childrenon work hours was twice as large for women as it wasfor men. And, one study found that once family circum-stances were accounted for, the gender of the physicianhad no significant effect on hours worked [17].An important issue that was not covered in any of theliterature reviewed here is the balance between workand household responsibilities among physicians. Onestudy found that female physicians spent more time onunwaged childcare and household jobs than male physi-cians [41]. Once unwaged household responsibilitieswere accounted for, female PCPs who have childrenworked an average of 90.5 hours a week, compared with68.6 hours per week for males with children [41].Consistency of resultsResults were strongly consistent across some of the the-matic areas, and relatively less so in others. In particular,results relating to the hours and intensity of work wereconsistent across studies. In other areas, such as practicecharacteristics, results were highly variable.The results of this review demonstrate that the driversof observed differences between male and female PCPsare complex and nuanced. The size of an observed gen-der difference varied based on the characteristics of thehealth care system under study and on whether the pos-sible confounding effects of physician age, practice char-acteristics, and in particular, family characteristics andpart-time status were adequately controlled. There wereat least 36 different health care systems represented byHedden et al. Human Resources for Health 2014, 12:32 Page 7 of 11http://www.human-resources-health.com/content/12/1/32the studies included in this review. Inconsistent resultsacross studies may be caused by health care system dif-ferences including, but not limited to, physician remu-neration mechanisms and policies, the gatekeeping roleof general practitioners, and general employment pol-icies. An exploration of the role of such system differ-ences was well beyond the scope of this review, but is animportant area for future research.Inconsistent results could also be a function of meth-odological and measurement differences across studies,and whether the confounding effects of other physician,patient, and practice characteristics have been accountedfor. For example, gender differences in the number ofpatient contacts per day disappeared once full- versuspart-time status had been accounted for in work byBoerma and van den Brink-Muinen [18]. Differences inhours worked depended on whether auxiliary activitiessuch as on-call time were included as part of ‘hourworked’ [23]. Similarly, differences in care provisionwere attenuated once patient characteristics and practicelocation was accounted for (for example, [1,31]).Methodological issuesAs part of our qualitative assessment of study quality,we identified some significant methodological concernswith the studies included in this review. For the mostpart, they relied on cross-sectional retrospective surveys.Such surveys are always subject to recall bias, thoughunless there were systematic male versus female differ-ences in accuracy of recall, this may not be an issue inthis particular circumstance. But surveys do tend to pro-duce inflated estimates of hours worked for those whoreport high hours (more often male physicians) and de-flated estimates for those reporting low hours (moreoften female physicians), which may exaggerate any truegender difference [42]. Many studies relied on small,often unbalanced samples, raising concerns about selec-tion bias. All but one study failed to adjust statisticallyfor multiple comparisons, despite conducting as many as155 separate statistical significance tests [32].Perhaps even more concerning, however, is that 12(35%) studies presented only unadjusted, bivariate results,failing to control for the potential confounding effects ofother physician, patient or practice characteristics (forexample, [23,26,34]). Additionally 6 (18%) undertook onlyrudimentary stratification (for patient age and gender, forexample) (for example, [24,25,34,43]). Statistical methodscontrolling for confounders may not yet have been ac-cepted practice in this field when some of these earlier pa-pers were published, which may explain their limited use.Comparisons between adjusted and unadjusted resultssuggest that physician age, family characteristics and prac-tice location, at a minimum, can have important influ-ences on apparent male-female differences in key practiceand productivity indicators. For example, older physicians- who are more likely to be male - tend to see more olderpatients [18], and physicians who work in rural-basedclinics practice differently from physicians who practice inurban centres [32]. Thus the impacts of physician age andpractice location may be conflated with a gender effect inunadjusted analyses, since female PCPs tend to be youn-ger [31] and more likely to work in urban centres in somecountries [18].Gaps in knowledge and future researchGiven the reliance on cross-sectional and survey data,and the relative underutilization of longitudinal or ad-ministrative datasets in this area, there remains a needto critically examine activity levels, over time and at apopulation level, adjusting for the potentially confound-ing effects of age and cohort. The issue of retirementpatterns has also not been adequately examined with ref-erence to the effects on time spent working. It is pos-sible, for example, that although female PCPs work less,especially around childbearing years, they may retirelater than their male counterparts, reducing or eveneliminating a career difference in time spent working.While historically this may not have been true, trendsover time suggest that it might become so in future. Thekey point is that differences in retirement patterns be-tween male and female physicians may partially orwholly offset other trends in service provision, whenviewed over an entire life-cycle. Leaves of absence takenfor parental or other reasons should also be examinedfor their effects on both time and intensity of working.No studies included in this review examined absencesfrom practice.To date, the literature examining other practice differ-ences between male and female physicians that couldhave an important impact on health human resourcesplanning has been limited. More studies comparing thepatient populations of male and female PCPs - beyondsimple gender concordance and patient age - are cer-tainly warranted. Specifically, very little work has beendone examining differences in patient morbidity levels,or chronic disease burdens. Additionally, more nuancedinvestigations of service mix, problems seen, and caredelivered would address currently unanswered, but im-portant, questions bearing on the future provision ofphysician services. For example, differences in practicestyle between male and female physicians have currentlyreceived little attention beyond comparisons of timetaken for each appointment.Issues of work-life balance and childrearing and house-hold responsibilities are also under-researched, especiallygiven their observed impact on full- versus part-time jobstatus and working hours [17,41]. In the 2007 and 2010Canadian National Physician Surveys, the majority ofHedden et al. Human Resources for Health 2014, 12:32 Page 8 of 11http://www.human-resources-health.com/content/12/1/32respondents identified attaining balance between per-sonal and professional life as the most important factorfor a satisfying practice [44]. Physicians, regardless ofgender, are increasingly (and not unreasonably) seekinga work environment that provides this balance, withoutcompromising the quality of care they provide to theirpatients [45]. Secular trends in time made available forclinical practice obviously have direct implications forprojections of physician service provision.LimitationsThis systematic review used comprehensive search strat-egies encompassing multiple peer-reviewed and grey lit-erature sources to maximize capture of relevant articlesand minimize publication bias. The restriction of articlesto those published in English and within the last 23 yearsmay have eliminated some potentially relevant studies.Additionally, because the area of research is not yetwell-indexed and the specific topic area is broad, somestudies that would be relevant, but whose main compari-son was not male versus female PCPs, may have beenmissed.Our decision to include only those studies that focusedon PCPs, defined here as general practitioners or familymedicine specialists, (rather than also including otherspecialists like general internists or pediatricians - whomay practice like PCPs under certain circumstances)may limit the generalizability of our results, particularlywith respect to research from the US.An additional limitation is the decision not to elimin-ate studies that were deemed of poor quality. The meth-odologies employed in many of the studies is certainlyfar from ideal, with many relying on small, unbalancedsamples, retrospective surveys, and incomplete (or no)control for the impact of confounding factors. Thesestudies were, however, retained in the review since noneof the 30 included would have achieved the level of guid-ance required for formal guidelines (for example, thoseissued by the Cochrane Collaboration) and, thus, therewas no straightforward way to gauge methodologicalquality.Meta-analytic techniques could have been a useful wayto summarize the research within individual thematicand subthematic areas; however, small numbers and thevariance in outcome measures even within individualsubthemes were too great to allow for the use of thosetools.Implications for health human resource plannersProjections of physician supply must take into accountvariables other than estimated future physician head-counts. At a minimum, more robust measures that ac-count for gender differences in service volumes, but thatalso address the implications of the differences in patientmix, service mix, and practice style between male andfemale physicians need to be developed and used as evi-dence in these areas becomes available. Other demo-graphic and workforce factors, such as the impact ofphysician age and cohort - should also be considered.ConclusionsCompared with their male counterparts, female PCPsspend less time working, and deliver less care. Evidenceas to whether this gap is narrowing is mixed. The effectof childrearing is critically important, affecting femalePCPs far more than their male counterparts, in terms ofimpact on participation in clinical practice. Once the ef-fect of family characteristics has been accounted for, sexhas no effect on time spent working. Issues of work-lifebalance, caregiving and childrearing responsibilities war-rant attention in future research.The literature focuses heavily on differences in theamount of work done by female compared with male phy-sicians, and is almost exclusively based on retrospectivesurveys with some significant methodological limitations.These studies tell us nothing about differences in theappropriateness or quality of care. Also, more researchexamining differences in practice characteristics, and pa-tient/service mix, is warranted in order to support the de-velopment of robust forecasts of physician supply. Suchforecasts would ideally take into account sex-relateddifferences in volume, bct also the implications of the dif-ferences in patient/service mix and practice style, andtemporal trends in each of these. The extant literaturesuggests that secular trends in hours of work may domin-ate sex-related differences in service provision.Additional filesAdditional file 1: Medline Search Strategy.Additional file 2: Summary of Included Studies [1,2,8,14,17-32,34-40,43,46-51].AbbreviationsPCP: primary care physician; US: United States.Competing interestsThe authors declare that they have no competing interests.Authors’ contributionsAll authors contributed to the collection and interpretation of the data. LHdeveloped the search and abstraction tools. LH and KC selected studies forreview and conducted the abstraction and interpretation of the data. MB, IB,KM, and ML contributed significantly to the analysis and interpretation. LHwrote the first draft and managed subsequent drafts with revisions from allother authors. All authors give final approval of the publication of thisversion of the paper.AcknowledgementsWe wish to note, with thanks, the considerable time and effort that hasbeen committed to this paper by the two reviewers. The paper has beenconsiderably improved as a result.Hedden et al. Human Resources for Health 2014, 12:32 Page 9 of 11http://www.human-resources-health.com/content/12/1/32Author details1Centre for Health Services and Policy Research, University of BritishColumbia, 201-2206 East Mall, V6T 1Z3 Vancouver, BC, Canada. 2School ofPopulation and Public Health, University of British Columbia, 2206 East Mall,V6T 1Z3 Vancouver, BC, Canada. 3Telfer School of Management and Instituteof Population Health, University of Ottawa, 1 Stewart St, K1N 6 N5 Ottawa,ON, Canada.Received: 15 October 2013 Accepted: 26 May 2014Published: 4 June 2014References1. Harrison CM, Britt HC, Charles J: Sex of the GP - 20 years on. Med J Aust2011, 195(4):192–196.2. Dumontet M, Le Vaillant M, Franc C: What determines the income gapbetween French male and female GPs - the role of medical practices.BMC Fam Pract 2012, 13:94–101.3. American Medical Association: Table 1.7: female international medicalgraduates by age and self-designated specialty. In Physician CharactDistrib US; 2012.4. American Association of Medical Colleges: 2012 Physician Specialty DataBook. 2012.5. The Association of Faculties of Medicine of Canada: Enrolment in CanadianFaculties of Medicine by Sex, 1968/69 to 2012/13. 2013.6. Canadian Resident Matching Service: R-1 Match Reports, 2011. 2011.7. Harrison CM, Britt HC: General practice: workforce gaps now and in 2020.Aust Fam Physician 2011, 40:12–15.8. Watson DE, Slade S, Buske L, Tepper J: Intergenerational differences inworkloads among primary care physicians: a ten-year, population-basedstudy. Health Aff (Millwood) 2006, 25(6):1620–1628.9. Esmail N: Demographics and Canada’s physician supply. Fraser Forum2007, 16–19.10. Lomas J, Stoddart GL, Barer ML: Supply projections as planning: a criticalreview of forecasting net physician requirements in Canada. Soc Sci Med1985, 20:411–424.11. Birch S, Kephart G: Human resources planning and the production ofhealth: a needs-based analytical framework. Can Pub Pol 2007,33(Supplement):S1–S16.12. Dreesch N, Dolea C, Dal Poz MR, Goubarev A, Adams O, Aregawi M,Bergstrom K, Fogstad H, Sheratt D, Linkins J, Scherpbier R, Youssef-Fox M:An approach to estimating human resource requirements to achieve theMillennium Development Goals. Health Pol Plan 2005, 20:267–276.13. O’Brien-Pallas L, Baumann A, Donner G, Murphy GT, Lochhaas-Gerlach J,Luba M: Forecasting models for human resources in health care. J AdvNurs 2001, 33:120–129.14. Constant A, Leger PT: Estimating differences between male and femalephysician service provision using panel data. Health Econ 2008,17:1295–1315.15. Zelmer J, Leeb K: Health human resource planning: head counting is notenough. Hosp Q 2001, 5:51–54.16. Canadian Agency for Drugs and Technologies in Health: Grey Matters: APractical Search Tool for Evidence-Based Medicine; 2013.17. Gravelle H, Hole AR: The work hours of GPs: survey of English GPs. Br JGen Pract 2007, 57:96–100.18. Boerma WGW, Van den Brink-Muinen A: Gender-related differences in theorganization and provision of services among general practitioners inEurope: a signal to health care planners. Med Care 2000,38:993–1002.19. Norton PG, Dunn EV, Soberman L: Family practice in Ontario: howphysician demographics affect practice patterns. Can Fam Physician 1994,40:249–256.20. Slade S, Busing N: Weekly work hours and clinical activities of Canadianfamily physicians: results of the 1997/98 National Family PhysicianSurvey of the College of Family Physicians of Canada. Can Med Assoc J2002, 166:1407–1411.21. Atkin K: She Works Hard for the Money: Differences in the Income of Male andFemale General Practitioners. University of Toronto; 2000.22. McMurray J, Cohen M, Angus G, Harding J, Gavel P, Horvath J, Paice E,Schmittdiel J, Grumbach K: Women in medicine: a four-nation comparison.J Am Med Womens Assoc 2002, 57:185–190.23. Carek PJ, King DE, Hunter M, Gilbert GE: Practice profiles, procedures, andpersonal rewards according to the sex of the physician. South Med J2003, 96:767–771.24. Keane D, Woodward CA, Ferrier BM, Cohen M, Goldsmith CH: Female andmale physicians: different practice profiles: will increasing numbers offemale GPs affect practice patterns of the future? Can Fam Physician1991, 37:72–81.25. Cohen M, Ferrier BM, Woodward CA, Goldsmith CH: Gender differences inpractice patterns of Ontario family physicians (McMaster medicalgraduates). J Am Med Womens Assoc 1991, 46(2):49–54.26. Raymont A, Lay-Yee R, Pearson J, Davis P: New Zealand general practitioners’characteristics and workload: the National Primary Medical Care Survey.N Z Med J 2005, 118(1215):U1475.27. Aasland OG, Rosta J: The working hours of general practitioners 2000 to2008. Tidsskr Nor Laegeforen 2011, 131:1076–1080.28. Crossley TF, Hurley J, Jeon S-H: Physician labour supply in Canada: acohort analysis. Health Econ 2009, 18:437–456.29. Woodward CA, Hurley J: Comparison of activity level and service intensityof male and female physicians in five fields of medicine in Ontario.Can Med Assoc J 1995, 153:1097–1106.30. Canadian Institute for Health Information: The Practicing Physician Communityin Canada 1989/90 to 1998/99, Workforce and Workload as Gleaned throughBilling Profiles for Physician Services | Les Médecins En Exercice Au Canada, de1989-1990 à 1998-1999, Main-D’oeuvre et Charge de Travail Selon Les. 2001.31. Britt H, Bhasale A, Miles DA, Meza A, Sayer GP, Angelis M: The sex of thegeneral practitioner - a comparison of characteristics, patients, andmedical conditions managed. Med Care 1996, 34:403–415.32. Chaytors RG, Szafran O, Crutcher RA: Rural–urban and gender differencesin procedures performed by family practice residency graduates.Fam Med 2001, 33:766–771.33. Evans RG, Schnieder D, Barer ML: Health Human Resources Productivity: WhatIs It, How It’s Measured, Why (How You Measure It) Matters, and Who'sThinking about It. Ottawa, ON: Canadian Health Services ResearchFoundation; 2010.34. Bensing JM, van den Brink-Muinen A, de Bakker DH: Gender differences inpractice style: a Dutch study of general practitioners. Med Care 1993,31(3):219–229.35. Chan BTB, Austin PC: Patient, physician, and community factors affectingreferrals to specialists in Ontario, Canada - a population-based, multi-levelmodelling approach. Med Care 2003, 41:500–511.36. Bergeron R, Laberge A, Vezina L, Aubin M: Which physicians make homevisits and why? A survey. Can Med Assoc J 1999, 161(4):369–373.37. Brett T, Arnold-Reed D, Hince D, Wood I, Moorhead R: Retirement intentionsof general practitioners aged 45-65 years. Med J Aust 2009, 191(2):75–77.38. Leese B, Young R, Sibbald B: GP principals leaving practice in the UK:similarities and differences between men and women at different careerstages. Eur J Gen Pract 2002, 8:62–68.39. McKinstry B, Colthart I, Elliott K, Hunter C: The feminization of the medicalwork force, implications for Scottish primary care: a survey of Scottishgeneral practitioners. BMC Health Serv Res 2006, 6:56–64.40. French F, Andrew J, Awramenko M, Coutts H, Leighton-Beck L, Mollison J,Needham G, Scott A, Walker K: Why do work patterns differ between menand women GPs? J Health Organ Manag 2006, 20:163–172.41. Woodward CA, Williams AP, Ferrier B, Cohen M: Time spent onprofessional activities and unwaged domestic work: is it different formale and female primary care physicians who have children at home?Can Fam Physician 1996, 153(8):1928–1935.42. Williams R: Investigating hours worked measurements. Labour Mark Trends2004, 112:71–79.43. Weyrauch KF, Boiko P, Feeny D: HMO family physicians: men and womendiffer in their work. HMO Pract 1995, 9:155–161.44. College of Family Physicians of Canada, Canadian Medical Association, RoyalCollege of Physicians and Surgeons of Canada: Canada’s future physiciansmake choices to ensure personal - professional life balance. Coll FamlyPhysicians Canada News Events 2011.45. College of Family Physicians of Canada: Family Medicine in Canada: Visionfor the Future. 2004.46. Chambers R, Campbell I: Gender differences in general practitioners atwork. Br J Gen Pract 1996, 46:291–293.47. Cree MW, Yang Q, Johnson D: Gender-Related Differences in Practice Patternsin Alberta. 2001.Hedden et al. Human Resources for Health 2014, 12:32 Page 10 of 11http://www.human-resources-health.com/content/12/1/3248. Kazanjian A, Reid RJ, Pagliccia N, Apland L, Wood L: Issues in PhysicianResources Planning in B.C.: Key Determinants of Supply and Distribution, 1991to 96. Vancouver, BC: Centre for Health Services and Policy Research; 2000.49. Mayorova T, Stevens F, Scherpbier A, van der Velden L, van der Zee J:Gender-related differences in general practice preferences: longitudinalevidence from the Netherlands 1982 to 2001. Health Pol 2005, 72:73–80.50. Ubokudom SE: The effects of gender on primary care physician attitudesand practice orientations. J Heal Soc Pol 1997, 9:71–97.51. Weeks W, Wallace A: The influence of race and gender on familyphysicians’ annual incomes. J Am Board Fam Med 2006, 19:548–556.doi:10.1186/1478-4491-12-32Cite this article as: Hedden et al.: The implications of the feminization ofthe primary care physician workforce on service supply: a systematicreview. Human Resources for Health 2014 12:32.Submit your next manuscript to BioMed Centraland take full advantage of: • Convenient online submission• Thorough peer review• No space constraints or color figure charges• Immediate publication on acceptance• Inclusion in PubMed, CAS, Scopus and Google Scholar• Research which is freely available for redistributionSubmit your manuscript at www.biomedcentral.com/submitHedden et al. Human Resources for Health 2014, 12:32 Page 11 of 11http://www.human-resources-health.com/content/12/1/32


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items