|
TRENDSIntergenerational Differences In Workloads Among Primary Care Physicians: A Ten-Year, Population-Based Study
Analyses of population-based services and surveys in Canada from the early 1990s and early 2000s indicate that younger and middle-aged family physicians carried smaller workloads in 2003 than their same-age peers did ten years earlier and that older family physicians carried larger workloads in 2003 than their same-age peers did ten years earlier. Yet family physicians in all age groups worked similar numbers of hours in 2003. Intergenerational effects are similar for male and female physicians, although feminization of the workforce will affect supply, as a result of the falling service volumes delivered by women.
EVIDENCE INDICATES that strong primary care systems improve a population’s level of health, reduce health disparities, and buffer the health effects of socioeconomic circumstances at lower cost than in health systems that rely more extensively on secondary and tertiary care.1 Current health human resource forecasting models rely on evidence regarding changes in the primary care physician workforce. Some models attempt to account for temporal shifts in work-life balance, based on evidence that medical students increasingly desire more controllable lifestyles.2 Yet quantitative evidence about the magnitude of generational phenomena is lacking. This information is vital to understanding and projecting the impact of demographic shifts in the workforce on the current and future supply of physicians. This study quantifies temporal shifts in workload (number of hours and volume of services) among general and family practitioners (GP/FPs) in Canada. It also assesses the use of their services. The analysis is based on longitudinal administrative and survey data from Canada—a country that offers first-dollar, universal coverage for physician services and predominantly pays physicians fee-for-service (FFS) for all medically necessary care.
For most Canadians, primary care is most often delivered by a GP/FP who works in a solo or small group practice without an interdisciplinary team. There are no physician assistants and very few nurse practitioners in Canada. GP/FPs represent half of the medical workforce and are the principal point of access to specialist care. Canada’s GP/FP-to-population ratio stood at 98 per 100,000 people in 2004.3 In 2003, 86 percent of Canadians reported having a "regular" family doctor. However, in 2003, 16 percent of people requiring routine, nonurgent primary care reported difficulties getting access to these services.4 In 2002, a full 80 percent of Canadians believed that there was a shortage of "family doctors," and 97 percent considered the problem to be serious.5 In that year, 93 percent of Canadian physicians believed that shortages among their ranks were widespread.6 Paradoxically, these perceptions of shortages came close on the heels of apparent physician surpluses in the 1990s, at least in urban areas. As a result, medical school enrollment was reduced in Canada a mere ten years ago.7 Yet the ratio of GP/FPs to population in 2003 was virtually the same as in 1993 and dramatically higher than in prior decades.8 Moreover, the supply of physician services has kept pace with Canada’s aging and expanding population.9 Similar paradoxical perception shifts have been documented in other countries.10 If supply ratios and the volume of services provided have not changed substantively, what underlies the current belief in GP/FP shortages and concerns over access to primary care? In 2004, health human resource planning was identified as one of the top-priority themes for health services research in Canada. Current policies are designed to increase medical school enrollment; encourage more equitable recruitment and retention of physicians in rural and remote areas; increase international migration of medical graduates and GP/FP residency positions; and support transition to more interdisciplinary primary care teams.11 This comes at a time of emerging evidence of a "dispirited profession," far fewer medical students choosing family medicine, and 60 percent of GP/FPs either limiting the number of new patients or not accepting any.12
Our study used three data sources: national physician surveys, an observational study in six provinces, and a case study in one city. The use of all three data sources is meant to mitigate weaknesses in each and provide a comprehensive view of practice trends. Our analyses relied primarily on demographic and workload data derived from physician survey results collected by the Canadian Medical Association (CMA) and its partners. Samples of 1,632 GP/FP respondents in 1993 (46 percent response rate) and 1,196 in 2003 (28 percent response rate) were representative of the age and sex structure of the 29,361 and 31,503 GP/FPs in each period, respectively. We assessed workload by examining physicians’ reported hours spent in an average week on direct patient care, regardless of setting and excluding time on call and time spent on indirect patient care. Service volume, a second aspect of workload, was measured using population-based, FFS claims data for 1992–2001 from the Canadian Institute for Health Information’s (CIHI’s) National Physician Database (NPDB). Office assessments in six provinces for which fee codes were available were counted.13 Results were derived from data representing the activities of 10,361 GP/FPs in 1992 and 10,541 in 2001 (representing 90 percent and 85 percent of GP/FPs in these provinces, respectively). The six provinces represented approximately 45 percent of the Canadian population. Finally, a case study was conducted using population-based, FFS claims data for 1991–2000 for an entire GP/FP workforce in Winnipeg, Manitoba—a provincial capital with a population of approximately 650,000 in each period. These files enabled us to (1) minimize the potentially confounding effects of shifts in organization and funding, since there was little change over the period; (2) control for temporal shifts in fee codes; (3) limit the impact of temporal variation in practice across rural and urban areas; (4) calculate a broader array of workload measures; and (5) measure standardized population-based rates of GP/FP use among general and special populations. We report on the number of visits delivered by each GP/FP per year and GP/FP visits per person per year.14
Nationally, GP/FPs reported that they worked an average of 38.7 hours per week providing direct care in 1993 and 35.4 hours in 2003 (a relative decline of 8.5 percent). Across the six provinces studied, GP/FP-to-population ratios declined between 1992 and 2001 (a 3.5 percent relative decline). Within this workforce, the average annual number of office assessments per GP/FP declined from 3,665 to 3,509 (a 4.3 percent relative decline). Office assessments per capita declined from 2.88 to 2.66 (a 7.6 percent relative decline). In Winnipeg, the GP/FP-to-population ratio declined 5 percent between 1991 and 2000. Within the GP/FP workforce there in 2000, each physician delivered an average of 4,193 visits—a workload virtually unchanged since 1991 (4,198 visits). Each physician delivered an average of 27.8 visits in a full-time day of work, again virtually unchanged since 1991 (27.9 visits). The population’s use of services remained stable: Standardized per capita visit rates declined from 3.61 to 3.49 (a 3.3 percent relative decline). Visit rates for populations with higher-than-average needs for primary care (such as people with diabetes, hypertension, or mental health conditions) also remained relatively stable.15
Shifts in age and age-specific workloads.
The age structure of Canada’s physician workforce changed during the 1990s. Most notable were a decline in the proportion of GP/FPs under age thirty-five and an increase in the proportion of baby-boomer GP/FPs (ages 45–54) (Exhibit 1
These data should not be interpreted to mean that younger (or older) GP/FPs provided less (or more) service volumes in 2001 than they did ten years earlier; over time, their workload volumes have been relatively stable. They are simply not increasing (or decreasing) the volume of services they provide each year as they age. Consider GP/FPs ages 35–44 in 1992 who were in the 45–54 age group in 2001. At the beginning and end of the period, this age cohort delivered roughly the same number of office assessments per annum (about 4,000). Likewise, GP/FPs ages 55–64 in 2001 performed a similar number of office assessments as did those ages 45–54 in 1992 (Exhibit 2  ).
Similar age and cohort effects were evident among the Winnipeg workforce from 1991 to 2000 and held true irrespective of the metric used to measure workloads (full-time-equivalents, visits per annum, or visits per full-time day of work). These trends held true after accounting for inflation, which suggested corresponding shifts in real incomes.16
Across Canada, GP/FPs ages 35–44 in 2003 (29 percent of the GP/FP workforce) spent 2.9 fewer hours each week than their same-age peers did in 1993 (a 7.9 percent relative decline) (Exhibit 3
Although the relationship between physician age and workload volume remained bell-shaped at the end of the ten-year study period, the nature of that relationship changed in important ways. First, same-age cohorts tended to diverge over time in their workload volume. That is, younger and older GP/FPs increasingly work differently than prior generations did, in terms of service volumes and work time. Second, for the relationship between physician age and workload, the bell shape is more shallow in more recent periods, but the slope of the relationship differs by workload measure. In terms of workload volumes, there is divergence across age cohorts in more recent periods: Younger and older GP/FPs increasingly work differently. In terms of work hours per week, there is convergence across age groups among the current cadre of physicians. Younger and older GP/FPs work similar numbers of hours (Exhibit 4
The convergence in work hours across age groups was attained by decreases in hours worked among the young and increases among the oldest (Exhibit 4  ). GP/FPs under age forty-five in 2003 provided an average of 8.8 fewer hours of direct patient care each week than their same-age peers did in 1993 (a 21 percent relative decline). GP/FPs ages 45–54 provided 3.8 fewer hours by 2003, while those ages 55–64 and age 65 and older worked 0.4 and 4.7 more hours per week, respectively.17
Shifts in sex and sex-specific workloads.
During the 1990s, the increasing proportion of female GP/FPs resulted in important changes in the sex structure of the workforce. Although the majority of female GP/FPs are in younger age groups, they represent an increasing proportion within every age cohort (Exhibit 1
The nature and extent of the temporal shifts between same-age cohorts described above were essentially the same for middle-aged male and female physicians. But female GP/FPs under age thirty-five provided 22 percent fewer office assessments in 2001 than their same-age female colleagues did in 1992, while younger males provided 10 percent fewer assessments in 2001 than their same-age male peers did in 1992. Female GPs over age sixty-five provided 24 percent more office assessments in 2001 than their same-age female peers did in 1992, while older males provided 48 percent more (Exhibit 5
In terms of work hours, there are also important sex differences among different age groups of GP/FPs. For example, in 2003, male GP/FPs ages 35–44 spent approximately thirty-seven hours each week delivering direct patient care, while female GP/FPs in the same age cohort spent thirty hours (a work level 81 percent that of their male peers). In 2003 there were also substantive differences in direct patient care work hours per week among male and female GP/FPs who had children under age six (38.8 hours and 27.6 hours, respectively), while sex differences diminished among males and females who had no dependents (35.0 hours and 33.0 hours per week, respectively). Similar patterns were evident when workloads were measured using total hours.
Despite the relative stability in aggregate supply and use of GP/FPs in Canada between 1992 and 2003, perplexing shifts in perception from surplus to shortage occurred. Recent reports of strained access might relate to unmet need or demand for services, or both. The provision of stable levels of services to populations with chronic conditions at a time of increased prevalence and static supply could mean that those with acute episodic conditions experience difficulties obtaining services. It could also be that the public requires, but doesn’t receive, more frequent and longer visits in response to increasing complexity or greater expectations of care. Alternatively or simultaneously, shifts in public and provider perceptions might be driven by supply-side factors. For example, Canadians might believe that there is a shortage of GP/FPs, despite the fact that most of them have a family doctor and visit rates remain stable, because of the type or frequency of media reporting on health care issues, such as GP/FP discontent with workloads or the number of primary care practices that are closed to new patients.18 Age-specific workload effects. One of our most intriguing findings was evidence of substantial changes in age-related workloads. Younger and middle-aged GP/FPs, who represent 80 percent of the workforce, carried lower workload volumes and spent less time each week on direct patient care in 2003 than their same-age peers did only ten years earlier. Yet overall measures of service volume remained stable over the period: Older GP/FPs greatly increased their productive volume relative to that of their same-age peers a decade earlier, although they spent only slightly more time each week on direct patient care in 2003 than their younger colleagues did. These effects appear to trump the impact of an aging workforce and sex-related effects. An aging workforce should have increased the available supply, since the relationship between physician age and workloads is traditionally bell-shaped: Doctors are most productive in mid-career.19 Our analyses indicate that expected increases in aggregate service volumes as a result of workforce aging have been offset by workload reductions among younger physicians. The fact that older GP/FPs carry ever-increasing workloads is cause for concern, as more physicians retire each year than was the case in earlier periods, and each retiree carries a workload volume that is larger than that of any retiring GP/FP over the past ten years. In 1992, every soon-to-retire GP/FP (age sixty-five or older) carried a workload slightly smaller than that of each new physician (under age thirty-five years). But in 2001, each soon-to-retire GP/FP carried a workload 1.6 times larger than that of each new physician. This scenario becomes more dramatic if one considers potential early retirees (GP/FPs ages 55–64). In 1992, early retirees carried workloads that were 1.5 times larger than those of each new physician, but by 2001, early retirees held workloads 2.1 times larger than those of each new physician. Consequently, the coming wave of physician retirements could cause unprecedented annual rates of shrinkage in GP/FP service volumes. Cohort effects. Baby-boomer GP/FPs themselves are not working less than they did at younger ages; over time, their workloads and real incomes stay relatively stable or increase slightly. They are simply not increasing the volume of visits they provide each year as they age. Also, GP/FPs are not slacking off. In 2001, GP/FPs in Canada reported working 52.7 hours per week on professional activities, excluding being on call.20 By comparison, U.S. GP/FPs reported working 55.3 hours per week on professional activities that year.21 Unfortunately, changes to the tables published in reports by the American Medical Association (AMA) do not allow one to monitor temporal shifts in age-related workloads and cohort dynamics among U.S. primary care physicians. Another interesting finding was that GP/FPs on average have reduced the time spent at work, which is in accordance with evidence emerging from other countries, including the United States and Australia.22 Our analyses illustrate important age-related trends that are masked by these averages. In Canada there has been temporal convergence in the amount of time spent delivering patient care, with younger GP/FPs working many fewer hours than their same-age peers a decade earlier and older physicians working slightly more. Thus, all age cohorts spend similar amounts of time at work each week, although younger GP/FPs carry lower workload volumes than their older colleagues do. Might this imply that opportunities exist for younger GP/FPs to learn efficiencies from the practice patterns of their older colleagues? Conversely, if younger GP/FPs provide more time-efficient care than older GP/FPs, then high-volume older GP/FPs who retire might be adequately replaced by physicians who treat their patients with fewer visits. We clearly need a better understanding of the relative cost-effectiveness of variation in office visits and time per visit. Although a recent systematic review found an inverse relationship between the years that a physician has been in practice and the quality of care he or she provides, the relationship between GP/FP age, quality of care, and patient outcomes is less clear.23 Age and sex effects. Over time, women represent an increasing proportion of the GP/FP workforce and of family medicine graduates.24 It is well recognized that the feminization of the GP/FP workforce will affect aggregate service volume. Our findings suggest that female GP/FPs represent an increasing proportion of the younger cadre and that each of these physicians increasingly carries service volumes that are lower than those of than their male counterparts. The combined variations in sex distributions across age groups and workload variations across sex groups support the notion that the sex structure of the workforce might influence available supply through multiple mechanisms. Policy implications. What do these findings mean outside the Canadian context, since there are important differences between countries in methods of funding and organizing primary care and in the supply of providers? The trends reported here held true in a city that underwent little change in funding and organization, which suggests that age and sex effects are independent of these factors. There is a relationship between remuneration and workload volumes, so we are uncertain whether differences between countries in methods of funding will accentuate or attenuate these effects.25 The GP/FP-to-population ratio is much higher, and the use of nonphysician primary care providers is much lower, in Canada than, for example, in the United States. This suggests that temporal trends documented here could have a more direct and noticeable impact on the accessibility of primary care in Canada than elsewhere.26 Our study underlines the necessity of combining data on age and sex determinants of GP/FP workloads with information about an aging workforce to more accurately incorporate supply-side factors into forecasts for future workforce requirements. There is an urgent need for a better understanding of the relative cost-effectiveness of practice styles among younger and older GP/FPs. The fact that the trends highlighted in this analysis have been rapid and not been previously quantified, coupled with our estimates of the magnitude of change in the Canadian context, points to the need for similar analyses elsewhere, new forecasts in Canada, and subsequent policy action. GIVEN THE LONG LAGS between shifts in workforce policy and downstream changes in supply, current perceptions of GP/FP shortages and complaints of high workloads in Canada are likely, on the basis of this analysis, to get worse before they get better. Pending retirements among highly productive GP/FPs are likely to result in substantive shrinkage in aggregate service volumes in the coming years, particularly if patterns of practice among younger GP/FPs do not change. Until we incorporate estimates of intergenerational workload shifts and understand their effects on clinical outcomes, we will be inadequately prepared to diagnose and treat issues of accessibility and work-life dissatisfaction that have arisen despite stable aggregate supply and visit rates in Canada.
Diane Watson (dwatson{at}chspr.ubc.ca) is on the faculty of the Centre for Health Services and Policy Research at the University of British Columbia in Vancouver. Steve Slade is a consultant at the Canadian Institute for Health Information in Toronto, Ontario. Lynda Buske is associate director of research at the Canadian Medical Association in Ottawa, Ontario. Joshua Tepper is assistant deputy minister in the Health Human Resources Strategy Division, Ontario Ministry of Health and Long-Term Care, in Toronto. This work was supported in part by Manitoba Health, the Canadian Institute for Health Information, and the Canadian Medical Association. The results and conclusions are those of the authors.
This article has been cited by other articles:
Comments:Read all Comments
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Public, Policy, And Practice...
Study Data And Methods
