How long does it take patients to find a new primary care physician when theirs retires: a population-based, longitudinal study

There are longstanding concerns about the potential impact of primary care physician retirements on patient access to and continuity of care, and on service supply more broadly [1‐8]. While considerable research has been done on the services that older physicians provide as they age and transition out of the workforce [1, 2, 9‐13], very little attention has been given to how physician retirement affects patients [7]. Existing work suggests that disruption to longitudinal relationships between patients and physicians has been associated with difficulty accessing care [14], emotional distress [14, 15], poorer health outcomes [8], and decreased use of downstream primary care services [7, 8].

In 2019, eighteen percent of British Columbians and 15% of Canadians overall stated that they did not have a regular source of primary care [16]. Existing research suggests that patients without a regular source of primary care are more likely to rely on walk-in clinics or emergency departments [17‐19]. This has a negative effect on care continuity and coordination [19‐23], and access to preventive services [24‐26].

Physician retirements add an additional challenge, as they mean that some patients who have a regular source of care then lose it. In a recent survey of patients who were actively seeking a primary care physician, 20% of respondents reported that they were looking because their physician had either retired or moved [27]. While finding a family physician may not be a high priority for young, healthy patients, it is critical for many, particularly those who have chronic conditions that require regular monitoring [27]. Like the broader population of unattached patients, patients of retired physicians reported using walk-in clinics or emergency departments more frequently [28]. They also reported concerns about poor continuity and poor access to preventative care [28].

Many provincial medical associations and state medical boards have published guidelines and recommendations for physicians who are considering retiring from clinical practice (e.g., [15, 33‐36]) These guidelines include notifying patients of their leave date, whether another physician is taking over their practice, or providing a list of physicians taking new patients in the clinic or community [15, 35, 36]. Physicians are concerned about the negative impact their retirement may have on their patients. Several studies have reported that a feeling of responsibility for their patients (concerns about what will happen to the patients after the physician retires) keeps many physicians in practice longer than might otherwise have been the case [37‐42].

The possible effect of physician retirement on access to primary care services across Canada in general, and in British Columbia (BC) in particular, has received some academic and media attention [1‐8]. This paper expands on this existing literature by examining the use of primary care services by patients whose primary care physician retires. We describe both the periods preceding and following the retirement. Our research objectives are: (1) to measure the amount of time it takes for patients to find a new primary care physician; and (2) to assess whether time to finding a new primary care physician is associated with specific clinical and sociodemographic characteristics.

We identified 463,044 patients who were continuously registered with MSP between 2005/2006 and 2011/2012 and who had the same MSOC for 2005/2006 and 2006/2007 (Fig. 1). Of these, 12,013 patients had an MSOC who was over age 50 at the beginning of the study period, and who ceased clinical practice in either 2007/2008 or 2008/2009 (N = 68 unique physicians) (Table 1).¹ A majority of the patients of retiring physicians were over age 50 (69.3%). Just over half were women (54.5%). Patients were equally distributed across the five income quintiles. Half had no major ADGs (52.1%) while 4.8% had three or more.

The vast majority (94.3%) of the patients found a new MSOC between when their original MSOC retired (in 2007/2008–2008/2009) and the end of the study period (2011/2012) (Table 2). However, 45.8% took longer than one year, and 5.7% still had not found one 36 months after their original MSOC retired (Fig. 2).

Compared to patients who successfully found a new MSOC, those who did not were more likely to be in the younger age categories (for example 18.7% versus 3.6% under age 20, p < 0.0001), male (51.7% versus 45.1%, p < 0.0001) and have fewer major ADGs (for example 63.2% versus 51.4 had no major ADGs, p < 0.0001). They were also more likely to be in the lowest income quintile (28.8% versus 20.5%, p < 0.0001).

In the time-dependent analysis and controlling for the effect of other factors, men had lower adjusted rates of finding a new MSOC compared to women (HR 0.92, 95% CI 0.88–0.95) (Table 3, Fig. 3). We observed a clear trend by age group, with individuals in the oldest age categories having the highest likelihood of finding a new MSOC compared to those in the lowest age category (e.g., individuals 70–79 HR 3.26, 95% CI 2.91–3.65). A patient’s regional health authority of residence was also associated with the odds of finding a new MSOC, with individuals located in two of the health authorities having less success with finding a new physician (Vancouver Coastal Health: HR 0.73, 95% CI 0.67–0.78; Vancouver Island: HR 0.86, 0.80–0.93); however regional primary care physician supply had no effect (HR 1.0, 95% CI 0.98–1.01).

Rates of finding a new MSOC were positively associated with morbidity, with individuals having 3, or 4+ ADGs in particular having higher rates (3 Major ADGs: HR 1.20, 95% CI 1.08–1.33; 4+ Major ADGs: HR 1.19, 1.00–1.41). Compared with the lowest income quintile, patients in the highest income quintile had small but statistically significantly greater likelihood of finding a new MSOC (HR 1.07, 95%CI 1.00–1.13).

Physician characteristics also influenced time to finding a new MSOC. Patients whose original (retired) physician was male (HR 0.90, p = 0.001) had lower adjusted rates of finding a new MSOC, while patients whose retiring physicians were in any but the youngest age category had significantly higher rates (e.g., age 75 + HR 1.35, p < 0.0001). We observed no difference between physicians who trained internationally and those who trained in Canada.

To our knowledge, this is the first study that examines patients’ primary care interactions in the period following the retirement of their primary care physician, using a population-based approach. We found that while the vast majority of patients do go on to find a new MSOC, that process can take considerable time. Six percent of our sample had not found a family physician 36 months after the retirement of their original MSOC. Eighteen percent took between 18 and 36 months.

Older patients and those with more significant morbidity had higher rates of finding a new primary care physician, and found one more quickly, consistent with the suggestion that younger, healthier patients are less likely to make seeking out a new, consistent, source of primary care a priority, even within this cohort of patients with three or more visits per year [52]. Individuals in the lowest income quintile were somewhat less likely to find a new MSOC. While at first glance this may seem to be an artifact of fewer physicians providing longitudinal primary care in low-income areas, this relationship persisted in the Cox regression where regional primary care supply was accounted for. This finding raises questions about whether marginalized populations are less likely to seek out a new family physician, or less successful at doing so. There is evidence suggesting that the likelihood of having at least one primary care visit is independent of income [53, 54]; however, individuals who have low socioeconomic status are more likely to face discrimination when seeking out an appointment with a new primary care physician [55], suggesting that they still have relatively more challenges with accessing primary care may compared to individuals in higher income groups.

We found significant differences in the rates of individuals finding a new MSOC across regional health authorities, even when we accounted for differences in regional primary care capacity. Individuals in the largest (by population) and most urban health authority had the lowest time-adjusted rates of finding a new MSOC. On the surface, this finding seems contrary to the body of evidence suggesting that accessibility challenges are more salient in rural areas compared to urban ones; however, it may relate to greater access to walk-in clinics in urban centers in particular, and to the role these clinics play in shaping access to longitudinal primary care.

Several organizations have tools and programs designed to assist physicians with the process of retirement, one element of which is ensuring a smooth transition for their patients (for example [56, 57]). For example the B.C. General Practice Services Committee, through their broader GP for Me Initiative, offers a program that attempts to match new-to-practice physicians with retiring general practitioners, while also providing business coaching and support [57]. To our knowledge, however, neither this program nor any of the other supports offered, have been evaluated in terms of their ability to assist physicians with retirement, or ensure an easy transition for their patients.

Furthermore, given the more general ‘taut’ primary care supply situation in the province, retiring physicians’ ability to find younger primary care practitioners who are able to take on large numbers of new patients is likely to be an incredible challenge. Connection to a centralized waiting list is one potential solution that is being trialed in a number of Canadian provinces [58]. While there are wide differences in implementation of these lists, a consistent challenge among them is extensive waiting periods due to demand for attachment that far outstrips supply [59]. Lower rates of attachment of vulnerable patients, such as those with harmful substance use issues, mental health concerns, or chronic diseases are also a known concern with some implementations of this type of management strategy [59, 60].

Primary care physician retirements represent a potential challenge to accessibility and continuity of care for patients, particularly in light of apparent systemic shortages in many provinces. We find that, on average, patients take more than a year to find a new regular source of primary care following the retirement of their original physician. Existing programs intended to smooth the retirement transition for physicians and for patients should be evaluated; however, we expect their efficacy is likely to be hampered by broader accessibly challenges. These can only be addressed through initiatives designed to increase capacity for longitudinal, community-based primary care.