Distance From Family Doctor Tied to More Fragmented Care in the Last Year of Life

For hundreds of thousands of Canadians, the final year of life unfolds largely outside the hospital, in the offices of family physicians who coordinate symptoms, medications, and referrals. A new population-level study drawing on the health records of more than 400,000 people who died in Ontario between 2021 and 2024 suggests that how far a patient lives from their family doctor leaves a measurable imprint on how coherent that care turns out to be. The findings, published in PLOS One, indicate that greater driving distance to a family physician is associated with modestly lower relational continuity of outpatient care during the last twelve months of life.

Relational continuity refers to the degree to which a patient sees the same physician, or a small and stable set of physicians, over time rather than encountering a rotating cast of providers. Researchers consider it a cornerstone of good primary care, particularly near the end of life, when patients typically accumulate complex symptoms, polypharmacy, and frequent transitions between home care, clinics, and emergency departments. Prior work has linked higher continuity with fewer hospitalizations, better symptom control, and even survival advantages, but the role of geography in shaping continuity has been difficult to quantify at scale.

To address that gap, a research team led by Shuaib Hafid of McMaster University and colleagues assembled a retrospective cohort using Ontario’s linked health administrative data, which capture physician billing claims, hospital encounters, and vital statistics for the province’s population. The cohort included every adult who died in Ontario from 2021 to 2024, yielding 400,499 decedents with usable records. The mean age at death was 76.0 years, and 52.2 percent of the cohort was male, roughly mirroring the demographic profile of mortality in Canada’s most populous province.

The study’s central methodological innovation lay in how it measured geographic access. Rather than relying on straight-line distance, which can be misleading in a province crisscrossed by lakes, highways, and rural road networks, the researchers calculated the shortest driving distance in kilometers between each patient’s residential address and their family physician’s practice location. They then divided the cohort into quartiles of this distance. The mean distance was 25.5 kilometers, but the distribution was heavily skewed: the standard deviation of 92.0 kilometers reflects a small population of patients who travel very long distances, while the top quartile of patients lived more than 14.2 kilometers from their physician.

Continuity itself was assessed with established measures that capture three distinct dimensions: how concentrated outpatient visits were with a single usual physician, how dispersed care was across multiple providers, and how temporally regular the visits were over the final year. The researchers then modeled the association between distance quartile and continuity using proportional-odds regression, a statistical approach suited to ordered outcomes, adjusting for demographic characteristics, socioeconomic status, and clinical factors such as comorbidity burden. This adjustment matters, because patients in rural and remote areas tend to differ systematically from urban dwellers in age, income, and health status, and the analysis was designed to isolate the independent contribution of distance itself.

The results were consistent and statistically robust. Each one-quartile increase in driving distance was associated with lower odds of higher relational continuity with the usual family physician, with a common odds ratio of 0.88 and a 95 percent confidence interval of 0.87 to 0.88. In practical terms, a patient in the farthest quartile, living more than 14.2 kilometers from their doctor, had meaningfully reduced odds of experiencing concentrated, regular care from a familiar physician compared with someone in the nearest quartile. The effect is modest at the individual level but consequential at the population level, given that hundreds of thousands of Ontarians die each four-year window and most of them depend on family physicians for outpatient care in their final year.

To test whether the finding was an artifact of how distance was measured, the team ran sensitivity analyses substituting estimated travel time and Euclidean, or straight-line, distance for driving distance. Both alternatives produced consistent results, strengthening the conclusion that the association reflects genuine geographic barriers rather than the quirks of any single metric. This kind of triangulation is important in health services research, where the choice of a spatial measure can sometimes drive the conclusions, and here the signal persisted across all reasonable operationalizations of proximity.

Why would distance erode continuity? The authors point to several plausible mechanisms. Patients who live farther away face longer travel times for each appointment, which raises the effective cost of a visit and may lead them to defer or skip routine follow-ups, or to seek care from whichever nearby provider can see them sooner. Emergency departments and walk-in clinics in the patient’s own community may absorb visits that would otherwise have gone to the usual physician, fragmenting the record and diluting the relationship. Weather, seasonal road conditions, and transportation limitations compound these effects in rural and northern regions. Over a full year, small frictions accumulate into measurably less regular and less concentrated care.

The clinical stakes are considerable. Fragmented end-of-life care is associated with more emergency department visits, more hospitalizations, poorer symptom management, and lower concordance with patient preferences for where and how they receive care. Family physicians who know a patient well are better positioned to manage pain and other symptoms, to align treatment with goals of care, and to coordinate with palliative and home-care services. When continuity slips, those functions weaken, and the burden often shifts to hospitals and to families.

The authors argue that the findings call for system-level strategies rather than individual effort. Improving local access to primary care, for example by expanding community-based and interprofessional clinics in underserved areas, could shorten the distance between patients and a stable medical home. Enhanced virtual-care models, which expanded rapidly during the COVID-19 pandemic, offer another route: telephone and video visits can sustain the relational bond even when in-person travel is burdensome, although the study’s data predate a full accounting of how virtual care reshapes continuity. The researchers also suggest that continuity metrics could be incorporated into regional planning, so that jurisdictions with large distances between patients and physicians receive targeted support. As populations age and more deaths occur outside hospitals, the geography of primary care, the study concludes, is not a background detail but a significant determinant of how well the health system accompanies people through the end of life.

Subject of Research: Geographic proximity to family physicians and continuity of outpatient care in the last year of life

Article Title: Measuring geographic proximity to family medicine and continuity of care in the last year of life: A population-level retrospective cohort study using Canadian health administrative data

Article References: Hafid, S., Mbuagbaw, L., Newbold, B., Costa, A., Gayowsky, A., Gozdyra, P., Jones, A., Isenberg, S. R., Gallagher, E., & Howard, M. (2026). Measuring geographic proximity to family medicine and continuity of care in the last year of life: A population-level retrospective cohort study using Canadian health administrative data. PLOS One, 21(10), e0360638. https://doi.org/10.1371/journal.pone.0360638

Image Credits: AI Generated

DOI: 10.1371/journal.pone.0360638

Keywords: continuity of care, family medicine, end-of-life care, geographic access, health administrative data, Ontario, primary care, retrospective cohort, rural health, PLOS One, Measuring, geographic