Abstract
Purpose: Continuity is a core component of primary care and known to differ by patient characteristics. It is unclear how primary care physician payment and organization are associated with continuity.
Methods: We analyzed administrative data from 7,110,036 individuals aged 16+ in Ontario, Canada who were enrolled to a physician and made at least 2 visits between October 1, 2017 and September 30, 2019. Continuity with physician and practice group was quantified using the usual provider of care index. We used log-binomial regression to assess the relationship between enrollment model and continuity adjusting for patient characteristics.
Results: Mean physician and group continuity were 67.3% and 73.8%, respectively, for patients enrolled in enhanced fee-for-service, 70.7% and 76.2% for nonteam capitation, and 70.6% and 78.7% for team-based capitation. These differences were attenuated in regression models for physician-level continuity and group-level continuity. Older age was the most notable factor associated with continuity. Compared with those 16 to 34, those 80 and older had 1.45 times higher continuity with their physician.
Conclusion: Our results suggest that continuity does not differ substantially by physician payment or organizational model among primary care patients who are formally enrolled with a physician in a setting with universal health insurance.
Introduction
The ongoing relationship between a family physician and a patient developed over time is one of the central tenets of primary care.1 Relational continuity with a single provider is highly valued by both patients and providers2⇓⇓⇓–6 and has also been consistently associated with better outcomes.7⇓–9 High relational continuity has been associated with improved preventive care,10 better chronic disease management,11⇓–13 fewer emergency department visits,14⇓⇓–17 fewer hospitalizations,14,18⇓–20 and even lower mortality21,22 and costs.19,23,24
Given its strong association with better population health, it is important that health system leaders understand factors influencing relational continuity in primary care. Studies to date have highlighted relevant patient and physician factors. Research from a range of jurisdictions has found that continuity is higher for people who are older in age and have more chronic conditions.14,25⇓⇓–28 Physicians with larger panel sizes, who are newer to practice, and who provided fewer hours on-call had lower continuity.27,29⇓–31 However, little research has been done to understand how physician payment and team-based care influence continuity—2 types of reform gaining more traction and that are under the influence of health system leaders.32
We sought to understand factors influencing continuity in a large jurisdiction where visits to a primary care provider are fully insured and free at the point of care. We hypothesized that patients of physicians paid primarily through blended capitation would have higher levels of continuity than those paid largely through fee-for-service and that team-based care may lead to reduced physician-level continuity but better group-level continuity.
Methods
Context
In Canada, health care is publicly administered and funded through tax revenue. Ontario is Canada’s most populated province, with approximately 14.7 million residents in 2020. All permanent residents are eligible to receive medically necessary hospital and physician services free at the point of care through the Ontario Health Insurance Plan (OHIP). Most physicians bill the government for patients seen or cared for; the type and amount of remuneration varies based on the practice model. In the early 2000s, most primary care physicians were paid fee-for-service, worked in their own office, and did not have the support of a team. By 2011, approximately 84% of comprehensive primary care physicians in Ontario practiced in a patient enrollment model where physicians work in an administrative group with joint responsibility for after-hours care, formally enroll patients, receive blended payments, and are eligible for financial incentives from government for specific chronic disease and preventative care services.33,34 Patient enrollment models differ based on the proportion of payments that are by capitation versus fee-for-service and by whether they include funding for an interprofessional team. Patients are enrolled to a physician who in turn belongs to a group. Joining a patient enrollment model was voluntary for both physicians and patients, so their distribution across geographic regions of Ontario is variable; for example, there is a higher proportion of team-based models in rural areas.35
Study Design and Patient Population
We conducted a cross-sectional analysis using administrative health data to evaluate the association between continuity of care and primary care model and patient characteristics for Ontario residents, aged 16 years or older, over the 2-year period from October 1, 2017 to September 30, 2019. We included residents with a valid OHIP number who were alive on September 30, 2019 and attached to a general practitioner or family physician who practiced in an office location. We excluded physicians who had a focused practice designation. Our analysis only included patients who were enrolled to the physician and had at least 2 primary care visits between October 1, 2017 and September 30, 2019. Other studies of continuity in primary care have also limited analysis to patients with 2 or more visits over 2 years.15,18 We excluded patients admitted to long-term care; those who attended a community health center, a salaried model serving less than 2% of Ontario’s population; and those enrolled to a number of small nonstandard enrollment models.
We limited our analysis to patients who were formally enrolled because we wanted to measure continuity to the enrolled physician or group rather than to the most frequent provider. Enrollment denotes formal responsibility for patient care and is a core component of Ontario’s new models incorporating teams and capitation. Our study was not designed to understand continuity for patients who were not enrolled to a physician and who we know have more gaps in care.36
Data Sources
We analyzed data held at ICES, which houses administrative health service records for the population of Ontario. All datasets were linked using unique encoded identifiers and analyzed at ICES. We assigned patients to primary care physicians, and physicians to groups, using the enrollment database that is maintained by the Ministry of Health and Long-term Care to pay physicians. Physician billings were used to assess the number of outpatient visits and calculate continuity. We used the database for all residents registered with OHIP to examine patient age, sex, and postal code. We used postal code to derive neighborhood income quintile using a conversion file provided by Statistics Canada that uses 2016 census data (quintile 1: poorest to quintile 5: wealthiest).37 We used postal code and the Rurality Index for Ontario to determine whether patients resided in a rural area (40+), small town (10 to 39), or urban area (<10).38 We examined whether patients registered for the first time with OHIP in the last 10 years, a common proxy for immigration. The Johns Hopkins Adjusted Clinical Group (ACG) method39⇓–41 was used to measure comorbidity and morbidity (ACG System Version 10). Comorbidity was assessed using the ACG System Aggregated Diagnosis Groups (ADGs) categorized as 0 to 4 (no or low comorbidity), 5 to 9 (moderate comorbidity), and 10+ (high comorbidity). Morbidity was measured using the ACG System Resource Utilization Bands (RUBs) categorized as 0 to 1 (nonuser/healthy user), 2 (low morbidity), 3 (moderate morbidity), and 4+ (high morbidity).
Measures of Continuity of Care
Our primary analysis included 2 measures of continuity: continuity to the enrolled physician and continuity to the enrolled group. We assessed group continuity because physicians in the same group are supposed to share responsibility for after-hours care, and many have systems to support informational continuity and daytime cross-coverage between physicians.
We quantified continuity of care using a modified usual provider of care (UPC) index,42 defined as the fraction of a patient’s visits to the primary care physician or group out of all outpatient visits; we were only able to include in-person visits to physicians as phone calls and visits to other health professionals were not captured by administrative billing data at the time. Essentially, the UPC measures the extent to which visits are concentrated with a single physician or group of physicians. Typically, the UPC measures the concentration of visits to the most frequently seen physician or group; in our study, we measured the concentration of visits to the enrolling physician or enrolling group. The UPC index was calculated for the 2-year period from October 1, 2017 to September 30, 2019 for both visits to the physician and practice group:
The index ranges from 0 to 100% with 0 indicating no visits to the enrolling physician or group and 100 indicating highest continuity with all visits made to the same enrolling physician or group. The denominator included all outpatient visits made in an office location to any general practitioner or family physician who was not designated as focused practice (eg, because they are exclusively practicing sports medicine, addiction medicine, palliative care, or psychotherapy).
Patient Enrollment Models
We categorized patient enrollment models into 3 categories based on the predominant type of payment and whether there was government funding for nonphysician team members.34,43 In enhanced fee-for-service (Comprehensive Care Model, Family Health Group), physicians receive approximately 80% of remuneration through fee-for-service billings, 15% from capitation payments per enrolled patient adjusted for age and sex, and 5% from financial incentives and bonuses with no additional funding to hire nonphysician health professionals. In nonteam capitation (Family Health Network, Family Health Organization), 70% of funding is from capitation, 20% from fee-for-service billings, and 10% from incentives and bonuses with no additional funding to hire nonphysician health professionals. In team-based capitation, the type of payment is the same as nonteam capitation, but physicians are part of a group that receives funding to hire health professionals such as nurses, nurse practitioners, social workers, pharmacists, and dietitians. Patients were assigned to the physician and group they were enrolled to on September 30, 2019.
Statistical Analysis
For descriptive statistics we calculated the mean value of UPC index for physician continuity and group continuity across sociodemographic groups and patient enrollment models. To measure the association between patient enrollment model and the continuity index, as risk ratios, we used log-binomial regression with continuity entered in the model as a proportion. The risk ratios (RRs) and 95% CIs were adjusted for age, sex, rurality, income quintile, recent immigration, comorbidity (ADG), and morbidity (RUB). A significance level of 0.05 was used in all analyses. Analyses were conducted in SAS Enterprise Guide v9.4 (SAS Institute, Cary, North Carolina).
Additional Analyses
Health care utilization patterns and the distribution of enrollment models are known to vary by sex and by urban-rural location. We conducted additional sensitivity analyses where we stratified patients by (1) rurality and (2) sex and then measured the association between the continuity index and the enrollment model within each stratum.
Ethics
This project has been approved by the Research Ethics Board at Sunnybrook Health Sciences Centre, Toronto, Canada.
Results
We analyzed data for 7,110,036 Ontarians enrolled with a physician who made 2 or more visits to their provider in the period between October 1, 2017 and September 30, 2019 (Figure 1). Patient demographic characteristics varied by enrollment model (Table 1). A higher percentage of patients in an enhanced fee-for-service model were younger, were recent registrants, lived in urban areas, and lived in a neighborhood in the lowest income quintile. Patients in an enhanced fee-for-service model had a higher mean number of primary care visits in the last 2 years.
Mean continuity was 69.4% and 76.0% to the enrolling physician and group, respectively (Table 2). In unadjusted analyses, continuity varied by patient characteristics, with higher levels of physician and group continuity among older age groups, among patients living in rural areas and those who were long-term residents. Patients in enhanced fee-for-service models had the lowest levels of physician and group continuity (67.3% and 73.8%, respectively), whereas patients in team-based capitation models had the highest levels of group continuity (78.7%).
Figure 2 shows the percentage of patients by decile of continuity, stratified by rurality and enrollment model. The level of continuity was not normally distributed; between one third and one half of patients had continuity levels of 90% to 100%, and around one tenth or fewer patients had continuity levels of under 10%.
Table 3 presents results of the regression analysis. After adjustment for patient characteristics, there were only small differences by enrollment model for physician continuity (enhanced fee-for-service 67.3%; nonteam capitation 70.7% aRR 1.003, 95% CI, 1.003-1.003; team-based capitation 70.6% aRR 0.980, 95% CI, 0.979-0.980) or group continuity (enhanced fee-for-service 73.8%; nonteam capitation 76.2% aRR 0.994, 95% CI, 0.994-0.995; team-based capitation 78.7% aRR 1.003, 95% CI, 1.002-1.003). The most notable persistent association was between age and continuity. Compared with patients 16 to 34 years of age, those 80 years and older had 1.45 times higher continuity with their physician (57.0% vs 79.6%, RR: 1.447, 95% CI, 1.446-1.448) and 1.34 times higher continuity with their group (64.8% vs 84.7%, aRR: 1.337, 95% CI, 1.336-1.338). Patients living in rural areas and those with higher morbidity also had higher levels of continuity.
Sensitivity Analyses
We found no major differences in the association between continuity and enrollment model when regression analyses were stratified by (1) rurality and (2) patient sex (Appendix Table 1).
Discussion
Continuity is a crucial component of high-quality primary care, and it has been unclear how it is influenced by physician payment and organization. We conducted a population-based study of more than 7 million adults who were enrolled to a primary care physician and made at least 2 visits over a 2-year period. We found that there was little difference in continuity between patients of doctors paid by fee-for-service or capitation and between doctors who did and did not work in a team both before and after we accounted for differences in patient characteristics. The biggest differences in continuity were related to patient characteristics themselves, specifically age, with continuity increasing substantially with older age.
Our findings related to the association between continuity and patient characteristics are consistent with other studies. Studies from many jurisdictions have found that older age and higher morbidity are both factors related to higher continuity. However, only a handful of studies have assessed the association between physician payment, team-based care, and continuity, and findings have been mixed. For example, Hickson and colleagues44 found that salaried payment was associated with lower continuity. In contrast, Kristjansson and colleagues27 found that a capitation model, now being phased out, had higher levels of continuity probably because of physician financial penalties when patients saw another physician from outside the group. Physicians in capitation models included in our study also faced similar financial penalties,45 but this did not seem to influence overall levels of relational continuity.
Overall levels of continuity in our study were relatively high and likely relate to health system factors that influence primary care delivery broadly. We limited our analysis to patients attached to family doctors, and, in our setting, visits to family physicians and nurse practitioners are fully insured for permanent residents and free at the point of care. There may also be cultural factors that influence care-seeking behaviors, with most patients understanding the role of primary care providers as the first point of contact in the health care system. In theory, patients in Ontario can choose to see any family physician, even those different from their enrolling physician or group; however, physicians act as gatekeepers and usually only walk-in clinics or covering physicians agree to see a patient that is not part of their existing panel. It is heartening that we found minimal differences in continuity of care by neighborhood income quintile or between new and long-term registrants, a proxy for immigration. These findings suggest that patients who are attached to a family physician in a setting where primary care services are fully insured have high levels of continuity regardless of socioeconomic position or physician payment and organization.
Strengths and Limitations
Major strengths of our study include that it was population based, analyzing data for all enrolled patients in a jurisdiction, and that we included sensitivity analyses using stratifications for rurality and patient gender. There were also limitations. First, we assessed relational continuity based on office visits to physicians captured using billing data and were unable to capture visits to nurse practitioners or visits involving phone, video, or secure messaging; however, these other visit types were in the minority in our setting during the time of study. Second, there is heterogeneity between physicians and groups who share the same practice model, and our study was not designed to understand this variation or related context. Finally, we intentionally limited our analysis to patients formally enrolled to a primary care physician but are planning future analysis to understand differences in continuity between those who are and are not enrolled. Other work we have done has highlighted that those left behind from enrollment models experience more gaps in care.36
Conclusion
Overall, our results suggest that among patients who have a primary care provider and insurance coverage for physician visits, the primary care practice model does not have a major impact on relational continuity.
Acknowledgments
We are grateful to Maryam Daneshvarfard’s support with preparing the manuscript for publication and Rahim Moineddin’s advice on the regression analysis.
Appendix Table 1. Risk Ratios and 95% CIs for the Associations Between Patient Enrollment Model and Continuity Stratified by Rurality (Urban, Small Town, Rural) and Sex (Male, Female) for Patients with 2 or More Primary Care Visits, October 1, 2017–September 30, 2019, Ontario, Canada
Notes
This is the Ahead of Print version of the article.
This article was externally peer reviewed.
Funding: This study was supported by ICES, which is funded by an annual grant from the Ontario Ministry of Health (MOH) and the Ministry of Long-Term Care (MLTC). This study was funded by the Toronto Central Local Health Integration Network, the Canadian Institutes of Health Research funding reference number SOP 162662, and the INSPIRE Primary Health Care Research Program, which is funded through the Health Systems Research Program of the Ontario MOH and the MLTC. Parts of this material are based on data and information compiled and provided by Ontario MOH and the Canadian Institute for Health Information. The analyses, conclusions, opinions, and statements expressed herein are solely those of the authors and do not reflect those of the funding or data sources; no endorsement is intended or should be inferred. TK and RHG are supported as Clinician Scientists by the Department of Family and Community Medicine at the University of Toronto and at St. Michael’s Hospital. TK is the Fidani Chair of Improvement and Innovation in Family Medicine at the University of Toronto. MEG is supported by the Brian Hennen Chair in Family Medicine at Queen’s University.
Conflict of interest: None.
To see this article online, please go to: http://jabfm.org/content/00/00/000.full.
- Received for publication July 7, 2022.
- Revision received October 3, 2022.
- Accepted for publication October 4, 2022.