Abstract
PURPOSE Primary care access is a key health system metric, but little research has compared models to provide primary care access when one’s regular physician is not available. We compared health system use after a visit with a patient’s own family physician group (ie, within-group physician who was not the patient’s primary physician) vs a visit with a walk-in clinic physician who was not part of the patient’s family physician group.
METHODS We conducted a population-based, retrospective cohort study using administrative data from Ontario, Canada, including all individuals formally enrolled with a family physician, from April 1, 2019 to March 31, 2020. We compared those visiting within-group physicians to those visiting walk-in clinic physicians using propensity score matching to account for differences in patient characteristics. The primary outcome was any emergency department visit within 7 days of the initial visit.
RESULTS Matched patients who visited a within-group physician (N = 506,033) were 10% less likely to visit an emergency department in the 7 days after the initial visit compared to patients who saw a walk-in clinic physician (N = 506,033; 20,117 [4.0%] vs 22,320 [4.4%]; risk difference [RD] 0.4%; 95% CI 0.4-0.5; relative risk [RR] 0.90; 95% CI, 0.89-0.92). Restricting to visits occurring on weekends, the observed association was stronger (7,964 [3.7%] vs 10,055 [4.7%]; RD 1.0%; 95% CI 0.9-1.1; RR 0.79; 95% CI, 0.77-0.82). Those accessing after-hours within-group physician visits were more likely to have ≥1 additional virtual or in-person within-group physician visit within 7 days (virtual RR 1.86, in-person RR 1.87).
CONCLUSIONS Compared to visiting a walk-in clinic physician, seeing a within-group physician after hours might decrease downstream emergency department visits. This finding could be explained by better continuity of care and can inform primary care service models and the policies that support them.
INTRODUCTION
Balancing timely access with care continuity remains an important challenge in primary care. Although 90% of the population in Ontario, Canada’s most populous province, report having a family physician or primary care provider,1 less than one-half are able to make a same- or next-day appointment.2 Walk-in clinics provide episodic care without an appointment and are available to all patients, whether they are attached to a primary care physician or not.3-6 Attached patients sometimes visit walk-in clinics when their perceived need for care falls outside regular office hours such as during evenings or weekends,7 if they have poor same-day or next-day appointment access with their regular physician,8 or if they perceive the walk-in clinic to be more easily accessible.3,9-11 An estimated 30% of Ontarians visit a walk-in clinic each year.12
Continuity of care is associated with greater patient satisfaction, better health outcomes, and decreased subsequent health care utilization and costs.7,13-15 Informational continuity exists if information about the patient, including their previous health care use, is available at the point of care.16 Management continuity captures “the extent to which services delivered by different providers are timely and complementary such that care is experienced as connected and coherent.”17 Walk-in clinics can decrease care continuity.4,18,19 In addition to introducing a new physician, there is no expectation of an ongoing relationship between the patient and the walk-in clinic physician, who typically does not have access to patients’ existing health records, nor is the walk-in clinic record shared with patients’ usual family physicians.
In an attempt to offer an alternative that ensures access while preserving informational and management continuity, some health care funders have introduced requirements and incentives for after-hours coverage within each group of primary care physicians.20,21 Our aim was to compare 2 alternatives for rapid access when a patient’s own physician is not available. We hypothesized that compared to visits with a walk-in clinic physician, after-hours visits with a within-group physician would be associated with lower subsequent usage of the emergency department (ED).
METHODS
Setting
This population-based, retrospective cohort study used administrative claims data from Ontario, a province with 14.5 million residents and approximately 14,000 practicing family physicians in 2019.22 Ontario provides permanent residents with universal health coverage of medically necessary physician and hospital services, without copayments or deductibles, and it does not restrict patients’ choice of physician. The single payer is the Ontario government.
Most (81%) Ontarians are linked to a family physician who has signed on to their care via a patient enrollment model.23 Patient enrollment models are group practices with blended remuneration including capitation and fee-for-service payments.24 Primary care group practices are staffed primarily by family physicians who refer patients to specialists outside the practice as needed. In 2004, the Ontario government introduced after-hours premiums and minimum after-hours time requirements for physicians who practiced in patient enrollment models.25-27 Physicians in primarily capitation-based groups (>40% of physicians)28 of ≥3 physicians were required to provide at least one 3-hour block of evening or weekend coverage per week.20,21 The number of required after-hours blocks differed by group size and group type. In addition, physicians received an access incentive, which was decreased if enrolled patients visited outside physicians (eg, at a walk-in clinic).29,30 In Ontario, walk-in clinics are typically unaffiliated with primary care clinics, and patients can choose to visit any walk-in clinic without an appointment at no cost to them.
Data Sources
Datasets were linked using unique encoded identifiers and analyzed at ICES, an independent nonprofit research institute whose legal status under Ontario’s health information privacy law allows it to collect and analyze health care and demographic data, without consent, for health system evaluation and improvement. A description of the ICES databases we used in this study can be found in Supplemental Table 1.
We acquired data from the College of Physicians and Surgeons of Ontario (CPSO) annual license renewal survey31 via a data-sharing agreement. Using physicians’ self-reported practice settings and hours worked in each setting per week, the CPSO provided variables indicating whether a walk-in clinic or episodic care clinic outside of a hospital was a setting in which a physician worked the majority of the time.19 These questions were mandatory and therefore were not missing for any physician.
Study Population
The study population included all Ontario residents who were formally enrolled with a family physician as of April 2019,32 and had ≥1 family physician encounter from April 1, 2019 to March 31, 2020 (see Supplemental Figure 1 for study population flow chart).
Exposure
The principal exposure was assessed at the first eligible visit (index visit) during the study period and consisted of the type of family physician office encounter. The exposure groups were defined as follows.
Visit With a Within-Group Physician
We included all office encounters with physicians who belonged to the same group as the patient’s enrolling physician (but not with the patient’s own family physician) within the study period. In this group, we restricted time of visit to after hours or weekends—times when acute care might be more likely and more similar to care provided at a walk-in clinic (see Supplemental Table 2 for visit type definitions).
Visit With a Walk-In Clinic Physician
We identified all office encounters with walk-in clinic physicians using a definition of walk-in clinic physician that we have applied previously.7,19,33 Because our aim was to compare 2 alternatives for care when a patient’s own physician was not available, we excluded encounters between patients and their own enrolling physician. Visits could be at any time including after hours and weekends. Because physicians often work in multiple settings, if a walk-in clinic physician encounter also met criteria for a within-group physician visit, the visit was classified as the latter.
Outcomes
The primary outcome was any ED visit within 7 days of the index visit (including the day of the index visit). Secondary outcomes included time to ED visit (up to 30 days); low-acuity ED visit (Canadian Triage and Acuity Scale level 4 or 5) within 7 days; ED visit within 30 days; virtual or in-person visit with any physician, the patient’s own physician, or a physician from the patient’s group within 7 days (excluding the day of the index visit); and the number of all family physician visits within 30 days.
Other Variables
Other patient demographic and health care utilization characteristics included calendar year quarter; patient age and sex; neighborhood income quintile; recent health insurance registrant status (a proxy for recent immigration); rurality (as measured by the Rurality Index for Ontario)34; count of family physician visits in previous 2 years; count of ED visits in the previous year; patient enrollment model type; size of enrolling physician’s group; distance from patient’s residence to enrolling physician’s practice; diagnosis of an upper respiratory tract infection at index visit; and whether the index visit occurred on a weekend day (operational definitions in Supplemental Table 3). We used The Johns Hopkins ACG System (version 10) to derive counts of comorbidities based on Aggregated Diagnosis Groups and prior health care utilization based on Resource Utilization Bands.35 Missing data are reported in the supplemental data tables.
Statistical Analysis
We described the cohort using mean (SD), median (interquartile range), and count/frequency. We compared visit characteristics between exposure groups using standardized mean differences (SMDs) such that differences >10% (SMD 0.1) were considered potentially meaningful.36
To ensure comparability between patients having visits with a within-group physician and those having visits with a walk-in clinic physician, we derived a logistic regression–based propensity score.37 Variables in the propensity score model included all those listed in “Other Variables,” with the exception of rurality. Index visit diagnosis was categorized as any of the top 20 diagnoses or as “other.” We included missing values for income quintile in the regression model as their own level. Restricted cubic splines were used to model the association between the log-odds of visiting a walk-in clinic physician and the following continuous variables: age, count of own family physician visits in previous 2 years, count of ED visits in the past year, distance from patients’ residence to own physicians’ practice address, and size of own physicians’ group.38 Individuals from both groups were then matched 1:1 based on the logit of the propensity score within a caliper distance of 0.2 of the SD of the logit of the propensity score.39 We also hard matched on the following key variables: age group (0-17 years, 18-64 years, ≥65 years) and rurality (Rurality Index for Ontario score; 3 categories: 0-9 large urban, 10-40 small urban, >40 rural).
For binary outcomes, we reported relative risk (RR) and risk difference (RD) with 95% CI, estimated using methods that accounted for the matched sample.40,41 For the time to ED visit outcome, we reported the hazard ratio using a Cox proportional hazards model robust variance estimator to account for clustering within matched pairs.42 We conducted all analyses using SAS statistical software (version 9.4; SAS Institute).
Subgroup and Sensitivity Analyses
For age and rurality subgroups, we reported the RR of having an ED visit within 7 days of the index visit. To further compare outcomes after visits that were most likely to be acute in nature, we conducted a subgroup analysis with weekend-only visits. For both groups, we restricted the subanalysis to visits that occurred on a Saturday or Sunday. In the propensity score model for this subgroup analysis, whether the visit was on a weekend was not included, and patients were only hard matched on rurality.
Ethics Approval
This study was approved by the Women’s College Hospital Research Ethics Board (REB 2020-0095-E) with a waiver of patient consent.
RESULTS
Patient Characteristics
Of the 1,701,381 individuals in the cohort, we identified 607,166 (35.7%) whose index visit was with a within-group physician after hours or on a weekend and 1,094,215 (64.3%) whose index visit was with a walk-in clinic physician they were not enrolled with. Compared to patients who had a visit with a walk-in clinic physician, patients who saw a within-group physician were older (mean [SD] = 40.7 [23.4] years vs 38.3 [20.7] years; SMD 0.11) and less likely to live in a large urban area (74.7% vs 88.5%; SMD 0.36) (Supplemental Table 4). Patients who saw a within-group physician had more own family physician visits in the prior 2 years (mean [SD] = 5.2 [5.7] vs 4.5 [5.6]; SMD 0.13) despite having similar prior health care utilization and numbers of comorbidities (Aggregated Diagnosis Groups).
Patients who visited a within-group physician more often belonged to team-based models of primary care compared to those who saw a walk-in clinic physician (35.3% vs 16.5%; SMD 0.44). Within-group patients were more often enrolled with physicians who belonged to smaller physician groups (mean [SD] = 43 [69] physicians vs 64 [109] physicians; SMD 0.24) and lived an average 10 km closer to their enrolling physician’s practice (mean [SD] = 10.6 (31.0) km vs 20.6 (57.3) km; SMD 0.22).
A total of 83% of patients who saw a within-group physician were matched. After matching, there were 506,033 patients in each group. There were no differences in measured characteristics exceeding 10% (SMD 0.1), with the exception of visits on a weekend (34.1% vs 28.1%; SMD 0.13) (Table 1). The top 20 visit diagnoses are listed in Supplemental Table 5.
Characteristics of Patients in Matched Cohort (April 1, 2019-March 31, 2020)
Outcomes
Within 7 days of the index visit, patients who saw a within-group physician after hours or on weekends were 10% less likely to visit an ED than those who saw a walk-in clinic physician (RD 0.4 [95% CI, 0.4-0.5]; RR 0.90 [95% CI, 0.89-0.92]) (Table 2). They also had a decreased hazard of ED visit (hazard ratio 0.94 [95% CI, 0.93-0.95]), with early separation of time-to-event curves, suggesting that much of the difference in risk was front loaded (see Supplemental Figure 2a for Kaplan-Meier curves).
Propensity Score–Matched Outcomes
Patients who had a visit with a within-group physician were more likely to have either an in-person or virtual visit with their own family physician or a within-group physician within 7 days of the index visit compared to after a walk-in clinic visit, although RDs were small.
The decrease in 7-day ED visits after an encounter with a within-group physician was driven by individuals who lived in large urban areas (RR 0.86 [95% CI, 0.84-0.88]) and children/adolescents (aged <18 years; RR 0.87 [95% CI, 0.83-0.91]) (Supplemental Table 6). In contrast, those who lived in rural areas and saw a within-group physician had an increased likelihood of an ED visit in the next week (RR 1.26 [95% CI, 1.13-1.41]).
Weekend-Only Visits
Of the 545,352 individuals in the weekend-only visit cohort, 279,119 (51.2%) had an index visit with a within-group physician, and 266,233 (48.8%) had an index visit with a walk-in clinic physician. Group differences in characteristics were similar to the overall cohort analysis (Supplemental Table 7).
After matching, there were 213,190 patients in each group—76% of the patients who saw a within-group physician were matched (Supplemental Table 8). The groups were well matched including the top 20 visit diagnoses (Supplemental Table 9). Outcomes after the index visit were similar to those for the main analysis, with the exception that differences were larger (Supplemental Table 10). Compared to patients who had an index visit with a walk-in clinic physician, those who had a visit with a within-group physician were 21% less likely to have an ED visit within the following 7 days (RD, 1.0% [95% CI, 0.9-1.1]; RR, 0.79 [95% CI, 0.77-0.82]) (Supplemental Figure 2b). Those who had a visit with a within-group physician were more likely to have an in-person or virtual visit with their enrolling physician or group within 7 days.
DISCUSSION
In this population-based matched cohort study, we found that patients with a family physician who saw a colleague from their family physician’s group after hours were 10% less likely (0.4% lower absolute risk) to visit an ED within 7 days compared to those who saw a walk-in clinic physician. This small difference in risk was front loaded in the first few days after the initial visit and was driven by those who lived in large urban areas. Compared to patients who had a visit with a walk-in clinic physician, those who saw a within-group physician were also more likely to have a visit with their regular family physician or a within-group physician in the following 7 days.
This study builds on a body of literature linking access to high-continuity, after-hours primary care with decreased use of ED visits.26,43-46 The mechanism behind our present findings is uncertain—fewer ED visits could be the result of more appropriate care, greater access to one’s own primary care clinic, or improved patient satisfaction owing to greater informational or management continuity. In addition to timely access and continuity, patients’ trust and physicians’ confidence in the advice, treatment, or follow-up plan could also be contributing factors.
A 2005 Ontario-based study found walk-in clinic visits to have a greater rate of 3-day health reutilization for minor illnesses compared to a family physician office visit, with an effect that might have been mediated by differences in patient satisfaction.47 In the United States, visits to walk-in clinics (ie, retail clinics unaffiliated with longitudinal primary care) for low-acuity conditions were associated with increased health care utilization and costs.48 Opening more walk-in clinics in an area was also not found to decrease the rate of low-acuity ED visits.49 In a recent study of virtual family physician visits, we found that within-group virtual visits were also followed by fewer trips to the ED than virtual visits with an outside physician,50 suggesting that care continuity might be valuable for decreasing repeat utilization for both virtual and in-person care.
In primary care practice and systems, there is an ongoing tension between timely access and care continuity.51 Relationship continuity is associated with better health outcomes and decreased health care utilization and costs14,52-54 and is valued by patients.55,56 Yet, many patients might choose or feel forced to trade off continuity if they are unable to access timely care for an issue they perceive to be urgent. Convenience and distance from where patients reside are additional factors influencing patient choice.7,8,27 Family physicians might strive to have more availability to be able to provide both continuity and timeliness, but it is not a requirement for primary care physicians to be available 24/7, and many physicians are choosing part-time work.57 In this context, timely care within a group is likely a second-best alternative to timely care with one’s own clinician. Findings from the present study support this approach.58,59 Many groups share electronic medical records, thus ensuring informational and management continuity for patients presenting out of hours. Yet, within-group after-hours care might be underutilized because many patients do not know these options exist. In a 2012-2019 population-based survey, approximately 60% of Ontarians were unaware that their family physician’s clinic offered after-hours services8,60 Improving awareness of after-hours offerings is essential to ensure that policies around coverage achieve their desired effect.
Limitations
There are several limitations to this study. First, our findings might not apply broadly to all settings. Indeed, we observed that the association between within-group visits and 7-day ED visits in rural and small urban settings was opposite to that found in large urban areas. This might be because EDs serve a different function in rural areas, with local family physicians also providing ED coverage, available to see their own patients in this setting. Since most walk-in clinics are located in large urban areas, our findings are relevant to the majority of walk-in clinics in Ontario. Second, it is possible that certain groups face barriers related to racism that we could not investigate in this study because our datasets do not have information on race. Third, we do not know with certainty that all within-group physicians had access to a shared electronic medical record, assuring informational continuity. Future research should confirm or refute whether and how continuity is a main driver of our findings, while also elucidating how certain practices can sustain accessible shared after-hours care. Relatedly, we did not seek to describe or compare the quality of care received in these different care settings. Fourth, owing to the nature of the data, we cannot know the reasons why patients went to the ED after their initial visit. In some cases, the physician might have recommended it.21,61 Indeed, it is important to consider that an increase in ED visits does not necessarily imply worse quality of care62 because these visits might have been warranted.21,61 An initial walk-in visit might also have averted an ED visit for patients who could not access their regular family physician in a timely manner. Fifth, the decision as to where to seek care might be influenced by patients’ health care–seeking behavior, perceived urgency of their condition, and/or tolerance for waiting.63 In other words, the patients more likely to seek access at a walk-in clinic might also be more likely to visit the ED for reasons beyond what is captured in health administrative data. As a result, there might be residual confounding between the type of index visit and the decision to visit the ED. Finally, the present study might not be generalizable to settings in which patients must pay for outside-of-group visits or where walk-in clinics are integrated with longitudinal primary care, which is not the case in Ontario.
Conclusions
We found that compared to having a visit with a walk-in clinic physician, having an after-hours visit with a physician who belongs to one’s own enrolling group was associated with a small decrease in the risk of an ED visit in the following week. More research is needed to understand the underlying reasons for our observation, particularly the role that informational or management continuity might play in mitigating health system use.
Acknowledgments
We thank patient partners Krysta Nesbitt and Patrick Roncal for their contributions to the analytic plan and results interpretation. Patient partners were provided with an hourly honorarium for reviewing study materials and attending project meetings.
This document used data adapted from the Statistics Canada Postal Code Conversion File, which is based on data licensed from Canada Post Corporation, and/or data adapted from the Ontario Ministry of Health Postal Code Conversion File, which contains data copied under license from Canada Post Corporation and Statistics Canada. Parts of this material are based on data and/or information compiled and provided by the College of Physicians and Surgeons of Ontario (CPSO), the Canadian Institute for Health Information, Ontario Health, and the Ontario Ministry of Health. The data provided by the CPSO were acquired via a data-sharing agreement between ICES and the CPSO. 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.
Footnotes
Conflicts of interest: authors report none.
Data-Sharing Statement: The dataset from this study is held securely in coded form at ICES. Whereas legal data-sharing agreements between ICES and data providers (eg, health care organizations and government) prohibit ICES from making the dataset publicly available, access may be granted to those who meet prespecified criteria for confidential access, at www.ices.on.ca/DAS (das{at}ices.on.ca). The full dataset creation plan and underlying analytic code are available from the authors on request, understanding that the computer programs might rely on coding templates or macros that are unique to ICES and are therefore either inaccessible or might require modification.
Funding support: This study was supported by ICES, which is funded by an annual grant from the Ontario Ministry of Health and the Ministry of Long-Term Care. This study also received funding from a Canadian Institutes of Health Research project grant awarded to L.L-S. and N.M.I. (grant #175285).
- Received for publication May 1, 2024.
- Revision received August 9, 2024.
- Accepted for publication August 14, 2024.
- © 2024 Annals of Family Medicine, Inc.
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.