Abstract

Objective To identify social and health determinants affecting wait times for primary care in Canada.

Design Secondary analysis of national survey data.

Participants Canadian Community Health Survey participants aged 18 years or older during the 2015, 2016, and 2019 cycles.

Main outcome measures Weighted summary statistics and a partial proportional odds regression model were used to assess relationships between patient-level social and health determinants and wait times for primary care (ranging from same-day appointments to waits of 1 month or longer).

Results A weighted sample of 9,380,662 participants across the 3 survey cycles was used. Participants had a mean age of 49.7 years (standard deviation=17.5), 55.7% were female, and 74.7% were white. Approximately 57.4% of participants waited less than a week for care while 11.0% waited 1 month or longer. Factors associated with higher odds of waiting at least 1 month for a consultation were attaining only a high school diploma (adjusted odds ratio [aOR]=1.12, 95% CI 1.01 to 1.23) compared with having postsecondary education; identifying as Asian (aOR=1.42, 95% CI 1.20 to 1.67) or Black (aOR=1.57, 95% CI 1.16 to 2.13) compared with identifying as white; having “excellent” self-reported health status (aOR=1.10, 95% CI 1.02 to 1.19) compared with “good” self-reported health; lacking a regular primary care provider (aOR=1.47, 95% CI 1.27 to 1.71) compared with having a regular provider; or being classified as having multimorbidity (aOR=1.11, 95% CI 1.01 to 1.21) compared with having no chronic condition.

Conclusion Wait times for primary care were found to be associated with patient-level social and health determinants including race, education level, absence or presence of a regular health care provider, self-reported health status, and multimorbidity. Future research could investigate health system determinants of wait times to inform specific policy measures designed to reduce disparities in access to care.

The potential impact of primary care wait times (PCWTs) on patients with urgent health issues is immense. For example, previous studies have reported patients living with chronic conditions (eg, mental health issues, asthma) often experience declines in quality of life when access to care is challenging; prolonged PCWTs may also lead to lost opportunities for effective treatment.1,2 PCWTs are not only inversely associated with patient satisfaction but are also associated with increased risks of mortality.1,3-5 Health care systems may incur higher costs and experience decreased efficiency if limited access to primary care results in patients being redirected to hospitals inappropriately.6

A 2020 survey conducted in 11 developed countries reported only 41% of adults in Canada could get a same- or next-day primary care visit the last time they were ill.7 This proportion was below the Commonwealth Fund average of 57% in the report, and the only country included where fewer patients were able to access same- or next-day appointments was Sweden (38%). Prolonged PCWTs are concerning as primary care providers (PCPs) are important conduits to specialist care referrals when indicated.

Being of lower socioeconomic standing,8-12 being female,10 having poor health status,9,11 or having recently migrated to Canada9,13-15 are significantly associated with prolonged PCWTs in national and provincial cross-sectional studies. This study addresses previous limitations by using a national-level survey that accommodates respondents with varying language competencies and disabilities and by adjusting for health conditions.16 The objective of this paper is to use Canadian Community Health Survey (CCHS) data to investigate possible associations between social and health determinants and wait times experienced between requesting an appointment with a PCP and the appointment date.

METHODSStudy design, data source, and study population

A cross-sectional analysis of secondary data from the 2015, 2016, and 2019 cycles of the CCHS (weighted N=20,823,198) was conducted. Figure 1 summarizes participant exclusion criteria.17 Our reporting follows STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines (Appendix A, available from CFPlus*).17 The CCHS includes people in Canada aged 12 years or older in 2016 and 2019 and people aged 18 years or older in 2015, residing in any of the 10 provinces or 3 territories. People outside the CCHS’s catchment include those living on First Nations reserves or other Indigenous settlements, personnel serving full-time in the Canadian Armed Forces, and individuals who are institutionalized, with exclusions representing less than 3% of the population.16 Supplementary information regarding the design and implementation of the CCHS has been published elsewhere.18-21

Figure 1.

STROBE diagram for inclusion of CCHS participants in analysis

The sample was drawn from the 2015, 2016, and 2019 cycles of the CCHS as these were the only years where the Patient Experiences (PEX) module had been implemented nationally. Pooling cycles allowed for increased statistical power and more robust estimates versus a cross-sectional analysis of individual years. The PEX module of the survey collects information regarding respondents’ most recent PCWTs. PCPs are defined as health care professionals who routinely provide preventive care and health advice to patients (ie, general practitioners, nurse practitioners, registered nurses, or other allied health care professionals under supervision). Survey respondents from the 3 territories were excluded because income data had not been collected. This resulted in an effective weighted sample size of 9,380,662.

The main outcome variable was the PCWT between a patient contacting their PCP to book an appointment and the appointment taking place. PCWT was measured as an ordinal variable in the CCHS with the following response options: on the same day, the next day, in 2 to 3 days, in 4 to 6 days, in 1 to 2 weeks, between 2 weeks and 1 month, or 1 month or more.

Independent variables were respondents’ social and health determinants. We used the Public Health Agency of Canada’s definition of multimorbidity, which is the cooccurrence of at least 2 of the 5 categories of major chronic diseases: cancer (current or previous), diabetes, cardiovascular disease (heart disease or stroke), chronic respiratory disease (asthma or chronic obstructive pulmonary disease), and mood or anxiety disorders.22 Documentation regarding CCHS coding and variables used is provided in Appendices B and C, available from CFPlus.*

Data analysis

A partial proportional odds model was used to examine the link between PCWTs and reported social and health determinants. This model is more parsimonious than multinomial logistic regression and less restrictive than ordinal logistic regression.23 A literature review was conducted a priori to identify theoretically relevant independent variables. The final model was fit using Stata’s autofit command, which uses the Wald test to iteratively constrain variables that meet the proportional odds (parallel lines) assumption until only variables that violate the assumption remain unconstrained.23,24 The result is a partial proportional odds model that relaxes the parallel line assumption only for variables that do not satisfy it.

Statistical tests for regression assumptions, goodness-of-fit tests, and regression diagnostics for potential models are provided in Appendices D to H, available from CFPlus.* Multicollinearity was assessed using variance inflation factors, with no covariates exceeding the threshold (variance inflation factor >10; see Appendix H for details).25 Sampling weights provided in the CCHS were used to allow for inference of findings about the underlying population. Sample weights from the CCHS were adjusted by dividing each by the number of cycles used to ensure population-level inference from the pooled sample.18 Model coefficients were reported as adjusted odds ratios (aORs) with 95% confidence intervals (CIs). All analyses were conducted in Stata 17.0 with P values less than .05 considered statistically significant (sex-stratified results are provided in Appendix I, available from CFPlus*).

Approval for the use of CCHS data is covered by the Western University Health Sciences Research Ethics Board.

RESULTSSample characteristics

Table 1 provides descriptive statistics of CCHS participants. In the weighted population the mean age was 49.7 years (standard deviation=17.5), 65.6% possessed a postsecondary education, 44.3% were male, 74.7% were white, and the fourth- and fifth-highest income quintiles had the largest proportions of participants (20.8% and 22.2%, respectively). Ontario residents composed the largest proportion of participants (39.5%) followed by those in Quebec (19.6%). Reported PCWTs ranged from receiving care within 1 week (57.4%), between 1 and 2 weeks (20.0%), between 2 weeks and 1 month (9.1%), to 1 month or more (11.0%). Before regression analysis, 708,506 (7.6%) incomplete cases were removed, with 8,672,156 cases remaining. A supplementary analysis of potential bias from complete case analysis is provided in Appendix J, available from CFPlus.*

Table 1.

Descriptive statistics of weighted population (N=9,380,662) of CCHS participants in cycles 2015, 2016, and 2019

Associations between social and health determinants and PCWTs

Results of the regression analysis are presented in Table 2. Same-day appointment was used as the reference category for the outcome variable; hence, all levels are interpreted in relation to this level. Higher odds of waiting 1 to 6 days for consultation compared with those being seen the same day were found in respondents of female sex and among those who received care from a health care professional other than a family doctor or nurse. On the other hand, lower odds of waiting 1 to 6 days for a consultation were found among landed immigrants compared with non-immigrants. Higher odds of waiting 1 month or longer compared with those who received same-day consultations were found among those who attained only a high school diploma (versus those with postsecondary education); identified as Asian, Black, or “other” race (versus those who identified as white); self-reported a very strong sense of community belonging (versus a good sense of community belonging); or had no regular health care provider.

Table 2.

Weighted partial proportional odds regression of patient social and health determinants on primary care wait times for CCHS participants in cycles 2015, 2016, and 2019

Respondents who reported having excellent health or were categorized as having multimorbidity had 10% and 11% increased odds, respectively, of waiting 1 month or longer for care compared with those who received a same-day consultation relative to the covariates’ respective reference level and holding all other variables constant. Income, mental health status, and sexuality were not found to be significant predictors of wait times. Sex-stratified results showed consistent association directions with the total sample (Appendix I), differing only in covariate significance and constraint status. Absolute effects and descriptive statistics for attached versus unattached patients are provided for context beyond relative measures (Appendices K and L, available from CFPlus*).

DISCUSSION

Long wait times for health care in Canada continue to drive unmet needs, with Statistics Canada data from 2022 indicating 9.2% of people in Canada report unmet health care needs.26 Commonwealth Fund data27 show the proportion of adults in Canada who report having access to a regular primary care provider dropped from 93% in 2016 to 86% in 2023. Our study underscores persistent disparities in access to primary care and leverages the only national wait time data available.15 Although likely conservative, our findings help identify care gaps and can be used to inform equitable policy-making.

Our cross-sectional findings align with literature from Canada showing longer PCWTs for patients with complex conditions or multimorbidity,12,28,29 low levels of education,8,30 no regular PCP,31 or racialized identities.13,32 Evidence related to patient age, income, mental health status, and sexuality is limited, with our model showing minimal effects. Existing studies from around the world report mixed results: some found no age-related differences in PCWTs33,34 while others found older patients experienced longer PCWTs.35,36 Some studies report income-based disparities with longer PCWTs among less affluent populations.9,10

This is the first study we know of to examine mental health status and sexuality in relation to primary care wait times. While our findings showed no statistically significant differences in PCWTs, they differ from existing studies that found sexual minorities and those with poor mental health face greater access barriers.37,38 Our results also contrast with literature linking better self-reported health to shorter PCWTs.39

Pooling CCHS cycles enabled the assessment of whether associations with PCWTs remained stable across time points and helped control for temporal variation in health care access within repeated cross-sections, which is common in PCWT studies. Our study is the first of which we are aware to operationalize a multimorbidity measure to examine trends in PCWTs in Canada.

This study highlights disparities in primary care access among people in Canada with adverse social and health determinants, underscoring the need for equity-focused interventions. Our findings can be used to inform evidence-based wait time targets, social and health determinants–related policy performance metrics, and funding models that holistically account for patient complexity. Similar frameworks have guided resource allocation in England,40 and targeting PCP distribution in socially marginalized areas has reduced disparities in PCP access in England.41 In Canada, health system navigators have improved access to care for refugees,32 and similar strategies (combined with case managers) may effectively improve care access for patients with adverse social and health determinants.42 Value stream mapping was found to reduce PCWTs for patients with chronic conditions in Nova Scotia without added costs43 while in other places telehealth has been shown to reduce barriers for those with mobility, language, or PCP attachment challenges.44-47 Additionally, more practice-specific interventions, such as access bonuses for PCPs who deliver timely care and expanded team-based care models, have also been shown to decrease PCWTs in some settings.48,49 Scaling these interventions and policies to address populations highlighted in our study would better equip the Canadian health care system to reduce PCWTs.

Strengths and limitations

The CCHS offers key strengths, including a large sample size, detailed respondent data, first-person health accounts, and primary care access data from before the COVID-19 pandemic for future comparisons. Limitations include reliance on self-reported data, use of older cycles, lack of linkage to administrative health records, residual confounding, and the inability to infer causality due to the cross-sectional design. The type of regular PCP may influence wait times due to the substitution effect when a nonphysician serves as the primary provider. However, our wait time data reflect the reality of routine care delivery, which includes a range of providers. Another limitation was the lack of data on rural or urban place of residence in the CCHS cycles used for the current analysis. Previous research found mixed results regarding place of residence on wait times and access indicators.50-56 The most notable limitation is the limited number of CCHS cycles that could be included in the analysis, as the PEX module had been implemented nationwide only in 2015, 2016, and 2019.

Conclusion

Continuity of care, provider type, and patients’ health status, sex, and racial identity were found to be strongly associated with PCWTs. Social and health determinants need to be addressed in primary care and in the broader health care system, where primary care serves as the foundation. Future studies reporting primary care access metrics disaggregated by social and health determinants would continue to enhance accountability and support evidence-based policy decisions.

Footnotes

Contributors

Bill Le, Dr Saverio Stranges, and Dr Shehzad Ali were involved in the formulation and design of the study. Bill Le drafted the initial manuscript and interpreted the data, and Drs Stranges and Ali contributed to revising the drafts up until the finished manuscript.

Competing interests

None declared

This article has been peer reviewed.

Cet article a fait l’objet d’une révision par des pairs.

Copyright © 2025 the College of Family Physicians of Canada