A retrospective population-based cohort study was conducted using administrative healthcare data from Ontario, Canada's most populous province. The data were securely linked using unique encoded identifiers and analyzed at ICES, a not-for-profit research institute funded by the Ontario Ministry of Health (http://www.ices.on.ca/).

The Ontario Health Insurance Plan (OHIP) Registered Persons Database (RPDB) was used to determine health insurance status, demographic characteristics, and residential postal code. Physician service claims were obtained from OHIP Claims History Database. Cancer history was collected from the Ontario Cancer Registry (OCR). Breast, cervical, and colorectal cancer screening and influenza vaccination were identified from a combination of records in the Ontario Breast Screening Program (OBSP), Primary Care Population Database (PCPOP), and OHIP Claims History Database. These databases were securely linked at the individual level and analyzed in an anonymous, coded form at ICES under Section 45 of Ontario's Personal Health Information Protection Act.

The study included two cohorts of Ontario adults aged 18 and older who were alive on January 1, 2020. The first cohort comprised individuals residing in social housing buildings, which were sites identified in the 2023 cycle of the Social Housing of Ontario Registry (16). The second, or complement, cohort consisted of individuals in the general population who were not residing in social housing. Detailed cohort derivation methods have been described previously (6, 8, 11).

The study measures were influenza vaccination and screening for cervical, breast, and colorectal cancer. Eligibility for the “Pre” COVID-19 period was determined as of January 1, 2018, and for the “During” COVID-19 period was as of January 1, 2020. All study measures except colorectal cancer screening were based on administrative data between January 1, 2018, and December 31, 2019, for the “Pre” period and between January 1, 2020, and December 31, 2021, for the “During” period. Individuals were categorized as having received at least one influenza vaccination if they had an OHIP billing code for influenza vaccination in each 2-year period (17). Cervical cancer screening (Pap test) eligibility included females aged 21–69 years, according to the RPDB, with no history of cervical cancer or hysterectomy based on OHIP billing codes and OCR entries. Pap test screening was identified using specific OHIP billing codes. Eligibility for breast cancer screening (mammography) included females aged 50–69 years in the RPDB, with no history of breast cancer or mastectomy, based on OHIP billing records and OCR entries. Mammography screenings conducted as part of the OBSP, as well as those performed outside the program, were included (18). Colorectal cancer screening eligibility included individuals aged 52–74 years, excluding high-risk individuals with a history of colorectal cancer or inflammatory bowel disease. Screening status was determined using the PCPOP database, a population-level ICES database that includes all Ontario residents eligible for OHIP and provides information on health services use, including cancer screening status. Individuals were considered up-to-date if they had at least one fecal immunochemical test (FIT) or fecal occult blood test (FOBT) in the previous 2 years, or a flexible sigmoidoscopy or colonoscopy in the past 10 years (19).

The characteristics of the cohorts were summarized using descriptive statistics and grouped according to public health unit regions used by a provincial surveillance program at the time of the data (2020) (20). Influenza vaccination and breast, cervical, and colorectal cancer screening rates were estimated for the social housing and complement cohorts during the “Pre-COVID-19” (2018–2019) and “During COVID-19” (2020–2021) time periods. Age- and sex-standardized rates for the social housing population were calculated using the direct method of standardization (21), with the age and sex population structure of the complement cohort serving as the reference. The rates were estimated overall and across age and sex groups, where applicable. Differences in vaccination and screening rates between the two cohorts, and their 95% confidence intervals (CIs) during “Pre-COVID-19” and “During COVID-19” time periods, were calculated using the method recommended by Altman et al. (22) and presented as percentages.

Across all age groups and both sexes, the social housing cohort had lower rates of receiving at least one dose of influenza vaccination within the respective 2-year time period compared to those in the complement cohort (Table 3). This disparity became more pronounced during the COVID-19 period. The overall age- and sex-standardized difference between the social housing and the complement cohorts increased from −0.30% (95% CI: −0.42 to −0.19%) to −1.84% (95% CI: −1.95 to −1.74%). In the “During COVID-19” period, the social housing cohort had consistently lower sex-standardized vaccination rates across all age categories compared to the complement cohort (range: −1.13 to −3.20%), with the largest difference observed in the 80+ age group (−3.20%, 95% CI: −3.68 to −2.71%). The degree of disparity during COVID-19 was similar for males and females between the social housing and complement cohorts with age-standardized differences of −1.85% (95% CI: −2.00 to 1.69%) for males and −1.83% (95% CI: −1.98 to 1.68%) for females.

Across both time periods, individuals in social housing had consistently lower rates of breast cancer screening (mammography) compared to those in the complement cohort (Table 4). Among females aged 50–69 years in the “Pre-COVID-19” period (2018–2019), the age-standardized mammography rate for breast cancer screening was 46.37% in the social housing cohort and 56.95% in the complement cohort, representing a difference of −10.58% (95% CI: −11.02 to −10.14%). This difference was similarly reflected across the 50–59 and 60–69 age groups, at −10.23 and −11.0%, respectively.

During the COVID-19 period (2020 to 2021), age-standardized mammography rates declined in both cohorts. The decline was larger in the complement cohort, falling from 56.95 to 21.87%, compared with a decline from 46.37 to 17.79% among the social housing cohort, representing a narrowed disparity of −4.08% (95% CI: −4.42 to −3.74%). Similar patterns were observed across the 50–59 and 60–69 age groups, with differences of −4.18 and −3.95%, respectively.

Cervical cancer screening (Pap test) showed persistent disparities between the social housing and complement cohorts, with consistently lower rates observed in the social housing cohort across both time periods and all age groups (Table 4). In the “Pre-COVID-19” period (2018–2019), the overall age-standardized difference in Pap test was −6.09% (95% CI: −6.36 to −5.83%). In “During COVID-19” period (2020–2021), rates decreased in both cohorts, from 39.72 to 30.16% in the complement cohort and from 33.63 to 25.92% in social housing cohort, resulting in a narrowed difference of −4.24% (95% CI: −4.49 to −3.99%). The greatest difference between cohorts was reported in the 60–69 age group, with a difference of −8.62% (95% CI: −9.19 to −8.05%) in the “Pre-COVID-19” period and −5.11%, (95% CI: −5.60 to −4.62%) in the “During COVID-19” period.

Overall, the disparity in age-standardized colorectal cancer screening rates between the social housing and complement cohorts increased from −7.42% (95% CI: −7.72 to −7.12%) in the “Pre-COVID-19” period to −9.69% (95% CI: −10.0 to −9.38%) in the “During COVID-19” period (Table 5). Disparities were greater among males than females in both periods: −10.32 vs. −6.61% in the “Pre-COVID-19” period and −11.52 vs. −7.94% in the “During COVID-19” period. Among females, the largest disparity was observed in those aged 50–59 years (−7.47% Pre-COVID-19 and −8.71% During COVID-19), whereas among males, the most pronounced gap was in those aged 60–74 years (−10.71% Pre-COVID-19 and −12.07%, During COVID-19).

During the COVID-19 pandemic, influenza vaccine uptake patterns shifted across populations, revealing growing disparities between individuals living in social housing and those in the complement cohort. While public health campaigns initially promoted both influenza and COVID-19 vaccinations, our findings emphasize that compared to the general population, residents of social housing were less likely to receive the influenza vaccine during the pandemic, potentially due to reduced access to in-person health services, competing health priorities (e.g., prioritizing COVID-19 vaccinations), and a lack of targeted outreach (23–25). The transient nature of COVID-19 focused vaccine campaigns may have further widened existing disparities, reinforcing the need for sustained, community-based interventions to support influenza vaccination in marginalized populations (26). This differs from the literature which demonstrates that influenza vaccination rates did not change significantly before, during, and after the pandemic in retirement residences or older adults-care centres (23). Among adults aged 18–64 without chronic medical conditions, the decline in vaccination was most pronounced during the pandemic, with an odds ratio of 0.82 (95% CI: 0.69–0.98) in the 2021–2022 season compared to 2018–2019 (23).

In Ontario, organized breast cancer screening services were paused in early 2020, resulting in a substantial backlog. Although services resumed later that year, the recovery in screening participation was uneven (27). The COVID-19 pandemic significantly disrupted breast cancer screening programs across Canada, with a pronounced impact on vulnerable populations, including residents of social housing. Our findings show that the disparities in mammography screenings between social housing residents and the general population were already substantial before the pandemic. The overall mammography screening rates for the social housing cohort were 46.37% before the pandemic, dropping to 17.79% during the pandemic. While this appears to be comparable to general population's drop in mammography screening from 56.95 to 21.87%, the narrowing gap is likely due to a “floor-effect” since the pre-pandemic screening rate among social housing residents was already low. Fu et al. (18) reported a 93.4% decrease in weekly bilateral screening mammography rates among average-risk women aged 50–74 in Ontario at the onset of the pandemic, followed by a gradual weekly increase of 8.4% until the end of 2020. However, screening rates among women who were immigrants or had limited access to material resources remained low before and during the pandemic, indicating that existing disparities had persisted during the recovery phase (18). Similarly, Lofters et al. (27) found that the pandemic exacerbated pre-existing disparities in breast cancer screening among individuals living in lower-income neighborhoods and for immigrants significantly widened during the pandemic. The proportion of women in the general population up to date on breast cancer screening in Ontario decreased from 61.1% before the pandemic to 51.7%, showing almost a 10% decrease, compared to our study which shows a 35% decrease (27). This may be because our study only covers the time period 2020–2021 whilst the Lofters et al. (27) study captures later data up to May 2022, including post-pandemic recovery. The study emphasized that policy makers should prioritize improving access to team-based primary care for people who are immigrants and/or living with low income to address these screening disparities (27). A population-based cohort study by Agarwal et al. (8) showed that, prior to the pandemic, women aged 50–69 living in social housing in Ontario had mammography screening rates that were 10%−11% lower than non-residents (8). Our findings confirm that this gap persisted during the pandemic, underscoring the need for targeted interventions to improve screening uptake in social housing populations.

Cervical cancer screening rates declined in both cohorts during the pandemic, with consistently lower rates among residents of social housing. However, we observed a slightly narrower gap in screening rates between the social housing and general populations. Routine cervical cancer screening programs were suspended or adapted in several provinces during the pandemic (28, 29). In the first 6 months of the COVID-19 pandemic, Ontario experienced a 63.8% decrease in Pap tests compared to the same period in 2019 (30). A pan-Canadian survey further reported that 15.8% of healthcare professionals observed a 75% or greater reduction in Pap tests during the pandemic and 56.9% reported delays in scheduling these tests (31).

Screening for colorectal cancer is itself a challenging test, requiring an understanding of the health consequences and overcoming certain barriers in order to actually have the test, and low socioeconomic status is associated with poorer uptake (32, 33). Therefore, it is not surprising that residents of social housing, who often face intersecting barriers such as lower health literacy, limited access to primary care, and socioeconomic instability experienced disproportionate declines in colorectal cancer screening participation. In contrast, while the general population also experienced disruptions, the rate of decline in screening is disproportionately higher in the social housing cohort. This raises important concerns for post-pandemic recovery efforts, which must ensure that all individuals receive the screenings they missed (19, 34). Individuals in social housing are likely to remain at a disadvantage due to structural and systematic barriers, whereas individuals in the general population, particularly those with a higher income, may be better positioned to resume or access screening services as they become available. A population-based study in Ontario observed a 3.9% decline in overall colorectal cancer screening from 65.9% in March 2019 to 62.0% in March 2022, with larger declines in low-income neighborhoods (27). Recovery was also slower among low-income populations, raising concerns about delayed diagnoses and poorer long-term health outcomes (6, 19, 27, 35).

Preventive health services remain inequitable for residents of social housing, who often face barriers such as limited access to care, low health literacy, and a lack of tailored outreach (6–8, 11, 36). These disparities worsened during the COVID-19 pandemic and underscore the need for equity-focused policies, community-based delivery models, and robust data systems to monitor outcomes by housing status. Integrating preventive health services into routine community care through mobile clinics and trusted, culturally tailored outreach is essential, given that vaccination and screening data are not consistently tracked by housing status (37, 38). Preventive care should also include embedded navigation services within primary care to support uptake. Given the low cervical cancer screening rates before the pandemic, expanding human papillomavirus self-sampling, increasing access to community-based services, and strengthening infrastructures are necessary during public health emergencies (39, 40). For colorectal cancer screening, recovery strategies such as mailed fecal immunochemical test (FIT) kits and community outreach are critical to engage individuals who have missed screening (41, 42).

A key strength of this study is the use of large, population-based, linked administrative health datasets from Ontario, which allowed for comprehensive and robust comparisons between individuals living in social housing and the general population. The study captured real-world healthcare utilization across multiple preventive services including influenza vaccination and cancer screening over both pre-pandemic and pandemic periods, enabling a nuanced assessment of trends and disparities. The linkage of multiple high-quality datasets, including those from the OHIP, OBSP, and OCR, provided reliable identification of preventive service uptake and eligibility.

The study also has limitations. This analysis does not include data beyond 2021 and therefore may not reflect full post-pandemic recovery trends in preventive care. Minor inaccuracies may exist in the social housing cohort due to residents' movement in and out of social housing; however, the impact is likely very minimal, as relocation often occurs between buildings. Rigorous methods were employed to ensure accurate identification of the social housing cohort. Our vaccination data are incomplete, as they capture only physician-administered influenza vaccines and not those provided by pharmacists. This may introduce bias if social housing residents were more likely to obtain vaccines from pharmacists. The use of administrative data limits the ability to capture individual-level factors such as social determinants of health, gender, race, and prevalence of mental health conditions which may influence healthcare utilization. Despite these limitations, the study provides important evidence of the inequitable impact of the COVID-19 pandemic on preventive health service use among a socially and economically disadvantaged population.