Primary care utilization in people who experience imprisonment in Ontario, Canada: a retrospective cohort study

Worldwide, more than 10.3 million people are in prison at any given time [1], and an estimated 30 million people move through prisons annually [2]. On an average day, 37,864 persons in Canada and 2,217,000 persons in the USA are imprisoned in jails and prisons [1].

Given the increased morbidity and mortality experienced by this population [3], primary care in prison and after prison release offers an important opportunity to improve health [4, 5]. Further, States have obligations regarding the provision of health care to persons in prison, and to attend to aftercare at the time of prison release [6]. Comparing the health care utilization rates for people who experience imprisonment with rates for the general population is a strategy to assess health care accessibility as an indicator of health equity. For example, relatively low rates of primary care utilization would suggest worse access to care and highlight the need to facilitate primary care access to improve health status.

Studies from several countries have found increased primary care utilization for people while in prison and after release [5, 7‐13]. In contrast, studies in the USA have identified relatively low use of primary care in prison and on release [14‐17]. Current evidence is limited in internal and external validity, however, by small or select samples [15], the lack of general population comparator groups, [5, 8, 9, 11‐17] use of self-reported data, [5, 12‐17] and the age of data reported [8, 10]. Overall, there is a paucity of data on primary care use for this population, especially for the period after prison release, and no longitudinal studies have examined primary care utilization in prison and on release. In addition, we lack information on the ways in which factors such as comorbidity may contribute to increased health care utilization.

In the setting of a publicly funded universal health insurance system and a publicly funded and administered prison system, we aimed to describe the utilization of primary care for persons released from provincial prison in Ontario in 2010, and to compare primary care utilization for this group with the general population.

Of 53,955 persons released from provincial prison in Ontario in 2010, we linked 52,546 (97.4%) (Additional file 1). We excluded 233 persons who had a release period of 1 day or less in 2010, 2,178 persons transferred to federal custody on release, 7 persons whose reason for release was death, and 1,267 persons whose reason for release was related to immigration, leaving 48,861 persons in the prison release group. We identified four age and sex-matched general population controls for each person in the prison release group, for a total of 195,444 general population controls.

A larger proportion of those in the prison release group were in lower neighbourhood income quintiles and were from rural areas or small towns, compared to general population controls (Table 1). The median number of ADGs was significantly greater for the prison release group compared to controls. Persons in the prison release group also had a significantly higher prevalence of all conditions examined, except for hypertension.

For the prison release group, primary care utilization rates were highest while in prison (Fig. 1), with a substantial decrease in use at the time of release.

The rate of health care utilization was greater for persons in the prison release group than general population controls, both in prison and in the post-release periods, as shown in Fig. 2.

As shown in Table 2, adjustment for ADGs resulted in a substantial decrease in the ratio of primary care utilization rate for the prison release group compared to general population controls. However, there remained a significant positive association across time periods between imprisonment status and primary care utilization.

The proportion of people who accessed any primary care was significantly higher for those in the prison release group compared to general population controls for the periods in prison and post-release days 0-6, 7-29 and 30-89, as shown in Table 3. In contrast, the proportion that accessed any primary care was significantly greater for the general population compared to the prison release group for days 180-364 and 365-730 post-release.

Figure 3 shows the time to first use of primary care after release from prison to two years. By one month after release, 66.3% of women and 75.5% of men had not yet accessed primary care and by three months after release, 50.5% of women and 62.9% of men had not yet accessed primary care. By two years after release, 16.8% of women and 28.2% of men had not accessed primary care.

This study identifies very high absolute and relative rates of primary care utilization in people who experience imprisonment, with over six-fold the primary care utilization rate in prison and more than twice the utilization rate post-release compared to general population controls. The associations did not change substantially after adjusting for neighbourhood income quintile and rurality, and the positive associations persisted after adjusting for comorbidity. In prison and in the months after release, a higher proportion of people in the prison release group accessed care compared to the general population controls for the same period. We note, however, that a substantial proportion of people in the prison release group did not access primary care in the months after release; more than half of women and more than 60% of men had not accessed primary care by three months post-release.

In prison, the rate of primary care utilization in our study was similar to rates in other studies, which were between 5 and 20 visits per year [7‐12]. Regarding use of primary care after prison release, the proportion accessing primary care was higher in our study than in a sample of women leaving jail in New York City between 1997 to 2001, in which 47% reported primary care use in the year post-release [15], and lower than in a study of persons released from prison in Australia, in which 43% of men and 58% of women accessed care in the month post-release [5]. These differences may reflect differences in the health care system, for example the lack of a universal health insurance system may have contributed to lower primary care utilization in the New York City study. There is a paucity of available and comparable data on the proportion of persons accessing primary care in prison or on rates of primary care utilization post-release [14].

This study has limitations that may affect internal and external validity. We were not able to discern whether primary care visits in prison were for administrative reasons only, e.g. for ministry-required physician assessments, or whether ministry-required physician assessments were included in OHIP billings and therefore represented in health administrative data. We think it is unlikely that these issues would substantially alter our findings. We did not include primary care encounters in Community Health Centres, which provide primary care for an estimated 4% of the Ontario population [38] and about 5% of persons who experience imprisonment (data available from author on request). This exclusion would likely have led to a conservative bias in the relative rate ratios. A limitation due to the longitudinal nature of the study is that a substantial proportion of persons who were released in 2010 were not included in some follow up periods in the months after release due to right censoring as described in the analysis plan, such that by 6 months after release, only 68.6% of persons were still in the community (as per Table 3), largely due to readmission to prison. If utilization of primary care were associated with risk of readmission to prison, this could contribute to apparent differences in use over time. However, as the largest differences in rates of use occur between the time in prison and the immediate post-release period, by which time very little right censoring had occurred, we think it is unlikely that this would explain the general pattern of use by period post-release. Finally, the study results may not be generalizable to other jurisdictions, for example to jurisdictions with different healthcare systems (including those with no universal health insurance program), criminal justice systems, or prison environments, or in which patterns of morbidity and mortality differ in prison or general populations.

Various factors may contribute to the high rate of primary care utilization in this population, including increased comorbidity burden, increased injury and acute illness, and a low threshold for seeking care in prison. Data from our study and from previous reviews [3, 39] reveal high comorbidity in this population, which could lead to greater primary care use. However, even after controlling for ADGs, the positive association between imprisonment status and primary care utilization persisted. While ADGs have been shown to predict morbidity [40], mortality [33], and health care utilization, [41, 42] their validity as a comorbidity indicator has not been assessed in people who experience imprisonment; even after controlling for ADGs, there may be a residual effect of comorbidity. More complex models and more detailed data would be valuable to explore specific mechanisms in greater detail. The social and physical environment, as well as risk behaviours in prison and in the community may also increase the risk of injury or acute illness and stress in this population [8, 39, 43‐46], which could contribute to primary care use. These factors would not be reflected in ADGs, which capture only current diagnoses and their sequelae. Other studies have suggested that increased primary care use in prison may reflect care seeking for issues that would be managed in the community without seeing a physician, for example through discussion with friends, family or a pharmacist [7], and that the low opportunity costs of seeing a physician in prison may encourage care use [10]. These factors may explain some of the difference in use in prison and post-release, but would not explain the increased use post-release compared to the general population.

While relative rates of health care utilization were consistently higher for the prison release group compared to general population controls, we note that only a minority of people in the prison release group accessed any primary care in the weeks after release. This is concerning given the high morbidity identified in this population in this study and others [3, 39], for which patients may need and want ongoing medical care, and given the high risk of adverse outcomes in the immediate post-release period [20, 34‐36]. Of note, the increased rate of primary care utilization for the prison release group in the week after release compared to later post-release periods could reflect people appropriately accessing primary care for continuity of care for ongoing medical and social issues or to address new issues associated with the transition back to the community. However, it could also signal a lack of attention to health needs prior to release, such as providing bridging medications or prescriptions. Other types of data, such as data from medical charts and interviews, would be required to elucidate the reasons for the difference in primary care utilization rates over the post-release period.

Further research should explore primary care use in this population, including reasons for seeking care in prison and after prison release, and care experiences including access to preventive care and quality of care. This work should be done in collaboration with affected populations including people with a lived experience of imprisonment. In parallel with research, we should advance work to improve appropriate access to and acceptability of primary health care for this population. Building on the limited evidence to date [4, 13, 47‐50], interventions should focus on improving access to care and quality of health care in prison and post-release.

This large population-based study shows that people who experience imprisonment in Ontario, Canada have very high rates of primary care utilization compared to the general population. This association is only partly explained by increased comorbidity. Of note, a substantial proportion of people in the prison release group did not access primary care in the months after release. Research is required to understand the reasons for increased primary care utilization in this population and to support access to care for persons who are not accessing care post-release.