Abstract
Continuity of health care among the formerly incarcerated is an emerging public health challenge. We used data from the San Francisco County Jail to determine whether discharge planning improves access to care on release. Inmates who were HIV positive and received discharge planning were 6 times more likely to have a regular source of care in the community compared with inmates with other chronic medical conditions, and they were as likely to have a regular source of care compared with the general San Francisco population.
Inmates with chronic medical conditions experience discontinuity of health care on release despite the risk of serious health outcomes.1,2 Of those released from jail, 90% lack insurance or financial resources for medical care.3–5 Although incarceration presents an opportunity to link inmates to health care on release, most are released without medical appointments. This discontinuity may lead to poor health outcomes, duplication of health care services, and recidivism.6 We sought to compare inmates' access to care with and without discharge planning. Because the San Francisco County Jail offers discharge planning for inmates who are HIV positive, including coordination of primary care and social services, we hypothesized that persons with HIV were more likely to identify a regular source of care compared with those without this service.
The ACCESS study is a cross-sectional study of people incarcerated in San Francisco County Jail, which averages a daily population of 2000 inmates and 55 000 bookings per year.7 From March 2005 to January 2006, we interviewed 347 English-speaking adults. Because ACCESS focused on health care use by inmates who were HIV positive, they were oversampled. Every inmate who was identified as HIV positive within the study time frame, identified through the San Francisco County Jail electronic database, was asked to participate. A systematic sample of inmates not identified as HIV positive was recruited as a comparison group.
Trained personnel conducted private interviews in San Francisco County Jail. Participants were asked about sociodemographic variables; history of homelessness, substance abuse,8 and incarceration; health status9; and history of chronic disease, including hypertension, cardiovascular disease, diabetes, emphysema, asthma, hepatitis, kidney disease, cancer, and seizure disorder. Participants were categorized as (1) HIV positive, (2) having another chronic disease and no known HIV infection, or (3) having neither known HIV infection nor another chronic disease. Because being HIV positive entitled inmates to discharge planning, those with HIV were analyzed together regardless of their chronic disease status. Participants were categorized as having a regular source of care if they answered: “I go to 1 place and see the same provider(s).”
We performed bivariate and multivariate analyses to investigate characteristics associated with having a regular source of care. The sample was weighted according to the 2001 prevalence of HIV in San Francisco County Jail.
We compared age-, gender-, ethnicity-, and insurance-specific proportions of inmates with a regular source of care for each of the 3 disease categories with the proportions reported in the California Health Interview Survey, a telephone survey of the Californian civilian population.10 We used indirect methods to calculate standardized morbidity ratios of the number observed compared with the number expected if our sample had the California Health Interview Survey rates of access to care.11
Of the 451 inmates approached for participation, 347 were enrolled (Table 1). Inmates who were HIV positive had a 6-times greater odds of identifying a regular source of care compared with inmates with other chronic medical conditions (odds ratio [OR] = 6.38; 95% confidence interval [CI] = 3.65, 11.14) and a 10-times greater odds of having a regular source of care compared with inmates with neither HIV nor another chronic disease (OR = 10.61; 95% CI = 4.61, 24.4). This association persisted in multivariate analysis after we adjusted for factors associated with having regular care, including age, marital status, insurance, and health status.
TABLE 1 Sample Characteristics of Jail Inmates, by Chronic Disease Status: ACCESS Study, San Francisco County, March 2005
| HIV | Chronic Disease Other Than HIV | No Chronic Disease | |
| Total, No. (%) | 181 (100) | 102 (100) | 64 (100) |
| Age, y, mean ± SD (range) | 41.5 ± 8.5 (18–63) | 38.2 ± 11.2 (18–63) | 34.2 ± 8.0 (18–50) |
| Gender, No. (%) | |||
| Men | 162 (90) | 83 (81) | 58 (91) |
| Women | 19 (10) | 19 (19) | 6 (9) |
| Race/Ethnicity, No. (%) | |||
| White | 52 (29) | 17 (17) | 15 (23) |
| Black | 87 (48) | 56 (55) | 29 (45) |
| Non-White Hispanic | 8 (4) | 4 (4) | 8 (13) |
| Other | 31 (17) | 24 (24) | 12 (19) |
| Education, No. (%) | |||
| Less than high school | 56 (31) | 31 (30) | 15 (23) |
| High school graduate | 80 (44) | 41 (40) | 31 (48) |
| Some college or more | 44 (24) | 30 (29) | 18 (28) |
| Marital status, No. (%) | |||
| Married or living with a partner | 24 (13) | 15 (15) | 15 (23) |
| Divorced, separated, or widowed | 55 (30) | 33 (32) | 15 (23) |
| Single | 102 (56) | 55 (54) | 24 (38) |
| Income the month prior to incarceration, $, median (IQR) | 860 (500–1200) | 1000 (422–2500) | 878 (255–2000) |
| Insurance, No. (%) | |||
| Medi-Cal | 104 (57) | 18 (18) | 7 (11) |
| Medicare | 28 (15) | 4 (4) | 4 (6) |
| Other (Department of Veterans Affairs | 15 (8) | 13 (13) | 16 (25) |
| benefits, private) | |||
| None | 49 (27) | 53 (52) | 39 (61) |
| Ever homeless, No. (%) | 160 (88) | 81 (80) | 43 (67) |
| History of drug abuse,a No. (%) | 163 (90) | 89 (87) | 46 (72) |
| Days in jail this year, median (IQR) | 60 (15–180) | 60 (17–180) | 50 (8–180) |
| Times in jail, No. (%) | |||
| 1–5 | 51 (28) | 30 (29) | 24 (38) |
| ≥6 | 115 (64) | 67 (66) | 37 (58) |
| Cannot remember | 7 (4) | 5 (5) | 3 (5) |
| Health status | |||
| Physical Functioning Scale,b Score ± SD | 66 ± 31 | 84 ± 24 | 95 ± 12 |
| Regular source of care, No. (%) | 152 (84) | 46 (45) | 24 (38) |
a Positive responses were those who responded “yes” to 2 or more questions on the CAGE questionnaire, adapted to include drugs.
b Measured with the RAND Corp SF-36, Physical Functioning Scale. The scores range from 0 to 100, and a higher RAND Corp SF-36 physical functioning score correlates with improved health status.
In comparison with the California Health Interview Survey data on the 2005 general San Francisco population, inmates with chronic medical conditions other than HIV were less likely to have a regular source of care (Figure 1). No significant differences were seen in access to care for inmates who were HIV positive compared with the general population, with the exception of the uninsured. Uninsured inmates with HIV were 1.3 times as likely to have a regular source of care compared with the general uninsured population.
FIGURE 1 Access to community care of San Francisco County Jail inmates, by chronic disease status, compared with the general population by selected sociodemographic characteristics: ACCESS Study, San Francisco, Calif, 2005
Note. Access to care of the general San Francisco population is represented by the dashed black line, which is assigned a value of 1.0 for comparisons with inmates' access to care. Standardized morbidity ratios for sociodemographic groups (older than 50 years, male, Black, insured, and uninsured) compare jail inmates' access by chronic disease status with that of the general San Francisco population, with interpretations similar to those for odds ratios.
We found that San Francisco County Jail inmates who received discharge planning—namely, inmates who were HIV positive—were more likely to have a regular source of care than were inmates who did not receive this service. Although comprehensive health care for inmates who are HIV positive is considered a matter of public health, providing discharge planning to inmates with other chronic conditions may improve their health care access as well. HIV discharge planning exemplifies how jailed adults can be successfully connected to care in the community with targeted efforts.
Our study had several limitations. Because it was cross-sectional, no conclusions could be made about cause and effect. We relied on self-report for measurement of chronic disease and did not obtain information on nonresponders, increasing the potential for recall and volunteer bias. By limiting enrollment to English-speaking adults, we could not generalize study results to non–English-speaking people.
Despite having a regular source of care, 42% of the inmates in this study reported interruptions in care when transitioning between jail and the community. Although human factors may contribute to discontinuity, it raises the question of whether other health care system strategies in addition to discharge planning would be more successful in ensuring continuity of care. Health care models in which providers follow-up their patients through the correctional system or in which jail health is integrated into the public health care system may lead to improved health outcomes for patients.12,13
Acknowledgments
Research was supported by the National Institutes of Health (grant RO1 DA13892) and the National Research Service Award Research Training Grant in General Internal Medicine to University of California, San Francisco (grant T32 HP 19025).
We appreciate the assistance of the San Francisco County Jail Health Services and the Sheriff's Department deputies and staff.
Human Participant Protection
This study was approved by the University of California's institutional review board.
