Skip to main content
. Author manuscript; available in PMC: 2020 Apr 15.
Published in final edited form as: Am J Epidemiol. 2012 Feb 13;175(6):519–526. doi: 10.1093/aje/kwr327

Table 4.

Predicted Mean Time Between Release and Death Among Formerly Incarcerated People Released From Jails Who Died of Drug-Related Causes, by Time Spent in Jail During the Last Incarceration, New York City, 2001–2005a

Time Spent in Jail During the Last Incarceration, days Predicted Mean, daysb 95% Change Confidence in Predicted Interval for Mean Change in Predicted Mean P Valuec
<4 (n = 74) 425.1 0 Reference
4–14 (n = 53) 330.8 −22.2 −48.5, 17.6 0.234
15–30 (n = 29) 241.8 −43.1 −65.2, −7.0 0.024
31–60 (n = 23) 246.4 −42.1 −66.2, −0.7 0.047
61–90 (n = 16) 235.3 −44.7 −71.1, 5.9 0.074
≥91 (n = 24) 188.7 −55.6 −74.3, −23.4 0.004
a

Results were from the negative binomial regression model, with days between release and death as the dependent variable. Covariates included time spent in jail during the last incarceration, sex, age at death, race/ethnicity, and the number of times incarcerated. Among all covariates, time spent in jail during the last incarceration and number of times incarcerated were significantly associated with the dependent variable.

b

Predicted mean was computed holding the other covariates at their mean.

c

Two-sided chi-squared test for the difference in mean days between release and death compared with the reference.