Table 2:

Within-county associations between a 1 per 1000 change in county jail incarceration rate and change in county mortality rates for nine common causes of death with 1-year, 5-year, and 10-year lags, 1987–2017

	1-year lag	5-year lag	10-year lag	% attenuation from 1-year to 10-year lag
Infectious disease	1·065 (1·061–1·070)	1·038 (1·034–1·043)	1·011 (1·007–1·015)	5·1%
Chronic lower respiratory disease	1·049 (1·045–1·052)	1·041 (1·038–1·045)	1·028 (1·024–1·032)	2·0%
Substance use	1·026 (1·020–1·032)	1·020 (1·014–1·026)	1·008 (1·002–1·014)	1·8%
Suicide	1·025 (1·020–1·029)	1·008 (1·003–1·012)	0·999 (0·995–1·004)	2·5%
Heart disease	1·021 (1·019–1·023)	1·016 (1·014–1·018)	1·012 (1·010–1·014)	0·9%
Unintentional injury	1·015 (1·011–1·018)	1·012 (1·008–1·016)	1·008 (1·004–1·012)	0·7%
Malignant neoplasm	1·014 (1·013–1·016)	1·008 (1·007–1·010)	1·004 (1·002–1·006)	1·0%
Diabetes	1·013 (1·009–1·018)	1·001 (0·996–1·006)	0·997 (0·992–1·002)	1·6%
Cerebrovascular disease	1·010 (1·007–1·013)	1·007 (1·003–1·010)	1·005 (1·001–1·009)	0·5%

Data are RR (95% CI) unless otherwise stated. All models included the jail incarceration rate, county and year fixed effects, and the county poverty rate, county crime rate, county Black resident population, county unemployment rate, state incarceration rate, and party control of state legislature lagged 1-year before the jail incarceration rate for 1094 unique counties. RR=risk ratio.