Restrictive immigration policy and self-reported health among policy-vulnerable adults in the United States: A difference-in-differences study.
We evaluated associations between implementation of Florida Senate Bill 1718 and self-reported physical and mental health among Hispanic and non-Hispanic Black adults, and examined whether these associations differed by cancer survivorship.
Using 2022-2024 Behavioral Risk Factor Surveillance System data, Hispanic and non-Hispanic Black adults in Florida were compared with those in Southern control states; cancer survivors were propensity-matched 1:2 to adults without cancer. Survey-weighted difference-in-differences and triple-difference models estimated changes in self-reported poor physical and mental health days. Parallel pre-policy trends were assessed using event-study models on an extended 2020-2024 panel, and p values for state-comparative models were obtained using a wild cluster bootstrap.
Among 5991 respondents, the Florida-versus-control increase in poor physical health days among cancer survivors alone was imprecisely estimated (difference-in-differences, 4.97 days; 95% CI, -0.59 to 10.52; bootstrap p = 0.137). In the survey-weighted analysis, implementation of SB 1718 was associated with a differential increase of 7.09 physically unhealthy days per month among cancer survivors relative to adults without cancer (triple difference, 95% CI, 3.63-10.54). However, this estimate was substantially attenuated and no longer statistically significant in the unweighted analysis (1.34 days; 95% CI, -0.19 to 2.87). The estimate also varied when the largest survey weights were excluded or winsorized and was based on small Florida cancer-survivor samples. No robust differential association was identified for mentally unhealthy days.
Although the survey-weighted analysis suggested an increase in physically unhealthy days among cancer survivors following implementation of Florida SB 1718, the magnitude and statistical significance of this estimate were not robust to the weighting specification. Small Florida cancer-survivor samples, unequal survey weights, and limited compositional comparability between Florida and the control states preclude a definitive conclusion regarding the policy's association with self-reported health.
Using 2022-2024 Behavioral Risk Factor Surveillance System data, Hispanic and non-Hispanic Black adults in Florida were compared with those in Southern control states; cancer survivors were propensity-matched 1:2 to adults without cancer. Survey-weighted difference-in-differences and triple-difference models estimated changes in self-reported poor physical and mental health days. Parallel pre-policy trends were assessed using event-study models on an extended 2020-2024 panel, and p values for state-comparative models were obtained using a wild cluster bootstrap.
Among 5991 respondents, the Florida-versus-control increase in poor physical health days among cancer survivors alone was imprecisely estimated (difference-in-differences, 4.97 days; 95% CI, -0.59 to 10.52; bootstrap p = 0.137). In the survey-weighted analysis, implementation of SB 1718 was associated with a differential increase of 7.09 physically unhealthy days per month among cancer survivors relative to adults without cancer (triple difference, 95% CI, 3.63-10.54). However, this estimate was substantially attenuated and no longer statistically significant in the unweighted analysis (1.34 days; 95% CI, -0.19 to 2.87). The estimate also varied when the largest survey weights were excluded or winsorized and was based on small Florida cancer-survivor samples. No robust differential association was identified for mentally unhealthy days.
Although the survey-weighted analysis suggested an increase in physically unhealthy days among cancer survivors following implementation of Florida SB 1718, the magnitude and statistical significance of this estimate were not robust to the weighting specification. Small Florida cancer-survivor samples, unequal survey weights, and limited compositional comparability between Florida and the control states preclude a definitive conclusion regarding the policy's association with self-reported health.