Table 4

Sensitivity analyses* evaluating associations between 6-week postpartum healthcare use in NY and homelessness

Adjusted OR95%CIP value
Delivery-LOS was included as a log-transformed continuous variable†
Postpartum readmission
 Homeless0.860.75 to 0.990.04
 Non-homelessReference
 Postpartum ED visit
 Homeless0.830.77 to 0.9<0.001
 Non-homelessReference
Only women with a delivery-LOS between 2 and 4 days‡
Postpartum readmission
 Homeless0.90.78 to 1.030.12
 Non-homelessReference
 Postpartum ED visit
 Homeless0.820.76 to 0.89<0.001
 Non-homelessReference
11-Month follow-up§
Postpartum readmission
 Homeless0.830.62 to 1.110.21
 Non-homelessReference
 Postpartum ED visit
 Homeless0.850.69 to 1.030.1
 Non-homelessReference
Hospital fixed effect¶
Postpartum readmission
 Homeless0.850.7 to 1.040.11
 Non-homelessReference
 Postpartum ED visit
 Homeless0.820.75 to 1.89<0.001
 Non-homelessReference
Only 2014 records**
Postpartum readmission
 Homeless0.820.67 to 1.010.06
 Non-homelessReference
 Postpartum ED visit
 Homeless0.740.62 to 0.88<0.001
 Non-homelessReference
Non-homeless income quartiles††§
Postpartum readmission
 HomelessReference
 1st (lowest) income quartile1.181.02 to 1.360.02
 2nd income quartile1.110.96 to 1.280.15
 3rd income quartile1.140.98 to 1.310.08
 4th (highest) income quartile1.040.9 to 1.210.57
Postpartum ED visit
HomelessReference
 1st (lowest) income quartile1.261.17 to 1.37<0.001
 2nd income quartile1.191.1 to 1.29<0.001
 3rd income quartile1.111.03 to 1.210.009
 4th (highest) income quartile1.030.95 to 1.110.53

  • *The regression models were adjusted by age, race/ethnicity, insurance type during delivery, delivery length of stay, maternal comorbidity index score, other pregnancy complications, neonatal complications, delivery characteristics, that is, caesarean delivery, year fixed effect and a birth hospital random effect.

  • †Delivery-LOS data were positively skewed. To correct this, delivery-LOS was log-transformed and included in the model as a continuous variable.

  • ‡Only women with a delivery-LOS between 2 and 4 days were included in the model. Delivery-LOS was included as a categorical variable.

  • §Women who delivered in January during the study period (ie, 2009–2014) were followed for 11 months to examine their postpartum readmissions and ED visits.

  • ¶Hospital IDs were included as a fixed effect, rather than a random effect, in the model.

  • **Since personal identification numbers do not carry across calendar years, we included only records of women who delivered in 2014, to avoid potentially counting individuals who delivered multiple times over the study period (ie, 2009–2014) more than once.

  • ††§§Non-homeless individuals were divided into income quartiles based on zip code-level median income information.