Preventive Medicine br Table br Community health

 Preventive Medicine 124 (2019) 91–97
Table 1
Community health center and eligible patient characteristics, 2012–2015.
Expansion states (9 Non-expansion
N primary care CHCs 131
Last payer mix (% Medicaid/
Medicare/Other Public/
Private/Uninsured)
Patient demographics
patient
Note: ACA = Affordable Care Act; CHC = HA 1077 health center;
FPL = Federal poverty level; SD = standard deviation; Q = quarter.
This study was approved by the Oregon Health & Science University Institutional Review Board. Clinical trial registration: NCT02936609.
3. Results
We observed 329,126 patients eligible for screening with > 2.8 million ambulatory visits to expansion state CHCs and 295,475 patients with > 1.7 million visits in non-expansion state CHCs in the four-year study period (Table 1). The visit payer mix followed expected patterns pre- to post-ACA: in expansion states, Medicaid visit rates increased from 33.8% in 2012–2013 to 52.7% in 2014–2015, while uninsured visit rates decreased from 37.3% to 18.0%. In non-expansion states, uninsured visit rates decreased from 51.3% to 40.0% and privately-insured visit rates increased from 7.5% to 22.3% while the proportion of Medicaid visits stayed constant. Compared to the expansion state sample, patients of non-expansion state CHCs were more commonly
Table 2 Cervical cancer screening prevalence, pre- and post-ACA by Medicaid expansion status overall and stratified by insurance group and race/ethnicity.
Expansion states
Non-expansion states
DID (post/pre, expansion/
Pre-ACA Post-ACA Odds of screening,
Pre-ACA Post-ACA Odds of screening,
Overall
Hispanic
NH black
NH white
Note: ACA = Affordable Care Act; NH = Non-Hispanic; DID = difference-in-difference.
Eligible population for cervical screening: women age 21–64 throughout period, without history of hysterectomy, due for cervical cancer screening (no doc-umentation of pap received in past 3 years or pap + HPV testing in past years if age 30–64). Odds ratios and DID ratios were obtained from logistic generalized estimating equation models with a robust sandwich variance estimator specifying an independent working correlation structure of health systems nested within states and adjusted for age, race/ethnicity (for the overall sample), federal poverty level, payer type, urban/rural status, and number of ambulatory visits.
female, older, non-white, lower income, and more urban (Table 1).
3.1. Cervical cancer screening
Among female patients aged 21–64, the adjusted prevalence of cervical cancer screening increased pre- to post-ACA in both expansion and non-expansion state CHCs. Patients had 19% increased odds in expansion states and 23% increased odds in non-expansion states of receiving cervical cancer screening post-ACA relative to pre-ACA (Table 2). Prevalence of screening also increased in each insurance subgroup (Medicaid, private insurance, uninsured) in both expansion and non-expansion states. The greatest increase was among uninsured patients in expansion states (aOR = 1.36, 95% CI = 1.16–1.59) and privately-insured patients in non-expansion states (aOR = 1.43, 95% CI = 1.11–1.84). There was no significant relative difference in pre- to post-ACA changes for cervical cancer screening in expansion vs non-expansion states (Table 2).
For results stratified by insurance status and race/ethnicity, non-Hispanic white patients in both expansion and non-expansion states had lower rates of screening relative to Hispanic or non-Hispanic black patients (Table 2). Overall, all race/ethnicity groups in both expansion and non-expansion states improve their odds of screening from pre- to post-ACA with the greatest change among non-Hispanic blacks in ex-pansion states (aOR = 1.36, 95% CI = 1.08–1.71). In expansion states, the odds of screening were higher among non-Hispanic white patients with Medicaid coverage or no coverage post-ACA compared to pre-ACA while in non-expansion states all insurance groups increased their odds of screening. For Hispanic patients, the odds of cervical cancer screening increased from pre- to post-ACA for uninsured patients in expansion states (aOR = 1.48, 95% CI = 1.18–1.84) and Medicaid-in-sured patients in non-expansion states (aOR = 1.29, 95% CI = 1.01–1.65). Among non-Hispanic black patients, patients with private insurance or no coverage in expansion and Medicaid coverage or private insurance in non-expansion states showed increased odds of receiving cervical cancer screening post- compared to pre-ACA. There 
was no relative difference in pre- to post-ACA changes for cervical cancer screening in expansion vs non-expansion states by race/ethni-city.
3.2. Colorectal cancer screening
Colorectal cancer screening increased from 10.9% to 17.7% pre- to post-ACA in expansion states (aOR = 1.76, 95% CI = 1.41–2.18) and from 12.9% to 21.0% in non-expansion states (aOR = 1.79, 95% CI = 1.28–2.53) (Table 3). Screening improved for patients with all insurance types: the greatest increase was among patients with Medi-caid coverage in expansion states and Amber mutation with private coverage in non-expansions states. For the overall study sample, no significant changes in screening prevalence pre- to post-ACA between expansion groups were observed.