Table 2.

Characteristics associated with having any medical spending & associated with the amount of spending on care occurring outside of the ACO (among those with spending)

	All individuals			Among those with any medical spending
	Odds of having any medical spending			% of spending on care delivered inside of the ACO
	Odds Ratio	95%	CI	Coefficient	95%	CI
24+ months (vs <24 months)	1.17	1.16	1.18	4.40	4.17	4.62
Female	1.66	1.64	1.67	−1.41	−1.61	−1.21
Aged <18 years	0.71	0.70	0.72	7.74	7.37	8.10
Aged 18–25 years	0.55	0.54	0.56	−6.27	−6.65	−5.88
Aged 26–35 years	0.61	0.60	0.62	−4.71	−5.04	−4.38
Aged 36–45 years	0.73	0.72	0.74	−3.53	−3.86	−3.20
Aged 46–55 years	0.80	0.78	0.81	−2.74	−3.05	−2.42
Aged 56–64 years	reference			reference
Mean HHS-HCC risk score in ACO entry month	1.10	1.10	1.10	0.03	0.00	0.05
Plan type – HMO	reference			reference
Plan type – PPO	1.14	1.12	1.15	3.03	2.78	3.29
Plan type - POS	1.14	1.11	1.17	−1.37	−1.96	−0.79

Notes: ACO: accountable care organization; CI: confidence interval; HHS-HCC: refers to a measure of predicted medical spending, i.e., the Department of Health and Human Services-Hierarchical Condition Categories; HMO: health maintenance organization; PPO: preferred provider organization; POS: point of service. We used logistic regression models to examine the odds of having any monthly medical spending, adjusting for insurance plan, insurance plan type, year of entry, month, and primary physician practice group, plus individual-level random effects. We then used linear regression models to examine the percentage of monthly spending on care occurring outside of the ACO among those with any spending, adjusting for insurance plan, insurance plan type, year of entry, month, and primary physician practice group, plus individual-level random effects. These two models are analogous to the components of a two-part model.