Expanded Mental Health Benefits and Outpatient Depression Treatment Intensity Anthony T. Lo Sasso University of Illinois at Chicago Richard C. Lindrooth Medical University of South Carolina Ithai Z. Lurie John S. Lyons Northwestern University Research supported by NIMH grant R01MH62114 Goal of the Research To study how a mental health benefit expansion affected outpatient depression treatment utilization Particular emphasis will be placed on examining how the benefit change affected the “dose” of outpatient psychotherapy received by patients Initiation → OP visits → sufficient # of OP visits Setting & Benefit Design Large Fortune 50 electronics manufacturing company Between 1995 and 1996, the company initiated a new approach in health benefits for mental health with three components: – reduced copayments for mental health treatment – the creation of a network of select mental health providers – de-stigmatize the treatment of mental disorders Details on the Benefit Change Selective contracting with mental health providers High cost sharing & utilization limits for out-ofnetwork care No limits are placed on in-network mental health services Coinsurance rates on inpatient care were changed from 80% to 90%, 80% coinsurance for outpatient services was replaced with a flat $10 copayment Data Claims and enrollment data from intervention company for years 1995-1998 – for all persons with a mental health diagnosis or procedure code suggestive of mental health treatment – Data only for continuously enrolled members of the company’s self-insured health insurance plan Comparison group comprised of random sample of employed persons from Medstat Marketscan database, 1995-1998 Methods We compare outpatient treatment in the intervention group to the comparison group over time: – Outpatient treatment initiation for depression – Outpatient treatment visits – An indicator variable for more than 8 visits for psychotherapy Basic model: OP Visits = α + βPost + γIntervention + δPost×Treatment + θX + ε Methods II By restricting the regression to the withinprovider type impact, we can isolate the impact of the copayment reduction & destigmatization: OP Visits = α' + β'Post + γ'Treatment + δ'Post×Treatment + λ1 Non-MD Specialist + λ2 MD-Specialist + θ'X + ε Where δ' represents the remaining effect of the benefit change controlling for the effect of provider choice Methods III Using estimates of the elasticity of demand for mental health treatment from published studies, we can decompose δ' into the expected impact of the cost sharing reduction and destigmatization Descriptive Statistics of Depression Treatment Initiation and Utilization Intervention Group Control Group Initiation % OP visits % > 8 Initiation visits % OP visits %>8 visits 1995 2.2 5.6 28.7 2.5 5.6 21.7 1996 2.6 7.5 41.0 2.3 6.0 25.2 1997 2.7 6.8 38.0 2.0 5.8 27.4 1998 2.7 6.8 37.3 2.1 6.3 26.9 Regression Results, Basic Specification Treatment× Post (δ) Initiation (Logit-OR) OP Visits (OLS) 8 or More Visits (Logit-OR) 1.33** 1.25** 1.25 N=310,882 for Initiation regression, n=7,560 for other regressions * Indicates 0.01<p<0.05, ** indicates p<0.01 Descriptive Statistics of Specialist Use & Percentage > 8 Visits 1995 1996 1997 1998 Intervention Group Comparison Group Generalist Non-MD specialist MD Specialist Generalist Non-MD specialist MD Specialist 511 385 143 157 226 171 22.5% 35.8% 31.5% 22.9% 21.2% 21.1% 530 765 150 135 198 132 31.1% 46.9% 46.5% 18.5% 27.3% 28.8% 609 771 124 150 270 142 36.3% 39.4% 37.9% 16.0% 34.1% 26.8% 647 775 83 118 274 117 34.9% 39.0% 39.8% 7.6% 37.2% 22.2% Regression Results, WithinProvider Type Treatment× Post (δ') OP Visits (OLS) 8 or More Visits (Logit-OR) 1.03** 1.22 N=7,560 * Indicates 0.01<p<0.05, ** indicates p<0.01 Summary of Findings Our prior work has shown that the selective contracting network was not defined restrictively enough to increase distance to providers, but the emphasis on specialist access (particularly non-MD) had a significant impact on utilization: 1.25-1.03 = +0.23 visits Based on the RAND HIE elasticity of -0.17, we calculate the predicted effect of the copayment reduction to +0.42 visits The remainder then is attributed to destigmatization: +0.6 visits Under Different Assumptions… If instead we use a higher elasticity measure from the literature, -0.32, the predicted impact of the copayment reduction is +0.79 visits Under this assumption, we are left with a de-stigmatization effect of +0.23 visits Discussion Our empirical results suggest that the combination of benefit changes instituted by the company was successful in not only encouraging more employees to enter depression treatment, but also in enabling more patients to stay in treatment longer increasing the likelihood that they receive the appropriate minimum adequate “dose” of 8 outpatient psychotherapy visits Discussion II Our results suggest that each aspect of the benefit change—promotion of innetwork specialty providers, reducing copayments, and a company-wide effort to reduce the stigma associated with mental health treatment—played an important part in increasing the number of outpatient visits per episode of treatment conditional upon treatment initiation for depression Conclusion Starting treatment is not what matters in behavioral healthcare Engaging in a successful episode of care is a far more relevant indicator The results point to the potentially important role that corporate benefit designers and planners can play in improving access to mental health care for employees