|
UPDATE: SPECIAL REPORTThe Impacts Of Mental Health Parity And Managed Care In One Large Employer Group: A Reexamination
Although the impacts of carve-outs to managed behavioral health care organizations (MBHOs) and parity mandates on costs are largely settled in the literature, their impacts on access are less clear. Here we reexamine a study published by Samuel Zuvekas and colleagues in this journal, which found that the number of people receiving mental health/substance abuse treatment increased by almost 50 percent after the introduction of mental health parity and an MBHO. Using multivariate panel data methods, we now suggest that secular trends were largely responsible for this increase.
Numerous case studies demonstrate that managed behavioral health care organizations (MBHOs) reduce mental health specialty provider treatment costs, even when mental health benefits are greatly expanded.1 Less clear is how access to mental health treatment changes in response to changes in coverage and the introduction of an MBHO carve-out. In this paper we reexamine a finding from a previously published case study by Samuel Zuvekas and colleagues, which showed an almost 50 percent increase in the number of people using outpatient specialty mental health and substance abuse (MH/SA) treatment services after the simultaneous introduction of mental health parity and an MBHO in one large employer group.2 In that paper we were unable to determine whether this observed increase was attributable to the mental health parity benefit, the introduction of the MBHO carve-out, or secular increases in the use of MH/SA treatment. Since the publication of that paper in 2002, a number of descriptive studies have documented substantial nationwide increases in the use of MH/SA treatment from the late 1980s onward.3 These papers suggest that the increases are likely driven by the availability of newer medications with fewer side effects, direct-to-consumer advertising, and possibly reduced stigma. These findings motivated us to reexamine the data using multivariate panel data methods to determine the extent to which secular trends might explain the increases. Methods. Our case study comprises a single large employer group and a collection of medium-size and small employers with coverage through the same fee-for-service insurer. The insurer implemented a carve-out to an MBHO for all of its groups after the large employer was required under legislative mandate to provide parity coverage for severe mental disorders. The large group also chose to reduce consumer cost sharing (but not visit limits) for other MH/SA disorders. The midsize and small groups also had their MH/SA services carved out to the MBHO but without parity. These groups also had less generous coverage and lower baseline utilization. We included them because relatively little is known about smaller groups’ experience under MBHOs, and we wanted to see whether the same secular trends were present for all employer groups.4 Instead of yearly observations as before, for this update we used quarterly observations to examine trends in the use of ambulatory specialty MH/SA treatment. Specifically, we compared trends in the four quarters before the implementation of the MBHO and parity with the twelve quarters afterwards. The sixteen quarters of observations for each person yielded 1,205,760 person-quarter observations for the large group and 147,648 observations for the midsize/small groups. We used multivariate panel data methods to test whether the quarterly trend changed from the pre period to the post period. We used a probit model to describe the probability of any ambulatory specialty mental health use during a quarter. To account for the within-person correlation (sixteen quarterly observations for each person), we used the generalized estimating equations (GEE) approach to estimate the probit model.5 We also controlled for age, sex, and type of enrollee (employee, spouse, or dependent). To these we added county-level geographic and provider supply measures from the Area Resource File (ARF). Finally, we added controls for provider type change associated with the MBHO.6 Our specification allowed for a different, linear quarterly trend in the pre and post periods, providing a direct test of whether the trend changed. We also estimated separate trends for the large and medium/small employer groups. We present estimates of the quarterly trends and change in trend computed from the probit model results according to the method described by Chunrong Ai and Edward Norton.7
Results. The percentage of people using ambulatory specialty mental health treatment increases steadily for both the large and midsize/small groups (Exhibit 1
We formally tested whether there was a change in trend after the introduction of the MBHO and parity using the multivariate panel data techniques. We estimate that the percentage of enrollees using ambulatory MH/SA treatment increased, on average, by 0.14 percentage points per quarter in the pre period for both the large and the midsize/small employer groups (Exhibit 2  ). The trend appears to attenuate in the post period. Average quarterly increases in the percentage of enrollees using ambulatory MH/SA treatment in the post period were 0.07 percentage points for the large group and 0.08 percentage points for the mid-size/small groups. However, the difference in the pre and post period trends is statistically significant for the large group only.
Discussion. We return to our central question of whether the increases in the number of people getting specialty treatment were attributable to reduced consumer cost sharing under parity, the introduction of the MBHO, or an underlying secular trend of increased demand for treatment. In our previous paper we suggested that parity probably did not have a substantial impact but that the MBHO might have increased the number of people receiving specialty care. We now suggest that it was the underlying secular trend that was largely responsible for the increase. We found no evidence that the MBHO increased the number of people in treatment. If anything, the smaller increases observed in the post period in both the large and the midsize/small groups after the introduction of the MBHO suggest that it might have moderated the underlying secular trend. Similarly, the smaller increases in the number of people in treatment after the simultaneous introduction of parity along with the MBHO in the large group suggest that parity is not responsible for the increases in treatment use, either. The limited scope of the parity mandate likely reduced its potential impact. It is important to note that because we cannot disentangle the effects of the MBHO from the effects of parity in the large group, reduced consumer cost sharing from parity might have partially offset the incentives to reduce costs in the MBHO. Nevertheless, the net effect of parity and the MBHO appears not to have increased the number of people in specialty treatment. This is not to say that parity or carve-outs to MBHOs do not affect the number of people receiving treatment. We also note that as does most of the published literature, we lack good comparison groups to fully control for secular trends. We must rely instead on differences in the pre and post trends to examine the potential impacts of the MBHO and mental health parity. As a result, we cannot be certain that the lower growth rates observed in the post period are attributable to the MBHO and not some change in underlying secular trends. However, the use of MH/SA treatment increased throughout the three-year post period, which suggests the continued presence of an upward secular trend.
Sam Zuvekas (szuvekas{at}ahrq.gov) is a senior economist at the Center for Financing, Access, and Cost Trends, Agency for Healthcare Research and Quality (AHRQ), in Rockville, Maryland. Agnes Rupp is chief of the Mental Health Economics Research Program at the National Institute of Mental Health in Bethesda, Maryland. Grayson Norquist is professor and chairman of the Department of Psychiatry and Human Behavior, University of Mississippi Medical Center, in Jackson. The authors thank Steve Hill, John Fleishman, Chad Meyerhoefer, and Jessica Banthin for their insightful comments; Donald Rae for programming support; and Darrel Regier, who made this research possible. An earlier version of this paper was presented at the Sixth Workshop on Costs and Assessment in Psychiatry, Venice, Italy, March 2003. The authors thank the workshop participants and two anonymous reviewers for their helpful comments. The views expressed in this paper are those of the authors, and no official endorsement by the Agency for Healthcare Research and Quality, the National Institute of Mental Health, or the Department of Health and Human Services is intended or should be inferred.
This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||
Top
NOTES
