Utilization and Cost
Descriptive Analysis of All Enrollees
Seven of the nine selected FEHB plans (FFS-NAT and HMO-NE were the two exceptions) increased in membership between 1999 and 2002. There was considerable variability in both probability of use and expenditures for MH/SA services. Probability of use for MH/SA services ranged from 11.7% to 16.8% in 1999 to 14.0% to 19.7% in 2002 with pre- to post-parity increases in all nine FEHB plans. Total expenditures for MH/SA services per user ranged from $647 to $1,390 in pre-parity 1999 to $755 to $1,306 in post-parity 2002, showing total spending increases in seven of the nine examined FEHB plans.
The probability of SA service use alone was extremely low throughout the period, ranging from 0.4% to 2.0% in 1999 to 0.5% to 2.3% in 2002, with little or no rise in any of the nine FEHB plans. Expenditures for SA services per user ranged from $16 to $74 in 1999 to $30 to $75 in 2002 with all but one plan experiencing an increase. Out-of-pocket expenditures for MH/SA services per user ranged from $60 to $403 in 1999 to $95 to $319 in 2002, with six of the nine plans showing a decline in out-of-pocket expenditures--even though most plans experienced little or no significant change in probability of use.
The parity policy was associated with an increase in per user total spending on medications for MH/SA disorders. Each of the nine selected plans exhibited a spending increase that ranged from 10.6% to 41.7% from 1999 to 2002. While per user medication spending ranged from $266 to $519 in 1999, in 2002 it increased to a range of $377 to $632.
Although the analysis of all enrollees is important for obtaining an overview of all use for the period from 1999 to 2002, it is problematic for assessing most of the questions about the impact of the parity policy directive. The changes observed between 1999 and 2002 can be accounted for by multiple factors. It is difficult to attribute them solely to the implementation of the parity policy directive. Observed changes could have been due to effects such as secular trends in MH/SA use experienced in the private insurance market more broadly or to the implementation of managed care or alterations in existing managed care practices. There may have been different patterns of out-of-plan use over this period or changes in the plans’ enrollment due to enrollees’ decisions to change plans during each year’s open enrollment period. The population of “all enrollees” in these analyses includes a mix of enrollees who have moved in and out of plans over the period.
It is better to focus questions of impact on the sub-population of individuals who were continuously enrolled in each of the nine FEHB plans for the entire 1999 to 2002 period. Focusing the analyses on the continuously enrolled population controls for the characteristics of the mix of individuals over the 4-year period, other than changes they undergo as a group during that period. The remaining analyses also make comparisons to continuously enrolled populations in a matched set of plans that did not experience the FEHB parity policy to control for secular trends and isolate the effect of the implementation of the FEHB parity policy directive.
The remaining summary findings report on the continuously enrolled population.
Before-after-parity Descriptive Analyses for MH/SA Utilization and Total Spending
There was considerable variability in both the probability of use and total spending per user for MH/SA services. Increases in the probability of use ranged from 0.5 to 3.0 percentage points, representing percentage increases ranging from 3.2% to 16.6%. Spending varied more dramatically with one plan experiencing an actual decline in spending of $45.32 per user (a decline of 6.4% from 1999 to 2002). The rest of the plans experienced increases in spending, ranging from $11.99 to $238.85 and from 1.0% to 46.9%. HMO-W1 had about twice the absolute increase in spending compared to the next highest plan, and its 46.9% increase was much higher than the rest of the plans that topped out at an increase of 19.2%.
Difference-in-differences Estimates for MH/SA Utilization and Total Spending
Only one of the nine FEHB plans had a significant post-parity increase in the probability of MH/SA service use after taking into account secular trends. FFS-MA2 experienced a 0.78 percentage point increase in the probability of MH/SA use. The estimated probability of use for the remaining eight plans was either not significantly different from zero or was negative and significant, as in the case of 2 plans. The difference-in-differences analysis for total spending showed a decline in expenditures of between $5.50 and $201.99 in seven of the nine plans; four of these differences were statistically significant. The two other plans showed non-significant increases.
Before-after-parity Descriptive Analyses for SA Utilization and Total Spending
Probability of use and total spending for SA services was a very small component of total MH/SA utilization, so the absolute impact of changes between 1999 and 2002 was small. The estimate of the before-after change in the probability of use of SA services, however, was positive and significantly different from zero for all nine plans, ranging from an increase of 7.2% to 61.1%. Most of the estimates, however, fell in the 23% to 38% range. The estimates for SA services spending showed a statistically significant reduction in spending in two plans and non-significant changes in six plans. HMO-W1 showed a dramatically different result; spending increased $1,130.60 (1,245%). (This result may be attributable to very small numbers.)
Difference-in-differences Estimates for SA Utilization and Total Spending
The difference-in-differences analysis showed that the probability of use for SA services increased between 0.01% and 0.25% in all of the nine selected FEHB plans, in relation to the matched comparison plans, and the increase was significant for four of the nine plans. The picture was somewhat more varied for total spending with a mix of positive and negative difference-in-differences results. Only one plan showed a significant change and that was a decrease of $288.41 per user for FFS-NAT when compared to a matched plan that did not experience parity.
Difference-in-differences Estimates for MH Utilization and Total Spending
As expected, the results of the difference-in-differences analysis for MH utilization and total spending were identical to the difference-in-differences analyses for MH/SA services combined. The level of SA use was so small (in all cases except for HMO-NE it was less than 1% of beneficiaries) that its removal from the combined utilization and total spending data had a negligible effect. There was only one significant increase and two significant decreases in probability of MH use, as well as significant decreases in total spending in four of the nine FEHB plans. All other results were non-significant.
Before-after-parity Out-of-pocket Spending for Adults
The parity policy was associated with a significant reduction in MH/SA out-of-pocket spending in six plans, ranging from 7.8% to 32.9%. Out-of-pocket burden, measured as the proportion of spending due to out-of-pocket expenses, was also reduced for these six plans. The three remaining plans, however, experienced significant out-of-pocket spending increases of 7%, 68%, and 141%. These results indicate that the parity policy increased financial protection for beneficiaries in most but not all plans.
Difference-in-differences Estimates for Out-of-pocket Spending for Adults
The difference-in-differences estimate for MH/SA out-of-pocket spending for adults showed a significant decline in out-of-pocket spending for five of the nine plans. For a sixth plan, out-of-pocket spending also declined, but the result was not significant. Out of-pocket spending significantly increased in three plans. As with the before-after analysis, the results indicate that the parity directive expanded financial protection for most but not all plans.
Before-after-parity Descriptive Analyses for Children’s MH/SA Utilization and Total Spending
All the plans experienced an increase in probability of child MH/SA use, ranging from 0.8 to 3.0 percentage point increases and 15.7% to 41.6% increases from pre-to-post parity implementation. Total spending also significantly increased in eight of nine plans--only FFS-S showed a non-significant increase. These spending increases ranged from 10.1% to 53.6%.
Difference-in-differences Estimates for Children’s MH/SA Utilization and Total Spending
The difference-in-differences analysis showed significant changes in probability of use, relative to secular trends, for children in only two plans -- FFS-NAT experienced an increase and HMO-NE a decrease. Results for the rest of the plans were not statistically significant. These findings indicate that, in general, the post-parity increases in MH/SA utilization observed in the before-after-parity analysis were on par with secular trends and thus unlikely a direct result of the parity policy.
Regarding total spending per user on children’s MH/SA, the difference-in-differences analysis comparing FEHB plans with non-parity comparison plans showed significant decreases in spending for four plans, ranging from $174.04 to $353.37, after secular trends were taken into account. The other impact estimates were also negative but were not significantly different from zero. These results parallel the differences-in-differences analysis in adults.
Adult High Service Utilizers of MH/SA Care--Inpatient Users
There was no significant change in inpatient utilization for eight of the nine plans from before to after the parity policy implementation. This result is likely due to small impact estimates. Only HMO-W1 experienced a rather large increase in hospital use during the post parity period, probably due to general expansion of services observed in other areas of study.
Comment on Utilization and Spending Findings
All of the analyses on MH/SA utilization and spending show an upward trend over this period, except for out-of-pocket spending. This upward trend is matched--and at times exceeded--by the pattern of utilization and spending in the comparison plans. As a result, when the secular trend is taken into account there was little overall increase in utilization and spending as an impact of the parity policy. However, out-of-pocket spending on MH/SA care actually declined in most plans.
Quality of MH/SA Care
Quality of Care for SA Treatment
Measures of quality included five process indicators based on the Washington Circle Group’s recommendations, including rates of utilization, identification of individuals with substance use disorders, and engagement in treatment. Except for a small increase in rates of identification, there was no evidence of significant change over the period of parity implementation from 1999 to 2002.
Quality of Care for Major Depressive Disorder Treatment
Measures of quality for treating major depressive disorder (MDD) either did not change or improved only slightly from pre-parity 1999 to post-parity 2002 in all but one selected FEHB plan. In each year from 1999 to 2002, approximately 90% of those diagnosed as having MDD received at least one therapy session and/or antidepressant prescription. Quality improvement was more notable in the use of medication than for psychotherapy in the treatment of MDD. Three of the selected FEHB plans improved from pre- to post-parity in meeting a minimal standard for the duration of MDD follow-up; two plans improved only slightly in meeting a minimal standard for the intensity of MDD follow-up. None of the selected FEHB plans experienced quality improvement in meeting minimal standards of psychotherapy intensity or duration, or in antidepressant duration.