The parity policy in the Federal Employees Health Benefits program began on Jan. 1, 2001, and offers comprehensive insurance coverage for mental disorders, including substance use disorders, on terms that are identical to the coverage of general medical conditions when the treatment is provided by in-network providers.
We compared seven Federal Employees Health Benefits plans with a matched set of plans that did not change benefits or management and did not have parity. We compared use and spending by enrollees in these plans for the 2 years before parity (1999 and 2000) and for the 2 years after parity began (2001 and 2002). We observed 1) the proportion of federal employees, retirees, and their dependents who used behavioral health services; 2) how much they spent for behavioral health services; and 3) how much of the spending was out of their own pockets (1).
Parity for behavioral disorders covers all use of health care services for any of the disorders (including substance use disorders) in DSM or the mental disorders chapter in the ICD. It includes specialty mental health services, such as psychotherapy, as well as visits to a general medical provider when a mental disorder diagnosis is recorded. It also includes the use of all medications for which behavioral health conditions are an indication. When medications might be used for a mental disorder or a general medical condition, use and spending were included only if accompanied by a mental disorder diagnosis in the record. This is the broadest definition of use and spending, designed to capture the impact of parity.
The study found that the policy was implemented smoothly and without plans dropping out of the Federal Employees Health Benefits program. There was a significant decline in out-of-pocket spending by people who used behavioral health services in the Federal Employees Health Benefits plans compared to the nonparity plans. This indicates that parity coverage resulted in improved insurance protection against financial risks—the principal objective of health insurance. This savings to Federal Employees Health Benefits plan members was not associated with significant increases in use and spending attributable to parity. In fact, for the most part, increases in use and total spending in the Federal Employees Health Benefits plans were no greater than use and total spending increases in the comparison plans (1). This was true for adults as well as for children and adolescents (2).
We concluded that “parity of coverage of mental health and substance abuse services, when coupled with management of care, is feasible and can accomplish its objectives of greater fairness and improved insurance protection without adverse consequences for health care costs” (1, p. 1386).
The parity policy performed just as insurance should: it reduced costs from out-of-pocket payments with a small increase in plan payments (3). This could result in very small increases in insurance premiums without leading to an increase in the use of services. The Congressional Budget Office estimates a premium impact for group plans of a 0.4 percentage point increase (4), a figure that is identical to our estimate based on the Federal Employees Health Benefits experience.
We also looked at indirect measures of quality of behavioral health care in the Federal Employees Health Benefits plans during this same period. Parity was accomplished without increases in the hospitalization of patients and without a decline in the measures of quality of care that we studied, such as the likelihood of receiving follow-up care for depression or being referred for substance abuse treatment.
There was no use of or spending for (oft-parodied) trivial behavioral conditions under managed care plans. It is worth noting that the ICD contains a wide range of general medical conditions, such as scrapes and bruises, rashes, sprains, and the common cold, just as it includes sleep disorders, mild phobias, and mild learning problems. Managed care arrangements and “medical necessity” criteria control unnecessary use and spending for trivial cases of general medical conditions and mental disorders alike.
In response to concerns raised about a mandated benefit, we conclude that by reducing financial risk, parity improves the well-being of insured people without distorting the market for mental health services.
Legislation is the way to achieve this social good because parity coverage offered by only one or two plans would result in those plans probably attracting a disproportionate share of people with persistent mental illness (3). This is what is referred to as “adverse selection.” In fact, legislation of parity provides the best protection for insurers and self-insured companies from experiencing adverse selection. When they offer parity benefits at the same time as mandated by the legislation, they can avoid a shift of high-cost individuals into their plans. A mandate thereby promotes market efficiency—or at least avoids the market failure associated with adverse selection. Ironically, a mandate may help insure employers and plans against financial risks when they try to offer better benefits to their employees.
For decades, advocates for parity relied only on an argument of fairness to gain support for their cause. Now they can argue that parity promotes social well-being and economic efficiency in the form of better insurance benefits for all of us.