Chapter 2 - Case Studies
Mental health and substance abuse (MH/SA) parity may affect insurers, employers, state agencies, and consumers. We used case study methods to more fully understand the effects of parity from the perspectives of these groups.
Case study analyses of parity describe informants' judgments about the effects of parity and provide information more quickly than statistical analyses. But unlike statistical analyses, they cannot be used to estimate the costs of parity or provide the percentage of organizations that are affected. Statistical studies require detailed health expenditure data before and after parity is implemented. In addition, for comparison purposes, statistical studies require health expenditure data from employers that have implemented parity and those who have not. To date, published statistical studies of the insurers' or employers' experiences with parity are limited (National Advisory Mental Health Council, 1997).
Study Methods
To explore the effects of MH/SA parity on key stakeholders, we conducted case studies in 5 of the 12 states that have parity laws. We wanted each state to have at least one year of experience with parity. Therefore, we chose the four states that required parity in coverage for private-sector employees on or before 1995 (Maryland, Minnesota, New Hampshire, and Rhode Island), and Texas, which is one of two states with a law applying to coverage for public employees. In addition, we contacted five employers who voluntarily adopted parity for MH/SA insurance benefits, and two agreed to participate in our study.
From May through July 1997, we contacted 47 organizations in the study states, including insurers, employers, employer associations, and public officials, such as representatives from the state departments of mental health and substance abuse. We also spoke with provider associations, such as state psychiatric associations, and consumer and family advocates. Informants from the 5 states included representatives from 9 insurers, 6 employers and unions, 3 associations of small employers, 18 public agencies, 6 provider associations, and 5 consumer and family advocacy groups.
We spoke with more organizations in states with broader laws. These included 18 organizations in Minnesota, 13 in Maryland, 8 in Rhode Island, 6 in New Hampshire, and 2 in Texas. We did not contact any insurers in New Hampshire because Lewin (1997) had recently done so. Finally, we spoke with benefits staff from the two employers that voluntarily adopted parity. Some of the information shared with us is confidential. We therefore keep the organizations and individuals that participated in this study anonymous.
Study Findings
Impact of Parity on Premiums
We asked 20 insurers, employers, and insurance regulators about the effects of parity on MH/SA expenditures and premiums. We spoke with at least one insurer, employer, or regulator in each study state and many in Minnesota. All nine of the insurers in our case studies used managed care for MH/SA treatment services. In the discussion below, we emphasize the reports from the most knowledgeable informants.
For two reasons, many informants could not say exactly if, or by how much, parity raised MH/SA costs or service use. First, data on the subject were sometimes confidential. Second, because MH/SA expenditures are generally a small portion of a health insurer's total premium, many insurers do not allocate resources to collect these data. According to William M. Mercer (1997), MH/SA expenditures constitute about 4 percent to 7 percent of total health care expenditures in a highly managed preferred provider organization (PPO) plan.
Even when changes in MH/SA expenditures were known, informants reported that other factors, such as competition, could have larger effects on premiums than parity. These other factors made it difficult to precisely determine the role of parity.
Informants knew the trends in total premiums, but most relied on their judgment and experience to decide whether any change in premiums was due to MH/SA parity laws. Informants provided more detailed information on the MH/SA expenditures or utilization for only six plans. Given the small number of informants and the variable quality of the information they provided, our answers to the research questions are qualitative, rather than precise.
Most insurers, especially managed care plans, experienced small increases in total premiums. Most, but not all, insurers in Maryland, Minnesota, New Hampshire, and Rhode Island reported small increases in total premiums due to MH/SA parity laws. Representatives from two managed care companies in Maryland stated that premiums increased by 1 percent or less due to the 1995 MH/SA parity law. After this initial increase, total premiums generally "leveled out."
However, the two Maryland managed care plans had different experiences with MH/SA inpatient and outpatient expenditures and service use. While one plan observed a "slight" increase in outpatient service use by enrollees as a result of parity, the second reported a "significant" increase in the use of outpatient and partial hospitalization services. Inpatient MH/SA hospital use rose for enrollees in the first plan but dropped slightly for enrollees in the second. Hospital use was measured by the number of days per thousand enrollees.
For the Maryland insurer with the largest increase in outpatient care, expenditures increased by 22 percent during the first 6 to 8 months after parity was implemented. Because MH/SA expenditures represent about 5 percent of total health care costs, the total premium increase due to parity was just over 1 percent. However, this informant told us that a 22 percent increase in MH/SA expenditures is a large increase in MH/SA costs.
In Minnesota, informants said that after the 1995 MH/SA parity law was enacted, health care premiums did not increase or did so by only a few percentage points. Relying on their judgment and experience, six informants said they believed that premium increases due to parity were small. Two of these informants quantified premium increases as 1 percent or 2 percent. These findings are similar to those reported by Koyanagi (1996) and Blewett (1997).
In New Hampshire, informants used data on total premium increases along with their judgment and experience to assess the increase in premiums due to the parity law. These informants estimated that the state's 1995 serious mental illness (SMI) requirement led to premium increases of 5 percent or less. These findings are consistent with the findings from the Lewin Group (1997).
Informants in Rhode Island generally did not attribute premium increases to the newly established parity for SMI benefits. The increases they did mention were very minor. One informant from a managed care company observed no impact on costs but noted that it may take some time for the law to have an effect. Since the law was passed, no enrollees in this company's plan have used more than 90 days of inpatient care, which is the minimum benefit under the law.
Another Rhode Island managed care company with information based on health expenditure data observed a premium increase of less than 1 percent. One informant said that the state's SMI parity law did not lead to large premium increases because it covers only medication visits and hospitalization.
Expenditures for MH/SA treatment dropped if managed care was introduced at the same time as parity. The SMI parity mandate for state employees in Texas became effective in September 1992. At that time, a managed behavioral health care plan replaced the fee-for-service (FFS) plan that had been offered to state employees. State employees could enroll in this managed care plan offered by Blue Cross/Blue Shield of Texas or in one of several HMOs.
From 1992 to 1995, the cost of MH/SA care for state employees enrolled in the Blue Cross/Blue Shield plan dropped by 47.9 percent (National Advisory Mental Health Council, 1997). By 1995, all claims for SMI amounted to about $2.40 per member per month. This is equal to about 1.5 percent of all claims costs2. HMO data were unavailable.
2 These costs are based on data in a letter from Sheila W. Beckett, Executive Director, Employees Retirement System of Texas, to Texas Rep. Garnett Coleman, February 27, 1997.
Of the two employers that voluntarily adopted parity for MH/SA services, one had extremely high MH/SA costs in a FFS plan before offering parity. To contain costs, the company decided to offer employees a managed care plan with mental health parity, in part to make the plan more attractive to employees who were opposed to managed care. Costs for MH/SA care fell dramatically because of the increased utilization management.
Employers who did not have such high costs before parity likely would not realize comparable cost savings, our informant noted. These findings and those from Texas are consistent with findings of another case study conducted by William M. Mercer, an employee benefits consulting firm, which found that two employers that implemented parity lowered their MH/SA expenditures by using managed care firms (Mercer, 1997).
Reasons Why Premium Increases Were Small
Case study informants identified two main reasons for the small total premium increases after MH/SA parity laws were passed. First, as mentioned, managed care played a key role in containing expected cost increases. Second, parity represented only a small increase in MH/SA benefits for some states.
Managed care and competition constrained premium increases. Managed care played a role in containing expected cost increases after MH/SA parity laws were passed. As noted above, MH/SA treatment dropped dramatically for employers who introduced managed care for MH/SA services when parity was adopted.
In Maryland and Minnesota, informants noted that MH/SA care was moderately to tightly managed before parity. This resulted in either no premium increase or an increase of a few percentage points or less. Several informants said that competition in the health care market in Minnesota also constrained premium increases. For example, a purchaser changed health plans because premiums rose by 2 percent. (Informants said the increase was not due to the parity law.) According to one informant, some health plans in Minnesota had been heavily discounting their premiums. However, insurers might have more heavily discounted premiums if the state had not mandated parity.
Some states already required broad MH/SA coverage, or they legislated limited parity. Parity resulted in only a small increase in MH/SA benefits in New Hampshire and Rhode Island, in part because the 1995 laws apply to a small number of people (those with SMI), and New Hampshire had generous benefits before parity was implemented.
In New Hampshire, our informants explained, the 1995 parity law applies only to people who have serious mental illnesses, who represent a very small percentage of the population. Less than 3 percent of U.S. adults have serious mental illnesses (National Advisory Mental Health Council, 1993). In addition, case study informants told us that few people with serious mental illnesses work for large employers that have health insurance plans. In Rhode Island, the new law applies only to medical treatment for serious mental illnesses.
Further, two informants noted that New Hampshire had passed MH/SA mandates as early as 1975. At that time, the state required group insurers and HMOs to provide mental health hospital services on the same basis as services for other illnesses under major medical care. The law also required group insurers and HMOs to cover at least 15 hours of outpatient treatment after two visits (The Lewin Group, 1997).
Impact on Employers and Insurers
MH/SA parity laws increase the incentive for insurers and employers to minimize the cost of MH/SA treatment. They may reduce these costs in three main ways. First, employers can offer coverage through managed care rather than FFS plans. Second, they can drop insurance coverage or become self-insured. State parity laws do not apply to self-insured employers. Finally, they can pass on the costs of parity to employees by raising employee contributions to health insurance or by paying lower wages.
Employers and insurers used managed care to contain costs. Parity was related to increased enrollment in managed MH/SA care plans and tighter utilization management for MH/SA treatment. Public employers in Texas and one employer who voluntarily adopted parity replaced their FFS MH/SA plans with managed behavioral plans. In addition, some informants reported that parity increased the intensity of utilization management.
In New Hampshire, managed care plans responded to parity by tightening the criteria they use to determine medical necessity. Some insurers in New Hampshire reported expanding their case management protocols for SMI (The Lewin Group, 1997). The managed care market was already strong in Minnesota, Maryland, and Rhode Island, so informants did not report much increase in managed care in those states.
Employers did not become self-insured or decide to pass on the full cost of parity to employees. None of the insurers or associations of small employers in our study identified MH/SA parity laws as a main consideration in a decision to self-insure. However, they may have made the decision for other reasons. For example, in Minnesota, some employers decided to self-insure to avoid a 2 percent premium tax.
None of the employers we spoke with mentioned changing employee contributions to reflect changes in MH/SA benefits. For example, two employers continued to use the same premium contribution formula, so employees paid the same percentage of their health insurance costs. A few informants said that the low costs of adopting the law might be responsible for the lack of employer response. In addition, employers may have decided to keep employee contributions the same in order to sustain employee morale and good will.
In Minnesota, small employers may offer their employees a basic benefit plan instead of a plan that provides full MH/SA parity. A case study participant told us that most small employers, however, chose a plan subject to the state laws because the total costs are about the same. Under Rhode Island's Small Employer Health Insurance Availability Act, employers with fewer than 50 employees may purchase a standard or economy benefit plan with less extensive coverage than full parity. However, an informant told us that these benefit packages are so limited that employers never purchase them. Instead, they choose coverage subject to the law.
Most employers did not collect and analyze data to measure changes in productivity or employee absenteeism due to parity. It could be argued that employees who use the additional MH/SA benefits may function better in the workplace and become more productive. Employers would benefit from less absenteeism and lower employee turnover. However, case study informants thought that parity did not affect productivity, or they did not know whether it did.
Two informants said that MH/SA parity had no effect on employee absenteeism or turnover. Several informants told us that because productivity is difficult to measure, they could not determine the extent to which more generous MH/SA benefits increased productivity. One informant credited increased productivity to the company's employee assistance plan.
These findings are consistent with those of William M. Mercer (1997), which conducted a telephone survey of benefit managers at 24 companies. Of the six companies that provided the most comprehensive MH/SA coverage, none had measured changes in productivity or absenteeism due to MH/SA parity.
Impact on Public MH/SA Expenditures
State and local governments traditionally have financed a substantial portion of MH/SA services. Some actuaries have assumed that MH/SA parity laws would shift the provision of MH/SA care from the public to the private sector (Bachman, 1996b; Rodgers, 1996).
This reasoning is based on two assumptions. The first assumption is that before parity was legislated, the public sector was financing MH/SA care for people who were privately insured but whose benefits did not cover necessary MH/SA services. Under parity, these people should be able to bill their private insurers for more MH/SA care instead of using publicly funded services. The second assumption is that parity mandates must increase the amount or type of services that private insurers cover. Neither of these assumptions was borne out by our study.
Nearly all case study informants reported they had seen no changes in state spending on MH/SA as a result of parity. These informants included state program officials from Maryland, Minnesota, New Hampshire, and Rhode Island, as well as insurers, MH/SA provider associations, and consumer advocates. Furthermore, detailed expenditure data from New Hampshire's public mental health system show no evidence of a decline in public mental health expenditures resulting from parity.
Our informants gave two key reasons why parity did not seem to affect public MH/SA expenditures. First, publicly financed MH/SA services are provided primarily to people who have serious mental illnesses or severe substance abuse disorders. Because most of these individuals are unable to work or can work only part-time, they have no access to private insurance. Therefore, they are not affected by parity.
Second, the public system finances many services that private insurers do not cover, even under parity, because they are not considered medically necessary. Such services include psychosocial services (such as psychosocial rehabilitation and life-skills training) and services requested by a third party (such as court-ordered services). People needing these services must seek care from publicly financed providers.
In addition, ERISA and small employer exemptions reduce the number of people affected by parity laws. State MH/SA parity laws do not affect most people in employer-sponsored health plans because they are enrolled in a self-funded plan (exempt under ERISA), or because they are in a small employer plan (exempt from the state law).
More specifically, we found that state MH/SA parity laws affect only about 30 percent of people with health insurance in Maryland, Minnesota, and Rhode Island (Lipson and De Sa, undated; Maryland Insurance Administration, undated). Informants in New Hampshire had no data or estimates on the number of people enrolled in self-funded plans.
In Maryland, Minnesota, and Rhode Island, small employers are exempt from MH/SA parity laws. Under the Maryland Health Care and Insurance Reform Act of 1993, all insurance contracts sold to employers who have 2 to 50 employees eligible for coverage must provide the benefits required in a standard comprehensive benefit plan. This small group policy limits inpatient MH/SA hospital days to 25 per person per year. As we noted above, small employers in Minnesota and Rhode Island have not used the small employer exemptions in those states.