Report on an Actuarial Review of the HealthChoices Program in Southeastern Pennsylvania
House Resolution 207 called for an independent study of the actuarial soundness of the methodology and assumptions used by the Department of Public Welfare in setting rates for the physical health component of the HealthChoices program. HealthChoices is the mandatory Medicaid managed care program which began in Philadelphia and the four surrounding counties on February 1, 1997. The Legislative Budget and Finance Committee contracted with Arthur Andersen LLP to conduct this study. Arthur Andersen found:
William M. Mercer, Inc., DPW’s actuarial consultant, used reasonable methodologies to develop the cost and utilization data in the Data Book provided to the health plans in May 1996 to assist them in developing their proposals.
Mercer and DPW provided sufficient documentation to allow conclusions as to the appropriateness of some but not all of the adjustments used in developing the target premium rate ranges (DPW only accepted bids within the target ranges). Where sufficient information was provided, Andersen found that two adjustments (for antiselection and the managed care discount) appeared to be overly aggressive for certain categories of beneficiaries (i.e., the projected savings were overly optimistic).
The methodology used to develop target premium rates should have incorporated existing managed care Medicaid program cost information, not just fee-for-service information. The managed care data suggests that higher rates may have been appropriate for several classes of individuals, such as those receiving SSI without Medicare.
The discount used to reduce payments to the plans during the voluntary HealthChoices program appears inappropriate.
The health plans must pay for most behavioral health drug treatments, but have little control over the providers who prescribe these drugs. This creates an inappropriate match between treatment decisions and financial responsibility.
A recipient’s age, sex, and special needs status have not been considered in the rating cell structure. Although the lack of age and sex distinct rates does not appear to have resulted in adverse consequences to the health plans to date, a more refined rating system should be considered.
DPW did not establish separate premium rates for special needs recipients, thus placing the health plans at high risk of receiving a disproportionate share of such recipients. (DPW later allowed plans to submit contract adjustments to help mitigate this problem with respect to AIDS.) Also, DPW assumed the largest savings could be achieved for the SSI population, a population for which there may be the least opportunity for such savings.
- DPW’s 11% allowance for administrative costs, profit, and contingencies for health plans is also low. However, Mercer overstated the adjustment for inflationary trends by 5%The net effect of all these factors is that premium rates for the Supplemental Security Income (SSI) categories may have been understated by at least 6% (with Medicare) to at least 3% (without Medicare), while rates for other categories may have been reasonable.
Arthur Andersen recommends the Department of Public Welfare consider:
- Incorporating historical Medicaid managed care data from the HealthChoices program when prospectively developing
future target premium rates.
- Increasing the plans’ allowance for administrative expense, profit, and risk to 13% - 15% of medical expense.
- Making behavioral health providers financially accountable for the prescriptions they authorize.
- Modifying the rating process to compensate for key risk factors, such as AIDS cases and institutionalized recipients.
- Incorporating age, gender, and health status into the auto-enrollment process to reduce the plans’ unmanaged exposure
to uncompensated, random fluctuations in enrollment distributions.
- Retroactively eliminating the voluntary adjustment for General Assistance and SSI groups.
- Moderating the antiselection and managed care adjustment factors for the SSI eligibility groups to correct for
a net understatement in rates of at least 3% to 6%.