A Brief History…

A PROPOSAL FOR THE STUDY OF HEALTH OUTCOMES AND FUNCTIONAL STATUS OF MEDICAID RECIPIENTS ENROLLED IN A MANDATORY MANAGED CARE PROGRAM

December 21, 1996,

The initial attempt to enroll Medicaid (MediCal) recipients in managed care occurred in the 1970s, not long after the passage of laws creating the federally subsidized HMO. The results in California were not auspicious as a number of scandals ensued due to marketing abuses and poor financial planning by the entrepreneurial HMOs.

In 1982 medically indigent adults were removed from the MediCal program if they were not eligible for Federal assistance programs. Responsibility for these low income patients was transferred to the counties. In Los Angeles county, patients eligible for general relief were given identification cards to obtain free care at county clinics. The other medically indigent adults who were not on general relief could receive care but were charged for visits to county facilities unless an emergency existed. A study of the health outcomes in these patients, as compared to a control group with insurance coverage, showed statistically significant deterioration in health status after six months.

In Orange county a Medically Indigent Adult program was developed by the County Health Care Agency and the private doctors and hospitals in the county. Using careful utilization review and fee for service payment, the patients dropped from Medi-Cal were absorbed into this system. This program continues as the Indigent Medical Services program to this day. Initially the reimbursement levels for providers approximated Worker’s Compensation rates and participation was nearly universal. Recently reimbursement levels have fallen below MediCal rates and access to care for this patient group has suffered. No similar study of the health outcomes of the Medically Indigent Adults has been carried out in Orange county so we cannot know how the results compared to Los Angeles.

The Health Insurance Study by the RAND Corporation suggested that poor patients with medical problems had worse health outcomes in a prepaid, HMO style, delivery system than in a Fee For Service system. This study suggested that these patients were more susceptible to access barriers in the HMO model than the middle class members. In fact, such patients also had increased bed days and clinic visits compared to similar patients in a FFS plan with free care (no deductible) for low income members.

As part of a pilot program implementing managed care in the Medicaid program, Orange county has established a mandatory Medi-Cal managed care program. CalOPTIMA began January 1, 1996. All 350,000 Medi-Cal beneficiaries were informed of the details of the new program and encouraged to choose a provider for primary care. If they did not choose such a provider by February 1, they were assigned a primary care provider. CalOPTIMA contracts with providers to care for the Medi-Cal beneficiaries of Orange county. The program supervises the care of the 240,000 covered beneficiaries and provides information and capitated payment to providers.

A subgroup of the 240,000 beneficiaries, comprising about 30,000 individuals, is “CalOPTIMA Direct”. This is a fee-for-service group which includes Medicare-MediCal eligible, for which MediCal is a secondary payer, beneficiaries in transition from one provider to another and foster care children. Some of the Medi-Medi patients are Medicare HMO members with high pharmacy costs due to chronic illness. If these patients exhaust the HMO pharmacy benefit MediCal pays the balance of pharmacy charges while the patient remains in the HMO for other health care.

Collection and analysis of data is part of the supervision function of the administrator of the program. The information base includes encounter data from ambulatory care, hospital discharge data and social service agency reports. Since eligibility for Medi-Cal benefits is based on economic factors and disability or age status, such information is available for inclusion in a profile of the health status of these individuals. A limitation in this data is the fact that economic factors, on which eligibility may be based, can change from month to month and, therefore, this population, unlike Medicare, is not a constant one.

The CalOPTIMA program includes approximately 50 provider groups qualified to provide capitated care to beneficiaries. These include traditional HMOs, the University Medical Center which has in the past provided a large share of Medi-Cal care in this county and a group of large hospital- physician consortia and several smaller hospital-physician groups. Included in these provider groups are nearly all the existing “safety net” providers who have cared for Medi-Cal patients in the past.

What we propose is that the University of California, Irvine departments of Family Practice, Medicine, Surgery plus the administration of CalOPTIMA form a joint program to assess the health status and the effect of managed care on this population. The program will conduct a continuing assessment of the health outcomes of patients with specific indicator conditions. These indicator conditions will be used to measure the quality of health care for the entire population much as the canary monitors the air in the coal mine for the miners. A second goal will be to assess differences in outcome attributable to the type of provider. Prior research noted above has suggested that poor, sick patients do not do as well in the HMO environment. This study will allow us to use the three provider types to see if any difference can be noted in different models within a prepaid system.

Indicator conditions in adults could be hypertension and diabetes, which were the diagnoses used by Nicole Lurie in her study of the medically indigent adults dropped from Medi-Cal in 1982. Other adult conditions could include congestive heart failure and coronary artery disease. Breast and uterine cancer are also prominent in poor populations, suggesting high priority for inclusion. Psychiatric conditions may be prominent in the disabled group and, if prominent in this group, should be included. Depression has been found to be a significant share of primary care and, in the prepaid primary care setting, is often not recognized.

Pediatric indicators would include low birth weight babies as well as those without prenatal care. Child asthma patients would be a good indicator and another miscellaneous group which has had a prior hospitalization for any reason could be included. Two other high risk pediatric groups are foster care children and children with a report of abuse. Children eligible due to disability should also be studied.

The existence of several distinct ethnic groups in Orange county, Asian, Hispanic and white in particular, offers another opportunity to compare health outcomes with these groups, as well. Some of these groups have different health belief models. Alan Hubbell MD recently described a belief of Hispanic women that cervical cancer is God’s punishment for promiscuity. This belief may reduce the use of Pap smears in this community. There are other beliefs which may affect health status in ethnic communities and which might be identified in such a study.

The methodology suggested is similar to the Medical Outcomes Studies of the RAND Corporation. An indicator condition is chosen and a sample of patients is selected from the CalOPTIMA database using claims data from the year prior to the beginning of the managed care program. A survey of health status is conducted on a sample of this population using the SF36 or the Dartmouth COOP method. The population is followed for a year and another study of health status is administered. Encounter data may be usable for some follow-up on indicator conditions and to screen for new problems in the entire population of beneficiaries. The quality of the database from encounter records will determine the usefulness of such data. Claims data has been shown to be quite useful in the analysis of quality in the Medicare program. Diastolic blood pressure, Hemoglobin A1C, and an abbreviated SF36 would be useful indicators obtained on a sample of the indicator groups. In addition, analysis of encounter data with respect to hospital admissions, emergency room visits, and deaths would provide more objective evidence of health status. If funding were available, the study should be continued for a longer period, especially with the pediatric group. School performance might become another measure of health outcome among children with indicator conditions.

Results should be analyzed for trends including:
1. The effect of managed care on this population.
2. The effect of the three (or four) provider models in the CalOPTIMA program. Is one model of provider better at providing the health needs of a particular patient population or even all the different beneficiary subgroups ?
3.The effect of self-selection which may introduce differences in risk, or adverse selection. For example patients with conditions requiring intensive or more expensive care may select the University Hospital. Pregnant women beneficiaries may select classical HMOs and elderly poor with chronic illness may select the PHO which includes their long-term provider.
4. Ethnic and language considerations may affect risk and outcomes. One ethnic group may be more represented in a lower risk group such as AFDC.
5. Geographic factors may affect outcome due to provider-specific performance or difficulty in access due to transportation considerations. Are providers distributed geographically in rough symmetry with beneficiaries ?

Results of a meeting with members of the CalOPTIMA staff on December 20, 1996:

1. They are interested in cooperative studies of the patient population.
2. They do not yet have any encounter data from the providers to work with and it will be some time before it is available.
3. They have begun a small study of pediatric asthma patients similar to my suggested inclusion of such cases as an indicator group in pediatric patients.
4. They are quite interested in a study of the disabled population and accepted my suggestion of obtaining more information about this population including age and diagnosis data. The diagnosis data may come from fee-for-service claim forms or from pharmacy data.
5. They are very interested in getting information from the UCI Faculty Medical Group data about the care of AIDS patients and about a recent study of diabetic patients’ care. CalOPTIMA has been considering a study of HbA1C data on diabetics to determine the quality of diabetic care in the program. They accept the premise, recently reported in the NEJM, that intensive care of diabetics, while more expensive in the short run, is cost effective in that it prevents expensive complications in the long run.
6. They recommended several consultants on data sources including MediCal paid claims data and UHDDS hospital data. I will contact these consultants and inquire about what data is available and what the costs are. This may be a source to replace my collaborator in New Hampshire who is very busy now.
7. The CalOPTIMA person who handles the disabled beneficiaries records will contact me with more information on this population.
8. We will meet again in a month or two. They are not yet ready to plan a grant proposal. There is no grant cycle for California Endowment and a new sister foundation is coming on line which may be more appropriate for this type of proposal.
9. They are very interested in having a liaison with UCI and are very interested in getting data on the AIDS cases treated by the UCI medical group. This is the largest MediCal AIDS population in the county. We talked about the new emphasis on funding pharmacological treatment rather than social services for AIDS patients which is creating some tension in the AIDS community. As AIDS patients show marked improvement with drug therapy there is more pressure for the funding to go to drug procurement rather than hospice-type services. The Wall Street Journal had an article on this subject the day of the meeting.
10. I am willing to function in the liaison role if the UCI medical group and the other UCI departments are willing to cooperate. This could lead to several worthwhile projects.