|
Preface Over the last five years the Iowa Department of Human Services (IDHS) has incorporated outcome measures from the Healthplan Employer Data and Information Set (HEDIS) as part of the quality assurance activities within the Iowa Medicaid program. The University of Iowa Public Policy Center has helped to identify, adapt, and determine the rates for these HEDIS outcomes measures annually. This report provides information regarding annual Medicaid outcomes for the period 1998-2000. There are 14 measures across the three years, although not necessarily in each year. Measures that were used in more than one year allow for year-to-year comparisons. By comparing rates over time we should be able to determine whether the outcomes of care are improving for the Medicaid population. In particular, with intra-HMO comparisons we can determine whether the managed care plans are improving their care over time. For some measures we also have national level data for comparison. The American Public Human Services Association (APHSA) undertook a project funded by the Commonwealth Fund to analyze data within the National Committee for Quality Assurance to determine rates for specific HEDIS measures for the Medicaid population. These analyses provide national benchmarking data that allow Iowa insights into how the stateís Medicaid program compares with programs in other states. The APHSA Medicaid HEDIS Database Project, Report for the Third year by Lee Partridge was released in December 2001.Over the last five years the Iowa Department of Human Services (IDHS) has incorporated outcome measures from the Healthplan Employer Data and Information Set (HEDIS) as part of the quality assurance activities within the Iowa Medicaid program. The University of Iowa Public Policy Center has helped to identify, adapt, and determine the rates for these HEDIS outcomes measures annually. This report provides information regarding Medicaid rates for the following HEDIS measures, ages and calendar years: Information and conclusions presented in this report are the responsibility of the authors and do not represent the views of the Iowa Department of Human Services, The CMS, the health plans or the University of Iowa. MethodsFollowing established HEDIS criteria, only those enrollees eligible for at least 11 months of the year are included in these analyses. It is expected that enrollees eligible for the entire year have the greatest opportunity to utilize services as compared with those eligible for only part of the year. It is also expected that we will be capturing the entire health care utilization experience for those who were eligible for at least 11 months. Those who were eligible for a shorter time period may have sought and received care that was not recorded in the claims and encounter databases during the months they were not enrolled in the program. From a performance measurement and quality assurance perspective, this provides utilization rates that are most fair when holding health plans accountable for the care provided to their covered populations.
Measures that were used in both years allow for year-to-year comparisons. By comparing rates over time we should be able to determine whether the outcome of care is improving for the Medicaid population. In particular, with intra-HMO comparisons we can determine whether the managed care plans are improving their care over time. For some measures we also have national level data for comparison. The American Public Human Services Association (APHSA) undertook a project funded by the Commonwealth Fund to analyze data within the National Committee for Quality Assurance to determine rates for specific HEDIS measures for the Medicaid population. These analyses provide national benchmarking data that allow Iowa insights into how the stateís Medicaid program compares with programs in other states. The APHSA Medicaid HEDIS Database Project, Report for the Third year by Lee Partridge was released in December 2001.The measures within the report that can be used for comparisons are:
These national benchmarks allow us to determine how Iowaís Medicaid outcomes compare to other states. However, before Iowa is compared to national benchmarks the populations that are included in the benchmarking data must be understood. The national benchmarks for the HEDIS data contain a more urban population than Iowa, making comparisons difficult. Also, should Iowa rates for HEDIS outcomes be more favorable than national benchmarks, it is not reasonable to assume that the desired rate as been reached. For example, the national benchmark for well child visits in children three to six years old within Medicaid programs was 52%. Iowaís rate is higher, however, programmatically100% of Iowaís children in Medicaid should obtain a well child visit annually between the ages of three and six. Although the outcome measures utilized for this report are based on the HEDIS measures adopted nationally for quantifying the outcomes of care in managed care plans, some are modified for use with the data available through the Iowa Medicaid program. The use of administrative data and the need for adjustment to the measures leads us to outline the limitations of the information contained within this report. Since we have access to administrative data only and are not able to augment this data with chart review, we are unable to adjust some measures for information that would be contained within medical charts. For example, HEDIS allows for the exclusion of some enrollees based on prior medical information (e.g., women who have had a double mastectomy may be excluded from the breast cancer screening rates). For the outcome analyses, these enrollees are included in the rates because we are unable to review the chart to determine whether a mastectomy has been performed. In addition, due to varying lengths of time enrolled in the Medicaid program, the administrative data available for each person often does not cover a sufficient period of an individualís health service experience (in this example, the time when a woman may have had a double mastectomy) to exclude such enrollees from the analyses. There are other limitations inherent in using claims and encounter data for outcomes based research. First, the health services data from the HMOs (encounter records) have a significant lag time between the date of service and the date they are paid. Generally, 95% of claims are adjudicated and paid within 3 months; however, Iowa Health Solutions adjudicates and pays only 90% of claims within 3 months of the date of service. Second, all administrative data contains coding errors and may not have procedure codes or diagnoses that correctly reflect what happened during a given contact with the health system. However, this problem should be minimized as the measures within the HEDIS set utilize widely accepted, well-defined protocols. Finally, HMOs may have differing code sets for the data that we are unable to interpret. In analyzing data for this report we have had problems locating over 2,000 claims for newborn care within specific HMOs. This may be related to individual HMOs coding tendencies that are unknown to both the IDHS and to us as we analyzed this data. Future investigation into this issue by the state is encouraged. The administrative data, encounter, claims and eligibility files, are furnished to the Public Policy Center under contract with the Iowa Department of Human Services (IDHS) for the completion of these analyses. The University of Iowa houses over 10 years of Medicaid data within a readily accessible dataset. All research has been approved by the University of Iowa Institutional Review Board to ensure that the privacy of all involved is maintained |
