Challenges in measuring nursing home and home health quality: lessons from the First National Healthcare Quality Report.

BACKGROUND: The availability of patient assessment data collected by all Medicare- and Medicaid-certified nursing homes (NHs) (the Minimum Data Set [MDS]) and home health agencies (HHAs) (the Outcome and Assessment Information Set [OASIS]) provides an opportunity to measure quality of care in these settings. OBJECTIVE: The objective of this study was to examine methodologic issues encountered as these datasets are used to report the nation's health care in the National Healthcare Quality Report (NHQR) at national and state levels. FINDINGS: Although the reliability of most data elements from MDS and OASIS are considered acceptable in research studies, mixed evidence exists for the reliability and validity of the quality measures themselves. Detection bias can affect the quality measures, particularly for pain and pressure ulcers. Although risk adjustment is used for all measures, effectiveness varies among measures and methods. Additional quality measures such as patient satisfaction, quality of life, and structural measures would be desirable but will require additional data collection efforts. Although the NH measures represent most NH residents, the HHA measures only apply to Medicare and Medicaid patients served by Medicare-certified agencies. Finally, the absence of clinical benchmarks limits the interpretation of the NHQR HHA and NH measures. CONCLUSIONS: Further developmental work is needed to address many of these issues to improve the usefulness of these quality measures in future NHQR reports.

A Study of Reliability and Burden of Home Health Assessment Using OASIS.

The Outcome and Assessment Information Set (OASIS) is used for outcome reporting, quality improvement, and case mix adjustment of per-episode payment for home health care. The research described here addresses interrater reliability of OASIS items and compares clinician time required to complete patient assessment with and without OASIS. Interrater reliability for OASIS data items was estimated using independent assessments by two clinicians for a sample of 66 patients. Incremental assessment time due to OASIS was estimated using interview data from two agency-matched groups of clinical care providers-one group who used OASIS in the assessment and a second group whose assessment did not include OASIS items. Interrater reliability is excellent (kappa > .80) for many OASIS items and substantial (kappa > 0.60) for most items. The reported time required to complete an assessment with OASIS did not differ from the time required for a comparable assessment without OASIS. The results of this study are being used to guide developmental efforts to improve OASIS items. They can also be informative to home health care agencies when interpreting OASIS-based outcome and case mix reports.

A study of reliability and burden of home health assessment using OASIS.

The Outcome and Assessment Information Set (OASIS) is used for outcome reporting, quality improvement, and case mix adjustment of per-episode payment for home health care. The research described here addresses interrater reliability of OASIS items and compares clinician time required to complete patient assessment with and without OASIS. Interrater reliability for OASIS data items was estimated using independent assessments by two clinicians for a sample of 66 patients. Incremental assessment time due to OASIS was estimated using interview data from two agency-matched groups of clinical care providers--one group who used OASIS in the assessment and a second group whose assessment did not include OASIS items. Interrater reliability is excellent (kappa > .80) for many OASIS items and substantial (kappa > 0.60) for most items. The reported time required to complete an assessment with OASIS did not differ from the time required for a comparable assessment without OASIS. The results of this study are being used to guide developmental efforts to improve OASIS items. They can also be informative to home health care agencies when interpreting OASIS-based outcome and case mix reports.

Improving patient outcomes of home health care: findings from two demonstration trials of outcome-based quality improvement.

OBJECTIVES: To evaluate effects on patient outcomes of Outcome-Based Quality Improvement (OBQI), a continuous quality improvement methodology for home health care (HHC). DESIGN: A quasi-experimental design with prospective pre/post and study/control components within two multiyear demonstration trials (occurring from 1995 to 2000) in which 73 home health agencies implemented OBQI, receiving several annual cycles of outcome reports to evaluate and enhance patient outcomes. SETTING: New York and 27 other states. PARTICIPANTS: The study involved 157,548 predominantly older adult patients admitted over 3 years to 54 OBQI agencies from 27 states in the National Demonstration Trial, 105,917 patients admitted over 4 years to 19 OBQI agencies in the New York State Trial, and 248,621 patients admitted over 3 years to non-OBQI control agencies in the 27 demonstration states. INTERVENTION: As a clinical management and administrative intervention, OBQI involves collecting, encoding, and transmitting patient-level health status data to a central source that provides each OBQI agency with a risk-adjusted outcome report comparing the agency's patient outcomes with those from a reference population and with its own outcomes from the prior period. Target outcomes are selected and focused plans of action implemented to change care behaviors. Outcome changes are evaluated through the next report cycle. MEASUREMENTS: Outcome measures include hospitalization rates and improvement and stabilization outcome rates in functional, physiological, emotional/behavioral, and cognitive health. RESULTS: For the National and New York State Demonstration Trials, the risk-adjusted relative rates of decline in hospitalization of 22% and 26%, respectively, for OBQI patients over the 3-year and 4-year demonstration periods were significant (P <.001) and unparalleled by considerably smaller rates of decline for the non-OBQI patients in the 27 states. The risk-adjusted rates of improvement in OBQI target outcome measures of health status averaged 5% to 7% per year in both demonstration trials and were significantly greater (P <.05) than analogous improvement rates for nontarget comparison outcomes, which averaged about 1% per year. CONCLUSION: It is feasible to integrate the programmatic, data collection, data transmission, and outcome enhancement components of OBQI into the day-to-day operations of home health agencies. The aggregate findings and the agency-level evidence available from site-specific communications suggest that OBQI had a pervasive effect on outcome improvement for home health patients. OBQI appears to warrant expansion and refinement in HHC and experimentation in other healthcare settings.