I have been an OASIS booster since 1996, and still am. But I am increasingly uneasy about the quality of data collected, aggregated, and reported. As uses of OASIS data have expanded, data quality and accuracy are increasingly important. The importance of inter-rater reliability was initially stressed as critical to the success of a uniform data collection tool such as OASIS and is also critical to other OASIS uses as:

* an assessment tool used to plan care and communicate the patient's condition among and across disciplines,

* the foundation for performance improvement activities,

* a means of determining how much reimbursement an agency will receive based on the initial assessment.

If we can't be sure that every one understands and uses it the same way, few structures built on the foundation of OASIS will withstand scrutiny.

Unfortunately, inter-rater reliability still appears to be a problem. Inconsistent OASIS ratings across staff were cited in the CMS OASIS Town Hall Meeting in July. My contacts with clients, providers, and consultants support that OASIS data accuracy concerns are immense.

Inconsistent responses and definition confusion continue to plague many OASIS data collectors. Even more disturbing, while many agencies seem to be concerned about inter-rater reliability, only a few seem to understand how to assure it in their agency.

Although providers often blame OASIS data items, we must realize OASIS data are everywhere, and is or will be used in ways that have potential negative impacts for agencies and patients. Some problems that keep me awake at night:

CMS planned release to the public of local agency-specific OASIS information as part of its Quality Improvement Outcome (QIO) initiative. Currently being piloted in long-term care, CMS plans to use OASIS data to compare local agencies in 2003 and make it available to consumers. The goal of enhancing consumer decision making is good; however, publication of inaccurate data can cause unfounded problems for agencies and raise unnecessary consumer concerns.

Agency and public responses to Adverse Event Outcome Reports. OASIS data quality is critical to produce accurate Adverse Event Outcome Reports. A leading problem with this data is the Unexpected Death adverse event related to M0280 (Life Expectancy), since prognosis responses are typically conservative. The results are agencies with higher "unexpected death" rates than the norm.

As agencies work to improve their adverse event outcomes, they rely on OASIS data to identify areas for improvement and clues for improvement. If the data are inaccurate, how can resulting PI initiatives be productive and positive?

Home care agency reliance on OASIS data as the basis for Outcome-Based Quality Improvement (OBQI) activities. As noted above, reliance on poor quality OASIS data can significantly misdirect an agency's OBQI activities. Agencies may waste precious PI resources on problems that don't exist, overlook problems that are present, make changes in care processes that are not in the patient's best interests, or make changes in care processes that actually harm patients.

Speaking out on the quality and appropriate use of OASIS data is important. Industry leaders are advocating for changes in the OASIS data items that can help improve data quality. However, clinicians and managers also must join together to improve the quality of OASIS data.

The impending OASIS changes will not be enough to protect the industry and our patients from the consequences of poor quality data. If significant improvements in OASIS data collection are not achieved soon, home care's reputation and viability will suffer immeasurably.