Who’s Talking Turkeys? Crafted in Response to the CARE Tool Debate

Barbara J. Gage

recent blog suggested that CMS’ efforts to standardize assessment data was based on a goal of “….creating a functional measurement tool that could be used throughout the industry.” In fact, CMS has been working since 2005 to meet the Congressional directive to standardize assessment information at hospital discharge, and post-acute care (PAC) admission and discharge for payment and quality reporting purposes (Deficit Reduction Act of 2005). The CARE tool was developed as part of the national Post-Acute Care Payment Reform Demonstration (PAC PRD). The conceptual domains and items were selected with the input of the wide range of stakeholder communities working with PAC populations. Clinicians from acute hospitals and each of the four PAC settings, including long term care hospitals (LTCHs), inpatient rehabilitation facilities (IRFs), skilled nursing facilities (SNFs), and home health agencies (HHAs) identified items to test in four areas: medical status, functional status, cognitive status, and some social support factors. Input was given by physicians, nurses, physical therapists, occupational therapists, speech and language pathologists, social workers and case managers working in each of the different levels of care. Initial item selection was based on a review of existing assessment items, including those in the three Federally-mandated instruments, (the IRF-PAI, MDS 2.0, and OASIS-B which were in effect at this time) and the input of each of the scientific communities working in these areas.

Developers of proprietary systems such as the UDS-MR©, Inter-RAI ©, and AM-PAC ©, as well as public domain items tested in clinical trials such as the PROMIS items, were all reviewed as part of this process. The selected items needed to be in the public domain so the measures could be modified as science advanced practice.

Over 200 providers participated nationwide to submit over 53,000 CARE assessments over the course of the PAC PRD. Participating clinicians also provided feedback during training and exit interviews. In general, positive feedback was provided on most items. Feedback showed that almost all items were commonly collected on existing instruments in hospitals and PAC providers, although some of the information may have been informally noted in charts rather than provided in the structured form of the CARE items.

The items were tested for reliability so they could be applied consistently across populations and settings. Most of the items were previously tested and found reliable in at least one of the five levels of care. Two types of reliability tests were conducted on the final CARE tool item set used in the PAC PRD. The results showed that most items when applied to the other four settings were at least as reliable as the existing Federal assessment items (Kappa scores of 0.6 or better) ensuring their reliable use in future quality measures or payment models would reach consistent results. Complete reports on item reliability and PAC PRD results can be found here.

Data standardization is critical to allow providers to exchange information as they follow the patient. The Deficit Reduction Act of 2005 directed CMS to use standardized assessment items at acute hospital discharge and PAC admission and discharge to allow for empirical comparisons of key questions arising out of changing incentives in the Medicare payment policies. The standardized CARE items are consensus-based versions of the items already collected by clinicians. These and additional items being incorporated into CMS’ assessment item library represent the “best in class.” The team developing the CARE item set represented the leading experts in each of the areas – Dr. Margaret Stineman of the University of Pennsylvania, developer of the function-related groups associated with the proprietary FIM©, Dr. Deborah Saliba, UCLA, lead developer of the MDS 3.0, and Dr. Chris Murtaugh of the Visiting Nurse Service of New York. Team members included Drs. Anne Deutsch and Trudy Mallinson of the Rehabilitation Institute of Chicago. Input was also given by Dr. David Hittle, of the University of Colorado who has worked closely with the OASIS tool, Dr. Samuel Markello, formerly of the UDS-MR©, and Dr. Patrick Murray of Case Western University.

The blog suggested that, “the early reviews of the CARE tool have been poor.” While this clearly is not true, it is worth pointing out that the author owns one of the key proprietary assessment instruments. The CARE items have been evaluated for reliability and they meet the national standards; they allow providers and others the opportunity to download the e-specification of the items without charge and to have the clinicians trained for free under CMS’ regular assessment training initiatives. CMS is currently developing quality measures using the “best in class” assessment items which all meet scientific standards. The quality measure development process already requires CMS to submit measures for endorsement by the National Quality Forum. The “loophole” identified by the UDS-MR© author is non-existent. The Measures Application Partnership is part of the existing NQF process included in the IMPACT legislation. Further, use of uniform data elements across settings, such as those used in the currently collected pressure ulcer measure, allows for exchangeability and improves communication across the system, finally creating a “data follows the person” system.