Evaluating a national assessment strategy for urinary incontinence in nursing home residents: reliability of the minimum data set and validity of the resident assessment protocol

Evaluation of 1 million incontinent American nursing home residents is hampered by both failure to detect incontinence and logistical barriers to diagnostic testing. The nationally mandated Minimum Data Set (MDS) and Resident Assessment Protocol (RAP) were devised to address these deficiencies. Although both instruments are also used in at least 18 other countries, neither has been evaluated. Our goal was to determine the reliability of the MDS and the accuracy of the RAP in predicting the lower urinary tract cause of incontinence. We determined interrater reliability for the 13 MDS items related to urinary incontinence in 123 randomly selected residents of 13 nursing homes in 5 states; forms were completed blindly by 2 nurses from each facility who were trained for a day. The RAP was assessed in 102 representative institutionalized women by blinded evaluation of its diagnostic accuracy compared with the multichannel videourodynamic criterion standard. For the MDS, interrater reliability for incontinence of all grades was excellent (weighted kappa correlation coefficient = 0.90), although reliability was greater at the extremes of measurement than for incontinence of intermediate severity. With the exception of delirium, correlations for the 11 MDS items related to incontinence were 0.65-0.96; for 6 items, correlations were > or = 0.8. The diagnostic accuracy of the RAP, successfully administered to 80% of women, was 70%. The accuracy of the nearly identical algorithm that formed the basis for the RAP was 84%. Importantly, serious misclassifications were not observed for either the RAP or the algorithm. Although its definitions should be modified slightly, the MDS appears to be feasible and reliable when administered by trained staff. In women, the diagnostic accuracy and safety of the RAP are good-particularly when administered as instructed-but the original, sex-specific algorithm is preferable. Together, the MDS and modified RAP provide a useful, stepwise, and non-urodynamically based strategy to guide evaluation and therapy of incontinence in this setting.