OBJECTIVE: To determine the accuracy of rapid urine screening tests in detecting bacteriuria among incontinent nursing home residents. SETTING: Six nursing homes. PARTICIPANTS: 214 chronically incontinent, but otherwise asymptomatic, nursing home residents who were enrolled in a clinical intervention trial for urinary incontinence. MEASUREMENTS: 684 urine specimens were collected, the majority (76%) by a clean catch technique and the remainder by in-and-out catheterization. Each specimen underwent dipstick testing for nitrite and leukocyte esterase, a rapid, enzyme-based screening test for bacteriuria, and a quantitative urine culture using standard laboratory techniques. RESULTS: No one screening test or combination of tests had adequate sensitivity and specificity for clinical purposes. However, using all three tests, the sensitivity increases to 97% in females and 92% in males when any one of the tests is positive, and the specificity increases to 95% in females and 97% in males when all three tests are negative. Among nursing home residents suspected of having a symptomatic urinary tract infection, the prevalence of bacteriuria is probably higher than in our study population (e.g., 60-70%, compared with 32%). At these prevalence rates, the positive predictive value of any of the three tests being positive is 93% and higher, and the negative predictive value of all three tests being negative is 80 to 90%. CONCLUSIONS: Our data must be interpreted cautiously because of the specimen collection methods, the definitions we used, and the fact that we studied a population who did not have symptoms of an acute urinary tract infection. In addition, we did not examine the absolute accuracy of the screening tests, but their accuracy as a clinician might use them in a nursing home. Despite these caveats, our data suggest that a combination of these screening tests could be useful in the initial assessment of nursing home residents for bacteriuria, and result in considerable cost savings. Studies are needed to replicate our findings among nursing home residents with symptomatic urinary tract infections.
OBJECTIVE: To determine the accuracy of rapid urine screening tests in detecting bacteriuria among incontinent nursing home residents. SETTING: Six nursing homes. PARTICIPANTS: 214 chronically incontinent, but otherwise asymptomatic, nursing home residents who were enrolled in a clinical intervention trial for urinary incontinence. MEASUREMENTS: 684 urine specimens were collected, the majority (76%) by a clean catch technique and the remainder by in-and-out catheterization. Each specimen underwent dipstick testing for nitrite and leukocyte esterase, a rapid, enzyme-based screening test for bacteriuria, and a quantitative urine culture using standard laboratory techniques. RESULTS: No one screening test or combination of tests had adequate sensitivity and specificity for clinical purposes. However, using all three tests, the sensitivity increases to 97% in females and 92% in males when any one of the tests is positive, and the specificity increases to 95% in females and 97% in males when all three tests are negative. Among nursing home residents suspected of having a symptomatic urinary tract infection, the prevalence of bacteriuria is probably higher than in our study population (e.g., 60-70%, compared with 32%). At these prevalence rates, the positive predictive value of any of the three tests being positive is 93% and higher, and the negative predictive value of all three tests being negative is 80 to 90%. CONCLUSIONS: Our data must be interpreted cautiously because of the specimen collection methods, the definitions we used, and the fact that we studied a population who did not have symptoms of an acute urinary tract infection. In addition, we did not examine the absolute accuracy of the screening tests, but their accuracy as a clinician might use them in a nursing home. Despite these caveats, our data suggest that a combination of these screening tests could be useful in the initial assessment of nursing home residents for bacteriuria, and result in considerable cost savings. Studies are needed to replicate our findings among nursing home residents with symptomatic urinary tract infections.