A long-term-care setting pilot study evaluating predictors of success in medication self-administration

To determine whether any specific patient variables or a short battery of neuropsychological tests of cognition and memory predict success in medication self-administration.

A prospective, single-blinded design.

An extended care center of a university-affiliated VA Medical Center.

Thirty predominately male, older veteran patients, mean age 70 +/- 4 years with a range of 63 to 79 years.

Neuropsychological testing [Mini-Mental Status Examination (MMSE), Delayed Word Recall Test (DWRT), Shipley Institute of Living Scale (SILS), Neurobehavioral Cognitive Status Examination (NCSE), Hopkins Verbal Learning Test (HVLT)], twice a week unannounced bedside medication counts and medication administration record inspections, and educational instruction, if needed, by nurses and pharmacists.

Patient characteristics such as age, number of medications, presence of a disorder that can alter cognitive or memory function, years of education, and results from the above listed neuropsychological tests. The dependent variable was successful or not successful as defined by whether the patient required a re-education intervention.

Fifteen patients required one or more re-education intervention(s) as a result of meeting the criteria for not being successful. The absence of major depression, stroke, or anxiety disorder did tend to predict success (P = 0.0716) in medication self-administration. The other patient specific characteristics did not predict success. Among the neuropsychological tests administered, only the Judgment Subtest of the NCSE tended to predict success (P = 0.098). The MMSE, DWRT, SILS,HVLT tests did not predict successful performance in the self-medication program.

Although the presence of a diagnosis that could potentially alter cognitive and memory function tended to predict success, no patient characteristics were found that predicted success independently. Among the neuropsychological tests, only the Judgment subtest of the NCSE tended to predict success in medication self-administration. We conclude that the NCSE and characterization of patient-specific factors, including diseases that may affect cognitive and memory function, seem to be the best predictors of success in medication self-administration in a long-term-care setting.