Documenting pathways to dementia care: relative validity of questionnaire, interview, and medical record formats.

Abstract

One of the shortcomings of the pathways-to-care literature is the lack of empirical support for the validity of the data collection methods. This study uses three common formats to collect retrospective pathways-to-care data for adults who have been diagnosed with possible or probable Alzheimer's disease (AD) and compares indicators to evaluate their relative validity. Forty family caregivers of adults diagnosed with possible or probable AD were recruited from the caregiver registry of the Boston University Alzheimer's Disease Core Center (BU ADCC). In each of three formats (questionnaire, structured interview, and medical record review), data were collected regarding four key events in the pathway to dementia care: first appearance of symptoms, first verbalized recognition of symptoms, first effort to seek professional help, and first diagnosis by a professional. In addition to the dates of these events, researchers attempted to determine: the first verbalized concern about the symptoms, who first sought professional help, what professional was first approached, and what professional made the first diagnosis. In a consensus meeting, data collected in all three formats were reviewed, and a consensus on the most likely answers to all questions was recorded and compared to data collected in each format. The results suggest that the three formats are not equivalent in terms of concurrent validity. While substantial agreement is found among data collection methods, the validity of the structured interview format and the medical record review is most consistently supported by the data in this study. Questionnaire data resulted in underestimates of delays and correlated poorly with other data sources, including the consensus judgment. Recommendations for pathways-to-care data collection procedures that maximize validity are discussed.