More efficient methods to screen and monitor elderly patients in clinical practice and research are needed, but visits to clinical offices are expensive and many older patient are restricted by mobility or transportation access. The Alzheimer's Disease Cooperative Study (ADCS) is evaluating several technology platforms for remotely monitoring patient status at home. Dr. Mary Sano leads this Home-Based Assessment (HBA) study, which completed patient recruitment several years ago and is now completing the final participant follow-up visits. All participants were comprehensively evaluated in diagnostic interviews by medical professionals at study baseline, and are completing similar evaluations at the end of the study (or when a change in clinical status is suspected). A speech-enabled, computer-automated telephone system using interactive voice response (IVR) technology, developed by Dr. Mundt's research team, is one component of the HBA study. Several of the IVR assessments record speech samples for linguistic analysis, acoustic characteristics of the speech patterns are not being analyzed and resources to do so are not included in the HBA study budget. Recent studies have demonstrated that analysis of vocal acoustic characteristics in speech can provide reliable, physiologically-based biomarkers of CNS functioning associated with major depression. Symptomatic similarities between clinical depression and early Alzheimer's disease have been noted for many years, but the extent of overlap and temporal sequencing of emergent symptoms remains unresolved. Objective, physiologically-based biomarkers of CNS dysfunction may provide new insights for diagnosing and managing Alzheimer's patients. The research proposed is to support the development and validation of potential screening measures that could be used for differential diagnosis in clinical practice, as well as provide a foundation for innovative assessment and management approaches for older persons with multiple chronic conditions. This application proposes to merge non-identifiable clinical outcomes measures and medical diagnoses obtained from HBA investigative sites across the nation with audio files of speech samples recorded by the IVR system developed by CPC. The speech samples will be analyzed by signal processing engineers at MIT's Lincoln Laboratory for acoustic properties reflecting physiologically-based biomarkers associated with CNS disorders such as Alzheimer's, Parkinson's, and depression. The clinical and diagnostic information available through the ADCS database will be used to develop and validate multivariate statistical models to improve diagnostic screening, noninvasive monitoring of disease progression, and/or differential diagnoses between conditions.
Restricted mobility of older patients limits research participation and access to treatment providers, so cognitive decline often goes undetected for longer periods than necessary. Efficient methods to remotely monitoring patients from home using automated telephone systems can improve assessment procedures, reduce access barriers, facilitate multicultural non-English speaking interactions, and enhance patient retention at minimal cost. The ADCS Home-Based Assessment Study has recruited a nationally-representative sample of 214 seniors and is monitoring them longitudinally for 4 years to observe emergence of amnestic MCI and conversion of MCI to mild dementia. An automated telephone system is used to record speech samples from study participants, providing a unique opportunity to identify and develop new, objectively- quantifiable biomarkers of CNS dysfunction reflected in the acoustic characteristic of the speech recordings. Such biomarkers would have the potential for population-based cognitive screening as well as remote longitudinal monitoring of patients being treated for memory impairment disorders.
