Using Consumer Credit Data to Identify Precursors and Consequences of Cognitive Impairment Rapid growth in the elderly population combined with a lack of effective medical treatments to reverse or delay Alzheimer's disease and related dementias are estimated to lead to over 12 million older adults living with dementia by 2050. One of the earliest signs of cognitive decline and dementia is impaired financial capacity, which can manifest as difficulties managing money and paying bills or making erratic and uncharacteristically risky financial decisions, heightening risks for financial fraud, inappropriate asset allocation, credit delinquency from unpaid bills and other losses. Earlier warning signs of cognitive decline may be observable through changes in routine financial behavior. Credit bureaus and other data aggregators collect vast quantities of high-frequency, real-time consumer spending information on the more than 80% of Americans regularly using credit products. 75% of adults age 50 and over use credit cards, 39% carry a credit card balance, and 60% of homeowners age 50 and above have mortgage debt. These data may help to identify specific financial predictors of cognitive decline, leading to new information sources that could help with clinical diagnoses and alert patients and their families about the need for assistance with financial decision-making. To date, the potential health uses of consumer financial data have largely been ignored, particularly for the older population. This exploratory project considers the utility of a big data resource from outside healthcare; consumer debt characteristics collected in credit reports, to predict new cases of cognitive impairment and dementia and healthcare utilization of cognitively impaired patients. We have 3 research aims: 1- to create datasets linking the Federal Reserve Bank of New York/Equifax Consumer Credit Panel to national Medicare claims and survey data without direct patient identifiers using probabilistic matching; 2- to assess the utility of consumer debt characteristics as predictors of cognitive decline; 3- to assess the utility of consumer debt characteristics as predictors of hospitalization and nursing home use among patients with dementia. We will study an estimated 1.4 million patients with up to 15 years of panel data follow-up to assess whether adverse credit events captured in credit report data reliably identify signs of early or advanced cognitive impairment among older adults. If they do, monitoring programs could be developed to warn patients and their families of the potential need for screening and assistance managing money. This type of surveillance tool can help to protect older adults from fraud and other financial risks and assist long-distance caregivers to know when to intervene. In addition to the potential benefits for cognitively impaired patients and their families, this study will be a proof-of-concept of the use of consumer big data to inform clinical diagnoses and patient management, which may have a number of important implications for researchers and ultimately the patients who benefit from future discoveries.

Public Health Relevance

Detecting early warning signs of cognitive impairment and associated healthcare utilization can help patients, their families and healthcare providers to make decisions that reduce the risk of financial ruin associated with impaired decision-making and optimize health and social care for patients with dementia and other forms of cognitive impairment. Findings have immediate relevance for banks and policymakers designing banking and money management solutions for the growing dementia population. This project also serves as a proof-of- concept for the utility of financial big data for clinical diagnoses and design of products to help cognitively impaired older adults.

Agency
National Institute of Health (NIH)
Institute
National Institute on Aging (NIA)
Type
Exploratory/Developmental Grants (R21)
Project #
1R21AG053698-01
Application #
9167606
Study Section
Social Sciences and Population Studies A Study Section (SSPA)
Program Officer
Bhattacharyya, Partha
Project Start
2016-09-01
Project End
2018-05-31
Budget Start
2016-09-01
Budget End
2017-05-31
Support Year
1
Fiscal Year
2016
Total Cost
Indirect Cost
Name
Johns Hopkins University
Department
Public Health & Prev Medicine
Type
Schools of Public Health
DUNS #
001910777
City
Baltimore
State
MD
Country
United States
Zip Code
21205
Dean, Lorraine T; Nicholas, Lauren Hersch (2018) Using Credit Scores to Understand Predictors and Consequences of Disease. Am J Public Health 108:1503-1505