Social Frailty Index Predicts Mortality in Older Adults
Key findings
- The purpose of this research was to create the Social Frailty Index, a short survey that uses social risk factors to estimate the four-year risk of mortality in adults age 65 or older
- Responses to just 10 items were able to discriminate individuals at high risk of death (AUC=0.73), new dependence in activities of daily living (AUC=0.72), and nursing home stays (AUC=0.74)
- In the development cohort, the factors most predictive of death were male gender, not working for pay at present, often feeling isolated from others, having children but meeting with them less than once annually, and doing no volunteering or charity work
- The model improves on the Charlson Comorbidity Index and the Lee Index
Social characteristics such as social support and social network strength are well-known predictors of mortality. Still, most predictive models rely solely on demographics, medical comorbidities, and function. The few existing social risk models are expansive inventories that are difficult to administer.
Subscribe to the latest updates from Geriatrics & Palliative Care Advances in Motion
Sachin J. Shah, MD, MPH, a physician–investigator in the Division of General Internal Medicine at Massachusetts General Hospital, and colleagues have created the Social Frailty Index, which predicts mortality in older adults based on just 10 factors. They introduce it in PNAS.
Methods
The team used data from the longitudinal, nationally representative Health and Retirement Study, in which more than 43,000 U.S. residents are being followed for changes in health, wealth, social networks, and function after age 50. A random half of the total cohort is interviewed every two years, with no overlap.
This study included participants aged 65 or older who completed the Psychosocial and Lifestyle Questionnaire in 2010 or 2012. Responses from 4,302 individuals in the 2010 cohort were used to develop the Social Frailty Index, and responses from 3,948 participants in the 2012 cohort were used to validate it. The median age of both cohorts was 75.
Two investigators independently reviewed each of the 472 predictors of mortality included in the Social Frailty in Older Adults framework, articulated by sociologists and gerontologists in 2017 in the European Journal of Ageing. They chose 183 candidate predictors to include in a logistic regression model.
Development Results
The primary outcome was all-cause mortality four years after the interview. The model development procedure yielded 10 predictors of that outcome:
- Age, per 1-year increase—OR, 1.12
- Male gender—OR, 1.74
- Local area is often to always full of rubbish and litter—OR, 1.42
- Low self-reported control over a financial situation—OR, 1.55
- Has children and meets with children less than once a year—OR, 1.64
- Engages in activities with grandchildren, niece/nephews or neighborhood children—OR, 1.11
- Not working for pay at the present time—OR, 1.74
- Does no volunteering or charity work—OR, 1.62
- Feeling isolated from others some of the time (vs. hardly ever or never)—OR, 1.29; feeling isolated from others often—OR, 1.67
- Self-report of being treated with less courtesy or respect than other people a few times a year or more often—OR, 1.34
Validation Results
The Social Frailty Index performed well in the validation cohort:
- Capacity to predict death—area under the receiver operating characteristic curve (AUC), 0.73 (95% CI, 0.75–0.78)
- To predict new dependence in activities of daily living—AUC, 0.72 (95% CI, 0.70–0.74)
- To predict nursing home stays—AUC, 0.74 (95% CI, 0.71–0.77)
Comparison to Existing Risk Models
The Social Frailty Index meaningfully risk-stratified participants beyond the Charlson Comorbidity Index and the Lee Index; the latter uses both comorbidities and function to predict mortality.
Using the Index
The Social Frailty Index is now an online calculator that clinicians can use to help older adults engage in advanced care planning and weigh medical interventions. It's important, though, not to use the tool to limit the care of patients at higher-than-average risk under the supposition that they are at risk of poorer outcomes.
For example, the index should not be used to recommend against a surgical intervention for a patient deemed "socially frail." Instead, the results should be used to improve access to high-quality care, such as preoperative referral to a social worker who can address any barriers to successful recovery.
In medical research, investigators could seek to understand whether the effect of an intervention differs according to Social Frailty Index results.
view original journal article Subscription may be required
Learn more about the Division of General Internal Medicine
Refer a patient to the Division of Palliative Care and Geriatric Medicine