Self-Reported Frailty and Health Care Utilization in Community-Dwelling Middle-Aged and Older Adults in the United States

To the Editor:

Frailty, a dynamic state, is a consequence of cumulative deficits in multiple physiological systems accompanied by an increased vulnerability to stressor events.¹ The aging of the population has increased the prevalence of frailty substantially, especially among the costliest patients;² however, the current healthcare response to frailty is mainly reactive and acute care-based. Some have proposed primary care approaches to frailty management in community settings.³ Although frailty occurs more often in the older population, it can occur at any time in a person’s life.¹ Thus, identifying frailty earlier is advantageous for its reversal or control. However, previous studies on frailty have mainly been conducted in people aged 65 years and above. It is well known that frailty in older adults is independently associated with health care burden across acute, post-acute care, and outpatient sectors.⁴ There is a knowledge gap in understanding the association between frailty and healthcare utilization in community-dwelling middle-aged adults. With a few exceptions,^5,6 most prior studies of frailty used nonrepresentative samples. With a nationally representative sample, we describe the prevalence of frailty in the United States and evaluate the associations between frailty and healthcare utilization in community-dwelling middle-aged and older adults. And these results are discussed to inform policymaking and quality improvements in primary care settings.

Self-reported frailty was measured using the revised FRAIL scale (Fatigue, Resistance, Ambulation, Illnesses, and Low BMI) and five questions in the 2019 National Health Interview Survey questionnaire.^7,8 Individuals were classified into different health status categories: robust, pre-frail, or frail. About healthcare utilization, emergency room usage was determined by asking participants, “During the past 12 months, how many times have you gone to a hospital emergency room about your health?” Similarly, overnight hospitalization, delay of care due to cost, and omission of medical care due to cost were also measured using self-reported data. All statistical analyses were performed using SAS and information on weights, strata, and clusters. Logistic regression analyses adjusted for survey design were conducted to determine the association between healthcare utilization variables and frailty status.

TABLE 1 shows that among respondents aged 45–64 years, about 9% were categorized as frail, 22% were pre-frail, and 70% were robust. The prevalence of frailty was about 21% in those aged 65 or older. In the 45–64 age group, the incidence of emergency room visits in the past twelve months was 14%, 26%, and 50% in the robust, pre-frail, and frail populations, respectively. The incidence of overnight hospitalizations was 5%, 11%, and 31%, respectively. After adjusting for covariates and survey design, the pre-frail and frail populations had two and five times the risk of an emergency room visit in the past twelve months, respectively, compared to the robust population; and the pre-frail and frail populations had two and seven times the risk of overnight hospitalization in the past twelve months, respectively, compared to the robust population. We observed a similar pattern in the 65+ age group, but the associations between frailty and healthcare utilization were not as strong as in middle-aged adults. Additionally, the frail populations had more than two times the risk of delayed and omitted care due to cost during the past 12 months compared to the robust population in both age groups.

TABLE 1.

Healthcare Utilizations in the Past 12 Months by Frailty Status and Age Groups: A 2019 Sample from National Health Interview Survey

Participant Characteristics (Unweighted N)	Weighted Frailty Prevalence	ER visit			Overnight hospitalization
		Weighted event rate	Unadjusted OR (95% CI)	Adjusted OR* (95% CI)	Weighted event rate	Unadjusted OR (95% CI)	Adjusted OR* (95% CI)
Middle-aged adults
Robust(N=6870)	69.7%	14.2%	1.00	1.00	5.1%	1.00	1.00
Pre-frail(N=2188)	21.7%	26.2%	2.144(1.850–2.485)	1.907(1.637–2.223)	10.5%	2.186(1.801–2.751)	2.053(1.660–2.539)
Frail(N=935)	8.6%	50.4%	6.123(5.116–7.327)	4.708(3.878–5.715)	30.6%	8.189(4.117–7.308)	6.970(5.398–8.998)
Older adults
Robust(N=4234)	49.20%	17.0%	1.00	1.00	9.0%	1.00	1.00
Pre-frail(N=2646)	30.30%	26.9%	1.799(1.563–2.069)	1.732(1.504–1.994)	17.5%	2.141(1.794–2.555)	2.166(1.810–2.591)
Frail(N=1792)	20.50%	46.3%	4.223(3.656–4.878)	3.872(3.338–4.491)	33.5%	5.094(4.298–6.038)	5.210(4.357–6.230)
Participant Characteristics		Delayed care due to cost			Didn’t get care due to cost
		Weighted event rate	Unadjusted OR (95% CI)	Adjusted OR* (95% CI)	Weighted event rate	Unadjusted OR (95% CI)	Adjusted OR* (95% CI)
Middle-aged adults
Robust(N=6870)		6.7%	1.00	1.00	5.7%	1.00	1.00
Pre-frail(N=2188)		16.4%	2.753(2.301–3.295)	2.057(1.691–2.502)	16.2%	3.175(2.625–3.840)	2.359(1.917–2.903)
Frail(N=935)		25.0%	4.662(3.795–5.728)	2.514(1.989–3.177)	23.9%	5.178(4.169–6.430)	2.712(2.122–3.466)
Older adults
Robust(N=4234)		1.9%	1.00	1.00	2.0%	1.00	1.00
Pre-frail(N=2646)		3.9%	2.077(1.380–3.125)	1.728(1.161–2.736)	3.0%	1.499(0.977–2.301)	1.251(0.802–1.952)
Frail(N=1792)		6.2%	3.509(2.340–5.263)	2.591(1.693–3.966)	6.9%	3.575(2.451–5.216)	2.575(1.732–3.826)

Note:

^*weighted logistic regressions controlling for Age, Sex, Race, Marital Status, Region, Family Income and Educational attainment.

^*Middle-aged means 45–64 years older; Older adult means 65years or older

Primary care providers in community settings need to identify frailty early.⁹ Many acute issues affecting frail people can be prevented or safely managed in primary care settings.³ As our health system transitions from prioritizing volume to value, a more proactive, value-based primary care response to frailty is required. Since health deficits accumulate over years, primary care can intervene upstream to prevent the development of these deficits.³ There are many ways to identify frailty in clinical practice or healthcare databases, which include gait speed, timed-up-and-go test, the PRISMA 7 questionnaire, claims-based frailty index,¹⁰ or natural language processing of clinical notes to identify common clinical presentations of frailty, if the medical informatics infrastructure permits. For frail middle-aged and older people, targeted allocation of resources, such as care management, geriatrician referral, and multi-disciplinary team approach, may slow and reverse frailty and optimize health status and quality of life.

Also, we need to focus on the accessibility of health care for the frail population to help them get better access to care. Frailty assessment and management in primary care will require personnel and technology. Financial incentives from Medicare and commercial payers are needed to promote these services. Proactive identification and prevention of frailty could result in significant cost savings.