Physical Function Assessment of Older Veterans with Serious Mental Illness

Abstract

Objectives:

To characterize the physical function of older veterans with serious mental illness (SMI) across endurance, strength, and mobility domains.

Design:

Retrospective analysis of clinical performance data.

Setting:

Gerofit program, a national outpatient supervised exercise program for older veterans, delivered in Veterans Health Administration sites.

Participants:

Older veterans aged 60 and older (n=166 with SMI, n=1441 without SMI) enrolled across eight national Gerofit sites between 2010 and 2019.

Measurements:

Performance measures of physical function covering endurance (six-minute walk test), strength (chair stands, arm curls), and mobility (10-meter walk, 8-foot-up-and-go), were administered at Gerofit enrollment. Baseline data from these measures were analyzed to characterize the functional profiles of older veterans with SMI. One sample t-tests were examined to compare functional performance of older veterans with SMI to age- and sex-based reference scores. Propensity score matching (1:3) and linear mixed effects models were used to evaluate differences in function between veterans with and without SMI.

Results:

Older veterans with SMI performed worse on all measures of function (chair stands, arm curls, 10-meter walk, six-minute walk test, 8-foot-up-and-go) compared to age- and sex-based reference scores with statistically significant differences present in the male sample. Functional performance of those with SMI was also worse compared to propensity-score matched older veterans without SMI with statistically significant differences on chair stands, six-minute walk test, and 10-meter walk.

Conclusions:

Older veterans with SMI have compromised strength, mobility, and endurance. Physical function should be a core component of screening and treatment for this population.

Introduction

Serious mental illness (SMI; i.e., schizophrenia, schizoaffective disorder, bipolar disorder, and recurrent major depression) is among the leading causes of disability¹ due to the substantial functional impairment caused by mental health symptoms. Individuals with SMI have a shortened life expectancy due to elevated rates of multiple comorbid medical conditions and low levels of physical activity.^2–5 The medical profile of those with SMI and its downstream consequences worsen over the lifespan: older adults with SMI have higher rates of chronic medical conditions, more falls and nursing home admissions, more medical emergency room visits, more extended medical hospitalizations, and spend more time being sedentary compared to their counterparts without SMI.^6–10 As such, the confluence of medical morbidity, medical health care utilization, and high sedentary behavior reduce the independence of older adults with SMI, contributing to exceptionally high costs for the health care systems caring for this population.^11–16

Physical function comprises multiple domains including endurance, strength, and mobility, all of which decline during aging, a process which is accelerated by a sedentary lifestyle.¹⁷ Each of these components is critical for older adults because better function promotes greater independence and quality of life, reduces the incidence of falls, and predicts survival.^18–20 Physical function is modifiable through exercise ^17,21–25 and, thus, offers a valuable target for intervention. Despite the compromised physical health of older adults with SMI, there is a lack of data on the multiple dimensions of physical function in this population. This information is needed to develop more targeted interventions.

To our knowledge, no studies have evaluated multiple domains of function in an older adult sample with SMI. The extent of the functional decline in the older SMI population is not well-understood, which limits opportunities for developing effective interventions for this group. Research on function in persons with SMI, which has largely focused on young and middle-aged adults, has shown substantive impairments in endurance compared to individuals without SMI.^26–28 A limited number of studies have explored other domains of function (i.e., strength and mobility) in young and middle-aged participants with SMI. Nygard and colleagues (2019) examined lower body strength, balance, and mobility in addition to endurance in young adults with schizophrenia (average age of 35 years) and found that performance in all domains was worse than age- and gender-matched references.²⁹ Zechner and colleagues (2021) examined lower body strength, endurance, and balance capabilities in middle-aged persons with SMI (average age of 50 years old) and showed that participants with SMI performed similarly on functional tests to non-SMI counterparts who were two to three decades older.³⁰ Together, the existing research suggests compromised physical function in young and middle-aged adults with SMI, with the most work done on endurance, and little to no physical function data on older adults with SMI.

The purpose of the present study was to characterize the physical function of older veterans (i.e., those who previously served in the active U.S. military)³¹ with SMI across endurance, strength, and mobility domains with performance-based assessments. The study aimed to compare functional performance of older veterans with SMI to (a) age- and sex-based reference scores and (b) older veterans without SMI. We hypothesized that older veterans with SMI would have lower physical function levels compared to both age- and sex-based reference scores and to older veterans without SMI.

Method

Study Design

This paper reports on a retrospective analysis of baseline data collected at enrollment in Gerofit,^32,33 a national facility-based outpatient clinical exercise program delivered in the Veterans Health Administration, between 2010 and 2019. Institutional Review Boards at The Durham VA Health Care System and Providence VA Healthcare System approved this retrospective analysis.

Participants

Inclusion criteria for Gerofit are: (1) veteran aged 65 or older (note: the lower age limit was reduced to 60 during periods of increased referrals from younger veterans), (2) medically stable per the veteran’s primary care provider, and (3) able to provide own transportation. Exclusion criteria are: (1) unable to independently perform activities of daily living (ADLs), (2) cognitive impairment preventing safe exercise, (3) medical conditions for which exercise is contraindicated (e.g., unstable angina, active proliferative diabetic retinopathy, oxygen dependence, frank incontinence, active open wounds), (4) active substance use disorder, (5) homelessness, (6) inability to successfully participate in a group setting, and (7) behavioral concerns impacting group participation.

Measures

Demographic, Mental Health, and Physical Health

Demographic, mental health, and physical health information were obtained from Department of Veterans Affairs electronic health records (EHR) for the year prior to Gerofit enrollment.

Demographics.

Age, sex, race, ethnicity, and marital status were examined for all veterans.

Mental Health.

SMI status was defined as the presence of any of the following mental health diagnoses: schizophrenia, schizoaffective disorder, bipolar disorder, or recurrent major depressive disorder.^14,15,34 International Classification of Disease (ICD)^35,36 codes for diagnoses were pulled from the EHR for the year prior to Gerofit enrollment. Based on prior research,³⁷ diagnoses were considered hierarchically in the following order: schizophrenia or schizoaffective disorder, bipolar disorder, and recurrent major depressive disorder, such that each veteran had only one primary SMI diagnosis. Psychiatric medication prescriptions in the year prior to Gerofit enrollment were obtained for three classes of medications: (1) antipsychotic, (2) mood stabilizer, and (3) antidepressants. Multiple medications were possible within and across classes.

Physical Health.

The Elixhauser comorbidity index³⁸ was used to evaluate physical health status. This measure is comprised of 30 possible conditions covering 26 medical conditions, two mental health conditions (depression and psychosis), and alcohol and drug use. Given the SMI focus of this analysis and that problematic alcohol/drug use is exclusory for Gerofit, only the 26 medical conditions were used in analyses. ICD codes ^35,36 for each of the 26 medical conditions were pulled from the EHR for one year prior to Gerofit enrollment. A sum of all conditions was calculated and categorized into three groups (0, 1–2, 3 or more) for analyses.

Physical Function

Physical function was assessed with a physical performance battery of well-established tests covering endurance, strength, and mobility domains.

Endurance.

Endurance was assessed with the six-minute walk test.³⁹ In this test, participants walk continuously for six minutes on a flat course covering as much distance as possible. The total distance (in yards) covered in six minutes was used in analyses.

Strength.

Upper body strength was assessed with the arm curl test.³⁹ Holding a weight in their dominant arm (five-pound weight for women, eight-pound weight for men), participants complete as many biceps curls as possible in 30 seconds. The total number of curls completed was used in analyses. Lower body strength was assessed with the 30-second chair stand test.³⁹ With arms across their chest, participants rise up out of a chair and back down as many times as possible in 30 seconds. The total number of chair stands completed in 30 seconds was used in analyses.

Mobility.

Mobility was assessed with two measures: (1) 10-meter walk test^40,41 and (2) 8-foot-up-and-go test.³⁹ In the 10-meter walk test, participants walk at their usual pace on a 10-meter flat course. The total time to complete the 10 meters is recorded in seconds. Usual walking speed (in meters/second) was used in analyses. In the 8-foot-up-and-go test, participants begin seated in a chair, then rise from the chair, walk eight feet at a rapid pace, and return to the chair. The total time to complete the test is recorded in seconds and was used in analyses.

Procedure

Veterans are referred to Gerofit by any Department of Veterans Affairs clinical provider but must be in stable health and have approval to participate by their VA primary care provider prior to enrollment. Gerofit program staff then perform a chart review and conduct an initial phone or in-person screen to assess interest and eligibility. If eligible, participants complete their initial Gerofit physical performance assessment. Gerofit staff uses the assessment results to develop an exercise prescription for each veteran that guides their participation in this facility-based outpatient clinical exercise program. This paper reports only baseline data before the initiation of exercise (A detailed description of the Gerofit program can be found elsewhere).^32,33

Data Analysis

Analyses were conducted using SAS (Version 9.4). Demographic and health characteristics of older veterans with and without SMI were compared using t-tests for continuous variables and chi-squares for categorical variables. For continuous variables, results of t-tests were compared to results of a non-parametric analog (Mann-Whitney test). Significance values between parametric and non-parametric tests were identical so for convenience, t-test p-values were reported.

Two types of analyses were conducted to characterize the physical function of veterans with SMI. In the first analysis, within the SMI group, baseline values of physical function measures (chair stands, arm curls, six-minute walk test, 8-foot-up-and-go, 10-meter walk) were compared to age- and sex-based reference scores^42,43 using one-sample t-tests. Specifically, one-sample t-tests were run separately for male and female veterans with SMI using reference scores for male or female individuals age 60 and older. ^42,43 In the second analysis, propensity scores^44,45 were generated by regressing SMI status on baseline demographic variables (age, sex, race, and marital status) using logistic regression. Propensity score matching (1 SMI :3 Non-SMI) was constructed using greedy nearest neighbor matching without replacement within 0.40 caliper width using SAS v9.4 PROC PSMATCH. Linear mixed effects models with SMI status as the predictor and a random intercept for matched set were conducted to evaluate propensity score-matched group differences in function (chair stands, arm curls, six-minute walk test, 8-foot-up-and-go, 10-meter walk) between veterans with and without SMI.

Results

Participants

The sample comprised 1607 older veterans enrolled across eight Gerofit sites between 2010–2019 of which 166 (10%) had SMI (Table 1). The SMI sample at baseline was, on average, 70 years old (SD=5.4), primarily male (91%), Black or African American (42%) or White (55%), not Hispanic or Latinx (95%), and not currently married (58%). The majority (60%) had three or more medical conditions. Except for race, all demographic and medical characteristics differed between the SMI and non-SMI samples. The SMI sample was younger, included more female veterans, included more Hispanic or Latinx veterans, was less likely to be married, and had more medical conditions than the non-SMI sample (all p-values <.05). Clinically within the SMI sample, the most common psychiatric diagnosis was recurrent major depressive disorder (75%), followed by schizophrenia or schizoaffective disorder (13%) and bipolar disorder (11%). Over three-quarters of the sample had been prescribed an antidepressant over the previous year (78%) with a smaller percentage having been prescribed mood stabilizers (30%) and antipsychotics (22%).

Table 1.

Baseline Characteristics of Older Veterans with and without SMI

Characteristic	Older Veterans with SMI (n=166)	Older Veterans without SMI (n=1441)	Test stasitic (t or χ2)	df	p-value
Demographic and Health Characteristics
Age, M (SD), range	70.3 (5.4), 60–92	76.0 (8.4), 63–97	8.5	1605	<.0001
Sex, n (%)			25.4	1	<.0001
Male	151 (91.0)	1410 (97.8)
Female	15 (9.0)	31 (2.2)
Race, n (%)			4.4	2	.11
Black or African American	69 (41.6)	496 (34.4)
White	92 (55.4)	916 (63.6)
Asian, American Indian or Alaska Native, or Native Hawaiian or Other Pacific Islandera	2 (1.2)	12 (.8)
Declined to Answer or Unknown	3 (1.8)	17 (1.2)
Ethnicity, n (%)			5.1	1	.04b
Hispanic or Latinx	6 (3.6)	19 (1.3)
Not Hispanic or Latinx	158 (95.2)	1403 (97.4)
Declined to Answer or Unknown c	2 (1.2)	19 (1.3)
Marital Status, n (%)			30.4	1	<.0001
Currently Married	70 (42.2)	924 (64.1)
Not Currently Married	96 (57.8)	517 (35.9)
Elixhauser Comorbidity Score, n (%)			36.6	1	<.0001d
0	13 (7.8)	354 (24.6)
1–2	53 (31.9)	533 (37.0)
3+	100 (60.2)	554 (38.4)
SMI Diagnosis, n (%) e			N/A	N/A	N/A
Schizophrenia or Schizoaffective Disorder	22 (13.2)	-	N/A	N/A	N/A
Bipolar Disorder	19 (11.5)	-	N/A	N/A	N/A
Recurrent Major Depressive Disorder	125 (75.3)	-	N/A	N/A	N/A
Psychiatric Medication Prescriptions, n (%)f			N/A	N/A	N/A
Antipsychotic	37 (22.3)	-	N/A	N/A	N/A
Mood Stabilizer	50 (30.2)	-	N/A	N/A	N/A
Antidepressant	129 (77.7)	-	N/A	N/A	N/A

Note. SMI = serious mental illness, df = degrees of freedom, N/A = not applicable (only included for SMI group). Percentages may not add up exactly to 100.0 due to rounding.

^aCategories are reported together given low frequencies and was combined with declined to answer/unknown for chi-square analysis given low frequencies.

^bFisher’s exact test p-value was used.

^cDue to low frequencies, this group was combined with not Hispanic or Latinx in group comparison analysis.

^dRow mean score test was used.

^eDiagnoses were considered hierarchically (schizophrenia or schizoaffective disorder followed by bipolar disorder followed by recurrent major depressive disorder). Thus, categories are not overlapping.

^fMultiple medications within and across categories were possible.

Function Compared to Age- and Sex-based Reference Scores: SMI Group Only

Older female veterans with SMI (n=15) performed significantly worse than reference scores on the six-minute walk test (t[14]=−2.5, p=.0253). Scores on chair stand, arm curl, 8-foot-up-and-go, and 10-meter walk did not significantly differ from reference scores (all ps>.05) (Table 2a).

Table 2a.

Baseline Physical Function of Older Female Veterans with SMI Compared to Age- and Sex-based Reference Scores

Physical Function Domain	Measure	Female
		Older Veterans with SMI M (SD) (n=15)	Reference Score	t	df	p-value
Lower Body Strength	Chair Stand (# in 30 seconds)41,a	10.9 (4.0)	12.7	−1.7	14	.1113
Upper Body Strength	Arm Curl (# in 30 seconds)41,a,b	15.8 (3.6)	14.3	1.3	9	.2183
Endurance	Six-minute Walk Test (yards)41,a	398 (206)	531	−2.5	14	.0253
Mobility	8-foot-up-and-go (seconds)41,a,c	8.4 (3.9)	6.2	2.1	13	.0564
Mobility	10-meter Walk (meters/second)42,a	1.05 (.31)	1.09	−0.5	13	.6101

Older male veterans with SMI (n=151) performed significantly worse than reference scores on the chair stand (t[146]=−9.1, p<.0001), arm curl (t[88]=−2.4, p=.0204), six-minute walk test (t[150]=−11.7, p<.0001), 8-foot-up-and-go (t[148]=6.9, p<.0001), and 10-meter walk (t[148]=−6.7, p<.0001) (Table 2b).

Table 2b.

Baseline Physical Function of Older Male Veterans with SMI Compared to Age- and Sex-based Reference Scores

Physical Function Domain	Measure	Male
		Older Veterans with SMI M (SD) (n=151)	Reference Score	t	df	p-value
Lower Body Strength	Chair Stand (# in 30 seconds)42,a	10.5 (4.9)	14.2	−9.1	146	<.0001
Upper Body Strength	Arm Curl (# in 30 seconds)42,a,b	16.4 (5.9)	17.9	−2.4	88	.0204
Endurance	Six-minute Walk Test (yards)42,a	431 (164)	587	−11.7	150	<.0001
Mobility	8-foot-up-and-go (seconds)42,a,c	8.5 (5.1)	5.6	6.9	148	<.0001
Mobility	10-meter Walk (meters/second)43,a	1.05 (.27)	1.2	−6.7	148	<.0001

Note. SMI = serious mental illness, df = degrees of freedom. Bolded p-values indicate p<.05. Reference scores for chair stand, arm curl, six-minute walk test, and 8-foot-up-and-go are the combined values for ages 60 and older reported in Rikli & Jones (1999).⁴² Reference scores for the 10-meter walk are a combined value for ages 60 and older from Hall et al. (2017).⁴³

^aSample size for Arm Curl was lower than other tests because it was only added to the Gerofit physical performance battery in more recent years. Other differences in sample sizes between tests reflect missing data. Sample sizes are given below:

Chair Stand: female =15, male=147

Arm Curl: female =10, male=89

Six-minute walk test: female =15, male=151

8-foot-up-and-go: female =14, male=149

10-meter walk: female =14, male=149

^bDominant arm score was used.

^cHigher scores indicate worse performance.

Function Compared to Older Veterans without SMI

On each of the demographic characteristics used in propensity score matching, there were no significant differences between SMI and non-SMI groups (all ps>.05). Older veterans with SMI performed significantly worse on the chair stand (F[1,491]=4.8, p=.03), the six-minute walk test (F[1,497]=6.5, p=.01) and the 10-meter walk (F[1,493]=6.0, p=.02) compared to propensity-score matched older veterans without SMI. There were no statistically significant differences between groups on arm curls or the 8-foot-up-and-go (ps>.05) (Table 3).

Table 3.

Baseline Physical Function of Older Veterans with SMI and Propensity-matched Older Veterans without SMI

Physical Function Domain	Measure	Older Veterans with SMI (n=166) M (SD)	Propensity-Matched Older Veterans without SMI (n=498) M (SD)	F	df (num, denom)	p-value
Lower Body Strength	Chair Stand (# in 30 seconds)	10.6 (4.8)a	11.5 (4.8)	4.8	1,491	0.03
Upper Body Strength	Arm Curl (# in 30 seconds)	16.4 (5.7)b,c	17.5 (4.9)b,c	3.6	1,186	0.06
Endurance	Six-minute walk test (yards)	428 (168)	463 (157)	6.5	1,497	0.01
Mobility	8-foot-up-and-go (seconds)	8.4 (5.0)d	7.9 (5.7)	1.5	1,492	0.23
Mobility	10-meter walk (meters/second)	1.05 (.28)d	1.11 (0.28)	6.0	1,493	0.02

Note. SMI = serious mental illness, df = degrees of freedom, num = numerator, denom = denominator. Bolded p-values indicate p<.05.

^an=162

^bn_SMI = 99, n_non-SMI = 237

^cSample size for Arm Curl was lower than other tests because it was only added to the Gerofit physical performance battery in more recent years.

^dn=163

Discussion

The present study characterized the physical function of older veterans with SMI across endurance, strength, and mobility domains. Results demonstrated compromised physical function in the older veteran sample with SMI compared to age- and sex-based reference scores and to older veterans without SMI. This study reveals a concerning functional profile of older veterans with SMI that includes substantially reduced endurance, strength, and mobility.

Older male and female veterans with SMI had significantly worse endurance, strength, and mobility compared to age- and sex-based reference scores; however, outcomes of the female veteran sample should be interpreted cautiously given the small sample size. Overall, these results extend prior research by illustrating that the compromised physical function observed in young and middle-aged adults with SMI^{29,30,46–48} persists into older adulthood. This is further emphasized by veterans’ relative performance on these tests with population norms. Older male veterans with SMI, which comprised 91% of the SMI sample, scored below the lowest quartile on the chair stand (<25^th percentile), 8-foot-up-and-go (<10^th percentile), and six-minute walk test (<10^th percentile),⁴² indicating substantial impairment in lower body strength, mobility, and endurance. These functional deficits observed in older veterans with SMI are especially concerning given that compromised function is associated with difficulty completing ADLs and greater risk of hospitalization and death.^49–52

Older veterans with SMI also performed worse than propensity-score matched older veterans without SMI on all physical function tests with performance on the chair stand, six-minute walk test, and 10-meter walk were statistically significant. These findings likely reflect the increased medical comorbidity of older veterans with SMI compared to their same age counterparts without SMI, which contributes to substantially reduced function.^14,53 Further, physical inactivity and sedentary behavior have been shown to reduce physical function performance in older veterans.⁵⁴ Individuals with SMI are, on average, less physically active, less likely to meet physical activity guidelines, and are more sedentary than those without SMI.^15,55 In turn, it may be that at Gerofit enrollment, the SMI sample was substantially more sedentary and inactive than the non-SMI sample, further exacerbating the health burden. The degree of physical functional deficits observed in those with SMI are further evidenced by the compromised function of the non-SMI group. In fact, older veterans without SMI performed below the 25^th percentile (considering male normative scores) on chair stands, six-minute walk test, and 8-foot-up-and-go.⁴²

The study had limitations that should be considered when drawing conclusions. First, the sample was comprised of veteran users of the Veterans Health Administration who enrolled in a clinical exercise program, which may not be representative of the broader population of older adults. Relatedly, survivor bias may have affected our sample of older veterans with SMI given the significant premature mortality in this population.^2,5 Second, although the sample was more racially diverse than many non-veteran studies, there were few female individuals, which is consistent with veteran studies but limits generalizability of these findings beyond older male veterans. Third, although propensity score matching was used, the SMI group was much smaller than the non-SMI group which may have impacted comparisons, particularly for the arm curl test (which was added to the Gerofit battery in more recent years). Fourth, due to the unbalanced composition of SMI diagnoses (i.e., majority had major depressive disorder), it was not possible to evaluate differences in function at the mental health condition level. Based on these limitations, this study should be replicated in a larger sample of older people (veterans and non-veterans) with SMI comprised of more female participants and those with schizophrenia/schizoaffective and bipolar disorders. Further, future work should consider the role of psychiatric medications on physical function in older persons with SMI.

Conclusions

Overall, this study comprehensively characterized physical function across multiple domains in a sample of older veterans with SMI. Results showed substantially compromised function across strength, mobility, and endurance domains. The compromised physical function profile of older adults with SMI and the downstream consequences (i.e., high nursing home admissions, medical hospitalizations) underscore an urgent need for effective interventions for this group that target whole-body health. Follow-up studies should consider evaluating the effects of multicomponent exercise, an established modality to improve function in older adults,^23,25 on strength, mobility, and endurance in older persons with SMI.