A total of 7,164 candidates completed the consent form; 347 (5%) did not consent to participate, and 124 (2%) provided incomplete answers to our survey. Our final sample included 6,693 participants, 1,332 in the LGBT + group (20%) and 5,361 in the non-LGBT + group (80%). There were answers from all the country's macro-regions (Fig. 1), although with a noticeable predominance from the Southeast.
Overall, the median age was 60 years, 68% were female, and 79% were white (Table 1). Participants had high literacy levels, with 79% having completed university or postgraduate courses. LGBT + participants were younger, more frequently male, single, and used the public healthcare system more often. They were also more likely to live in a rental (18% vs. 10%, p < .001) and to earn less than the minimum wage (10% vs. 6%, p < .001). LGBT + participants were more likely to report not having anyone to assist them if they became bedridden (22% vs. 15%, p < .001).
Table 1
Sample characteristics according to LGBT + group.
|
Total
|
Non-LGBT+
|
LGBT+
|
p-value
|
|
N = 6,693
|
N = 5,361
|
N = 1,332
|
|
Frail status (FRAIL Scale)
|
|
|
|
.07
|
Robust
|
3577 (53%)
|
2901 (54%)
|
676 (51%)
|
|
Prefrail
|
2742 (41%)
|
2170 (40%)
|
572 (43%)
|
|
Frail
|
374 (6%)
|
290 (5%)
|
84 (6%)
|
|
Sex assigned at birth
|
|
|
|
< .001
|
Male
|
2,115 (32%)
|
1,343 (25%)
|
772 (58%)
|
|
Female
|
4,578 (68%)
|
4,018 (75%)
|
560 (42%)
|
|
Cisgender
|
6,444 (96%)
|
5,361 (100%)
|
1,083 (81%)
|
< .001
|
Age (years)
|
|
|
|
< .001
|
50–59
|
3257 (49%)
|
2387 (45%)
|
870 (65%)
|
|
60–69
|
2490 (37%)
|
2112 (39%)
|
378 (28%)
|
|
≥ 70
|
946 (14%)
|
862 (16%)
|
84 (6%)
|
|
Race/ethnicity
|
|
|
|
< .001
|
White
|
5272 (79%)
|
4298 (80%)
|
974 (73%)
|
|
Black
|
357 (5%)
|
301 (6%)
|
56 (4%)
|
|
Other
|
1064 (16%)
|
762 (14%)
|
302 (23%)
|
|
Literacy
|
|
|
|
.24
|
College or more
|
5272 (79%)
|
4235 (79%)
|
1037 (78%)
|
|
High school
|
1201 (18%)
|
944 (18%)
|
257 (19%)
|
|
Middle school or less
|
220 (3%)
|
182 (3%)
|
38 (3%)
|
|
Macro-region
|
|
|
|
< .001
|
Southeast
|
5,186 (77%)
|
4,241 (79%)
|
945 (71%)
|
|
Southern
|
458 (7%)
|
350 (7%)
|
108 (8%)
|
|
Central-West
|
247 (4%)
|
171 (3%)
|
76 (6%)
|
|
Northeast
|
739 (11%)
|
553 (10%)
|
186 (14%)
|
|
Northern
|
63 (1%)
|
46 (1%)
|
17 (1%)
|
|
Public healthcare system utilization
|
1117 (17%)
|
788 (15%)
|
329 (25%)
|
< .001
|
Polypharmacy
|
2113 (32%)
|
1680 (31%)
|
433 (33%)
|
.41
|
Two or more chronic conditions
|
1651 (25%)
|
1349 (25%)
|
302 (23%)
|
.06
|
Geriatric Depression Scale score
|
1 (0, 2)
|
1 (0, 2)
|
1 (0, 2)
|
< .001
|
Comorbidities
|
|
|
|
|
Hypertension
|
2483 (37%)
|
2009 (37%)
|
474 (36%)
|
.20
|
Diabetes mellitus
|
1028 (15%)
|
827 (15%)
|
201 (15%)
|
.76
|
Cancer
|
357 (5%)
|
288 (5%)
|
69 (5%)
|
.78
|
Coronary disease
|
225 (3%)
|
177 (3%)
|
48 (4%)
|
.58
|
Heart failure
|
201 (3%)
|
156 (3%)
|
45 (3%)
|
.37
|
Cerebrovascular disease
|
98 (1%)
|
69 (1%)
|
29 (2%)
|
.03
|
Chronic obstructive pulmonary disease
|
234 (3%)
|
185 (3%)
|
49 (4%)
|
.69
|
Asthma
|
420 (6%)
|
352 (7%)
|
68 (5%)
|
.05
|
Chronic kidney disease
|
87 (1%)
|
57 (1%)
|
30 (2%)
|
< .001
|
Data are presented as median (interquartile ranges) for continuous measures and counts (%) for categorical measures.
In the LGBT + group, 816 (61%) identified as cisgender homosexual, 199 (15%) as cisgender bisexual, and 68 (5%) as cisgender pansexual or other sexual orientations. There were 249 (19%) participants who identified as transgender or other genders (29 transgender women, 3 transgender men, 6 travestis, and 211 non-binary or other genders).
In the LGBT + group, 816 (61%) identified as cisgender homosexual, 199 (15%) as cisgender bisexual, and 68 (5%) as cisgender pansexual or other sexual orientations. There were 249 (19%) participants who identified as transgender or other genders (29 transgender women, 3 transgender men, 6 travestis, and 211 non-binary or other genders).
We found 374 (6%) frail participants in our sample, of whom 84 (6%) were in the LGBT + group and 290 (5%) were in the non-LGBT + group. LGBT + participants were more frequently prefrail or frail than non-LGBT + participants (49% vs. 46%, p = .02). Frailty was more common in LGBT + females than non-LGBT + females (9% vs. 6%, p = .03), in LGBT + males aged ≥ 60 years than non-LGBT + males of the same age (8% vs. 3%, p = .004), and in non-cisgender participants (11% vs. 5%, p < .001) (Fig. 2).
In our multivariable analyses, LGBT + was not independently associated with frailty in our overall sample (OR = 1.24, 95%CI = 0.95–1.63, p = .1). However, we found that being LGBT + was independently associated with frailty in female participants aged ≥ 50 years (OR = 1.52, 95%CI = 1.08–2.13, p = .02) and in male participants aged ≥ 60 years (OR = 2.83, 95%CI = 1.41–5.69, p = .004) (Table 2). Older age, using the public healthcare system, and several comorbidities were also associated with frailty in our sample.
Finally, in a multivariable sensitivity analysis, we found that the non-cisgender group was independently associated with frailty (OR = 2.21, 95%CI = 1.42 = 3.42, p < .001). The association was confirmed both among females (OR = 2.11, 95%CI = 1.23 = 3.63, p = .007) and males (OR = 2.75, 95%CI = 1.30 = 5.85, p = .008) (Table 2).
Table 2
Generalized ordered logistic models examining the association between LGBT + groups and frailty, according to sex and age.
|
Prevalence
|
Unadjusted
odds ratios
(95%CI)
|
Adjusted*
odds ratios
(95%CI)
|
p-value
|
Overall
|
|
|
|
|
Age ≥ 50 years
|
|
|
|
|
Prefrailty or frailty vs. LGBT+
|
656/1332 (49%)
|
1.14 (1.01–1.29)
|
1.10 (0.97–1.25)
|
.14
|
Frailty vs. LGBT+
|
84/1332 (6%)
|
1.18 (0.92–1.51)
|
1.24 (0.95–1.63)
|
.11
|
Age = 50–59 years
|
|
|
|
|
Prefrailty or frailty vs. LGBT+
|
426/870 (49%)
|
1.12 (0.96–1.30)
|
1.04 (0.89–1.23)
|
.62
|
Frailty vs. LGBT+
|
42/870 (5%)
|
1.10 (0.76–1.59)
|
0.98 (0.66–1.44)
|
.92
|
Age ≥ 60 years
|
|
|
|
|
Prefrailty or frailty vs. LGBT+
|
230/462 (50%)
|
1.18 (0.97–1.44)
|
1.17 (0.96–1.44)
|
.13
|
Frailty vs. LGBT+
|
42/462 (9%)
|
1.51 (1.06–2.14)
|
1.56 (1.08–2.27)
|
.02
|
Sex assigned at birth = Female
|
|
|
|
|
Age ≥ 50 years
|
|
|
|
|
Prefrailty or frailty vs. LGBT+
|
320/560 (57%)
|
1.41 (1.18–1.69)
|
1.34 (1.12–1.62)
|
.002
|
Frailty vs. LGBT+
|
48/560 (9%)
|
1.44 (1.04–1.99)
|
1.52 (1.08–2.13)
|
.02
|
Age = 50–59 years
|
|
|
|
|
Prefrailty or frailty vs. LGBT+
|
194/340 (57%)
|
1.43 (1.13–1.80)
|
1.31 (1.03–1.67)
|
.03
|
Frailty vs. LGBT+
|
25/340 (7%)
|
1.59 (1.00-2.50)
|
1.45 (0.89–2.35)
|
.13
|
Age ≥ 60 years
|
|
|
|
|
Prefrailty or frailty vs. LGBT+
|
126/220 (57%)
|
1.41 (1.06–1.86)
|
1.34 (1.00-1.80)
|
.05
|
Frailty vs. LGBT+
|
23/220 (10%)
|
1.48 (0.93–2.35)
|
1.45 (0.88–2.37)
|
.14
|
Sex assigned at birth = Male
|
|
|
|
|
Age ≥ 50 years
|
|
|
|
|
Prefrailty or frailty vs. LGBT+
|
336/772 (44%)
|
1.26 (1.05–1.51)
|
1.21 (0.99–1.47)
|
.06
|
Frailty vs. LGBT+
|
36/772 (5%)
|
1.44 (0.92–2.26)
|
1.59 (0.94–2.69)
|
.08
|
Age = 50–59 years
|
|
|
|
|
Prefrailty or frailty vs. LGBT+
|
232/530 (44%)
|
1.23 (0.96–1.58)
|
1.15 (0.88–1.49)
|
.31
|
Frailty vs. LGBT+
|
17/530 (3%)
|
1.08 (0.54–2.20)
|
0.92 (0.42–1.98)
|
.82
|
Age ≥ 60 years
|
|
|
|
|
Prefrailty or frailty vs. LGBT+
|
104/242 (43%)
|
1.26 (0.94–1.68)
|
1.29 (0.95–1.76)
|
.10
|
Frailty vs. LGBT+
|
19/242 (8%)
|
2.37 (1.31–4.31)
|
2.83 (1.41–5.69)
|
.004
|
Sensitivity analysis
|
|
|
|
|
Age ≥ 50 years
|
|
|
|
|
Prefrailty or frailty vs. non-cisgender
|
143/249 (57%)
|
1.58 (1.22–2.03)
|
1.55 (1.19–2.02)
|
.001
|
Frailty vs. non-cisgender
|
27/249 (11%)
|
2.14 (1.41–3.23)
|
2.21 (1.42–3.42)
|
< .001
|
Age = 50–59 years
|
|
|
|
|
Prefrailty or frailty vs. non-cisgender
|
76/134 (57%)
|
1.51 (1.06–2.13)
|
1.48 (1.03–2.12)
|
.03
|
Frailty vs. non-cisgender
|
9/134 (6%)
|
1.56 (0.76–3.13)
|
1.43 (0.69–2.96)
|
.33
|
Age ≥ 60 years
|
|
|
|
|
Prefrailty or frailty vs. non-cisgender
|
67/115 (58%)
|
1.65 (1.13–2.41)
|
1.62 (1.09–2.41)
|
.02
|
Frailty vs. non-cisgender
|
18/115 (16%)
|
2.76 (1.64–4.66)
|
3.04 (1.74–5.33)
|
< .001
|
95%CI = 95% confidence interval; LGBT + = Lesbian, Gay, Bisexual, and Transgender; non-cisgender = transgender, non-binary, and other non-conforming genders.
All models were adjusted for age, race/ethnicity, public healthcare utilization, and comorbidities (hypertension, diabetes, cancer, coronary disease, heart failure, chronic obstructive pulmonary disease, asthma, cerebrovascular disease, and chronic kidney disease). Sensitivity analyses also adjusted for sex.