Nutritional, Cognitive, and Functional Deficits, Frailty, and Quality of Life Associated With Fall Risk in Community-Dwelling Older Adults: A Cross-Sectional Study Conducted in Brazil

Abstract

Introduction/Objective:

Falls affect approximately 30% of the older adult population. We aimed to compare the associations between fall risk and different multidimensional health aspects among older adults receiving care in the Brazilian Primary Health Care (PHC) system.

Method:

Cross-sectional, quantitative study involving older adults from PHC. The Fall Risk Score, Mini Nutritional Assessment, Mini-Mental State Examination, Edmonton Frail Scale, Barthel Index, Lawton & Brody Scale, and Medical Outcomes Study Questionnaire Short Form was used to measure the variables of interest. Correlation analyses and binary logistic regression were also employed.

Results:

A total of n = 257 individuals participated, of whom n = 102 (39.7%) were with risk for falls. Preserved cognition, absence of frailty, and better functionality levels were identified as protective factors against fall risk through association and correlation analyses. The binary logistic regression analysis found that the factors contributing most to the reduction of fall risk were higher nutritional scores, better cognitive function, preserved functionality (BADL and IADL), and the functional domain of quality of life (QoL).

Conclusion:

Better nutritional status, cognition, functionality, and QoL were associated with a lower risk of falls. Although frailty exhibited similar results, it did not stand out equally as a contributing factor to fall risk.

Introduction

Falls have become a significant health concern, occurring in approximately 30% of older adults worldwide and causing numerous physical and psychological impairments in this population. ¹ The most recent Brazilian National Health Survey, which investigated the incidence of falls among older adults, found that 7.5% of falls occurring in 2013 were severe. ² Authors reported a fall prevalence of 13.6% in 2022. ³

The primary risk factors for falls identified in the literature include a history of falls, impaired balance, mobility limitations, medication use, and environmental factors such as poor lighting and excessive household obstacles.^4,5 However, authors suggest that other aspects may also influence this context. For example, inadequate nutrition may reduce muscle strength, consequently impacting an individual’s balance and mobility. ⁶ This may also indicate functional and cognitive dependence, which affects autonomy and leads to frailty and reduced Quality of Life (QoL).^7,8

The early identification of fall risk is essential for implementing preventive measures. ¹ The World Health Organization (WHO) released the Guidance on Person-Centered Assessment and Pathways in Primary Care, which highlights as a key point the development of care plans aimed at reducing functional decline through physical exercise, nutritional supplementation, cognitive stimulation, and home adaptations in Primary Health Care (PHC). ⁹

Although numerous studies have investigated the main causes of falls among older adults, significant gaps remain in the literature. The interactions between different risk factors need to be evaluated not only individually but also collectively. This study, therefore, aims to advance knowledge by conducting a detailed and in-depth analysis of multidimensional factors, comparing their relative influence on the risk of falls within the same sample of older adults. By doing so, it seeks to identify pathways for more effective interventions, particularly in Primary Health Care (PHC), a setting dedicated to preventing diseases and conditions with a high potential for mortality, such as falls.

The objective of this study was to compare the associations between fall risk and different multidimensional health aspects among older adults receiving care in the Brazilian PHC system. The study hypothesized that changes in factors such as nutrition, cognition, functionality, and QoL exert varying levels of influence on fall risk in this population.

Method

This was an observational, cross-sectional, and quantitative study involving older adults receiving PHC in the municipality of Santa Cruz, located in the state of Rio Grande do Norte, Brazil. The study was conducted following the principles of the Declaration of Helsinki for good scientific practices and received approval from the Research Ethics Committee of the Universidade Federal of Rio Grande do Norte (approval number 4267762), in accordance with Brazilian ethical standards for human research. The institution coordinating PHC in the municipality granted prior authorization for conducting the study with its users.

Before data collection, participants provided written consent through the signing of an informed consent form. They were informed about the study’s objectives, the potential risks of discomfort related to personal questions, their right to withdraw from participation at any time, and the benefits of undergoing evaluations using scientifically validated instruments. To protect their personal data from breaches and ensure confidentiality, researchers entered data into digital databases stored on offline devices, replacing names with numerical codes corresponding to the physical records of each individual. No individual received any compensation, as participation was entirely voluntary, as stated in the signed informed consent forms. After the collected data were analyzed, participants were informed by phone about any identified changes and were referred for multidisciplinary healthcare services provided by the university involved in the study.

The study population consisted of older adults registered in PHC. According to government data, Santa Cruz had a population of 5,433 older adults during the study period. ¹⁰ Considering this data, a 95% confidence interval and a 5% margin of error, the sample size was calculated using an online finite population calculator (available at: https://calculareconverter.com.br/calculo-amostral/), which estimated a sample of 359 participants.

The sample recruitment was conducted with the assistance of PHC professionals, who provided the addresses of potential participants listed in the unit’s records, allowing researchers to visit their homes or schedule interviews at a day, time, and location of their preference. Despite these efforts, many individuals were not found at home or did not express interest in participating when contacted by phone. Consequently, this was not sufficient to reach the estimated sample size, and a total of 257 individuals completed the study. One possible factor contributing to this challenge is that the sample testing was based on the total elderly population of the municipality, whereas the PHC system does not have a registry covering the entire population. Additionally, the study focused on the city’s urban area, despite the municipality having a sizable rural population.

Inclusion criteria included individuals aged ≥60 years (the age criterion for older adults in Brazil, ¹¹ registered in one of the PHC units in the study setting, and residing in the municipality for at least 6 months at the time of data collection. Exclusion criteria included severe self-reported cognitive impairment (previous diagnoses of diseases with cognitive impairment reported by the participant, family member, or legal guardian) compromising response validity and reliance on non-oral nutrition, such as enteral or parenteral nutrition.

Data collection occurred at the researchers’ convenience through face-to-face interviews, using structured tools. Participants selected the options that best represented their responses, which were recorded by the interviewer. Data collection occurred between June 2023 and March 2024, with researchers consisting of 16 university students from various health disciplines. The control of variation in researcher observation was ensured through the methodology adopted in the training process for the researchers. All of them were trained to conduct full interviews using all instruments for each participant. The training was conducted by the coordinator in 3 stages. First, he provided the theoretical foundation through slide presentations in a closed room setting. In the second stage, each researcher conducted at least 2 pilot interviews with individuals who were not part of the study sample. In the third stage, the coordinator led a point-by-point discussion on the interpretation adopted by the researchers for each question to ensure consistency across all participants. Each interview lasted approximately 60 min and was conducted at PHC units, participants’ homes, or university spaces, depending on prior scheduling. There was no blinding of researchers or participants, and no rigorous randomization was performed for group formation.

To characterize the sample’s sociodemographic and health profile, the Older Adult Health Record Booklet, a tool provided by the Brazilian Ministry of Health for monitoring older adults in PHC, was used. This booklet includes information such as gender, age, marital status, self-reported skin color, education level, history of falls, polypharmacy (use of 5 or more medications), and self-reported chronic diseases such as hypertension and diabetes mellitus. ¹² Variables for sociodemographic and health characteristics were derived from this instrument.

Fall risk was assessed using the Fall Risk Score (FRS), which includes questions on fall history, medication use, sensory deficits, mental state, and gait difficulties. Scores range from 0 to 11 points, with individuals scoring >3 points classified as with risk for falls.^13,14 In this study, the sample was divided into 2 groups: With Risk and Without Risk, which were used as categorical variables to test the study hypothesis. For falls, we used the definition of an unintentional event that results in a change in the individual’s position to a lower level than the initial one, with or without impact against a solid surface. ¹⁵

Multidimensional aspects were evaluated using various instruments. Nutritional status was measured using the Mini Nutritional Assessment (MNA), a screening tool that classifies individuals as malnourished (0-16 points), at risk of malnutrition (17-23.5 points), or eutrophic (24-30 points) based on eating habits, weight changes, BMI, and food frequency.^16,17 For this study, categories were reorganized into “Impaired” (malnourished or at risk [0-23.5 points]) and “Eutrophic” (24-30 points).

Cognition was evaluated using the Mini-Mental State Examination (MMSE), which tests orientation, calculations, drawing, and writing skills, with scores ranging from 0 to 35 points. Scores below 17 indicate impaired cognition. ¹⁸ Participants were classified as having “Preserved” (≥17 points) or “Impaired cognition” (<17 points).

Frailty was assessed using the Edmonton Frail Scale (EFS), which includes physical and subjective health measures and classifies individuals as severely (≥11 points), moderately (9-10 points), or mildly frail (7-8 points), apparently vulnerable (5-6 points), or not frail (0-4 points) with a total score ranging from 0 to 22 points.^19,20 For this study, categories were reorganized as “With frailty” (≥5 points) and “Without frailty” (0-4 points).

Functionality was evaluated with 2 tools:

• The Barthel Index, which assesses independence in basic activities of daily living (BADL), such as bathing and dressing, with scores ranging from 0 to 100. Scores of 100 classified the individual as “Independent,” while lower scores (<100 points), as “Dependent.” ²¹

• The Lawton & Brody Scale, which assesses independence in instrumental activities of daily living (IADL). The original scale consists of 8 items related to skills in using the telephone, shopping, cooking, housekeeping, laundry, transportation access, self-managed medication, and financial management. It is scored trichotomously (1, 2, or 3 points per item), resulting in a total score ranging from 8 to 24 points.^22,23 We excluded the “laundry” item, as it is culturally a task predominantly performed by women in the study region. This adjustment standardized the evaluation and eliminated the need for gender-based weighting. With these modifications, the scoring range was adjusted to 7 to 21 points, classifying individuals as “Totally dependent” (7 points), “Partially dependent” (8-20 points), and “Totally independent” (21 points). In our study, we reclassified participants into 2 categories: “Dependent” (7-20 points) and “Independent” (21 points).

QoL was measured using the Medical Outcomes Study Questionnaire Short Form-36 (SF-36), which includes 8 domains and 2 dimensions (Physical and Mental health). Scores range from 0 to 100, with higher scores indicating better QoL. ²⁴ For this study, scores >50 were classified as “Better” QoL, while scores ≤50 were considered “Worse.”

Validated in Brazilian Portuguese instruments were used in the assessment. There were no missing data.

Data were tabulated using Microsoft Excel 2019 (Microsoft Corporation, Washington, WA, EUA) and analyzed with Statistical Package for the Social Science version 21.0 (IBM, Armonk, NY, EUA). Normality was tested using the Kolmogorov-Smirnov test, which showed a non-normal distribution. Pearson’s chi-square test was used to analyze categorical variables. Mann-Whitney U tests were applied to scalar analyses, with descriptive measures presented as means, standard deviation (SD), and percentiles. Spearman’s correlation test (r) assessed relationships between variables, with correlation strengths categorized as weak (r < .29), moderate (.29 < r < .49), and strong (r > .50). A 95% Confidence Interval (95% CI) and a 5% significance level (P < .05) was adopted. Binary logistic regression models was used to measure the prediction of study group variables and variables identified as potential confounding factors based on the independent variables. The data were synthesized, highlighting R² (Nagelkerke), the Model LR value, Wald value, Beta (β), standard error of Exp (β; SE), and Exp (β; OR). For all corresponding tests, the Odds Ratio (OR) was calculated, with an OR >1.00 indicating a positive association between the exposure and the outcome. ²⁵

The data supporting the findings of this study are openly available in the Mendeley Data repository at https://data.mendeley.com/datasets/pn8zj444jh.

Results

The total sample consisted of 257 participants, of whom n = 102 (38.9%) were classified as having some level of fall risk, while n = 155 did not present a fall risk. No individuals were excluded from the study.

Table 1 presents the sociodemographic and health profiles of the participants. The 60 to 79 years age group, absence of previous falls, absence of polypharmacy, and the presence of chronic diseases showed statistical significance with the Without Risk group. Their respective odds ratios (OR) were positive (>1.00) and were considered for adjustment analysis as potential confounding factors.

Table 1.

Sociodemographic and Health Characterization of Older Adults by Fall Risk (n = 257).

Sociodemographic and health profile	Fall risk (FRS)
	With risk		Without risk		Total		P a	OR (95% CI)
	(n = 102)		(n = 155)		(n = 257)
	n	%	n	%	n	%
Gender
Woman	71	69.6	103	66.5	174	67.7	.597	0.94 (0.77-1.16)
Men	31	30.4	52	33.5	83	32.3
Age range, yr
60-79	64	62.7	121	78.1	185	72.0	.007	1.38 (1.06-1.81)
>80	38	37.3	34	21.9	72	28.0
Marital status
Single/widowed/divorced	56	54.9	76	49.0	132	51.4	.357	0.91 (0.75-1.11)
Married/cohabitating	46	45.1	79	51.0	125	48.6
Skin color
White	45	44.1	66	42.6	111	43.2	.808	0.97 (0.79-1.19)
No white	57	55.9	89	57.4	146	56.8
Education
Literate	55	53.9	72	46.5	127	49.4	.241	0.88 (0.73-1.08)
Not literate	47	46.1	93	60.0	140	54.5
Fall history
No	16	15.7	82	52.9	98	38.1	<.001	1.82 (1.51-2.20)
Yes	86	84.3	73	47.1	159	61.9
Polypharmacy
No	73	71.6	144	92.9	217	84.4	<.001	2.41 (1.45-4.03)
Yes	29	28.4	11	7.1	40	15.6
Self-reported chronic diseases
No	5	4.9	46	29.7	51	19.8	<.001	1.70 (1.46-1.99)
Yes	97	95.1	109	70.3	206	80.2

Abbreviations: 95% CI, 95% confidence interval; FRS, fall risk score; OR, odds ratio for cohort “without risk.”

^aPearson’s Chi-Square Test.

Table 2 presents the categorical analysis of the evaluated multidimensional aspects. Impaired cognition and the presence of frailty emerged as significant risk factors for falls. Deficits in nutritional status, functional status (assessed through Instrumental Activities of Daily Living [IADL] and Lawton & Brody), and in the Physical Role Functioning and Physical Functioning domains of Quality of Life (SF-36) also indicated that lower scores in these aspects represent important risk factors. Similar results were observed in the Physical Health dimension.

Table 2.

Categorical Analysis of Multidimensional Aspects by Fall Risk (n = 257).

Multidimensional aspects		Fall risk (FRS)
		With risk		Without risk		Total		P a	OR (95% CI)
		(n = 102)		(n = 155)		(n = 257)
		n	%	n	%	n	%
Nutritional status (MNA)	Impaired	41	40.2	31	20.0	72	28.0	<.001	1.73 (1.30-2.30)
	Eutrophic	61	59.8	124	80.0	185	72.0
Cognition (MMSE)	Impaired	25	24.5	9	5.8	34	13.2	<.001	2.13 (1.62-2.79)
	Preserved	77	75.5	146	94.2	223	86.8
Frailty (EFS)	With frailty	72	70.6	54	34.8	126	49.0	<.001	2.49 (1.76-3.54)
	Without frailty	30	29.4	101	65.2	131	51.0
Functional status-BADL (Barthel index)	Dependent	28	27.5	36	23.2	64	24.9	.443	1.14 (0.82-1.59)
	Independent	74	72.5	119	76.8	193	75.1
Functional status-IADL (Lawton & Brody)	Dependent	90	88.2	110	71.0	200	22.2	.001	2.14 (1.26-3.61)
	Independent	12	11.8	45	29.0	57	77.8
QoL (SF-36)
Domains
Physical role functioning	Worse	52	51.0	40	25.8	92	35.8	<.001	1.86 (1.39-2.50)
	Better	50	49.0	115	74.2	165	64.2
Physical functioning	Worse	51	50.0	41	26.5	92	35.8	<.001	1.79 (1.34-2.40)
	Better	51	50.0	114	73.5	165	64.2
Pain	Worse	71	69.6	133	85.8	204	79.4	.002	0.59 (0.44-0.80)
	Better	31	30.4	22	14.2	53	20.6
General health perceptions	Worse	100	98.0	152	98.1	252	98.1	1.000	0.99 (0.33-2.93)
	Better	2	2.0	3	1.9	5	1.9
Vitality	Worse	76	74.5	121	78.1	197	76.7	.510	0.89 (0.63-1.25)
	Better	26	25.5	34	21.9	60	23.3
Social role functioning	Worse	72	70.6	116	74.8	188	73.2	.452	0.88 (0.64-1.22)
	Better	30	29.4	39	25.2	69	26.8
Emotional role functioning	Worse	63	61.8	103	66.5	166	64.6	.442	0.87 (0.65-1.20)
	Better	39	38.2	52	33.5	91	35.4
Mental health	Worse	48	47.1	83	53.5	131	51.0	.309	0.85 (0.63-1.16)
	Better	54	52.9	72	46.5	126	49.0
Total score	Worse	75	73.5	94	60.6	169	65.8	.033	1.45 (1.01-2.10)
	Better	27	26.5	61	39.4	88	34.2
Dimensions
Physical health	Worse	75	73.5	85	54.8	160	62.3	.002	1.68 (1.17-2.41)
	Better	27	26.5	70	45.2	97	37.7
Mental health	Worse	89	87.3	142	91.6	231	89.9	.257	0.77 (0.51-1.17)
	Better	13	12.7	13	8.4	26	10.1

Abbreviations: BADL, basic activities of daily living; Better QoL, individuals with SF-36 score > 50.0 points; 95% CI, 95% confidence interval; EFS, Edmonton Frail Scale; FRS, fall risk score; IADL, instrumental activities of daily living; MMSE, mini-mental state examination; MNA, mini nutritional assessment; OR, odds ratio for cohort “with risk”; QoL, quality of life; SF-36, medical outcomes study questionnaire short form; Worse QoL, individuals with SF-36 score ≤ 50.0 points.

^aPearson’s Chi-Square Test.

The association analysis of scalar variables for the evaluated aspects between the study groups is described in Table 3. Nutritional status, cognition, frailty, functional status for ADLs, and IADLs showed better scores in the Without Risk group.

Table 3.

Scalar Analysis of Multidimensional Aspects by Fall Risk (n = 257).

Multidimensional aspects	Fall risk (FRS)
	With risk (n = 102)		Without risk (n = 155)		P a
	Percentile	Mean (SD)	Percentile	Mean (SD)
	25-50-75		25-50-75
Nutritional status (MNA)	21.9-24.7-27.0	24.3 (3.4)	24.5-27.0-28.0	26.2 (2.6)	<.001
Cognition (MMSE)	16.7-21.0-27.0	21.0 (7.7)	21.0-25.0-29.0	24.7 (5.4)	<.001
Frailty (EFS)	4.0-6.0-8.0	6.1 (2.9)	2.0-4.0-5.0	3.9 (2.6)	<.001
Functional status—BADL (Barthel index)	90.0-95.0-100.0	89.4 (18.6)	95.0-100.0-100.0	96.6 (8.5)	<.001
Functional status—IADL (Lawton & Brody)	11.0-15.0-19.0	14.8 (4.2)	15.0-19.0-21.0	17.6 (3.5)	<.001
QoL (SF-36)
Domains
Physical role functioning	15.0-50.0-85.0	50.1 (33.9)	50.0-80.0-95.0	70.9 (28.9)	<.001
Physical functioning	0.0-50.0-100.0	54.4 (45.4)	50.0-100.0-100.0	75.2 (38.5)	<.001
Pain	10.0-30.0-60.0	35.7 (26.5)	0.0-20.0-40.0	24.8 (23.8)	.001
General health perceptions	15.0-20.0-31.2	23.6 (13.1)	10.0-15.0-25.0	18.5 (13.0)	.002
Vitality	40.0-45.0-51.2	46.7 (11.5)	40.0-45.0-50.0	43.9 (12.4)	.025
Social role functioning	38.0-50.0-50.0	49.5 (18.6)	50.0-50.0-50.0	48.4 (11.4)	.952
Emotional role functioning	40.0-48.0-56.0	46.9 (10.6)	44.0-48.0-52.0	47.1 (9.8)	.881
Mental health	44.0-52.0-56.0	50.1 (11.4)	44.0-48.0-56.0	49.1 (11.2)	.341
Total Score	39.7-46.0-51.0	44.7 (8.8)	44.0-50.0-52.0	47.2 (7.5)	.008
Dimensions
Physical health	35.0-43.0-51.0	42.1 (11.7)	42.0-50.0-53.0	46.6 (10.1)	.001
Mental health	39.0-44.0-47.2	43.4 (7.7)	38.0-41.0-46.0	41.3 (6.8)	.023

Abbreviations: BADL, basic activities of daily living; EFS, Edmonton Frail Scale; FRS, fall risk score; IADL, instrumental activities of daily living; MMSE, mini-mental state examination; MNA, mini nutritional assessment; QoL, quality of life; SD, standard deviation; SF-36, Medical Outcomes Study Questionnaire Short Form.

^aMann Whitney U test.

In the intergroup comparison of QoL, the domains Physical Role Functioning, Physical Functioning, Total Score, and the Physical Health Dimension exhibited higher scores in the Without Risk group. Conversely, the With Risk group exhibited higher scores in the Pain domain, General Health Perceptions, Vitality, and the Mental Health Dimension.

When analyzing the correlation between multidimensional aspects and fall risk (Table 4), we found a moderate correlation with nutritional status, frailty, functionality assessed through IADLs, and the Physical Role Functioning domain of QoL. The direction of these scales (negative or positive) indicated that higher fall risk was associated with poorer evaluations in these respective aspects.

Table 4.

Correlation Between Fall Risk Scores and Multidimensional Aspects Adjusted for Potential Confounding Factors.

Multidimensional aspects	Crude Fall risk (FRS) (n = 257)		Adjusted Fall Risk (FRS) By potential confounding variables
			By age range 60-79 year (n = 185)		By absence of fall history (n = 98)		By absence of polypharmacy (n = 217)		By presence of chronic diseases (n = 206)
	r a	P b	r a	P b	r a	P b	r a	P b	r a	P b
Nutritional status (MNA)	−.31	<.001	−.30	<.001	−.13	.210	−.25	<.001	−.27	<.001
Cognition (MMSE)	−.27	<.001	−.16	.030	−.20	.053	−.30	<.001	−.30	<.001
Frailty (EFS)	.39	<.001	.35	<.001	.28	.006	.34	<.001	.35	<.001
Functional status—BADL (Barthel index)	−.28	<.001	−.24	.001	.02	.827	−.27	<.001	−.29	<.001
Functional status—IADL (Lawton & Brody)	−.40	<.001	−.31	<.001	−.34	.001	−.42	<.001	−.34	<.001
QoL (SF-36)
Domains
Physical role functioning	−.37	<.001	−.36	<.001	−.22	.027	−.36	<.001	−.30	<.001
Physical functioning	−.23	<.001	−.29	<.001	.04	.723	−.17	.013	−.20	.003
Pain	.26	<.001	.29	<.001	.03	.792	.23	.001	.19	.005
General health perceptions	.24	<.001	.30	<.001	.26	.010	.24	<.001	.17	.014
Vitality	.14	.024	.11	.131	.03	.800	.14	.046	.15	.026
Social role functioning	.04	.504	.06	.448	.04	.705	.13	.046	.04	.609
Emotional role functioning	−.02	.792	.02	.738	.08	.428	.01	.858	.04	.563
Mental health	.04	.239	.11	.140	.12	.235	.11	.103	.07	.318
Total Score	.13	.036	−.11	.128	.03	.771	−.07	.292	−.10	.162
Dimensions
Physical health	−.20	.002	−.17	.017	−.02	.882	−.16	.021	−.17	.016
Mental health	.19	.002	.22	.003	.21	.042	.25	<.001	.16	.023

Abbreviations: FRS, fall risk score; MNA, mini nutritional assessment; MMSE, mini-mental state examination; EFS, Edmonton Frail Scale; BADL, basic activities of daily living; IADL, instrumental activities of daily living; QoL, quality of life; SF-36, medical outcomes study questionnaire short form.

Correlation levels: r < .29 (weak); .29 ≥ r < .49 (moderate); r > .50 (strong).

^aCorrelation coefficient.

^bSpearman’s correlation test.

The adjusted analyses for sociodemographic and health variables indicated that the 60 to 79 years age group increased the correlation strength between fall risk and the Physical Functioning, Pain, and General Health Perceptions domains of QoL. A similar pattern was observed with the absence of polypharmacy, which strengthened the correlation with cognition (MMSE). Likewise, after adjusting for the presence of chronic diseases, an increased correlation strength was noted with functionality assessed through BADL (Barthel Index).

Finally, we implemented a binary logistic regression model incorporating the independent variables to assess their representativeness in the outcome variable (Table 5). Most aspects exhibited an odds ratio (OR) >1.00 (95% CI), with notable contributions from nutritional status, cognition, basic activities of daily living (BADL), instrumental activities of daily living (IADL), and the physical role functioning domain of QoL.

Table 5.

Binary Logistic Regression Between Fall Risk and Evaluated Multidimensional Aspects.

Multidimensional aspects assessed and adjustments for confounding factors—Crude	Binary logistic regression—fall risk (FRS)
	R 2a	P b	β c (S.E)	Wald	OR (95% CI)
Nutritional status (MNA)	0.12	<.001	0.21 (0.05)	21.27	1.24 (1.13-1.35)
Frailty (EFS)	0.17	<.001	-0.28 (0.05)	29.70	0.76 (0.69-0.84)
Cognition (MMSE)	0.10	<.001	0.09 (0.02)	16.68	1.09 (1.05-1.14)
Functional status—BADL (Barthel index)	0.09	.001	0.05 (0.01)	11.70	1.05 (1.02-1.08)
Functional status—IADL (lawton & Brody)	0.15	<.001	0.18 (0.03)	26.81	1.20 (1.12-1.28)
QoL (SF-36)—Domains
Physical role functioning	0.13	<.001	0.02 (0.0)	23.45	1.02 (1.01-1.03)
Physical functioning	0.07	<.001	0.01 (0.00)	14.23	1.01 (1.00-1.02)
Pain—Crude	0.06	.001	-0.02 (0.00)	11.00	0.98 (0.97-0.99)
General health perceptions	0.05	.03	-0.03 (0.01)	8.58	0.97 (0.95-0.99)
Vitality	— d	—	—	—	—
Social role functioning	—	—	—	—	—
Emotional role functioning	—	—	—	—	—
Mental health	—	—	—	—	—
Total score	0.03	0.018	0.04 (0.02)	5.56	1.04 (1.01-1.07)
Dimensions
Physical health	0.05	0.002	0.04 (0.01)	9.86	1.04 (1.01-1.06)
Mental health	0.03	0.028	-0.04 (0.02)	4.80	0.96 (0.92-1.00)

Abbreviations: BADL, basic activities of daily living; 95% CI, 95% confidence interval; EFS, Edmonton Frail Scale; FRS, fall risk score; IADL, instrumental activities of daily living; MMSE, mini-mental state examination; MNA, mini nutritional assessment; OR, odds ratio; QoL, Quality of life; SF-36, Medical Outcomes Study Questionnaire Short Form.

^aR² de Nagelkerke.

^bModel (Forward LR).

^cUnstandardized coefficient.

^dThe variable was not eligible for the regression model and, therefore, was not calculated.

A similar analysis, adjusted for potential confounding factors, was conducted. No substantial variations were observed in the independent variables when exposed to adjustment variables. A detailed analysis can be found in the Supplemental Material of this study (Supplemental S1 Table).

Discussion

Our results highlighted significant associations between fall risk and multidimensional aspects. Among these, altered nutritional status, frailty, and low functionality were the most strongly associated with fall risk. We believe that these findings provide a detailed and in-depth assessment of factors that may influence the risk and occurrence of falls in community-dwelling older adults. By comparing the behavior of different factors, this study highlights which one exerts the greatest influence. Much of the literature attributes falls in this population exclusively to physical, functional, and environmental restrictions.^26,27 Our innovative focus in this research was on additional aspects that may contribute to this risk, as physical and functional states are often more impaired in older adults with nutritional and cognitive deficits, as well as social and economic frailties.^{28 -30}

Regarding the sociodemographic data, younger older adults (60-79 years) were predominant in both groups, with a greater presence among those without fall risk. Other Brazilian studies have found that older age, female gender, non-white skin color, low income, and low educational levels are associated with a higher prevalence of falls.^28,29 In our data, older age (≥80 years) showed statistical significance in the crude analyses between the study groups. However, when conducting the analysis with age adjustment, no significant change was observed in the fall risk results or the evaluated multidimensional aspects.

Health-related aspects, such as fall history and the presence of self-reported chronic diseases, were associated with fall risk. A Chinese study identified that a greater number of chronic diseases, smoking, alcohol consumption, and physical inactivity increased fall risk among older adults. ³⁰ Polypharmacy was uncommon in both groups but appeared more relevant among participants at risk of falling. Regarding the studied health aspects, the theoretical framework we adopted highlights fall history and medication use as key determinants of fall risk.^13,14 However, despite the associations found, the OR did not indicate a determinant potential for the presence of chronic diseases or polypharmacy in fall risk or the multidimensional aspects within our sample. Additionally, these variables did not show substantial changes in the controlled analyses.

The association and correlation analysis between fall risk and multidimensional aspects emphasized that better nutrition was associated with lower fall risk. This was also observed in a study conducted with older Chinese adults^31,32 and another study involving low-income Thai living in the community. ³³ The literature explains that one of the key nutritional factors influencing fall risk is the loss of muscle mass, which is further aggravated by the decline in nutrient and electrolyte absorption, a common occurrence in the natural aging process. ⁶ As a strategy to mitigate this type of decline, a randomized study conducted in New Zealand found that dietary changes involving calcium and protein intake through dairy foods over 3 months reduced fracture risk by 33% and fall occurrence by 11%. ³⁴

Individuals with preserved cognition were also less prone to falls, as observed in both categorical and scalar analyses. Although the crude correlation analysis was weak, the adjustment for age (≥80 years), history of previous falls, and presence of chronic diseases appeared to mediate this strength, making it moderate. Even so, the binary logistic regression did not indicate that this resulted in substantial changes. Previous research has reported similar findings, indicating that cognition should be monitored as an important factor in the occurrence of falls, with proposed preventive interventions. ³⁵ Authors explain that cognition can be improved through increased physical activity, as it helps maintain hippocampal volume, a brain region responsible for processing various memory mechanisms, thereby preventing or reducing the progression of neuropsychiatric disorders. ³⁶ However, a systematic review evaluating various guidelines on fall prevention and management noted a lack of studies on this aspect, despite strong recommendations for cognitive screening. ⁸

In addition to the association analyses showing higher levels of frailty in the fall risk group, the FRS demonstrated a moderate correlation in the same direction. Although previous studies have already demonstrated the association between frailty and fall risk, ³⁷ it is noteworthy that the assessment tool we adopted (EFS) also incorporates nutritional and functional status as key contributors to frailty. Regarding functionality specifically, the assessment includes questions about the individual’s ability to perform certain tasks, as well as direct observation of functional performance by evaluating their ability to walk a distance of 6 meters.^19,20 Authors explain that frail older adults have lower energy and physical reserves, as well as reduced physiological stress tolerance, which directly impacts their ability to safely perform daily activities. ⁷ A literature review highlighted that gait parameters are key determinants of frailty in the context of fall risk. The most relevant factors were gait speed, time, number of steps, and percentage of walking per day. ³⁸

Still on functionality, our findings indicate that functional assessment using ADLs (Barthel Index) and, more notably, IADLs (Lawton & Brody) highlights independence in activities requiring functional ability as a protective factor against fall risk. A similar outcome was reported in a population-based study of older Indian adults. ³⁹ To maintain functionality, muscle-strengthening exercises and regular physical activity are widely recognized as effective strategies, particularly for fall prevention. ⁴⁰ However, activities involving movement or walking in older adults paradoxically elevate fall risk. Researchers addressing this concern have investigated the link between fall risk and physically demanding exercises, finding that, although the risk increases during execution, the overall improvements in functional capacity are substantial and ultimately outweigh the risk of falls. ⁴¹

The association between QoL and fall risk was also evident, particularly in domains related to physical and functional aspects, supporting our more specific findings on ADLs and IADLs. Thus, individuals with higher fall risk displayed poorer QoL. However, correlation analysis of the total score (SF-36) revealed an apparently paradoxical finding that better QoL was associated with higher fall risk. Although we found a weak correlation in this analysis, it suggests that better physical and functional states among those without fall risk did not influence their overall perception of QoL. A possible explanation for these results is that these aspects may behave in a specific and distinct manner. Pain, for instance, can instill a sense of insecurity in individuals when mobilizing, leading them to avoid situations that might result in a fall, thereby acting as a protective factor. However, since it is a symptom that limits many movements, older adults often use analgesics for relief. Among the most used are opioids, which individually can also interfere with fall risk, as they may cause drowsiness and hypotension, for example. ⁴²

Similarly, the general perception of health might reflect an unspecific view of how an individual feels, without necessarily being related to fall risk. Vitality, representing the energy available for various activities, may have been perceived independently of fall risk. Likewise, the mental health dimension seems to have remained unaffected by this risk, as the Mental Health domain showed no association.

Other authors have noted that QoL can also be influenced by various factors not assessed by the SF-36, such as nutritional deficiencies. ⁴³ Another Brazilian study implemented interventions that improved QoL in older adults through various strategies, including social stimulation, better eating habits, technology interaction, cognition, and depression management. ⁴⁴

Regarding interventions aimed at reducing fall risk, innovative strategies have shown promising outcomes. A study conducted with older adults residing in Iranian nursing homes found that virtual reality-guided exercise interventions significantly improved balance and drastically reduced the fear of falling compared to a group receiving routine exercises such as walking, table tennis, and artistic activities. ⁴⁵ Another study demonstrated that a series of home visits focusing on furniture rearrangement, obstacle reduction, safe footwear, and strategies to eliminate slippery floors was decisive in reducing fall risk. ⁴⁶ Awareness and educational strategies on risk factors, although commonly discussed in the literature, continue to show success when implemented by professionals such as PHC nurses. ⁴⁷ Additionally, a systematic review highlighted that multifactorial interventions—such as exercise programs, assistive technology, assessments, environmental modifications, and improvements in basic aspects like medication review—have demonstrated greater success compared to single-approach strategies. ⁴⁸

Overall, our findings provided a broad and detailed analysis of fall risk in older adults. Focusing on nutrition, cognition, frailty, functionality, and QoL enables further reflection on underlying problems that may explain the alterations we observed. These findings may support the development of specific protocols for monitoring nutritional status, cognition, and functionality, highlighting the need to incorporate additional professional categories such as physiotherapists, nutritionists, and medical specialists into PHC. Currently, these resources are not widely available in Brazil or in many other settings. Such an approach could yield significant benefits by enhancing the capacity for resolving health issues and preventing falls and other complications at this level of healthcare delivery. The literature attributes fall risk to various aspects we did not evaluate in our research, such as medications causing drowsiness, impaired vision, and environmental factors like slippery floors, excessive obstacles, and poor lighting. ⁴⁹ However, our results highlight the need for a broader assessment of key aspects related to fall risk, which are often addressed individually—such as nutrition, cognition, frailty, and QoL—but are rarely included in a standardized evaluation within Brazilian primary healthcare (APS) and other settings. Therefore, studies on this topic require further exploration to comprehensively measure the multidimensional aspects involved in fall occurrence among older adults.

Our study has some limitations that may have influenced the results. The cross-sectional design prevents determining causal relationships between multidimensional factors and fall risk, as it is not possible to control for potential confounding factors that may change over time, such as health status, age, and nutritional status. The selection of participants may have influenced the results, as it was based on the researchers’ convenience, without management or randomization in the formation of groups according to their health status. Consequently, less communicative individuals or those with communication difficulties may not have been included in the sample. Similarly, healthier participants may be overrepresented in the study, and this may have resulted in a low comparative power between the formed groups. Another important point to consider is the use of dichotomous variables in many analyses, which may simplify the nuances present in the evaluations.

The final sample size, smaller than initially estimated, may have compromised the statistical power of some analyses, particularly because increased data variability makes sample parameter estimates less precise and weakens the adopted confidence interval. As a result, the findings have a reduced likelihood of detecting more detailed outcomes. Therefore, the generalization of our results should be made with caution, as it is not possible to establish a cause-and-effect relationship between the evaluated aspects and fall risk. This is justified both by the limitations imposed by cross-sectional design and by the characteristics of the studied population, which resides in a small town with unique cultural traits. The sample was obtained from a specific region of Brazil using a non-random selection method. This also limits the external validity of our findings. Therefore, our results can be compared with populations that share a similar sociodemographic and cultural profile, but not with populations from large urban centers or those living in significantly different conditions.

Conclusion

This study demonstrated that better nutritional status, cognition, frailty, functionality, and quality of life (QoL) outcomes were associated with a lower risk of falls in the sample studied, as evidenced by their levels of association and correlation. The findings also support the initial hypothesis of this research, highlighting those higher scores in nutrition, cognition, and functionality stood out compared to other evaluated aspects. Potential confounding factors were ruled out through adjusted analyses.

As its main innovative contribution, our findings can be compared to other PHC settings with similar sociodemographic characteristics and may suggest tools for more effective prevention. For example, populations similar to the one we studied could benefit from new protocols aimed at improving nutrition, enhancing cognitive stimulation, and promoting physical and functional fitness as strategies to reduce fall risk. Additionally, we suggest the development of more consistent interventions to improve these aspects through the implementation of a treatment plan.

Intervention studies aimed at improving nutrition, maintaining functional independence, and monitoring frailty and cognition may be effective strategies for evaluating their success in preventing falls and enhancing overall health status. Although these findings have significant implications for clinical practice, further longitudinal studies are needed to explore causal relationships and expand the applicability of these results to other older adult populations.