Hierarchical Model of Factors Associated with Falls in Older Brazilian Community-Dwelling Women

Keywords

Clavicle; Congenital pseudoarthrosis; Mandibular plate

Abstract

Objective: To estimate the prevalence of falls in a group of older women and to measure the influence of risk factors associated with age.

Methods: Longitudinal study with a representative probability sample of the AGEQOL study (Aging, Gender and Quality of Life). This article is based on 1226 older Brazilian community-dwelling women. Participants were interviewed on falls in past 12 months, demographic and socioeconomic characteristics, health status, functional ability and access to and use of health services. Poisson regression was used to confirm the association of decline in women with possible determinants, separated by age (60-74 years and ≥75 years).

Results: Overall, 250 women (54.2%) had a single fall, and the prevalence of falls was significantly different between age groups (p<0.001). Women aged less than 75 years old who smoked, drank, and reported nausea and imbalance had a higher prevalence of falls. Among the oldest women, a dose-response relationship was present between falls and functional capacity of ADL.

Conclusions: The prevalence of falls differed in each age group of women. For older women aged 60-74 years, the prevalence of falls was associated with self-reported health status and the type of health services used. In addition to performing ADL, worse health conditions, surgeries, and higher education were risk factors associated with a higher prevalence of falls in older women.

Introduction

Population aging is an important global phenomenon in public health. This new population profile requires the adoption of public and social solutions to develop policies [1]. In Brazil, there are approximately 25 million people aged 60 or older (10.8% of the population). Projections for 2030, estimate a life expectancy in Brazil around 77.4 years. In 2050 almost 30% of the Brazilian population will be 60 years and older, placing Brazil as one of the countries with the largest absolute number of older people worldwide [2].

However, ageing in Latin American countries is occurring even in the context of health inequalities, with high rates of poverty, relative low coverage and quality of health and pension systems [3]. In the older adults, health cannot be measured by the presence or absence of disease but by the degree of preservation of functional capacity (i.e., the ability to take care of oneself and to determine and perform Activities of Daily Living (ADL) with autonomy and independence, even with morbidities) [3]. Falls and fractures are major causes of morbidity and mortality in older adults [4-6]. Epidemiological studies show that approximately 30% of the population over the age of 65 suffer at least one fall per year [7,8]. This rate rises to 50% in people over 80 years old and those who are institutionalized [9].

The multifactorial etiology factors associated with falls, and recurring injuries, determined have been widely studied and reported in the literature [10-12]. Falling is not a normal event in the aging process. In the present, evidence-based interventions can be used in clinical or community settings to reduce fall risk [13,14]. Additionally, secondary prevention programs for falls and fractures are highly needed by identifying risks, performing environmental reorganization, and identifying functional rehabilitation factors [4,15].

In Brazil, most studies are epidemiological, observational, and few studies involved significant and representative sample of elderly living in communities. Risk factors for falls need to be further investigated so that preventive actions and public policies are proposed in the near future. Age and female gender are major risk factors for falls in the older adults [16], but they are not the only ones. We believe that association between falls and functional limitation, influenced by other factors is even more significant among older women. This study aimed to estimate the prevalence of falls in a group of older women and to examine the influence of factors associated with age.

Materials and Methods

Participants

 “Aging, Gender and Quality of Life (AGEQOL)” is a cohort study in Sete Lagoas, Minas Gerais and Brazil with a representative sample of 1226 community-dwelling women, between the ages of 60 and 106 years (mean age of 71.03 ± 8.35 years). A complex sampling design was adopted for this study and consisted of a combination of probabilistic sampling methods for selecting a representative sample of the population. For this sampling, the following two calculations were performed: an estimation of the number of older adults and an estimation of the number of households to be visited. The sampling process was conducted in two stages: in the first, census tracts were selected and in the second, households within each sector were selected.

In each household, all residents aged 60 years or more of both genders, regardless of your marital status or kinship were interviewed. Data collection was conducted in the homes of the older adults between January and July 2012 and involved household interviews and examinations conducted by three examiners and three annotators. All persons 60+ years in the selected households were informed of the study and were asked to sign an informed consent form that had been previously approved by the Ethical Committee of the Federal University of Minas Gerais (CAAE-0413.0.203.000-11).

The interviews lasted 40 to 60 minutes. At the end of the interviews, each subject in the city received guidance regarding health care and activity options as well as the personal contact information of the researcher responsible for the questionnaire. Additional details of the methodology are described in Campos et al. [17].

Measurements

The definition of a fall used in this study was “an unintentional event that results in the change of position of the individual to a lower level relative to its initial position, with inability to fix in a timely manner” [10]. The dependent variable was a dichotomous question about recent falls: “Last year, (a) Mr. (a) has fallen?” The group that did not report falls (falls = no) was defined as the reference category. For women who responded positively, we also asked about the recurrence (more than one fall) and the occurrence of bone fracture resulting from the fall. The independent variables were grouped according to an adapted form of the hierarchical model proposed by Cruz et al. [18], in which the levels of assessment of the relationships between groups of variables include distal, intermediate, and proximal factors (Figure 1).

Figure 1: Hierarchical theoretical model for factors associated with falls in older women living in the community. ADL: Basic Activities of Daily Living IADL: Instrumental Activities of Daily Living

The block was formed by the following distal variables of socioeconomic information: marital status (married, single/widowed/ divorced), self-reported color (white, black, other), retirement (yes, no), education (illiterate, literate), median income (≤US$300, >US$300), caregiver (yes, no), living arrangement (live alone, live together), time of residence in the same home in years (≤ 24 years, >24 years), and household (alone, relatives/other). The health of the women was assessed by subjective aspects (selfreported conditions and self-reported health) and goals (diagnostic tests and screening).

The self-reported conditions investigated were the following: smoking (yes, no), alcohol (yes, no), trouble sleeping at night (yes, no), physical activity (yes, no), use of corrective lenses (yes, no), use of hearing aids (yes, no), medication use (yes, no), use of controlled medication (yes, no), health perception (very poor/ poor/fair, good/very good), self-assessment of weight (underweight, normal, overweight), self-reported diagnosis of arthritis (yes, no), and self-reported diagnosis of osteoporosis (yes, no).

Common symptoms related to aging were investigated by self-report, such as tremors (yes, no), dizziness (yes, no), imbalance (yes, no), fainting (yes, no), tinnitus (yes, no), and difficulty walking (yes, no). To measure cognitive ability, the Mini Mental State Examination (MMSE) validated in Portuguese [19] was used. This instrument consists of questions grouped into seven categories, each designed with the objective of assessing orientation, immediate memory and recall, concentration, calculation, and language.

The Geriatric Depression Scale short validated in Portuguese version with 15 items (GDS-15) is specifically targeted toward women and was used to assess mood and level of depression [20]. The scores of each instrument were dichotomized to evaluate cognitive impairment (yes, no) and depression (yes, no). To measure access to and use of health services, we used the following variables: medical treatment (yes, no), medical consultation in the last 6 months (yes, no), health insurance (yes, no), type of health service most used (public/plan/private), hospitalization in the past year (yes, no), and surgery in the past year (yes, no).

In this study, functional ability was analyzed separately by two instruments. The Katz index is an instrument used to measure independence in the performance of the following six ADL: bathing, dressing, toileting, transferring, continence, and feeding [21]. The Lawton - Brody Index (ILB) evaluates the patient’s ability to perform instrumental activities of daily living (IADL) [22]. ADL and IADL dependence was classified as total dependence, moderate dependence, or independence.

Data Analysis

All analyses were performed separately by age group using a cutoff defined by the World Health Organization (WHO) for developing countries [23]. The sample was divided into two groups: 60-74 years (n=842) and ≥75 years (n=384). Descriptive analyses included the calculation of the prevalence of falls, recurrence of falls, and bone fracture by age as well as association with the chi-square test. We used a significance level of 5%. In the crude analysis, the prevalence of falls was calculated for each group of independent variables.

The selection of variables that constituted the blocks in the multivariate analysis was obtained by adopting a critical level of significance equal to 0.20. Multivariate analysis was performed using Poisson regression with robust variance calculation of adjusted odds ratios, 95% confidence intervals, and significance level using the Wald test for heterogeneity and linear trend. The effect of the complex sample design was considered in all analyses. Statistical analysis was performed using the SPSS statistical software version 19.0 (SPSS Inc., Chicago, USA).

Results

Overall, 250 women (54.2%) had a single fall, and the prevalence of falls was significantly different between age groups (p<0.001). The recurrence rate was higher in women under 75 years of age (46.5%) and older women with fractures (27.3%) (Table 1).

Table 1: Distribution of women residents in the community according to the occurrence of falls, recurrence and bone fracture by age group.

In the crude analysis for the group aged 60-74 years, only lifestyle (smoking, alcohol, and use of hearing aids), health (imbalance and fainting), and the most widely used type of health service were associated with the occurrence of falls, at a significance level of 5% (Table 2).

Table 2: Crude analysis of the prevalence of falls in women 60-74 years of age, associated with demographic and socioeconomic characteristics, lifestyle, health conditions, access and use of health services, and functional capacity.

Table 3 shows the crude analysis between falls and associated factors in women aged 75 years or older. The use of hearing aids, a diagnosis of osteoporosis, and a health plan were not associated with the outcomes. However, the variables achieved enough statistical significance (p≤0.20) to be included in the adjusted model.

Table 3: Crude analysis of the prevalence of falls in women aged 75 and older associated with demographic and socioeconomic characteristics, lifestyle, health conditions, access and use of health services, and functional capacity.