Functional status in relation to depression among elderly individuals in Indonesia: a cross-sectional analysis of the Indonesian National Health Survey 2018 among elderly individuals

The data used were obtained from the NHS and National Socio-Economic Survey (Susenas) in 2018, which are national surveys routinely carried out by the Ministry of Health (National Institute of Health Research and Development) and Central Bureau of Statistics [16]. Furthermore, samples with national representation, even at the district/city level, were used. These two surveys were carried out in all provinces in Indonesia (34 provinces), with approximately 300,000 selected households spread across 514 districts/cities.

The NHS sampling frame was obtained in a multistage manner. The first and second stages determine census blocks, namely, selecting 30,000 census blocks from 25% of the master frame blocks of the most recent population census (in 2010). The selection was carried out by systematic probability proportional to size (PPS) random sampling in every urban and rural strata. The second stage selects 10 households from each CB that have been updated by systematic sampling with implicit stratification based on the highest education level of the head of the household to maintain representativeness among household characteristics. This approach is consistent with a previous article on the determinants of prediabetes in Indonesia, and the difference in the sampling frame for prediabetes is only in 26 provinces [17]. In the 30,000 census block, 10 households were identified, and all members of the selected households were asked to participate in this survey voluntarily [16].

The inclusion criteria included respondents ≥60 years of age, and answers personally provided not by a representative but by the closest individual in the house. Subjects were excluded if they are unable to communicate or unable to speak Indonesian. In principle, the subjects have to be able to communicate, the closest person can only help in terms of finding the right word equivalent. The elderly individuals fully represented by someone were still interviewed but in the questionnaire were coded as represent and would not be interviewed for depression questions.

Depression was assessed according to a section in the Mini International Neuropsychiatric Interview (MINI) version 6. This version refers to the International Classification of Diseases, tenth revision (ICD-10), which is still used in the disease coding system in Indonesia. Furthermore, the questionnaire used was developed in 1997 [18]. According to MINI version 6, subjects were categorized as depressed if they answered at least 2 “yes” from the first 3 questions plus 2 “yes” to questions 4 to 10. The MINI is a recognized instrument for assessing depression. Before being applied to the NHS 2018, an assessment of its validity and reliability was carried out, and good results were obtained [19]. Functional status was assessed by using the Barthel index, which is a measurement tool developed by Mahoney and Barthel [20] in 1965. This tool is widely used to assess the functional status of patients after stroke, colorectal cancer, fractures and several other medical and even social conditions in nursing homes and the general population [5, 6, 21‐26]. A total of 10 questions were used to assess the respondents’ ability to perform their daily activities, including determining how much care and assistance was needed, and this assessment was carried out over a month. The criteria used included a score >/= 20 for the independent category, a score of 12-19 for the mild dependence category, a score of 9-12 for the moderate dependency category, a score of 5-8 for the severe dependence category and a score of 0-4 for the total dependence category. This stratification is based on the modified Barthel index [27]. Barthel index has been assessed for its validity and reliability in Indonesia [28].

Depression and functional status were assessed via face-to-face interviews by trained NHS enumerators. Interviews were conducted in the respondent’s house. Only elderly individuals residing at home were assessed; those living in nursing homes and those currently in hospitals were not included. The enumerators interviewed the respondents using a structured questionnaire in which there were questions, sociodemographic characteristics, history of chronic disease, depression and functional status. The Riskesdas report including questionnaire can be seen in Appendix 1.

Socioeconomic status was obtained from Susenas, which contains data describing household expenses. Susenas divides the socioeconomic status of Indonesian households into 5 groups. The expenditure quintile is the grouping of expenditures into five equal groups after being sorted from the smallest to the largest expenditure. The first quintile represents the poorest group, while the fifth quintile represents the richest group. Information about chronic diseases was obtained by via interviews. Participants were asked whether they suffer from hypertension, heart disease, or stroke; whether they have joints injuries confirmed by a doctor; and whether they suffered an injury that has interfered with daily activities in the last year.

The age variable was grouped into 5-year intervals, while the other variables included sex (male or female), activity (not working or working/retired), marital status (married, divorced or unmarried), education (graduated from diploma1/diploma2/diploma3/higher education, no school, did not complete elementary school, graduated from junior high school or graduated from senior high school), residence (rural or urban) and economic status (upper, upper-middle, middle, lower-middle or low class). Furthermore, all disease histories were considered a nominal variable type (yes or no), which included diabetes mellitus (DM), heart disease, stroke, hypertension, joint disorders/rheumatism, and injury.

The prevalence of depression was calculated by considering the weighted value of the population and analyzed using the SPSS 24 (IBM SPSS Statistics) tool with a complex sample method. Furthermore, the relationship between the main independent variable, namely, functional disorders, was assessed by using the Barthel index, and the depression outcome variable was assessed by using the chi-square test. The Barthel index was an ordinal variable, while depression was a nominal variable (yes or no). In addition, the relationships of other independent variables to depression in the early stages were also assessed using the chi-square test. Moreover, the relationships between the depression variable and other independent variables were assessed using multiple logistic regression analysis. In the final stage, an assessment of the interactions and/or confounders between the main and independent variables was carried out.

A provincial map of Indonesia is displayed to show the provinces that have a high incidence of elderly individuals with severe dependency using the spmap command from STATA15 version (Stata Corp LLC).

Ethical approval was obtained from the Research Ethics Commission of the NIHRD of Ministry of Health number LB 02.01/3/KE024/2018. All methods were carried out in accordance with the Helsinki Declaration. Subject participation was voluntary, and written informed consent was obtained from all participants prior to the interview.

The response rate of CB visited, household visited, household interviewed was 99.41, 98.57 and 95.58%, respectively. The response rate of interviewees is 93.20%. The initial sample comprised 97,407 respondents, but after the selection process was carried out according to the study criteria, the sample size was 93,829. A total of 3578 respondents were not analysed because they were not present at the time of the survey or refused to be interviewed. See Fig. 1.

Of the 93,829 respondents, 7491 or 7.7% (95% CI = 7.4-8.0) experienced depression, including 2829 males and 4662 females (two-thirds). As shown in Table 1, the marital status of the depressed respondents included over 4153 married, 3218 divorced (former spouses alive or dead) and 120 not married. The youngest respondent was 60 years old, and the average age was 68.48 years. Regarding residence area, the proportion of individuals residing in urban and rural areas was almost equal.

In terms of age range, the older the respondents were, the greater the proportion with depression. The highest proportion was within the age range of 90 years and above (12%). Regarding their activities, the proportion of individuals not working was greater than that of individuals working (Table 1). Moreover, the elderly individuals who were divorced, with their former spouses either alive or dead, comprised the largest proportion of depressed individuals compared to those still married to their spouse (husband/wife) or not married.

The proportion of elderly individuals who experience depression may increase with lower levels of education completed. It was discovered that depression is most common among those with no educational qualifications (8.6%). Regarding socioeconomic status, elderly individuals with a lower socioeconomic status comprised the largest proportion (8.8%) of depressed individuals compared to those with a higher socioeconomic status.

In the results section, the proportion of yes answers to the depression questions is also shown in Table 2.

The questions number 1 to 3 are filters therefore we focus on numbers 4 to 10. It shows the most symptoms that respondents experienced were sleep disturbances, disturbed appetite and limited movement.

Table 3 shows that the proportion of elderly individuals with an independent functional status comprised the smallest proportion of depressed (5.5%) individuals compared to those with functional disorders. In terms of functional disorders, elderly individuals with severe dependence represented the highest proportion of depression (36.3%). Furthermore, those experiencing chronic diseases such as diabetes mellitus, heart disease, stroke, hypertension, joint disorders/rheumatism and injuries comprised a greater proportion of patients with depression than those not experiencing such diseases.

Bivariate analysis with depression as the dependent variable, dependency as the independent variable and all covariates was carried out. The results show that all covariates were eligible for analysis in the multivariate stage because they were significant with p values < 0.25. In multivariate analysis, two covariates were not significant (p value > 0.05); therefore, they were excluded from the analysis. The final results are shown in Table 4.

In the multivariate analysis, it was observed that elderly individuals with severe dependence were more likely to be depressed than those without functional/independent disorders after adjusting for sociodemographic factors and disease history. Additionally, the area of residence and marital status did not have a statistically significant effect on the relationship between functional disorders and depression.

This result was then analyzed for confounding variables, but in fact, none of the variables were not confounding factors.

Figure 2 shows the provinces with elderly people with severe dependence on a map of Indonesia consisting of 34 provinces. On the map, several provinces need attention from the local government because they have a high rate of elderly individuals with severe dependence. The provinces in the darkest red color are North Sumatra, Riau, Jambi, West Kalimantan, North Sulawesi, Central Sulawesi, South Sulawesi, North Maluku and Maluku.

This is a national study with a large number of respondents covering all areas, including rural and urban areas; therefore, it has good representation. The limitation of this study was cross sectional design that can not explain real relationship of the variables. Another limitation is the fact that this study utilized only data available on from the NHS; therefore, several variables that can affect the output are not taken into account. The variables such family support, social participation, physical activity and other diseases or symptoms for example pain, cognitive impairment, peptic ulcer, migrain, asthma bronchiale .