Introduction

The dynamics of population aging worldwide has aroused interest in responding to the social and health needs of the elderly population due to the increased demand for health services ^1,2. According to ECLAC’s Centro Latinoamericano y Caribeño de Demografía (CELADE - Latin American and Caribbean Demographic Centre) and the World report on ageing and health released by the World Health Organization (WHO), the Latin American and Caribbean population will account for three to five times the global population between 2025 and 2050, and its population over 65 years of age will be larger than that of children under 14 ^3,4.

This process of population growth of the elderly is characterized by the increase in life expectancy, which gives rise to longer-lived populations who need more health and social care ², the generation of challenges in dealing with the elderly, as well as the creation of healthy conditions or environments ⁵. In the case of the Colombian population, according to the 2005 census conducted by Departamento Administrativo Nacional de Estadística (DANE National Administrative Department of Statistics), people over 65 years of age made up 6.5 % of the population, and an increase of 20 % is projected for 2050 ⁶.

The elderly population is characterized by a burden of chronic non-communicable diseases (NCDs) such as ischemic heart disease, cerebrovascular disease and chronic obstructive pulmonary disease ^7,8. The prevalent morbidity in the elderly includes mental disorders, as stated by the Encuesta Nacional de Salud, Bienestar y Envejecimiento (SABE - National Survey on Health, WellBeing and Aging), where 40 % of the population have hypertension and symptoms of depression, the latter represented by a burden of mental disorders secondary to the typical changes and adaptations of old age ^9-11.

Among the main phenomena in the elderly is loneliness, which is considered a multidimensional construct associated with mental health disorders. Similarly, pathologies related to the hypothalamic-pituitary-adrenal axis, such as cardiovascular risk, high blood pressure, high cholesterol, sleep disorders, migraine, altered immune function and effects on the transcription of some genes, can also be mentioned ^12,13. According to a systematic review conducted by Petitte et al., it was found that the feeling of loneliness has a prevalence of 20 % to 40 % in the elderly population ¹⁴.

Loneliness means feeling lonely, regardless of the number of social contacts ¹⁵. There are different types of loneliness and some authors identify unwanted loneliness ¹⁶, whether objective or subjective. Rubio and Aleixandre ¹⁷, as well as Yaben ¹⁸, draw a distinction between being alone and feeling lonely. Objective loneliness refers to being alone, while subjective or emotional loneliness refers to feeling lonely. For this reason, it is necessary to assess subjective loneliness using different measurement scales ⁵.

Currently, there are loneliness assessment scales, including the UCLA Loneliness Scale, the Social and Emotional Loneliness Scale for Adults (SELSA), the Emotional/Social Loneliness Inventory (ESLI) and the Philadelphia Scale ^19-21. The Spanish ESTE scale, already validated in Colombia with a Cronbach’s alpha of 0.90, consists of 30 items and measures family loneliness, marital loneliness, social loneliness and adaptation crisis. The research used selfadministered Likert-type questions. The administration of such scale within the gerontological assessment will allow identifying and preventing the feeling of loneliness ^19,20,22.

It is worth mentioning that loneliness in the elderly involves the intervention of different health disciplines which seek to promote people to be active subjects in their adaptation process during their lives, in order to contribute to reducing the burden of disease. It should be noted that loneliness has an impact not only on mental disorders but also on physical conditions ²³, since it has been shown that self-care and adherence to treatment are diminished in people who experience this feeling.

The present study aimed at identifying the factors associated with the categories of loneliness in the elderly population attending day care centers in Bogotá, Colombia, to promote activities for the prevention of loneliness. It should be emphasized that preventing loneliness supports the aging process proposed by the Colombian policy on human aging and old age, as well as the increase in wellbeing in all areas of health of elderly people ²⁴, and even more in the framework of the decade of healthy aging 2020-2030.

Materials and methods

Results

The research included 215 elderly people whose mean age was 70.5 years with a SD of 6.90 and 72 % of whom were females. In relation to the marital status, 38.60 % (n = 83) were single, followed by widowed with 21.40 % (n = 46). Fifty-six percent (n = 121) had primary education and 60.90 % (n = 131) belonged to the middle class (socioeconomic status 3). Regarding the number of comorbidities, a median of 1 (IQR 0-1) was found (Table 1).

According to the ESTE scale and the measurement of family loneliness, marital loneliness, social loneliness and adaptation crisis, 67 % (n = 143) of the study subjects showed a low level of family loneliness, 79 % (n = 170) a high and medium level of marital loneliness, 51 % (n = 108) a high and medium level of social loneliness, and 43 % (n = 92) a high and medium level of adaptation crisis (existential loneliness) (Table 2). Similarly, a data analysis by dimensions revealed that, in all dimensions, women experience higher levels of loneliness than men (Figure 1).

Figure 1 Dimensions of loneliness by sex 

For the bivariate analysis, a level of statistical significance (p < 0.05) was considered in the different dimensions of family loneliness, marital loneliness, social loneliness and adaptation crisis as shown in Table 3, which presents statistically significant variables in each of the dimensions included in the logistic regression models.

In relation to family loneliness, 28 % (n = 43) of the 155 participating women showed this type of loneliness with statistical significance (p < 0.01), 57.8 % (n = 41) had a history of chronic NCD (p = 0.03) and the number of people with whom the patient lives accounted for a median of 1 (IQR 0-1) (p < 0.01).

Concerning marital loneliness, 83.3 % (n = 130) of the participating women showed this type of loneliness with a statistical significance of p = 0.01. Regarding the marital status, statistical significance was found in all categories; however, they were not included in the regression model because of the collinearity effect. In relation to the number of people with whom the patient lives, p = 0.01 with a median of 1 (IQR 1-3) was found.

In the dimension of social loneliness, the statistically significant variables (p < 0.05) were female sex (p = 0.01), lower class (p = 0.01), history of chronic NCD (p = 0.04) and number of comorbidities (p = 0.03). Furthermore, in adaptation crisis, similar data was found in relation to sex and lower class.

In the logistic regression model, associations were found in the following variables according to each dimension: in marital loneliness, an adjusted OR = 3.15 (95 % CI: 1.55-6.39) related to female sex; in social loneliness, an adjusted OR = 1.95 (95 % CI: 1.08-3.52) related to lower class; and in adaptation crisis, an adjusted OR = 1.99 (95 % CI: 1.09-3.63) related to lower class (Table 4).

Discussion

The feeling of loneliness in the elderly population has an impact and a prevalence of 40 % and is related to the development of both physical and mental chronic NCDs ¹⁴. Therefore, the timely detection of loneliness will allow health teams to develop strategies for disease prevention and health promotion.

Through the ESTE scale-which has already been validated in the Colombian population-the study found that the most frequent dimension of loneliness was related to marital loneliness, followed by social loneliness, adaptation crisis and, finally, family loneliness. In addition, it was associated with female sex and lower class. In accordance with the research carried out by Garza-Sánchez et al. ²⁶, similar data was found for the dimension of marital loneliness in Spanish women.

As for sex, the results of this study agree with others’: women show higher levels of loneliness ²⁶. Regarding the age, the results of the bivariate analysis were significant in the dimensions of family and social loneliness, which were similar to those found in other populations ²⁷.

This research work also showed, in general terms, that the sample of older adults had a low educational level and, unlike previous studies, it was not associated with a higher level of loneliness in any of the dimensions ^26-28. Concerning the marital status, a statistically significant association was found in the dimension of marital loneliness, as expected ²⁷.

The low socioeconomic status showed a significant outcome in the dimensions of loneliness and adaptation crisis. Thus, it is important to examine and work on the adaptation crisis: a task that involves not only older adults but also the population that someday will become adults and should prepare themselves to reach this age from the approach of healthy aging ²⁹.

As for the number of people with whom the patient lives, the results of the bivariate analysis were not significant in the dimension of social loneliness. However, it was significant in the other dimensions, so that living commonlaw and having a relationship were important variables ³⁰.

Chronic NCDs were common in the study population and are generally frequent in older adults. It was significantly associated with the dimensions of family and social loneliness in the bivariate analysis but not significantly in the adjusted analysis. Furthermore, the number of comorbidities was low and showed a nonsignificant association in the regression model.

Regarding the limitations, as this was a cross-sectional study, only one measurement was taken. Therefore, the levels of loneliness and the associations resulting from the survey should be carefully examined and verified with another type of study, which will prevent a lower causality. Concerning the strengths, the ESTE scale, a validated tool in the Colombian population, was used to evaluate loneliness, and the expected sample size was reached. Since the population came from day care centers, a simple random sampling of all the people attending the center was performed to control the selection bias.

In conclusion, the feeling of loneliness is characteristic of women and single or widowed people. However, the results found in family loneliness should be highlighted, given that there is no evidence of high and/or medium levels.

Author contributions: CMCR and DACHC conceived the idea of the manuscript, collected the data, collaborated with the critical editing of the manuscript and approved the final version for publication. In addition, DACHC conducted the analysis of the study and CMCR wrote the first draft of the manuscript.