The interface of dementia and physical disability; causative factors and associations in the elderly population from Pakistan: evidence from the nationally representative cross-sectional Demographic and Health Survey (DHS, 2017-18) of Pakistan

Introduction

Cognitive impairment and physical disability are common occurrences in the elderly population.¹ Both conditions are partly driven by age-related degenerative processes of the nervous and muscular systems.² A potential coupling is suggested wherein the presence of dementia exacerbates muscle loss and physical disability and vice versa.² It is also well established in the Public Health research that these conditions increase the risks of morbidity and/or mortality in the elderly population.^3,4

The preclinical management of these conditions requires lifestyle modification and timely identification of socioeconomic and demographic factors dictating the cognitive and physical disability.^5,6 Several causative factors have been identified⁷; however, most of the data is from high-income countries and may not be relevant for low-income countries such as Pakistan.

The scarcity of data from low-income countries is further compounded by the conflicting results.^8–10 The discrepancy can at least partly be explained by the small sample size (< 1,000 participants), the differences in assessment tools for disability, geographical variations, and different socioeconomic, demographic, and lifestyle attributes of the target populations.^8–10 Furthermore, it is not known to which extent the shared risk factors drive the association between cognitive impairment and physical disability. These findings necessitate the concomitant cognitive and physical disability assessment in relevance to potential risk factors.

Pakistan is experiencing a rapid population growth. The number of older people above 60 years of age in Pakistan is estimated to be six per cent of the total population and is predicted to double by 2050.¹¹ The unplanned urban growth and overgrowing population have further compromised the insufficient health and civic services. The people living in the urban slums and rural areas have limited access to safe drinking water, sanitation, and other fundamental necessities. The elderly population finds it more challenging to access these limited facilities due to functional dependency and an age-unfriendly environment.¹²

Analysis of cognitive and physical disabilities is imperative in Pakistan, considering it is a highly populated country with limited availability of health care facilities and an under-appreciation of age-related degenerative diseases. However, to our knowledge, no previous study has investigated the concomitant prevalence of cognitive and physical disability in Pakistan. In addition, the multifactorial etiology of these conditions in the context of socioeconomic and demographic factors remains elusive in Pakistan.¹³ We aimed to bridge this gap by 1) reporting the prevalence and coupling of cognitive impairment and physical disability and 2) their associations with several socioeconomic and demographic factors in Pakistan. We used data from the demographic and health survey (DHS), a cross-sectional study involving rural and urban populations of six administrative areas of Pakistan.

Methods

This is an analysis of a cross-sectional survey. All data utilized in this paper stems from the Pakistan part of the DHS (https://dhsprogram.com/data/dataset/Pakistan), which are large representative harmonized cross-sectional face-to-face interview surveys of individuals and households covering multiple less developed countries worldwide, often for several years.¹⁴ The DHS Program collects data on various public health domains, demographic, and socioeconomic factors. This enables the research to draw a more holistic picture of society's most vulnerable groups.

Data was collected from urban and rural areas by creating sixteen strata from eight regions: Punjab, Sindh, Khyber Pakhtunkhwa, Baluchistan, Azad Jammu and Kashmir, Gilgit Baltistan, Islamabad and Federally Administered Tribal Areas (FATA). Cluster sampling was done by selecting households from each cluster within the sixteen stratum using probability sampling. 16,240 households were surveyed from a fixed number of twenty-eight households which were selected from 561 clusters (details on https://dhsprogram.com/data/dataset/Pakistan).¹⁵ The ethics approval for the original study was taken from the National Bioethics Committee, Pakistan Health Research Council and the International Review Board of ICF. Informed consent was obtained from the participants (https://dhsprogram.com/publications/publication-fr354-dhs-final-reports.cfm?csSearch=830937_1).¹⁵ The latest DHS conducted for Pakistan was from 22 November 2017 to 30 April 2018. We used the data based on the household questionnaire.¹⁵ According to the published country report for Pakistan, the sample size is as large as 67,586 individuals concerning disability analyses, while in the actual data set used here (PKPR71FL.DTA), we find 71,649 individuals. Although this seems to be a large discrepancy, the different disability estimates reported are usually the same or have negligible differences, e.g., the few differences were less than 0.5 % points.¹⁵ Since we here focus on the elderly aged 60 years and above, the sample size in this study is 5,183 individuals.

Data are weighted by applying the household weights (variable HV005) to make estimates representative, which is also the methodology used in the official publication.¹⁵ Although the population weights appear an obvious choice, estimates based on the two weights are highly correlated. Thus, we kept the household weights as the standard data collection tool in this study for consistency.

The central variables covering disability represent two domains, remembering and walking abilities, which are operationalized as following. For the first variable, the question asked by the interviewer (of the survey) to the respondent (in the survey) is: “I would like to know if (NAME) has difficulty remembering or concentrating. Would you say that (NAME) has no difficulty remembering or concentrating, some difficulty, a lot of difficulty, or cannot remember or concentrate at all?”.¹⁵ For the second concept, the asked question is: “I would like to know if (NAME) has difficulty walking or climbing steps. Would you say that (NAME) has no difficulty walking or climbing steps, some difficulty, a lot of difficulty, or cannot walk or climb steps at all?”.¹⁵ We applied Goodman and Kruskal's gamma with an estimated asymptotic standard error, ASE to analyze the two variables.¹⁶ All data analyses were carried out using the Stata 12 software.

Apart from the two disability variables, several socioeconomic and demographic characteristics are also included, as summarized in Table 1. The analyses were separately carried out for males (n = 2,899) and females (n = 2,284) aged 60+ years in this study.

Results

Around 38 per cent of the sample is 60-64 years, while only around 10 per cent are 80+ years. Among the elderly, 83 per cent of males and 49 per cent of females were married, reflecting the lower life expectancy of males. 84 per cent of females and 54 per cent of males were without education. The remaining variables exhibited a generally uniform distribution for both genders (see Table 1).

Relation between remembering and walking disability

An overview of the dataset regarding the relation between remembering and walking disability is presented in Figure 1. 61.1 per cent of the elderly population aged 60+ years who did not have remembering difficulty also did not have walking difficulty, and only 1.5 per cent could not walk. This marked correlation is also clear when looking at the elderly with some remembering disability since 53.2 per cent also had some walking difficulty. Similarly, 50 per cent of the elderly with a lot of remembering difficulty also had a lot of difficulty with walking. Finally, 53.6 per cent of those respondents who had complete difficulty remembering experienced complete difficulty with walking, while only 18.7 per cent had no difficulty with walking. In general, a greater difficulty in remembering was associated with a greater difficulty in walking, indicating the interface between the two disability conditions.

Figure 1. Relative contributions of participants with difficulty in walking/climbing vs. difficulty in remembering/concentrating.

Discussion

There were no previous reports of the potential interface between dementia and physical disability in Pakistan and the extent to which they are influenced by the demography and socioeconomic factors of the population. Using data from 5,183 participants of both genders, we report a robust interface between difficulty walking and difficulty remembering in the older population at least 60 years of age. The prevalence of both difficulties increased with age, albeit with higher proportions in females. The wealth status and education levels were negatively correlated with difficulty levels (i.e. on average, higher prevalence of difficulty was observed for people with lower wealth or lower education), but the area of residence and the access to safe drinking water exhibited no significant correlations with difficulty levels. Conversely, household size and number of children at home revealed gender-specific associations with disability levels. Among the geographical regions, Balochistan had the lowest prevalence of both disabilities.

As expected, we found a robust interface between dementia and physical disability in the elderly population. Additionally, with advancing age, the relative proportions of the participants with one or both disabilities increased. However, to our surprise, we found gender discrimination with an overrepresentation of females in both disability indexes. Globally, the incidence of dementia is higher in women than men.¹⁷ However, an increased incidence in women than men is only evident from the middle of the eighth decade of life.¹⁷ Contrary to the literature, there was higher preponderance of females starting at 60 years of age in the present study. While the causes can be multifactorial, dementia is associated with lower wealth status, as reported elsewhere.¹⁸ Specifically, we found that the effects of wealth status on the cognitive decline were more pronounced in women than men. While the causes may be speculative, women have greater household responsibilities in arranging domestic finances,¹⁹ which can induce financial stress. Women may also develop mental stress due to subordinate status in a patriarchal family system.²⁰ An association between stress and cognitive decline is well established²¹ and may account for the higher prevalence of dementia in women in our study.

We also report a higher prevalence of physical disability in women than men for all age groups. Age-related muscle loss, termed sarcopenia, is the primary driver of physical disability in old age.²² However, the literature is inconsistent about the gender-specific prevalence of sarcopenia in the older population. For example, a higher prevalence of sarcopenia is reported in women vs. men in the UK, USA, South Korea, and Brazil.²³ Conversely, men have a higher occurrence of sarcopenia than women in India and Taiwan.^23,24 Sarcopenia and physical disability are dictated by various factors with significant variations across populations and countries. However, when dissected according to the age-groups, women have a higher prevalence of sarcopenia and functional dependency than men, beginning at 60 years of age.²² Several factors can account for this, including menopause, which is associated with accelerated physical and cognitive impairment on top of gradual age-related degenerative processes.^25,26

We found a negative correlation between dementia and the levels of education, so higher education was associated with lower dementia irrespective of gender. The robust relation between dementia and education is more consistently found in developed vs. developing countries.²⁷ Multiple mechanisms have been proposed to explain the protective effects of education against dementia, including cognitive reserve theory, the “lose it or use it model”, and reduced brain battering.²⁸ However, the search for a definitive mechanism remains elusive. Contrary to dementia, the effects of education on physical disability were restricted to males only where higher education was associated with less physical disability. Conversely, we did not find an association between education and physical disability in females. This is probably due to a small sample size of educated females in the study cohort.

Interestingly, we found a higher representation of disability in people who were never married. The stress-buffering capacity of marriage and family support are previously highlighted and may assist in countering age-related debilitating conditions.²⁹ Specifically, the prevalence of physical disability, cognitive impairment, and frailty were lower in married vs. unmarried older people.^30,31 Consistent with this perspective, the married people exhibited lower disability in remembering and/or walking than the unmarried people, irrespective of the gender and type of disability. Similarly, the spouse availability also reduces the risk of going to the nursing home and requiring formal care in daily life activities.³² Our data indicate that the preventive effects of marriage on disability are long-lasting and persist after a partner dies, as the widowed people maintained a lower prevalence of dementia and physical disability than the unmarried people. However, the prevalence of disability was higher in widowed vs. married people, indicating an incomplete resurfacing of the hazards of single life. These findings are consistent with earlier reports suggesting the negative consequences of widowed status on disability in the older population.³³

The stress-buffering effects of family support were also evident in the association between the number of children and disability in elderly women.³² Accordingly, having four or more children in the family was associated with a lower prevalence of both disabilities in the women. However, this effect was not found in men. The parenting of children in the Pakistani culture is almost exclusively the responsibility of the women,³⁴ and this may have a protective effect against age-related disabilities. Consistent with this finding, older adults with children are less likely to develop physical limitations of daily life and go to nursing homes than adults without children.³²

To our surprise, we found Balochistan with the lowest prevalence of both disabilities among the six administrative areas of Pakistan. Balochistan is an underdeveloped region with limited access to health facilities.³⁵ However, it is possible that the people there live a physically active lifestyle, limiting the onset and/or phenotypes of various disabilities. On the other hand, the participants from different regions with relatively modern and sedentary lifestyles may be predisposed to developing disabilities.

The strengths of this study are standardized and validated tools for measuring disabilities and other variables. The simplicity of the tools ensures obtaining sufficient and reliable data from the rural and illiterate population. We obtained an adequate sample size representing different variables, which enhances the statistical and biological significance of our data. However, this study also has certain limitations. Our data is descriptive in nature as we intend to establish the prevalence of these disabilities according to different variables. Thus, the reader must be cautious in inferring direct causality. Like any cohort study on the ageing population, the selective survival of the participants must be considered. This is partly because several variables influence the survival and mortality in old age.^36,37

Taken together, we report an interface of dementia and physical disability in various geographical regions of Pakistan. Marital status, education, number of children at home, and wealth status have emerged as potential drivers of disabilities in old age. Our findings may be helpful in identifying and focusing on the potential negative modifiers of disability in preclinical settings.

Ethics approval

We analyzed secondary data. The ethics approval for the original study was taken from the National Bioethics Committee, Pakistan Health Research Council and the International Review Board of ICF. For more details please refer to the PDHS 2017–18 report.

Consent to participate

Informed consent was obtained in the original study.