CORRELATES OF DEPRESSION AMONG ELDERLY POPULATION RESIDING IN A COMMUNITY IN EASTERN NEPAL Affiliation

With a increase in life expectancy, demographic transi on has led to sharp rise in number of elderly popula on. The elderly popula on growth rate per year has always been more than the total popula on growth rate, signifying the start of the ageing dynamics in Nepal. Mental health is a neglected health problem of the elderly in Nepal, with depression topping the list. This needs to be addressed in the na onal programs for the elderly to bring about an overall improvement in their health status.


INTRODUCTION
Ageing is a normal phenomenon associated with physical, social and psychological changes.With declining fer lity and mortality rates and rising life expectancy, demographic transi on in the 21 century has led to a sharp rise in 1 number of older popula on.This increase has been noted 2 in both developed and developing countries.It has been projected that 80% of older people will be residing in low 3 and middle income countries by 2050.
The old age dependency rate calculated for different me periods shows an increasing trend from 7.5% in 1911 to 12.01% in 2001.The elderly popula on growth rate per year has always been more than the total popula on 4 growth rate in Nepal.With a rise in life expectancy, age structure of the popula on has been changing due to increase in propor on of people aged 60 years and above who cons tute 8.4% of the 5 popula on in Nepal.Though, the increase is not as high compared to those observed for developed countries (as high as 13%), it signifies the start of the ageing dynamics in Nepal, that will have adverse effects on Nepalese social structure and economy in the long run.
Ageing is associated with various physiological and pathological changes including Mental Health.Mental health is a neglected health problem in Nepal, especially 6 among elderly popula on.Limited researches have been conducted in the community as mental health ignites s gma za on, isola on and there have been lack of policies focusing on the needs of, and health measures provided for geriatric popula on in the health policies of 7 Nepal.Even the opera onal use of the exis ng policies regarding healthcare of elderly popula on has been 8 virtually ignored by current health system.This study was conducted in one such representa ve urban area of Duhabi Municipality with objec ves to iden fy the prevalence of depression and assess factors associated with depression among the geriatric popula on.

METHODOLOGY
A community based cross-sec onal study was conducted among elderly popula on residing in Duhabi Municipality of Sunsari District over a dura on of one month.This municipality is located in southeastern part of Sunsari district having a popula on of 25,545 with 5.75% elderly 5 subjects (≥ 60 years).
The required sample size was calculated to be 353 based on the findings from the available literature, a er adding for a 9 non-response rate of 10%.Data collec on tools included a semi-structured ques onnaire that was developed and pretested in Dharan Municipality on 10% of sample size, which was not included in the final analysis.The 15-item Geriatric Depression Scale (GDS) was translated into Nepali language, followed by back-transla on by independent experts.GDS has been used in Asian communi es for older adults and has been shown to have Cronbachs alpha of 10 0.80.The score of 5 or above was used as cut-off for detec ng depression as this cut-off has been reported to 11 have a high sensi vity (92%) and specificity (82%).
A er obtaining data on the popula on composi on of each ward of the municipality, the propor onate sample size for each ward was calculated.Data collec on for the ward was started from the ward office and the direc on was decided by rota ng a bo le.Following this the first house was visited, and if an elderly was present, interview was conducted.The second house was skipped and collec on was resumed in the third household.If there was no elderly in the household, data collec on was resumed in the next household.This was con nued ll the required propor onate sample size for the ward was obtained.
In each household the nature and purpose of study was clearly explained and an informed wri en consent was obtained.One elderly from each household was included as respondent (study subject) irrespec ve of gender and where two or more were present, lo ery was done to select one respondent.Each respondent was interviewed privately where s/he would feel comfortable to ensure privacy and confiden ality.Subjects absent at the me of visit were relayed informa on and asked to be present the next day to be assessed.Those absent even a er this visit were excluded from the study.
Opera onal defini on of a few terms used in this study have been included.Overcrowding was defined as less than 70 square feet of floor space area per person in the household.Floor space area refers to total floor space area of all the sleeping room in house divided by the number of family 12 members.Pucca houses are made from high quality materials, i.e. bricks, cement, concrete, including roof, walls and floor while katcha houses are made from mud and thatch or other low quality materials.Semi-pucca houses 13 are made from a combina on of the above types.Kuppuswamy socioeconomic score categories included lower socio-economic status defined as those having score of 0-10 andhigher socio-economic status as those having a 14 score of 11-29.
Odds Ra o (OR) with confidence limit was calculated for inferen al sta s cs.Logis c regression analysis was performed for all independent variables significantly associated in the bivariate analysis.This was done to iden fy associated factors for depression in elderly with the probability of significance set at 95% of Confidence Interval (CI).

Ethical permission was taken from the Ins tu onal Review
Commi ee (IRC) of BP Koirala Ins tute of Health Sciences before the study.

RESULTS
A total of 353 respondents aged 60 and above, residing in Duhabi Municipality par cipated in this study with a response rate of 94.13%.Majority (80.7%) of respondents belonged to age group of 60-70 years with a mean age of 66.4 years and standard devia on (SD) of 7.118.Among the par cipants, 81.6% were Hindu by religion and 59.2% Females.The Terai Janaja ethnicity (33.4%) dominated other ethnic groups and 70.8% of the respondents were married.

DISCUSSION
The prevalence of depression among elderly in this study was 65.2% using the Geriatric Depression scale -15, which is higher than that found by other studies conducted in community-dwelling elderly of three Asian countries (Indonesia: 33.8%, Vietnam: 17.2%, Japan: 30.3%) and and 25.45% in community.This wide varia on of prevalence could be a ributed to various factors, firstly, the use of GDS-30 scale in some studies and the use of different cut-off values for GDS-15 scale and secondly, smaller sample sizes of studies in sheltered homes and hospitals giving varied prevalence.However, this study included an adequate sample size and used a cut-off of 4/5 for the GDS-11 15 which has a high sensi vity and specificity.The high prevalence of depression in this se ng could be a ributed to socio-cultural factors like lack of health care facili es pertaining to elderly popula on, migra on of younger genera on for employment and lack of awareness le elderly popula on behind with no one to care for them.This could be one of the reasons for the high prevalence which needs to be studied further.
Variables that were significant in the bivariate analysis but not in the regression model were gender, marital status, type of house and living with/without spouse.Elderly female respondents were 1.8 mes more likely to have depression than their male counterparts (OR=1.8,CI: 1.2-2.9).Similar reports have been observed in studies conducted in Malaysia (OR = 2.87, 95% CI : 1.37-6.02)and Pakistan (OR=3.5,CI:1.5-8.1),where gender was a 2,20 significant predictors for depression.In contrast, a few 15,21 studies have not found gender to be a significant factor.Similarly in our study, gender was not significantly associated in the regression model.
A study showed that marital status had a direct role in 22 predisposing the elderly to depression.In this study respondents who were unmarried/widowed/divorced were found to be 1.9 mes likely to be depressed than those married (OR=1.9;CI: 1.1-3.1).However this was not significant in the regression model, similar to the findings of 23 a study done in rural Malaysia.
Poverty has been shown to be associated with depression in 21 various studies including this.Poor housing condi ons are present in underprivileged popula on that includes thatched houses, mud/wood walls which have been shown to be associated with depression in few studies, and was also observed in bivariate analysis in this study (OR=1.9;CI: 24 1.2-3.2).However, the regression model rendered this factor insignificant. Elderly popula on not living with their spouse were more likely to be depressed than those living with their spouse in the bivariate analysis in this study (OR=1.8;CI: 1.1-2.9).This observa on has also been seen in 25,26 meta-analysis done by Yan et al (OR=1.55)and Greece.
Among the findings significant in the regression model, this study showed that respondents aged 70 years or more were 2.1 mes more likely to have depression than those less than 70 years of age (adjusted OR=2.1;CI:1.1-4.2), a finding [27][28][29] that mirrors conclusions of other authors.However, few studies have not found any significant associa on of age 15,16,19 with depression.
The possible explana on could be that as a person grows older, the old age changes become more severe and incapacita ng leading to social exclusion and mental disorders.
This study also shows that elderly in the Lower Class category of the Kuppuswamy scale were 3.3 mes more likely to be depressed than their affluent counterparts (aOR=3.3;CI:2.0-5.4).This brings us to a point of view that being poor, lacking educa on and un/under-employment predisposes the elderly to depression as this scale is based 14 on income, educa on and employment.This is in agreement with study carried out on elderly persons from 28 two communi es in Kingston, Jamaica.Other studies done in rural south Indian Community and rural Malaysia also showed that respondents with low income were at high risk 21,23 for depression.
Elderly who were sa sfied by the respect they received from the community were less likely to be depressed in this study.Likewise those who were of the opinion that they were either avoided or perceived themselves as a burden to the family were 4.2 mes at more odds of being depressed (aOR=4.2;CI: 2.1-8.4).This is in line with the findings of a study where Care Recipients who perceived that they were afforded lower levels of respect were at greater risk of depression and another disserta on from the Loyola 30,31 University Chicago.Perceived respect from others is a posi ve support that keeps a person socially ac ve and mobile.This study also iden fied this aspect which was sta s cally significant a er adjus ng for other variables.

CONCLUSION
Prevalence of depression among elderly was found to be very high in this study and is of a major concern, especially in this age of increased migra on of the young of the household for employment, leaving the elderly behind.Depression was found to be associated significantly with age more than 70 years, low socio-economic status and perceived lack of respect for the elderly.

RECOMMENDATIONS
With the ini a on of special health programs for the elderly by the government in Nepal, depression is also an important factor to be addressed in those programs as its prevalence is very high.Social Organiza ons also need to focus on the problem of the elderly and address them.The recent trend of migra on for work is also partly responsible for the problem, for which long term solu ons needs to be iden fied by the Government.

LIMITATION OF THE STUDY
Addi onal measures like rou ne lab inves ga ons to assess the physical health would have provided a more comprehensive picture of the overall health of the elderly, but was not possible due to resource limita on.In addi on, the study would have been more consistent if the scales used had been validated in the local popula on.
Goodness of fit: Chi-square = 0.892, df = 3, p = 0.827 Variable(s) entered on step 1: Age, Gender, Religion, Marital status, Kuppuswamy socioeconomic scale, Type of House, Get Respect from others and Living with/without Spouse.Logis c regression analysis was used to assess the most significant factors associated with depression as iden fied by the Geriatric Depression Scale.A er adjus ng for other variables, age 70 years or more [adjusted OR (aOR)=2.4;(1.2-4.7)],lower class category in the Kuppuswamy socioeconomic scale [aOR=3.0;(1.8-5.0)] and those not sa sfied with the respect conferred by others [aOR=4.2;(2.0-8.7)] were significantly associated with depression.

7 regarding
mental illness.The migra on of the younger popula on for employment to Arabian and other countries Sharma

Table 1 .
Age distribu on of the study subjects (n = 353 ) About 64.9% of the respondents were illiterate and 68.3% unemployed.Only 57.5% of the respondents were in the normal BMI range.Based on the income, educa on and occupa on status, majority (60.6%) of households were in the Upper lower SES based on Kuppuswamy Scale as shown in Table1.Socioeconomic and demographic profile of respondents residing in Duhabi Municipality (n=353) Among the respondents, a slightly greater percentage (37.7%)lived in a Pucca house in comparison to katcha houses with overcrowding present in 61.8%.Only 9.1% of the respondents did not use the toilet regularly.(Table2)Among the respondents, only 66.3% s ll lived with spouse and children and 77.6% were sa sfied with the respect they

Table 2 :
Housing Profile and Social Characteris cs of the study popula on (n=353)

Table 3 .
Depression status among respondents residing in Duhabi Municipality using the Geriatric Depression Scale (n=353) Based on the Geriatric Depression Scale-Short Form, depression was seen among 65.2% of the respondents.(Table3)Bivariateanalysiswith Pearson Chi-square showed that subjects aged seventy years or more were 2.4 mes more likely to be depressed than subjects less than seventy years old (p=0.006)andfemaleshad 1.8 mes more chances of being depressed (p=0.007).(Table4)Similarly, not being married (p=0.015),living in a Kutcha house (p=0.010) and living without spouse (p=0.028) were significantly associated sta s cally with being more depressed.Sta s cal associa on was highly significant (p=0.000) for lower class elderly popula on based on Kuppuswamy socioeconomic scale and subjects not sa sfied with the respect bestowed, with depression.

Table 5 :
Logis c Regression Analysis