DEPRESSION IN ELDERLY:A CROSS-SECTIONAL STUDY IN RURAL SOUTH INDIA

JIMSA October - December 2007 Vol. 20 No. 4 259
DEPRESSION IN ELDERLY:
A CROSS-SECTIONAL STUDY IN RURAL SOUTH INDIA
Ankur Barua, Das Acharya, K Nagaraj, H Vinod Bhat, NS Nair
Department of Community Medicine, Kasturba Medical College, Manipal, India
Abstract : The objectives of the study were to determine the prevalence of depression in elderly rural population and also study the sociodemographic
correlates of the depressive disorders among the elderly in this community. It was cross sectional study performed on the
elderly subjects of rural area of Udupi taluk Karnataka in South India over 8 months period. A total of 627 elderly individuals of age group
of 60 years and above, were interrogated : results were subjected to statistical analysis i.e proportions and their 95% confidence intervals,
Chi-square test, multiple logistic regression and its 95% confidence interval. The prevalence of depression in elderly population was
determined to be 21 .7%. The prevalence in the age group of 80 years and above and those individuals who had a history of death in the
family within the last six months were found to be 34.4% and 52.4%, respectively. Multiple logistic regression analysis revealed that these
two correlates were independently associated with depressive disorders in elderly population.
Key words: Depression, Prevalence, Correlates, Elderly, Multiple Logistic Regression
INTRODUCTION
The Indian aged population is currently the second largest in the world.
The proportion of those who would be aged 60 years and above is estimated
to be 7.7% for the year 2000, and this proportion is expected to reach
12.6% in 2025.1 A high prevalence of mental disorders is seen in old age.
Predominant among these is depression1. The future projections of global
DALY’s in the year 2020 show that mental disorders are projected to
increase to 15% of the global disease burden and unipolar major depression
could become the second leading cause disease burden after ischemic
heart disease12 especially in high-income countries. The communitybased
mental health studies have revealed that the point prevalence of
depressive disorders among the geriatric population in India varies between
13 and 25 percent According to the observations made by the World
Health Organization, the correlates -= disorders in old age are reported
as genetic susceptibility chronic disease and disability, pain,, frustration
with limitations in activities of daily living - events (widowhood, separation,
divorce, bereavement, poverty, social, isolation ) and lake of adequate
social support. Though depression is the commonest mental health
problem in old age, very few community-based studies had been conducted
in India, to understand the problem. No such study had been conducted
in the past in Udupi taluk of Karnataka. Considering this background, a
community-based mental health study was conducted in the rural area of
Udupi taluk to determine the disease burden of depressive disorders and
to study the correlates of depression among the elderly in the community.
MATERIAL AND METHODS
The rural field practice area of the Department of Community Medicine,
Kasturba Medical College, Manipal is located in the coastal area of Udupi
taluk in Udupi District of the state of Karnataka in South India. The total
geriatric population (>60yrs) in the field practice area is approximately
10.5% of the total population covered by the rural field practice area.
Study period: 8 months ( March to October 2002).
Setting: Three villages i.e.—Udayavara, Kadekar, and Katapady.
Study Design : Cross-sectional study.
The sample size was estimated for finite population with the help of EPIinfo
version 5.0 statistical package. The total geriatric population
(>=60yrs.) covered by the 3 RMCW homes was estimated to be of 2259.
Here, the confidence level was taken as 95%, 11 .2% prevalence rate of
depression, required relative precision of the estimate was set at 20% and
a non-response rate of 10% was included; hence, the final sample size
was determined as 627.
Sample size : 627 people in the age group of 60yrs and above, who were
permanent members of their respective households, were selected for the
study.
Sampling method : Simple Random Sampling ‘without replacement
method using the Probability proportionate to size (PPS) technique was
used.
Sampling Procedure - Exclusion criteria: If a designated house was
found locked during the first visit and the eligible residents could not be
contacted and even after 2 successive revisits then they were all excluded
from this study. Criteria For Defining A Non-Respondent: If a designated
respondent was non-cooperative or had severe behavioural problems or
cognitive impairment, had severe hearing impairment or articulation
disorder, had any terminal illness or if he could not be contacted during
two separate revisits after the first, then he was considered a non-respondent
Selection Procedure: Due to some on-going projects in some of the field
practice areas, only 3 centres out of the total 6 RMCW (Rural Maternity
And Child Welfare) Homes were chosen for our project As all the villages
in the field practice area are culturally and sacio-demographically identical,
this selection bias had minimal effect on the results, Using PPS (probability
proportionate to size) method, the required number of parucipants from
each village was decided. Then the households and parcipants were
randomly selected from updated family folders in RMCW homes using
the random number table. All the eligible candidates of the selected
households were interviewed as it was presumed that the effect of genetive
susceptbility would be minimal because only 4% of our study population
had either 1st degree or 2nd degree relatives residing together in the same
household.
Study Instruments: A fact sheet consisting of information regarding the
household of the respondent was used for data collection. A semi-structured
proforma containing information regarding the soclo-economic status of
the individual that was later estimated by the modified Udai Pareek Scale8
was also used, Presence of depressive disorders was determined using
the instrument Mastering Depression In Primary Care Version 2.2: It had
two components: (a) WHO (five) Well-being Index (1998 version), (b)
Major (lCD-b) Depression Inventory. Cognitive impairment was estimated
by the 6CIT Dementia Test. Mastering Depression In Primary Care Version
2.2 and the 6CIT Dementia Test were translated into Kannada and Hindi
by the researchers and back-translated into English by another expert, not
acquainted with the original versions. The back-translation was
subsequently compared with the original version by a psychiatrist for
conceptual equivalence of the items.
Organization Of Field Work And Data Collection
The investigator, along with three field ANMs (auxiliary nurse mid-wives),
Correspondence: Dr. Ankur Barua, Assistant Professor, Department
of Community Medicine Sikkim – Manipal Institute of Medical
Sciences (SMIMS) 5th Mile Tadong, Gangtok – 737 102 India
Fax.: 03592-231496, E-mail : ankurbarua26@yahoo.com
ORIGINAL
260 JIMSA October - December 2007 Vol. 20 No. 4
were trained by the psychiatrists on how to administer the questionnaires.
All our study instruments were pre-tested to determine whether they
optimally suited our field conditions. At the beginning, officials of the
local panchayat office, village leaders, Anganwadi workers and the ANMs
were contacted and their help was sought to understand the geography of
the sites and to trace the households. After informed verbal consent was
obtained, the designated respondent(s) of a particular household was
administered the selected sets of questionnaires by the investigator along
with the help of the field ANMs. Care was taken to ensure privacy and
confidentiality of the interview as part of the study. A brief general health
check-up of the respondent was conducted at the beginning to establish a
good rapport with him and also to gain his confidence. Alt the
questionnaires administered in the field were evaluated and rated on the
spot, and If a respondent became positive in any of our screening or
diagnostic instruments he was immediately handed over a referral slip
and sincerely requested to visit the psychiatry OPD of Kasturba Hospital,
Manipal at the earliest for a free consultancy. The participants having
obvious medical disorders were referred to the nearest RMCW homes
for a free health check-up. The diagnoses generated by the instruments in
our study were strictiy kept confidential and were reconfirmed by
consulting a senior faculty member of the department of psychiatry of
KMC Hospital, Manipal before arriving at a final lCD-b diagnosis for
data analysis.
Data Analysis
The collected data was tabulated and analysed by using the statistical
package SPSS (Statistical Package For Social Sciences) version 10.0 for
Windows. Findings were described in terms of proportions and their
95% confidence Intervals. chi-square test was applied to study the
relationship between different variables and depression. To determine
the independent effect of various factors on depressive disorders, multiple
logistic regression was performed and their significance was estimated in
terms of adjusted OR and its 95% confidence interval. P value less than
0.05 was considered as significant.
RESULTS AND DISCUSSION
During our field survey, 487 households were visited and 627 individuals
in the geriatric age group of 60 years and above were contacted. Among
these 627 elderly people, we could interview only 609 individuals for the
assessment of depressive disorders (97.1%), The 18 individuals, whom
we could not interview due to various reasons, were categorized as nonrespondents
(2.9%). The baseline characteristics of the population
surveyed revealed that 36.0% were males while 64.0% were females.
Majority (52.6%) belonged to the age group of (60-69) years. Only 58.7%
of the elderly were literates. Majority (61.2%) belonged to the middle
socio-economic status and 56.3% of the individuals were married.
The overall prevalence of depressive disorders among the elderly of 60
years and above was found to be 21.7% (95%Cl18.4-24.9). Our study
findings were consistent with the observations made by Nandi et a14,
West Bengal, Ramachandran V. et a15 Madras and Tiwari S.C. Lucknow,3
who had determined the prevalence of depressive disorders in the geriatric
population to be 22.0%, 24.1% and 13.5% respectively. However, a high
prevalence of depressive disorders of 52.2% among the elderly 60 years
was observed in the study conducted by Nandi et a19 in the rural areas of
West Bengal. In contrast to these observations, Rao Venkoba A. et al10
Madurai had recorded the prevalence of depression to be as low as 6.0%.
Studies conducted by Newman at Canada, and Kennedy et al6, USA
reported prevalence of depression among the elderly to be 11 .2% and
16.9%; respectively. We had also assessed the status of positive weltbeing
by using the WHO (Five) Well-Being index (version 1998). We
had observed that the prevalence of depressive disorders was high among
individuals whose status of positive well-being was poor (75.9%) as
compared to those who were satisfactory (5.3%). Table 1 Shows the
prevalence of depressive disorders according to various sociodemographic
correlates.
In this study, the prevalence of depressive disorders was higher among
females (22.6%) than males (19.9%), but this difference was not found to
be statistically significant (x2= 0.616, dt=1, p= 0.433). Our study findings
are consistent with the study by Blazer 12 (1979, North Carolina), where
the prevalence of depression was similar in both sexes. However, the
studies conducted by previous workers 5,6,11,13 had documented a high
prevalence of depression among the elderly females. Higher standards of
living, matriarchal family system and a high female literacy rate (94.6%)
could explain a lower prevalence of depression among females in our
study.
The age of the respondents ranged between 60to93 years, while the
mean age was found to be 69.0 years (SD6.8). The revalence of depressive
disorders was highest (34.4%) in the age group of 80 years and above.
The difference in prevalence of depression between different age groups
was found to be statistically significant (x2 9.932, df2, p0.007). The
prevalence of depressive disorders showed a positive linear trend of
increase with the progression of age, which was also found to be statistically
significant. Majority otud’pop were Hindus (80.1%). The prevalence of
depressive disorders did not vary widely among the Hindus (22.5%),
Christians (17.3%) and Muslims (20.0%) and the difference was not
found to be statistically significant. Similar findings were reported from a
study conducted by Tiwari3. The prevalence of depressive disorders was
high among the individuals belonging to the low economic status (SES)
group (25.2%) and high socia economic status (13.6%) groups. But the
difference between these groups was not found to be statiscally significant.
Studies conducted by several worker 5,6,13 had observed the prevelance of
depressive disorders to be significantly higher among the elderly belonging
to the low SES group.The prevalence of depressive disorders was similar
among the unmarried widowed or separated individuals (23.2%) as
compared to their married counterparts (20.5%). Our study findings were
not consistent with the previous studies 56 who had documented a
significantly high prevalence of depressive disorders among the widowed
individuals, in this study, we had observed that majority of the unmarried,
widowed or separated individuals were women (92.1%) with only a few
staying alone (5.6%) and deprived of any living child (5.2%). Better
JIMSA October - December 2007 Vol. 20 No. 4 261
standards of living, a satisfactory level of family support systems network,
high female literacy rate (94.6%) and matriarchal family system could
explain a lower prevalence of depression among these individuals in our
study.
In this study we found that the prevalence of depressive disorders remained
similar in case of both nuclear (20.6%) and print/extended families (22.2%).
The respondents, staying alone 16 (2.6%), were not included under nuclear
family. - ln this study, only 16 (2.6%) of the individuals were living alone.
The prevalence of depression among those who were staying alone, living
only with their children or relatives or living with their spouse was found
to be 18.8%, 22.9% and 20.8% respectively. But the difference between
these groups was not found to be statistically significant These findings
were in contrast with the studies conducted by Ramachandran5.
Blazer Dan12 and Kennedy Gary J.6 who had observed a significantly
high prevalence of depression among those living alone.
The prevalence of depressive disorders among illiterates was higher (254%)
as compared to literates (19.0%). The difference between the two groups
was however, not found to be statistically significant. Ramachandran V.15
‘Madrasad also reported similar observations. Studies conducted by
Kennedy et al16 and Penninx et al also reported a significantly higher
prevalence of depression among individuals with lower level of education.
None of the respondents were unemployed in the past. The proportion of
housewives affected with depressive disorders was 20.1%. The prevalence
of depressive disorders was almost similar among the unskilled (23.1%)
and skilled (24.5) labourers. Some of the previous 5,6 had reported a higher
prevalence of depression among the unemployed individuals.
As compared to smoking and alcohol consumption (17.2%), tobacco
chewing (39.2%) and pan chewing (49.9%) habits were common among
the geriatric population in Udupi Taluk. In this study, the prevalence of
depression was found to be significantly high among the having pan
chewing individuals (25.7%), tobacco chewing (26.1%) and alcohol
consumption (29.1%) habits. In a study conducted by Hämäläinen J. et
al14 from Finland, it was found that cigarette smoking and alcohol
consumption were important risk factors for major depressive episode.
The prevalence of depression was similar among those who gave a history
of psychiatric illness (19.5%) as compared to those without family history
of psychiatric illness (21.8%) and the difference between the groups was
not found to be statistically significant. These findings are in contrast with
the observations by Ojen Van15 who reported a significantly high prevalence
of depression among those with a positive family history of mental
disorders. This difference from our study might be due to social stigma
resulting in considerable number of under-reported and undiagnosed cases
of mental illness.
The prevalence of depression was high among the individuals who had a
history of death in their family within the last 6 months. The difference
between the two groups was found to be statistically significant . Similar
observations were also noted by Kennedy Gary6.
Table 2 describes the association between correlates of depressive disorders
according to the univariate as well as the multivariate analysis.
It was observed by univariate analysis age group of 80 years and above
and a history of death in the family within last 6 months had strong and
significant association with depressive disorders. However, Multiple
Logistic Regression analysis revealed that age group of eighty years and
above and a history of death in the family within last six months had
independent significant association with depressive disorders in the
geriatric population. These findings are consistent with the observations
from the study conducted by Kennedy Gary6.
CONCLUSIONS
In this study, the prevalence of depressive disorders among the geriatric
population was determined to be 21.7%. The prevalence rates of depression
among the males and females were 19.9% and 22.6%, respectively.
Multiple logistic regression analysis revealed that age group of 80 years
and above and a history of death in the family within the last six months
were indepenc1eny associated with depressive disorders in the geriatric
population.
LIMITATIONS
Due to feasibility constraints, we could not interview the people who
lived in the open and were homeless. Due to the lack of practical skills in
communication, we could not interview the non-respondents who were
having severe hearing impairment and aphasia. Since the proportion of
non-respondents and the individuals who were homeless was very small
in our study population, we expect only a minimal effect on our prevalence
estimate.
ACKNOWLEDGEMENTS
The authors would like to express their deep sense of gratitude to Dr. R.S.
Phaneendra Rao, Dean, Professor of Community Medicine, Kasturba
Medical College, Manipal for his invaluable support, critical evaluation
and skilled guidance throughout the study. The authors are also indebted
to Dr. N. Kar, Associate Professor, Department of Psychiatry, Kasturba
Medical College, Manipal for his technical guidance and valuable advice
on various aspects of psychiatric evaluation.
REFERENCES
1. The World Health Organization. World Health Report Mental Health: New understanding
New Hope. Geneva: The institute; 2001
2. Wig NN. World Health Day, 2001. Indian Journal of Psychiatry 2001; 43(1): 1-4.
3. Tiwari SC. Gerfa —Naffic morbidfty in ru:a riocthern India: implications for tne futjre
International Psychogenatrics 2000 March: 12(i). 35-48
4. Nandi DN, Ajmany S, Ganguli H, Baneree G. Borai GC, Grosn A. et a Tre Incidence of
mentai disorders in one year in a rural community in West Bengal. Indian Journal of
Psychiatry 1976: 18: 79-87.
5. Ramachandran V, Menon Sarada M, Arunagirt S. Socio-cultural factors in late onset
Depression. Indian Journal of Psychiatry 1982; 24(3), 268-73.
6. Kennedy Gary J, Kelman R Howard, Thomas Cynthia, Wisniewski Wendy Metz eer
Polly. Hierarchy of characteristics associated with Depressive Symptoms in an urDan
eioery sample. American Journal of Psychiatry 1989 February; 146(2): 220-25.
7. Dey AB, editor. Handbook on Health Care of the Elderly: A manual for physicians and in
Primary and Secondary Health Care Facilities. New Delhi: The World Health Organization:
Regiona’ Office for Southeast Asia, Ministry of Health and Family Welfare, All India
Institute of Medical Sciences (india); 1999.
8. Pareek U, Trivedi G. Manual of Soclo-economic Status Scale (rural). Delhi: Manasayan,
1980.
9. Nandi PS, Baneijee G, Mukherjee SP, Nandi S, Nandi DN. A study of Psychiatric morbidity
of the elderly population of a rural community in West Bengal. Indian Journal of
Psychiatry 1997: 39(2): 122-9.
10. Rao Venkoba A, Madhavan T. Geropsi r a semi-urban area near Madurai. Indian Journal
of Psychiat-y I S2 :
11. Newman SC, Bland RC, Cm Hi The me elderly in Edmonton: a community survey using
GMS-AGEC a-a ra 1998; 43: 910—14
12. Blazer Dan, Williams CD. Epidemiology of Dysphoria and Depression in an elderly
population American Journal of Psychiatry I 980; 137(4):439-44.
13. Penninx Brenda WJH, Leveille S,Ferrucci L, Eijk JTM, Guralnik JM. Exploring the
effect of depression on Physical disability longitudinal evidence from the established
populations for epidemiologic studies of the eldery. American Journal of Public Health
I999;89: 1 346-52.
14. Hamalainen J, Kaprio J, Isometsa E, Heikkinen M, Poikolainen K, Lindeman S, et al .
Cigarette smoking , alcohol intoxication an major depressive episode in a representative
population sample. Journal of Epidemiology and Community Health 2001; 55:573-76.
15. Ojen Van R, Hooijer C, Bezeme D. late life depressive disorder in the community, early
onset and the decrease of vulnerability with increasing age. Journal of Affective Disorders
1995; 33:1 59-65.