Abstract
Background:
Longitudinal/follow-up studies of older adults are a tough task as sample attrition rates due to mortality and other factors may be high in this particular group. However, such studies are very much needed to assess the outcome of health status as well as explore preventive, protective, interventional aspects, as well as risk factors. Given this, a follow-up study was planned and carried out.
Aim:
To discuss the rate of sample loss as well as the reasons over 9 years.
Methods:
An Indian Council of Medical Research (ICMR) supported follow-up study of urban and rural elderly was done during June, 2016–May, 2017; these subjects were studied in 2007-09 through two independent ICMR supported studies. Similar methodology and assessment tools were applied in these studies. During follow-up a semi structured proforma was developed to get the information of study cohort, obtained data was analyzed and presented applying percentage statistics.
Results:
The sample attrition rate was reported to be comparatively high in urban 52.1% (n= 633) cohort than their rural counterparts 36.3% (n= 457).
Conclusion:
Over a period of 9 years chances of cohort loss due to mortality is about 32%–35%.
Keywords: Baseline, cause of death, elderly, follow-up, health status, sample attrition
BACKGROUND
Long-term follow-up studies on aging provide a chance to explore many issues related to healthy/unhealthy conditions, cognitive status, possible variations in the external milieu, etc. To identify risk factors for the assessment of negative health outcomes, one needs to carry out long-term follow-up; however, this leads to increased sample attrition rate due to noncooperation, nonavailability of the participants as well as mortality, and migration. Most of the longitudinal studies focus on one or other kind of problems or disease. Studies from abroad reveal that researchers seldom have information about how people who drop out would have responded if they had stayed in the study.[1] It is also reported that attrition can occur due to death or frailty, discontinued participation (withdrawal), and lack of success in recontacting the participant for a follow-up survey (no contact) or by nonreturn of a survey by a participant (nonreturn).[2] Another review reports the need for future work within six broad study topics, i.e., cognitive function, socioeconomic status (SES), health and physical performance, morbidity and mortality predictors, healthcare costs, and genetics.[3] The Australian Longitudinal Study of Aging reported that poor performance on nearly all cognitive variables is associated with mortality and cause of sample attrition.[4] Further, follow-up studies over 2–20 years have reported that poor self-reported health,[5,6,7] impaired cognitive function,[8] and depression[9] are some of the major risk factors for mortality in elderly. However, hardly Indian studies had provided a comprehensive analysis on healthy and unhealthy groups of elderly vis-a-vis cause sample loss. Such analysis provides clues and assists in estimating a usual pattern of mortality in elderly over a particular period.
The authors had an opportunity to carry out a follow-up study with the financial support of the Indian Council of Medical Research (ICMR) in which a total of 4786 (2283 urban and 2503 rural) elderly participants aged 64 years and above had to studied over 9 years. Nine years back, these participants were recruited and studied thoroughly in two independent ICMR-funded projects applying similar methodology and tools. These projects were “an epidemiological study of the prevalence of neuropsychiatric disorders with special reference to cognitive disorders among (Urban) Elderly (Lucknow urban elderly study)” and “an epidemiological study of the prevalence of neuropsychiatric disorders with special reference to cognitive disorders among (rural) elderly (Lucknow rural elderly study).” At the time of recruitment, these participants were aged 55 years and above and thus distinguished in two groups, i.e., preelderly (55–59 years) and elderly group (60 and above). Further, the present study titled “the Lucknow Urban and Rural Elderly Follow-up Study of Ageing, Neuropsychiatric, and Cognitive Disorders (Lucknow elderly follow up study)” was planned, developed, and submitted to the council which was approved and sanctioned financially in the recent past by the council in April 2016. The study aimed to follow-up both study groups applying similar methodology and tools as applied during baseline. Between baseline and follow-up study, there is a long gap and thus, before detailed assessment of the studied participants, a need for revival contacts was done to make a list of alive, migrated, and deceased participants of the area and to avoid inconvenience during detailed assessments.
The present paper aimed to provide the status of sample attrition rate of urban and rural elderly follow-up study pertaining to their age, sex and health status vis-a-vis migration, and mortality along with the reported cause of death.
MATERIALS AND METHODS
The article analyses the data of Lucknow Urban and Rural Elderly Follow-up Study of Ageing, Neuropsychiatric, and Cognitive Disorders. The study has been initiated with aims (i) to follow-up the identified urban and rural cohorts for their current status of neuropsychiatric, cognitive and medical disorders, and healthy aging; (ii) to study the association/effects of psychosociodemographic factors with/on neuropsychiatric, cognitive and medical disorders, and healthy aging; (iii) to study the quality of life of the study cohorts and factors influencing the same; and (iv) to study the role of life events in participants with neuropsychiatric, cognitive and medical disorders, and healthy aging. Here, we are sharing information collected during revival contact.
The revisit for follow-up to meet study participant was initiated from one of the urban areas of Lucknow on June 13, 2016, who were studied in 2007–2008 during Lucknow urban elderly study and details of 1216 urban participants were obtained; similarly, one rural area was also visited and information about 1260 participants were obtained, who were studied in 2008–2009 during Lucknow urban elderly study.
As the preliminary preparations, the data of Lucknow (urban/rural) elderly study were taken from the store and series wise its availability was assured. All assessment tools were collected for printing. Further, a brainstorming session with principal investigator, coinvestigators, and research team was organized to develop a field strategy. It was decided that sudden starting of the project with plenty of assessments after 9 years of time span may develop noncooperation. Thus, on first contact, a hi-hello session with a page of information regarding participants’ present status would be friendly and helpful. Moreover, a small and handy semi-structured pro forma was prepared. Before starting actual interaction with study participants, contacts were made with local influential people, namely, Sabhasad (Warden), volunteers in urban area and Village Pradhan and they were briefed about the project. With their consensus and assurance, the revival contact was started with the participants of the study.
The basic information, namely age, sex, SES, and health status of participants, was taken from the records of Lucknow urban elderly study so that during first contact subjects/caregivers/family members could recall the meeting and assessments done at baseline. Further, participants or their family members or familiar neighbor were contacted, and data of migrated, alive as well as deceased participants were recorded as of date. Moreover, on the basis of information and consent for the study, the sample attrition rate was analyzed and presented applying percentage statistics.
RESULTS
In 1 year of the study period, follow-up of one rural and urban area could be done and data of both study area (i.e., U = 1216; R = 1243) have been analyzed, which was obtained during revival contact before detailed follow-up assessment. A total of 583/1216 (47.9%) participants in urban area and 803/1243 (63.7%) participants in rural area could be contacted and studied. Details are summarized in Table 1.
Table 1.
Reasons of sample loss over 9 years by health status, gender, and study area (%)
| Descriptions (for A-Ha) | Urban-follow-up (year 2016) | Rural-follow-up (year 2017) | ||||||
|---|---|---|---|---|---|---|---|---|
| A | D | B, C, E, F, G, H | Total | A | D | B, C, E, F, G, H | Total | |
| Participants studied | ||||||||
| Baseline | 521 | 464 | 231 | 1216 | 342 | 631 | 287 | 1260 |
| Follow-up | 290 (55.7) | 209 (45.0) | 84 (36.4) | 583 (47.9) | 255 (74.6) | 403 (63.9) | 145 (50.5) | 803 (63.7) |
| Sample Loss Total | 231 (44.3) | 255 (55.0) | 147 (63.6) | 633 (52.1) | 87 (25.4) | 228 (36.1) | 142 (49.5) | 457 (36.3) |
| Death | ||||||||
| Male | 87 (16.7) | 95 (20.5) | 49 (21.2) | 231 (19.0) | 49 (14.3) | 114 (18.1) | 54 (18.8) | 217 (17.2) |
| Female | 62 (11.9) | 88 (19.0) | 47 (20.4) | 197 (16.2) | 31 (9.1) | 90 (14.3) | 68 (23.7) | 189 (15.0) |
| Total | 149 (28.6) | 183 (39.4) | 96 (41.6) | 428 (35.2) | 80 (23.4) | 204 (32.3) | 122 (42.5) | 406 (32.2) |
| Migration | ||||||||
| Male | 22 (4.2) | 19 (4.1) | 12 (5.2) | 53 (4.4) | 3 (0.9) | 6 (1.0) | 8 (2.8) | 17 (1.4) |
| Female | 31 (6.0) | 26 (5.6) | 21 (9.1) | 78 (6.4) | 4 (1.2) | 18 (2.9) | 12 (4.2) | 34 (2.7) |
| Total | 53 (10.2) | 45 (9.7) | 33 (14.3) | 131 (10.8) | 7 (2.1) | 24 (3.8) | 20 (7.0) | 51 (4.1) |
| Consent withdrew | ||||||||
| Male | 14 (2.7) | 13 (2.8) | 7 (3.0) | 34 (2.8) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) |
| Female | 15 (2.9) | 14 (3.0) | 11 (4.8) | 40 (3.3) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) |
| Total | 29 (5.6) | 27 (5.8) | 18 (7.8) | 74 (6.1) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) |
aA – Normal group (participants without discernable abnormality on physical, NP, or cognitive status); B – NP group-NP (participants having diagnosable NP disorders only other than cognitive disorders); C – Cognitive disorder group (participants having diagnosable organic disorders only); D – participants with any physical illness only; E – B+C; F – B+D; G – C+D; H – B+C + D. NP – Neuropsychiatric
Table 1 reveals that over 9 years’ sample attrition rate in urban population was 52.1% whereas, in rural population, it was slightly less, i.e., 36.3%. It is visible that the sample reduction rate is variable in all categories from 26% to 63%. The reported reasons of decrease in number and proportion of baseline study participants were death, migration, and withdrawal of consent by the study participants. Over 9 years, the death rate was found to be marginally high in urban areas (35.2%) as compared to rural (32.2%). In both areas, majority of participants died in the ill group. It also shows that slightly more death was in urban area than rural (U = 35.2%; R = 32.2%); migration was also more in urban area (10.2%) than rural (4.1%). However, consent was withdrawn only in urban area (6.1%). Comparatively, the low death rate was found in healthy group (26.5%) than unhealthy group (37.5%). Further, in deceased males outnumbered (U = 19.0%, R = 17.2%) females in both of the study areas (U = 16.2%, R = 15.0%).
It is also observable that relatively high proportion of deaths occurred in urban area both in physically ill and healthy group than rural. However, in the group of mentally ill elderly and elderly with two or more morbidity (B, C, E, F, G, H) condition, the death rate was found to be marginally high in the rural area.
Table 2 reveals stated causes of deaths in urban and rural areas. In both areas, reported cause for death was almost comparable. Majority of the family members reported that the death was natural (as the family members/neighbors reported that it was due to age) (U = 20.8%; R = 29.5%). The second leading cause of death was reported to be chronic obstructive pulmonary disease (COPD) (U = 17.5%; R = 15.8%) followed by cardiac arrest (U = 15%; R = 19.5%), brain strokes (U = 11.0%; R = 6.2%), and multiple organ failure (U = 10.5%; R = 4.2%). Psychiatric illness was reported to be the least accountable reason for death in both of the areas.
Table 2.
Reported cause of death in urban and rural areas (%)
| Causes of death | Urban area (year 2016) | Rural area (year 2017) | ||||||
|---|---|---|---|---|---|---|---|---|
| A (healthy) | D (medically ill) | B, C, E, F, G, H (NP/+org/+med) | Total | A (healthy) | D (medically ill) | B, C, E, F, G, H (NP/+org/+med) | Total | |
| Accident | 7 (1.6) | 9 (2.1) | 5 (1.2) | 21 (4.9) | 4 (1.0) | 5 (1.2) | 5 (1.2) | 14 (3.4) |
| Neurological/hemorrhagic | 18 (4.2) | 17 (4.0) | 12 (2.8) | 47 (11.0) | 3 (0.7) | 12 (3.0) | 10 (2.5) | 25 (6.2) |
| Cancer | 11 (2.6) | 10 (2.3) | 4 (0.9) | 25 (5.8) | 11 (2.7) | 13 (3.2) | 9 (2.2) | 33 (8.1) |
| Cardiac/heart attack | 24 (5.6) | 32 (7.5) | 8 (1.9) | 64 (15.0) | 18 (4.4) | 38 (9.4) | 23 (5.7) | 79 (19.5) |
| Diabetic + HBP/+ kidney | 4 (0.9) | 7 (1.6) | 3 (0.7) | 14 (3.3) | 3 (0.7) | 6 (1.5) | 3 (0.7) | 12 (2.9) |
| MOF | 11 (2.6) | 19 (4.4) | 15 (3.5) | 45 (10.5) | 6 (1.5) | 7 (1.7) | 4 (1.0) | 17 (4.2) |
| Natural death-aging | 31 (7.2) | 39 (9.1) | 19 (4.4) | 89 (20.8) | 18 (4.4) | 58 (14.3) | 44 (10.8) | 120 (29.5) |
| Psychiatric illness | 0 (0.0) | 1 (0.2) | 5 (1.2) | 6 (1.4) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) |
| Other illness | 5 (1.2) | 1 (0.2) | 1 (0.2) | 7 (1.6) | 8 (2.0) | 12 (3.0) | 9 (2.2) | 29 (7.1) |
| Not known | 16 (3.7) | 11 (2.6) | 8 (1.9) | 35 (8.2) | 5 (1.2) | 6 (1.5) | 2 (0.5) | 13 (3.2) |
| COPD | 22 (5.1) | 37 (8.7) | 16 (3.7) | 75 (17.5) | 4 (1.0) | 47 (11.6) | 13 (3.2) | 64 (15.8) |
| Total death (percentage of deaths in the group) | 149 (28.6) | 183 (39.4) | 96 (41.6) | 428 (35.2) | 80 (23.4) | 204 (32.3) | 122 (42.5) | 406 (32.2) |
DISCUSSION
This study reveals around 36%–52% of sample loss, the reason may be long gap between baseline and follow-up study. This is comparable to earlier studies.[1] Loss to follow-up found to be variable by demographic or other risk factors and little is known about the magnitude and direction of bias from loss to follow-up.[10] Further, death was the major cause of sample loss. Majority of death was reported from ill group of participants, which is comparable to the previous study; it is reported that age, sex, and functional disability are strong and consistent predictors of mortality in community-dwelling older adults.[9] Another study reports that respondents are reporting “fair” and “good” health also show elevated risks of mortality.[11] Self-perceptions of health status appear to be a factor of unique prospective significance in mortality studies.[12]
More males were found to be at risk for death than females, which favors longevity in female and favors feminization in elderly population.[13] It is reported that most of the states of India are realizing increased proportion of elderly with drastic increase in the proportion of elderly female population in comparison to their male counterpart.[14,15] In the year 2012, on the international day of older persons The Hindu reported that by 2050, women over 60 years would exceed the number of elderly men by 18.4 million, which would result in a unique characteristic of “feminization” of the elderly population in India.
Cause of death was reported to be natural (old age). The second leading cause of death was COPD (U = 17.5%; R = 15.8%), followed by cardiac arrest (U = 15%; R = 19.5%) supported by latest report,[16] brain strokes (U = 11.0%; R = 6.2%), and multiple organ failure (U = 10.5%; R = 4.2%). Psychiatric illness was reported to be the least accountable reason for death in both of the areas, which is in support of the earlier report.[9] Information also reveals that majority of death in males was due to heart attack, supports the Government of India report which reveals that in males most death occurs due to cardiac arrest than females.[17] Studies reported that unhealthy conditions as one of the significant reasons for mortality among hospitalized older patients.[18]
CONCLUSION AND IMPLICATIONS
This study reveals that over a period of 9 years chances of cohort loss due to mortality is about 32%–35% in community-dwelling urban and rural elderly. The risk of mortality in old age is more in males than females. Male is more at risk for developing cardiac problems, and majority females died due to brain hemorrhage/coma/paralysis.
Countrywide data and follow-up studies related to mortality enable to forecasts about death in elderly in an accurate, simple, and realistic way. Such studies will also enable to reveal age-specific mortality rate vis-a-vis cause and pattern of death over a particular period. It will also permit the researchers to develop follow-up studies on a particular sample/cohort with accurate calculations for financial support, which further will help in figuring out the resources too.
Limitations
This study has its own limitations such as whatever information was given by the informants (offsprings of the subjects/neighbor) is the only way to collect the details. No one wants to show any document related to their expired relative or wish to talk more on the issue; further, if a person died long back and the information is provided by the neighbor that may not be accurate. Sometimes, the family of deceased says that in old age everyone has to go, what will be the use of knowing it.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
Acknowledgment
The research is funded by the Indian Council of Medical Research and authors are grateful for the funding support given by the Council. The authors also wish to thank the entire research team, including Dr. Alok Kumar Chowdhury, Mr. Arunesh Tiwari, Dr. Pradeep, Ms. Chitra Srivastava, Ms. Ritu Shukla, Mr. Kalpjeet Sinha, Ms. Sarita Pal, and Mr. Shafeeq of the follow-up project, who put their full efforts for getting the data in hand. At last but not least, the authors wish to thank the participants of the study without whom the study could not be done.
REFERENCES
- 1.Gustavson K, von Soest T, Karevold E, Røysamb E. Attrition and generalizability in longitudinal studies: Findings from a 15-year population-based study and a Monte Carlo simulation study. BMC Public Health. 2012;12:918. doi: 10.1186/1471-2458-12-918. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Young AF, Powers JR, Bell SL. Attrition in longitudinal studies: Who do you lose? Aust N Z J Public Health. 2006;30:353–61. doi: 10.1111/j.1467-842x.2006.tb00849.x. [DOI] [PubMed] [Google Scholar]
- 3.Stanziano DC, Whitehurst M, Graham P, Roos BA. A review of selected longitudinal studies on aging: Past findings and future directions. J Am Geriatr Soc. 2010;58(Suppl 2):S292–7. doi: 10.1111/j.1532-5415.2010.02936.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Anstey KJ, Luszcz MA, Giles LC, Andrews GR. Demographic, health, cognitive, and sensory variables as predictors of mortality in very old adults. Psychol Aging. 2001;16:3–11. doi: 10.1037/0882-7974.16.1.3. [DOI] [PubMed] [Google Scholar]
- 5.Vuorisalmi M, Lintonen T, Jylhä M. Global self-rated health data from a longitudinal study predicted mortality better than comparative self-rated health in old age. J Clin Epidemiol. 2005;58:680–7. doi: 10.1016/j.jclinepi.2004.11.025. [DOI] [PubMed] [Google Scholar]
- 6.Beeharry G, Whiteford H, Chambers D, Baingana F. Outlining the Scope for Public Sector. 2002 [Google Scholar]
- 7.Thong MS, Kaptein AA, Benyamini Y, Krediet RT, Boeschoten EW, Dekker FW, et al. Association between a self-rated health question and mortality in young and old dialysis patients: A cohort study. Am J Kidney Dis. 2008;52:111–7. doi: 10.1053/j.ajkd.2008.04.001. [DOI] [PubMed] [Google Scholar]
- 8.Forcey DS, Walker SM, Vodstrcil LA, Fairley CK, Bilardi JE, Law M, et al. Factors associated with participation and attrition in a longitudinal study of bacterial vaginosis in Australian women who have sex with women. PLoS One. 2014;9:e113452. doi: 10.1371/journal.pone.0113452. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Ganguli M, Dodge HH, Mulsant BH. Rates and predictors of mortality in an aging, rural, community-based cohort: The role of depression. Arch Gen Psychiatry. 2002;59:1046–52. doi: 10.1001/archpsyc.59.11.1046. [DOI] [PubMed] [Google Scholar]
- 10.Tin Tin S, Woodward A, Ameratunga S. Estimating bias from loss to follow-up in a prospective cohort study of bicycle crash injuries. Inj Prev. 2014;20:322–9. doi: 10.1136/injuryprev-2013-040997. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Idler EL, Kasl SV, Lemke JH. Self-evaluated health and mortality among the elderly in New Haven, Connecticut, and Iowa and Washington counties, Iowa, 1982-1986. Am J Epidemiol. 1990;131:91–103. doi: 10.1093/oxfordjournals.aje.a115489. [DOI] [PubMed] [Google Scholar]
- 12.Menec VH, Chipperfield JG, Perry RP. Self-Perceptions of health. Prospect Anal Mortal Control Health. 2018;54:85–93. [Google Scholar]
- 13.Bhatnagar GV. India Needs to Start Addressing Issues Concerning Its Growing Elderly Population, Says UN. 2017. Jun 20, [Last cited on 2018 Aug 08]. Available from:https://thewire.in/politics/elderly-population-demographics-india .
- 14.Venkatesh S, Vanishree M. Title: Feminization among elderly population in India: Role of micro financial institutions. Glob J Finance Manag 2. 2014;6:897–906. [Google Scholar]
- 15.Chakrabarti S, Sarkar A. Pattern and trend of population ageing in India. Indian J Spat Sci. 2011;2:1. [Google Scholar]
- 16.Ayesta A, Martínez-Sellés H, Bayés de Luna A, Martínez-Sellés M. Prediction of sudden death in elderly patients with heart failure. Journal of geriatric cardiology: JGC. 2018;15:185–92. doi: 10.11909/j.issn.1671-5411.2018.02.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Office of the Registrar General of India. Report on causes of death in India 2001-2003. 2009. [Last retrieved on 2018 Dec 28]. Available from: http://www.cghr.org/wordpress/wp-content/uploads/Causes_of_death_2001-03.pdf .
- 18.Drame M, Jovenin N, Novella JL, Lang PO, Somme D, Laniece I, et al. Predicting early mortality among elderly patients hospitalised in medical wards via emergency department: The SAFES cohort study. J Nutr Health Aging. 2008;12:599–604. doi: 10.1007/BF02983207. [DOI] [PubMed] [Google Scholar]