Extended Abstract
1. Objectives
Because the elderly population is increasing in Iran, awareness of various causes of death in the elderly is necessary. Certainly, with an increase in the number of elderly people, the mortality rate in the community will be on the rise, followed by the increase in mortality rates in this age group [1]. 
The present study aims to identify the risk factors that reduce the survival time of the elderly so that preventive measures could be considered in the identification, clinical trials, and therapeutic measures to eliminate serious risks and increase survival time for the elderly. The main purpose of this study was fitting a semi-parametric blended model of survival analysis with competing risks for the elderly living in the nursing home and then estimating the parameters. Deaths in the elderly people due to cardiovascular diseases and other causes have been considered as competing risks.
2. Methods & Materials
The research method was retrospective. Analysis of data was performed by studying 510 elderly people over 60 years of age, who were admitted to Kashan’s Golabchi nursing home from 2000 to late 2012. Independent variables related to the time to death of the elderly included gender, age at the start of admission, blood pressure, blood lipids, mobility status, history of myocardial infarction, history of stroke, and kidney problems. The dependent variable was the length of stay at the nursing home, which was calculated from the difference between the admission date and the discharge date [2]. Heterogeneity of patients is often ignored in the analysis of medical data. On the other hand, individual treatment is often important in medical sciences; however, the use of blended statistical models to analyze data related to a sample of the heterogeneous population can lead us to a proper analysis [3]. Another feature of this method is the lack of the need for the presumption of the independence of competing risks and the simultaneous estimation of parameters [4-6]. A semi-parametric blended model of survival analysis with competing risks was used to analyze the data, and for this the expectation-conditional maximization (ECM) algorithm was applied to estimate the model parameters [7].
In the model, the decision criterion for the significance of indicators of odds ratio and risk ratio is confidence intervals. The collected data were first analyzed separately as single-variable for each independent variable and again as multivariate by entering all independent variables in the model. The target incident was the death of the elderly, and those elderly who had not died by the end of the study were considered as censored data based on the time variable. Deaths due to cardiac and non-cardiac causes were defined as competing risks. The present research was approved by the ethics committee of the University of Social Well-being and Rehabilitation Sciences (IR.USWR.REC.1394.241). Data analysis was performed using R 3-3-3 software.
3. Results
Of 510 elderly people living in Golabchi nursing home of Kashan, 29% (148 individuals) were men, and 71% (362 individuals) were women. Also, 19.4% (99 cases) of them had died of cardiac causes and 44.1% (225 cases) due to non-cardiac causes. The remaining 36.5% of the elderly (186 people) were either survived or discharged by the end of the study, which was considered as censored data in the study. The median age of the elderly was 11.18 years, which was used as the grouping level in the age variable. The mean age of the elderly was 80.35 years with a standard deviation of 18.19 years.
In the single-variable fitting of the models, factors such as blood lipids {(CI=1.00 / 1, 1.31); =1.04}, history of myocardial infarction {(CI=1.04, 1.10) =CI; 90.0=}, history of stroke {(CI=1.00, 1.14); =0.95} had a significant effect on the time to death of elderly with cardiovascular diseases. The coefficients derived from the univariate models in the estimation of the target incidence ratio, the chance of dying of cardiac diseases in the elderly men were 60% more than the elderly women (OR=0.66), with a confidence interval for this odds ratio CI=0.55, 1.03. The risk of death due to cardiac diseases in elderly patients with the history of myocardial infarction was 2.61 times more than others (2.61=OR experience of myocardial infarction). In a single-variable model, for an elderly person with a history of heart attack, the probability of death due to cardiac causes is 0.52, and the probability of non-cardiac death in an elderly person with a history of heart attack is 48.0.
In the fitting of the multivariate model (fitting a model with 8 independent variables simultaneously), renal problems have a significant effect (CI=1.77, 2.83; =1.58). The probability of cardiac and non-cardiac deaths was 0.17 and 0.83, respectively, in the elderly men under 81.11 years old who have high blood fat along with abnormal motor status, high blood pressure, and no history of myocardial infarction and renal failure. Also, the probability of death due to cardiac causes in elderly women with a minimum age of 81.11 years who have high blood fat along with abnormal motor status and high blood pressure and have no history of myocardial infarction and renal failure. It is while the probability of non-cardiac deaths in these women is 0.41. 
In the fitting of a multivariate model with the constant effect of other variables, the probability of cardiac deaths is lower in men than in women (OR=0.79), with a confidence interval of 0.84, 0.97. In addition, with the constant effect of other variables, the risk of cardiac deaths in elderly with renal problems is 1. 58 times greater than the elderly without renal problem (=1.88) with a confidence interval of 1.77, 2.83.
4. Conclusion
In single-variable fittings, the effects of the factors including age, history of myocardial infarction, history of stroke, and renal problems on time to death of the elderly were identified. Also, the results of multivariate analysis showed that with the constant effect of other variables, the renal problems variable has a significant effect on the time to death of the elderly living in nursing home. Therefore, it is recommended to consider preventive processes in the identification, clinical trials, and therapeutic measures to eliminate serious risks and to increase survival time for the elderly.
In the current aging society, proper planning for preventing renal and motor problems as well as good nutrition can help the quality of life of the elderly. One of the limitations of this study was the lack of accurate patient information, which might have resulted in the insignificant effects of some important clinical variables.
Acknowledgments
This research was extracted from the MSc. thesis of the first author in the Department of Biostatistics, University of Social Welfare and Rehabilitation Sciences, Tehran.
Conflict of Interest
The authors declared no conflicts of interest.
 

1. Ameri G, Govari F, Nazari T, Rashidinejad M, Afsharzadeh P. [The adult age theories and definitions (Persian)]. Hayat. 2002; 8(1):4-13. 
2. Farzianpour F. Arab M. Foroushani A. Morad Zali E. Evaluation of the criteria for quality of life of elderly health care centers in Tehran Province, Iran. Global Journal of Health Science. 8(7):68–76. doi: 10.5539/gjhs.v8n7p68 
3. Goharinezhad S. Maleki M. Baradaran HR, Ravaghi H. A qualitative study of the current situation of elderly care in Iran: what can we do for the future? Global Health Action. 2016; 9(1):32156. doi: 10.3402/gha.v9.32156
4. Noroozian M. The Elderly population in Iran. An ever growing concern in the health system. Iranian Journal of Psychiatry and Behavioral Sciences. 2012; 6(2):1-6.
5. Carroll W, Miller GE. Heart disease among elderly Americans. Estimates for the U.S. Civilian Noninstitutionalized Population, Statistical Brief 409. Rockville: Agency for Healthcare Research and Quality; 2013.
6. Farkhani EM, Baneshi MR, Zolala F. [Survival rate and its related factors in patients with acute myocardial infarction (Persian)]. Medical Journal of Mashhad University of Medical Sciences. 2014; 57(4):636–46.
7. Beyranvand MR, Lorvand A, Alipour Parsa S, Motamedi MR, Kolahi AA. [The quality of life after first acute myocardial infarction (Persian)]. Pajoohande. 2011; 15(6):264-72.
8. Andrawes WF, Bussy C, Belmin J.. Prevention of cardiovascular events in elderly people. Drugs & Aging. 2005; 22(10):859-76. doi: 10.2165/00002512-200522100-00005
9. Sharifirad GhR. Mohebbi S. Matlabi M. [The relationship of physical activity in middle age and cardiovascular problems in old age in retired people in Isfahan, 2006 (Persian)]. Ofogh-e-Danesh. 2007; 13(2):57-63.
10. Taghavi M. [Mortality rate in 23 province of the country in 2003 (Persian)]. Tehran: Ministry of Health; 2005.
11. Parkash R, Choudhary SK, Singh US. A study of morbidity pattern among geriatric population in an urban Area of Udaipur Rajasthan. 2004; 29(1):35-40.
12. Haller B. The analysis of competing risks data with a focus on estimation of cause-specific and sub-distribution hazard ratios from a mixture model [PhD thesis]. München: University of München; 2014.
13. Choi S, Huang X. Maximum likelihood estimation of semiparametric mixture component models for competing risks data. Biometric Methodology. 2014; 70(3):588-98. doi: 10.1111/biom.12167
14. McLachlan G, Peel D. Finite mixture model. Hoboken, New Jersey: John Wiley & Sons; 2000.
15. Abu Bakar MR, Daud I, Ibrahim NA, Rahmatina D. Estimating a Logistic Weibull Mixture Models with Long–Term Survivors. Jurnal Teknologi. 2006;45(1):57-66. doi: 10.11113/jt.v45.323
16. Azadchehr M, Rahgozar M, Karimloo M, Adib Haj Bageri M. [To identify some factors effective on survival of the elderly living in nursing home using Copula Competing Risk Model: Bayesian approach (Persian)]. Journal of Health Promotion Management. 2014; 3(4):46-55
17. Schlattmann P. Medical application of finite mixture model. Berlin: Springer; 2009.
18. Kuk AYC. A semi-parametric mixture model for the analysis of competing risks data. Australian Journal of Statistics. 1992; 34(2):169–80. doi: 10.1111/j.1467-842x.1992.tb01351.x
19. Ng SK, McLachlan GJ. An EM-based semi-parametric mixture model approach to the regression analysis of competing-risks data. Statistics in Medicine. 2003; 22(7):1097–111. doi: 10.1002/sim.1371
20. Thomas B, Matsushita K, Abate KH, Al-Aly Z, Ärnlöv J, Asayama K et al. Global cardiovascular and renal outcomes of reduced GFR. Journal of the American Society of Nephrology. 2017; 28(7):2167-79. doi: 10.1681/ASN.2016050562.
21. Mahjoub H, Rusinaru D, Soulière V, Durier C, Peltier M, Tribouilloy C. Long-term survival in patients older than 80 years hospitalised for heart failure. A 5-year prospective study. European Journal of Heart Failure. 2008; 10(1):78–84. doi: 10.1016/j.ejheart.2007.11.004 
22. Mosavy S, Soltanian AR, Roshanaei GH, Fardmal J. [Applying Aalen’s additive hazard model for analyzing 5-year survival of acute myocardial infarction patients in Bushehr (Persian)]. Journal of North Khorasan University of Medical Science. 2011; 3:161-69.
23. Nazari R, Haghdost AA, Rezaie R, Sa'atsaz S, Chan YH et al. Difference in clinical symptoms of myocardial infarction between men and women. Iranian Journal of Critical Care Nursing. 2011; 2011; 4(1):e7081.
24. Frost PH, Davis BR, Burlando AJ, David Curb J, Guthrie GP, Isaacsohn JL, et al. Coronary heart disease risk factors in men and women aged 60 years and older findings from the systolic hypertension in the elderly program. Circulation. 1996; 94(1):26-34. doi: 10.1161/01.cir.94.1.26
25. Frost PH, Davis BR, Burlando AJ, Curb JD, Guthrie GP, Isaacsohn JL, et al. Coronary heart disease risk factors in men and women aged 60 years and older: ّindings from the systolic hypertension in the elderly program. Circulation. 1996; 94(1):26-34. doi: 10.1161/01.cir.94.1.26