• Communications for Statistical Applications and Methods
• /
• 제27권6호
• /
• pp.675-688
• /
• 2020

In survival analysis of observational data, the inverse probability weighting method and the Cox proportional hazards model are widely used when estimating the causal effects of multiple-valued treatment. In this paper, the two kinds of weights have been examined in the inverse probability weighting method. We explain the reason why the stabilized weight is more appropriate when an inverse probability weighting method using the generalized propensity score is applied. We also emphasize that a marginal hazard ratio and the conditional hazard ratio should be distinguished when defining the hazard ratio as a treatment effect under the Cox proportional hazards model. A simulation study based on real data is conducted to provide concrete numerical evidence.

• 한국물환경학회지
• /
• 제23권1호
• /
• pp.87-96
• /
• 2007

In this paper a methodology of identifying individual pipes according to the internal and external characteristics of pipe is developed, and the methodology is applied to a case study water distribution pipe break database. Using the newly defined individual pipes the hazard rates of the cast iron 6 inch pipes are modeled by implementing the proportional hazards modeling approach for consecutive pipe failures. The covariates to be considered in the modeling procedures are selected by considering the general availability of the data and the practical applicability of the modeling results. The individual cast iron 6 inch pipes are categorized into seven ordered survival time groups according to the total number of breaks recorded in a pipe to construct distinct proportional hazard model (PHM) for each survival time group (STG). The modeling results show that all of the PHMs have the hazard rate forms of the Weibull distribution. In addition, the estimated baseline survivor functions show that the survival probabilities of the STGs generally decrease as the number of break increases. It is found that STG I has an increasing hazard rate whereas the other STGs have decreasing hazard rates. Regarding the first failure the hazard ratio of spun-rigid and spun-flex cast iron pipes to pit cast iron pipes is estimated as 1.8 and 6.3, respectively. For the second or more failures the relative effects of pipe material/joint type on failure were not conclusive. The degree of land development affected pipe failure for STGs I, II, and V, and the average hazard ratio was estimated as 1.8. The effects of length on failure decreased as more breaks occur and the population in a GRID affected the hazard rate of the first pipe failure.

• Communications for Statistical Applications and Methods
• /
• 제18권4호
• /
• pp.485-493
• /
• 2011

The proportional likelihood ratio order is an extension of the likelihood ratio order for the non-negative absolutely continuous random variables. In addition, the Lindley distribution has been over looked as a mixture of two exponential distributions due to the popularity of the exponential distribution. In this paper, we first recalled the above concepts and then obtained various properties of the Lindley distribution due to the proportional likelihood ratio order. These results are more general than the likelihood ratio ordering aspects related to this distribution. Finally, we discussed the proportional likelihood ratio ordering in view of the weighted version of the Lindley distribution.

• Communications for Statistical Applications and Methods
• /
• 제24권6호
• /
• pp.583-604
• /
• 2017

The most popular regression model for the analysis of time-to-event data is the Cox proportional hazards model. While the model specifies a parametric relationship between the hazard function and the predictor variables, there is no specification regarding the form of the baseline hazard function. A critical assumption of the Cox model, however, is the proportional hazards assumption: when the predictor variables do not vary over time, the hazard ratio comparing any two observations is constant with respect to time. Therefore, to perform credible estimation and inference, one must first assess whether the proportional hazards assumption is reasonable. As with other regression techniques, it is also essential to examine whether appropriate functional forms of the predictor variables have been used, and whether there are any outlying or influential observations. This article reviews diagnostic methods for assessing goodness-of-fit for the Cox proportional hazards model. We illustrate these methods with a case-study using available R functions, and provide complete R code for a simulated example as a supplement.

• Journal of the Korean Data and Information Science Society
• /
• 제17권4호
• /
• pp.1067-1076
• /
• 2006

Although the proportional hazards model is the most common approach used for studying the relationship of event times and covariates, alternative models are needed for occasions when it does not fit data. In the two-sample case, proportional odds models are useful for fitting data whose hazard rates converge asymptotically. In this thesis, we propose a new estimator of the relative odds ratio of the proportional odds model when two independent random samples are observed under uncensorship. We prove the asymptotic normality and consistency of the estimator by using martingale-representation. The efficiency of the proposed is assessed through a simulation study.

• Journal of the Korean Data and Information Science Society
• /
• 제12권2호
• /
• pp.125-143
• /
• 2001

We, in this paper, propose a kernel estimator of hazard ratio with censored survival data. The uniform consistency and asymptotic normality of the proposed estimator are proved by using counting process approach. In order to assess the performance of the proposed estimator, we compare the kernel estimator with Cox estimator and the generalized rank estimators of hazard ratio in terms of MSE by Monte Carlo simulation. Two examples are illustrated for our results.

• Journal of the Korean Data and Information Science Society
• /
• 제24권4호
• /
• pp.723-736
• /
• 2013

해약율의 추정은 최근 보험제도의 변화 (국제회계기준의 도입에 따른 현금흐름방식의 가격산출체계 시행, 무해약환급금 보험상품의 판매 허용 등)에 따라 보험료의 결정과 손익분석 그리고 리스크 관리 등에 있어서 중요한 요소로 부각되고 있다. 특히, 변액연금은 최저보증옵션으로 인하여 보험계약자의 해약요소가 중요시되고 다른 보험 상품에 비해 복잡하므로 차별성 있는 통계모형의 선택과 분석이 필요하다. 기존의 해약률 연구는 실태분석 또는 회귀분석을 위주로 모형화하는 것에 초점이 맞추어져 있었으나 본 연구에서는 변액연금 계약과 관련된 여러 변수와 최저보증옵션을 반영하기 위하여 생존분석기법 중 하나인 Cox 비례위험모형을 이용하여 해지율을 추정하였다. 변액연금 해지율에 영향을 미치는 주요변수로는 납입방법, 보험료, 보험기간 대비 유지기간, 계약자적립금 대비 최소보증금, 계약자연령이 있으며 본 연구에서는 이에 관하여 분석해보았다.

• 한국노년학
• /
• 제31권2호
• /
• pp.259-276
• /
• 2011

다른 사람의 도움을 받아야 하는 노인도 적절한 재가서비스를 받으면서 가능한 오래 가정에 머물 수 있다면 그들의 삶의 질을 향상시키고 비용을 절감할 수 있다고 알려져 있다. 이 연구는 노인장기요양보험제도를 시행한 후, 재가서비스 이용자가 시설서비스로 전환하여 입소하는 것에 대한 결정요인을 살펴보고자 한 종적 연구이다. 이 연구는 장기요양급여 이용자와 주수발자를 대상으로 한 '2009년 노인장기요양서비스 만족도 조사' 자료를 활용하였다. 이 연구의 분석대상자는 총 1,230명으로, 이들은 시도별, 장기요양등급별, 의료보장형태별로 비례할당으로 추출된 것이다. 종속변수는 연구대상자가 재가서비스를 이용하다가 시설로 입소하는 데까지 걸린 시간이고, 이것은 노인장기요양급여DB를 연계하여 파악하였다. 또한 이 연구는 Cox Proportional Hazard Model을 이용하여 hazard ratio를 산출하였다. 연구결과에 의하면, 다른 변수들을 보정한 후에도 이용자가 골절을 앓고 있는 경우에는 시설에 입소할 확률이 유의하게 높은 것으로 나타났다. 또한 통계적으로 유의하지는 않았지만 저하된 인지기능의 항목수가 많을수록 시설에 입소할 확률이 높은 것으로 나타났다. 이러한 영향요인들을 감안하여 향후에는 이용자의 욕구를 충족시키는 재가서비스 이용체계를 구축하고 보완하는 정책개발이 필요하다.

• 한국산학기술학회논문지
• /
• 제13권5호
• /
• pp.2219-2231
• /
• 2012

본 연구에서는 우리나라 30세 이상 성인들의 허혈성심장질환(ischemic heart disease, IHD) 발생률을 계산하고, 지질지표(콜레스테롤, 중성지방, 고밀도지단백콜레스테롤, 저밀도지단백콜레스테롤)가 허혈성심장질환 발생에 미치는 위험도를 파악하고자 19개 대학 및 종합병원의 건강검진센터에서 검진을 받은 417,642명을 대상으로 1993년 9월부터 2009년 6월까지 평균 8.5년동안 허혈성심장질환의 발생을 추적관찰하였다. 자료수집은 검진자들에 대한 설문조사지를 이용하였고, 허혈성심장질환의 발생여부는 국민건강보험공단의 데이터베이스에서 확인하였다. 발생률은 발생밀도로 계산하였고, 혈청지질 지표에 따른 허혈성심장질환의 발생 위험도는 콕스의 비례위험 회귀모형을 이용하여 연령, BMI, 생활양식을 보정한 상태에서 성별에 따른 위험요인별 위험비와 95% 신뢰구간을 계산하였다. 연구결과 TC/HDL 비의 증가에 따라 IHD의 발생 위험비는 남자에서 1.21배에서 1.84배까지, 여자는 1.26배에서 1.86배까지 증가하였으며, TG/HDL 비의 증가에 따른 IHD의 발생 위험비는 남자에서 1.17배에서 1.49배까지, 여자는 1.42배에서 1.97배까지, LDL/HDL 비에 따라 IHD의 발생 위험비는 남자에서 1.26배에서 1.82배까지, 여자는 1.26배에서 1.68배까지 증가하였다. 결론적으로 혈청지질지표는 심혈관질환의 중요한 위험요인으로 총콜레스테롤, 저밀도지단백콜레스테롤, 중성지방은 혈중 농도가 높을수록, 고밀도지단백콜레스테롤은 낮을수록 IHD의 위험이 높아지는 것으로 나타났고, TC/HDL 비, TG/HDL 비, LDL/HDL 비에서 단독의 지질지표보다 위험도가 더 높게 나타나는 경향이 있었다. 따라서 추후 허혈성 심장질환의 예방 및 관리에는 혈청지질지표의 비도 감안하여야 한다.

• 농촌의학ㆍ지역보건
• /
• 제32권1호
• /
• pp.13-26
• /
• 2007

Objectives: This study was conducted to investigate the association between health risk factors and mortality in Juam cohort. Methods: The subjects were 1,447 males and 1,889 females who had been followed up for 68.5 months to 1 January 2001. Whether they were alive or not was confirmed by the mortality data of the National Statistical Office. A total of 289 persons among them died during the follow-up period. The Cox's proportional hazard regression model was used for survival analysis. Results: Age, type of medical insurance, self cognitive health level, habit of alcohol drinking, smoking, exercise and BMI level were included in Cox's proportional hazard model by gender. The hazard ratio of age was 1.07(95% CI: 1.05-1.10) in men, 1.09(95% CI: 1.06-1.12) in women. The hazard ratio of medical aid(lower socioeconomic state) was 1.43(95% CI 1.02-2.19) in women. The hazard ratios of current alcohol drinking and current smoking were respectively 1.69(95% CI: 1.01-2.98), 1.52(95% CI: 1.02-2.28) in women. The hazard ratio of underweight was 1.56(95% CI 1.08-2.47) in men. The hazard ratios of underweight, normoweight, overweight, and obesity were respectively 1.63(95% CI: 1.02-2.67), 1.0(referent), 0.62(95% CI: 0.32-1.63), 1.27(95% CI: 0.65-3.06), which supported the U-shaped relationship between body mass index and mortality among the men over 65. Conclusions: The health risk factors increasing mortality were age, underweight in male, age, lower socioeconomic state, current alcohol drinking, current smoking in female. To evaluate long-term association between health risk factors and mortality, further studies need to be carried out.