Lifetime Data and Concepts
Prabhanjan Narayanachar Tattar, H. J. Vaman in Survival Analysis, 2022
With time to event data and indicators of whether the observations are complete or not, we have the useful nonparametric methods of inference concerning the survival curves. For instance, we can estimate the survival function or the cumulative hazard function using nonparametric methods. The survival function is estimated using the famous Kaplan-Meier estimator. Kaplan-Meier is often listed as one of the twenty important methods of the previous century. It provides an estimate of , denoted by and it is often reliable until the time point of the last complete observations. It is also known as the product-limit estimator. The justification of the name follows later.
Survival Analysis and Cox Regression
K. V. S. Sarma, R. Vishnu Vardhan in Multivariate Statistics Made Simple, 2018
Survival analysis is a statistical method of estimating the time-to- happening of an event. When the event of interest is dichotomous, we wish to estimate the hazard rate (instantaneous risk of event at a given time point) and the pattern of survival chance in terms of duration. It is used to predict the survival time of patients after a treatment or duration of disease-free survival etc. The Kaplan-Meier method is one popular tool to estimate the mean/median survival time and it is also used to compare the survival pattern between treatment groups. When more than one factor is likely to influence the survival, we can use the Cox regression (proportional hazard) model and estimate the resulting relative risk (of event). Both Kaplan-Meier and Cox regression are computer-intensive methods and software helps in quick and reliable results. Survival analysis is applicable even in the non-clinical context such as insurance, aircraft maintenance etc.
Multiple Measures of Outcome in Assessing a Prison-Based Drug Treatment Program
Nathaniel J. Pallone in Treating Substance Abusers in Correctional Contexts: New Understandings, New Modalities, 2012
For the variables “days to first illegal activity,” “days to first reincarceration,” and “days to first drug use,” we performed survival analyses using the Kaplan-Meier method. The advantages of this method are that it does not require a normal distribution (the time to the occurrence of an event is usually not normally distributed) and that it allows the inclusion of cases that did not engage in illegal activities or who were not reincarcerated. Scores for subjects reporting no illegal activity or drug use were censored at 365 days. Likewise, reincarceration data was censored at 365 days. These so-called “censored” cases are cases in which the event did not occur during the period of observation. The Kaplan-Meier method creates a survival curve for each group and tests whether the curves are significantly different from one another using the log-rank statistic. All survival analyses were performed using SPSS, version 10.1. As can be seen from the drug use findings that follow, subjects evidently did not include drug use as an illegal activity (i.e., subjects did not report the same amount of time to first drug use as they did for first illegal activity).
Poor performance status, urban residence and female sex predict inferior survival in pediatric advanced stage mature B-NHL in an Indian tertiary care center
Published in Pediatric Hematology and Oncology, 2018
Amol Patel, Meher Chand Sharma, Saumyaranjan Mallick, Manali Patel, Sameer Bakhshi
For baseline variables, descriptive statistics was used to know the distribution. Chi-square and Fischer's exact test were applied for determination of association between variable wherever required. OS was determined by calculating time from the date of diagnosis till the date of death due to any cause. Event free survival (EFS) was determined by calculating the time from the date of diagnosis till the date of event wherein event was defined as death due to any cause, relapse or progressive disease. The data were censored on December 31, 2016. Kaplan–Meier method has been used for analysis of survival. Log-rank test to evaluate the outcome differences between groups of patients. Cox regression analysis was used for univariate analysis. The significant univariate variables of value upto p < 0.10 were considered for multivariate analysis using Cox regression proportional hazard analysis. Patient database and tracking system was initiated after January 2011 and hence we did time trend analysis before and after this change to see if there was any difference in outcome. STATA software (STATACORP US, Texas) version 13 was used for analysis.
Hodgkin lymphoma in children, adolescents and young adults – a comparative study of clinical presentation and treatment outcome
Published in Acta Oncologica, 2018
Annika Englund, Ingrid Glimelius, Klaus Rostgaard, Karin E. Smedby, Sandra Eloranta, Daniel Molin, Thomas Kuusk, Peter de Nully Brown, Peter Kamper, Henrik Hjalgrim, Gustaf Ljungman, Lisa Lyngsie Hjalgrim
Variation in distribution of clinical characteristics was shown for patients in different age groups (children 0–9 years, adolescents 10–17 years and young adults 18–24 years) and assessed with Pearson’s chi2-test and Mantel–Haenszel trend test (ordinal variables). Treatment outcome analyses were stratified by treatment in pediatric (children: 0–14 years Denmark, 0–17 years Sweden) or adult departments (adults: 15–24 years Denmark, 18–24 years Sweden) and by country (Denmark and Sweden). The Kaplan–Meier method was used to estimate survival. Confidence intervals (CIs) for 5- and 10-year survival were calculated in SAS proc lifetest using the default complementary log–log transformation of survival estimates. Hazard ratios (HR) with 95% CIs were used to compare pediatric and adult departments, Denmark and Sweden, sex, stage, calendar period and first-line treatment concepts. HRs were presented with adjustments for sex, stage (limited/advanced), country and calendar period (before and after year 2000) and calculated using Cox regression with time since diagnosis as the underlying time scale. CIs were based on Wald tests. SAS version 9.4 (SAS Inc., Cary, NC) was used for all analyses.
Long term follow-up of frontline Dasatinib in older patients with chronic myeloid leukemia in chronic phase treated outside clinical trials: a real-life cohort observational study
Published in Acta Oncologica, 2021
Fabio Stagno, Massimo Breccia, Mario Annunziata, Malgorzata Monika Trawinska, Alessandra Iurlo, Nicola Sgherza, Carmen Fava, Antonella Gozzini, Luigiana Luciano, Ida Carmosino, Massimiliano Bonifacio, Federica Sorà, Sabrina Leonetti Crescenzi, Monica Crugnola, Gabriele Gugliotta, Sara Galimberti, Cristina Bucelli, Gioia Colafigli, Costanzo Feo, Mario Tiribelli, Endri Mauro, Antonella Russo Rossi, Attilio Guarini, Elisabetta Abruzzese, Gianantonio Rosti, Francesco Di Raimondo, Roberto Latagliata
Data were expressed as mean ± standard deviation (SD) (normally distributed data), median and interquartile range (IQR) (non-normally distributed data), or as percentage frequencies, and within-patient comparisons were made by unpaired t-test and χ2 test, as appropriate, at significance levels of p < 0.05. All the endpoints of treatment efficacy (CHR, CCyR, and MMR) were calculated as the best response rate at any time. Overall survival (OS) was calculated from the date of DAS treatment start to death related to any cause. Event-free survival (EFS) was calculated from the date of DAS start to any of the following events: primary hematologic or cytogenetic resistance to DAS, permanent DAS discontinuation due to toxicity or any other unrelated cause (excluding discontinuation for a treatment-free remission), secondary hematologic or cytogenetic resistance, CML-related (progression to blast phase or any other cause directly related to CML) and unrelated death (concomitant diseases or any other cause not directly related to CML). Cumulative incidence of progression was calculated from the date of DAS start to any of the following events: primary resistance, secondary resistance, evolution in accelerated/blast phase. Further, DAS discontinuation for toxicity and deaths for unrelated causes were considered as competing events. Survival probabilities were calculated using the Kaplan–Meier method. Survival comparisons were made by the log-rank test. All calculations were made using a standard statistical package (SPSS for Windows Version 15.0; Chicago, IL).
Related Knowledge Centers
- Failure Rate
- Logrank Test
- NONparametric Statistics
- Proportional Hazards Model
- Survival Function
- Censoring
- Empirical Distribution Function
- Medical Statistics
- Dependent & Independent Variables
- Nelson–Aalen Estimator