Revisiting methods for modeling longitudinal and survival data: Framingham Heart Study

Background Statistical methods for modeling longitudinal and time-to-event data has received much attention in medical research and is becoming increasingly useful. In clinical studies, such as cancer and AIDS, longitudinal biomarkers are used to monitor disease progression and to predict survival. These longitudinal measures are often missing at failure times and may be prone to measurement errors. More importantly, time-dependent survival models that include the raw longitudinal measurements may lead to biased results. In previous studies these two types of data are frequently analyzed separately where a mixed effects model is used for the longitudinal data and a survival model is applied to the event outcome. Methods In this paper we compare joint maximum likelihood methods, a two-step approach and a time dependent covariate method that link longitudinal data to survival data with emphasis on using longitudinal measures to predict survival. We apply a Bayesian semi-parametric joint method and maximum likelihood joint method that maximizes the joint likelihood of the time-to-event and longitudinal measures. We also implement the Two-Step approach, which estimates random effects separately, and a classic Time Dependent Covariate Model. We use simulation studies to assess bias, accuracy, and coverage probabilities for the estimates of the link parameter that connects the longitudinal measures to survival times. Results Simulation results demonstrate that the Two-Step approach performed best at estimating the link parameter when variability in the longitudinal measure is low but is somewhat biased downwards when the variability is high. Bayesian semi-parametric and maximum likelihood joint methods yield higher link parameter estimates with low and high variability in the longitudinal measure. The Time Dependent Covariate method resulted in consistent underestimation of the link parameter. We illustrate these methods using data from the Framingham Heart Study in which lipid measurements and Myocardial Infarction data were collected over a period of 26 years. Conclusions Traditional methods for modeling longitudinal and survival data, such as the time dependent covariate method, that use the observed longitudinal data, tend to provide downwardly biased estimates. The two-step approach and joint models provide better estimates, although a comparison of these methods may depend on the underlying residual variance.

A Longitudinal Measure of Medical Student Empathy at a Regional Campus

Introduction Empathy is an important characteristic of the ideal physician. Various quantitative measures of empathy have shown a steep decline during the third year of medical school. Methods We had 4 classes of medical students at our regional rural campus complete the Jefferson Scale of Empathy after each of the first 3 years. We report longitudinal results of 30 students, individually matched, including an analysis by gender. Separately, we report the cross-sectional results for 39 of our students as they began medical school. We compare our student scores to other allopathic and osteopathic student scores from large urban campuses. The Baptist Health Madisonville IRB approved the protocol as exempt. Results As they begin medical school, our students have similar scores to those at large urban campuses (difference of 1.1 points, p=.421). After the M-2 year, our students had significantly higher scores than those at urban campuses (5.7 points, p=.002) and after the M-3 year, they show an even larger positive difference (9.0 points, p<.001). As in previous publications, females had higher overall mean scores at each measure, but with our students this was only significant in post-M-2 measures (8.9 points, p=.01). Discussion We conclude that something about our students’ experience during their M-3 year is associated with a smaller decline in empathy measures than reported previously. We propose that some of this difference could be due to a formal professional identity curriculum we implemented recently during the M-3 year. However, without a concurrent or historical control group, we cannot be certain. We offer the concept of measuring empathy before and after curricular change as another useful evaluation tool for medical educators.

Revisiting Methods For Modeling Longitudinal and Survival Data: The Framingham Heart Study

Background: Statistical methods for modeling longitudinal and time-to-event data has received much attention in medical research and is becoming increasingly useful. In clinical studies, such as cancer and AIDS, longitudinal biomarkers are used to monitor disease progression and to predict survival. These longitudinal measures are often missing at failure times and may be prone to measurement errors. More importantly, time-dependent survival models that include the raw longitudinal measurements may lead to biased results. In previous studies these two types of data are frequently analyzed separately where a mixed effects model is used for the longitudinal data and a survival model is applied to the event outcome. Methods: In this paper we compare joint maximum likelihood methods, a two-step approach and a time dependent covariate method that link longitudinal data to survival data with emphasis on using longitudinal measures to predict survival. We apply a Bayesian semi-parametric joint method and maximum likelihood joint method that maximizes the joint likelihood of the time-to-event and longitudinal measures. We also implement the Two-Step approach, which estimates random effects separately, and a classic Time Dependent Covariate Model. We use simulation studies to assess bias, accuracy, and coverage probabilities for the estimates of the link parameter that connects the longitudinal measures to survival times. Results: Simulation results demonstrate that the Two-Step approach performed best at estimating the link parameter when variability in the longitudinal measure is low but is somewhat biased downwards when the variability is high. Bayesian semi-parametric and maximum likelihood joint methods yield higher link parameter estimates with low and high variability in the longitudinal measure. The Time Dependent Covariate method resulted in consistent underestimation of the link parameter. We illustrate these methods using data from the Framingham Heart Study in which lipid measurements and Myocardial Infarction data were collected over a period of 26 years.Conclusions: Traditional methods for modeling longitudinal and survival data, such as the time dependent covariate method, that use the observed longitudinal data, tend to provide downwardly biased estimates. The two-step approach and joint models provide better estimates, although a comparison of these methods may depend on the underlying residual variance.

Revisiting Methods For Modeling Longitudinal and Survival Data: The Framingham Heart Study

Background: Statistical methods for modeling longitudinal and time-to-event data has received much attention in medical research and is becoming increasingly useful. In clinical studies, such as cancer and AIDS, longitudinal biomarkers are used to monitor disease progression and to predict survival. These longitudinal measures are often missing at failure times and may be prone to measurement errors. More importantly, time-dependent survival models that include the raw longitudinal measurements may lead to biased results. In previous studies these two types of data are frequently analyzed separately where a mixed effects model is used for the longitudinal data and a survival model is applied to the event outcome. Methods: In this paper we compare joint maximum likelihood methods, a two-step approach and a time dependent covariate method that link longitudinal data to survival data with emphasis on using longitudinal measures to predict survival. We apply a Bayesian semi-parametric joint method and maximum likelihood joint method that maximizes the joint likelihood of the time-to-event and longitudinal measures. We also implement the Two-Step approach, which estimates random effects separately, and a classic Time Dependent Covariate Model. We use simulation studies to assess bias, accuracy and coverage probabilities for the estimates of the link parameter that connects the longitudinal measures to survival times. Results: Simulation results demonstrate that the Two-Step approach performed best at estimating the link parameter when variability in the longitudinal measure is low but is somewhat biased downwards when the variability is high. Bayesian semi-parametric and maximum likelihood joint methods yield higher link parameter estimates with low and high variability in the longitudinal measure. The Time Dependent Covariate method resulted in consistent underestimation of the link parameter. We illustrate these methods using data from the Framingham Heart Study in which lipid measurements and Myocardial Infarction data were collected over a period of 26 years.Conclusions: Traditional methods for modeling longitudinal and survival data, such as the time dependent covariate method, that use the observed longitudinal data, tend to provide downwardly biased estimates. The two-step approach and joint models provide better estimates, although a comparison of these methods may depend on the underlying residual variance.

Revisiting Methods For Modeling Longitudinal and Survival Data: The Framingham Heart Study

Background Statistical methods for modeling longitudinal and time-to-event data has received much attention in medical research and is becoming increasingly useful. In clinical studies, such as cancer and AIDS, longitudinal biomarkers are used to monitor disease progression and to predict survival. These longitudinal measures are often missing at failure times and may be prone to measurement errors. More importantly, time-dependent survival models that include the raw longitudinal measurements may lead to biased results. In previous studies these two types of data are frequently analyzed separately where a mixed effects model is used for the longitudinal data and a survival model is applied to the event outcome. Methods In this paper we compare joint maximum likelihood methods, a two-step approach and a time dependent covariate method that link longitudinal data to survival data with emphasis on using longitudinal measures to predict survival. We apply a Bayesian semi-parametric joint method and maximum likelihood joint method that maximizes the joint likelihood of the time-to-event and longitudinal measures. We also implement the Two-Step approach, which estimates random effects separately, and a classic Time Dependent Covariate Model. We use simulation studies to assess bias, accuracy and coverage probabilities for the estimates of the link parameter that connects the longitudinal measures to survival times. Results Simulation results demonstrate that Two-Step approach performed best at estimating the link parameter when variability in the longitudinal measure is low but is somewhat biased downwards when the variability is high. Bayesian semi-parametric and maximum likelihood joint methods yield higher link parameter estimates with low and high variability in the longitudinal measure. Time Dependent Covariate method resulted in consistent underestimation of the link parameter. We illustrate these methods using data from the Framingham Heart Study in which lipid measurements and Myocardial Infarction data were collected over a period of 26 years. Conclusions Traditional methods for modeling longitudinal and survival data, such as time dependent covariate method, that use the observed longitudinal data, tend to provide downward bias estimates. Two-step approach and joint models provide better estimates, although a comparison of these methods may depend on the underlying residual variance.

A Residency Professional Identity Curriculum and a Longitudinal Measure of Empathy

BACKGROUND AND OBJECTIVES Empathy measures were used before and after implementation of a structured professional identity curriculum to determine the effect among a group of family medicine residents. METHODS The Jefferson Scale of Empathy was completed by 18 residents at all three years of training before, immediately after a six month professional identity curriculum intervention, and six months after the curriculum was completed. The curriculum included one hour luncheon sessions on concepts of profession, burnout, and cynicism as well as thoughtful use of electronic medical records, prevention and management of burnout, mindfulness techniques and reflective writing and drawing. The Baptist Health Madisonville IRB approved the protocol as exempt and the authors have no conflicts of interests. RESULTS Similar to previous publications, a decline in empathy across the academic year was found, with a significant decline six months after the end of the curriculum. Residents who attended more sessions showed a non-significant smaller decline, and there were large standard deviations among each training level with some individual residents showing little change across the year. Evaluations of the curriculum were largely positive. CONCLUSIONS This professional identity curriculum in this group of residents may have temporarily mitigated the decline in measured empathy that has been described among residents. Results support some aspects of empathy as a trait in some residents rather than a state that is amenable to a training effect. Further study in this residency including longitudinal empathy measurements and focus groups is ongoing. Other programs’ experience with these issues is needed to add to sample size and diversity of training environments to discern which changes are significant and generalizable.

Should the Third Reminder be Sent? The Role of Survey Response Timing on Web Survey Results

Decreasing survey response rates are a growing concern as survey estimates may be biased by selective non-response. One method of assessing non-response bias is to examine the timing of survey response, specifically comparing those who respond late to a survey with those who respond early. This paper draws upon data obtained from multiple panel surveys conducted over a six-month period, and examines whether early, intermediate and late respondents differ significantly in demographics or in their responses to survey questions. By considering response timing as a repeated behaviour, or habit, spanning multiple surveys, a longitudinal measure of response timing is developed to identify the predictors of responding early to multiple surveys conducted over a period of time. Results indicate some directional differences in demographics and better data quality from early respondents, compared to their intermediate and late counterparts. We discuss the findings from the study and conclude with recommendations for future research.