Joint modeling of longitudinal outcomes and survival using latent growth modeling approach in a mesothelioma trial

Our plain-English summary of Joint modeling of longitudinal outcomes and survival using latent growth modeling approach in a mesothelioma trial, by P. Wang, W. Shen, M.E. Boye (Health Services & Outcomes Research Methodology, 2012).

The study

• Authors: P. Wang, W. Shen, M.E. Boye
• Journal: Health Services & Outcomes Research Methodology
• Published: 2012;12(2-3):182-199
• Cited by: 22 articles, as counted by NIH iCite
• Read it: PubMed · Publisher · Free full text

Abstract

Joint modeling of longitudinal and survival data can provide more efficient and less biased estimates of treatment effects through accounting for the associations between these two data types. Sponsors of oncology clinical trials routinely and increasingly include patient-reported outcome (PRO) instruments to evaluate the effect of treatment on symptoms, functioning, and quality of life. Known publications of these trials typically do not include jointly modeled analyses and results. We formulated several joint models based on a latent growth model for longitudinal PRO data and a Cox proportional hazards model for survival data. The longitudinal and survival components were linked through either a latent growth trajectory or shared random effects. We applied these models to data from a randomized phase III oncology clinical trial in mesothelioma. We compared the results derived under different model specifications and showed that the use of joint modeling may result in improved estimates of the overall treatment effect.

In plain English

This was a statistical methods paper. The authors, from Eli Lilly and Company, which funded the work, re-analyzed data from EMPHACIS, a randomized phase III trial in 448 patients with malignant pleural mesothelioma. Of these, 226 received pemetrexed/cisplatin and 222 received cisplatin alone. The paper did not study asbestos exposure. The authors tested "joint models" that analyze patient-reported symptom and quality-of-life scores (from the Lung Cancer Symptom Scale) together with time to progressive disease, meaning documented progression or death. Analyzed alone, the hazard ratio for pemetrexed/cisplatin versus cisplatin was 0.73 (P-value 0.001). In one joint model, patients on pemetrexed/cisplatin had significantly lower (better) scores at cycle 6 on dyspnea, pain and two combined symptom indices. Higher symptom scores were associated with a higher hazard of progression or death. Overall treatment hazard ratios from that joint model were consistently smaller than 0.73. The authors concluded that the treatment effect appeared larger in the joint models, and that joint modeling may improve such estimates. They noted that each symptom item was modeled one at a time, and that model assumptions may have been violated.

We wrote this plain-English summary from the full article with the help of AI, and a second AI model checked each sentence against the article. It is not the authors' wording; the abstract and the article are their own account.

What the study did and found

• Design: Statistical methods paper: secondary analysis of data from a randomized phase III clinical trial (EMPHACIS), comparing joint models of longitudinal patient-reported outcomes and time to progressive disease with separate analyses
• Who was studied: 448 patients with malignant pleural mesothelioma in the global EMPHACIS phase III trial: 226 given pemetrexed/cisplatin and 222 given cisplatin alone. Dates are not given in this article.
• Exposure: No asbestos exposure was studied. The comparison was treatment: pemetrexed/cisplatin versus cisplatin alone, given on day 1 of a 21-day cycle.
• Main results:
  • In the separate survival analysis, the hazard ratio for time to progressive disease for pemetrexed/cisplatin versus cisplatin was 0.73 (P-value 0.001).
  • In the separate analysis of symptom scores at cycle 6, the pemetrexed/cisplatin group had significantly or nearly significantly lower mean scores on dyspnea, ASBI5 and ASBI8 (P-values less than 0.1). For dyspnea the mean was 30.9 with pemetrexed/cisplatin versus 36.9 with cisplatin (P-value 0.023).
  • In the naive Cox model, with each symptom score as a time-varying covariate, the hazard of progression or death rose by about 9% (cough) to 23% (ASBI5) for each 10-point increase in score.
  • Among the joint models, those assuming a quadratic trajectory with a treatment effect on slope only fit best, having the smallest BIC.
  • In the trajectory joint model (trajq_s), the treatment effect on the slope of the score was significant only for dyspnea (P-value 0.039).
  • In the shared random-effects joint model (remq_s), there was a significant treatment effect on the slope of dyspnea, pain, ASBI5 and ASBI8, from which the authors inferred significantly lower scores at cycle 6 in the pemetrexed/cisplatin arm.
  • In the remq_s model, the overall treatment hazard ratios were consistently smaller than the 0.73 found without the symptom data.
  • In models that included other covariates, pemetrexed/cisplatin and tumor stage III or less were significantly associated with improved time to progressive disease.
• The authors' conclusion: The authors concluded that the treatment effect on time to progressive disease appeared larger in the joint models than in separate analyses, and that joint modeling may give improved estimates of the overall treatment effect.
• Limitations the authors state:
  • Each patient-reported outcome item was modeled with time to progressive disease one at a time, although treatment may affect several related dimensions at once.
  • Model assumptions may have been violated because the symptom data were skewed with many zero scores, or because hazards were not proportional.
  • No standard software exists to fit a wide range of joint models.

Topics

Medical terms in this study, explained in our Lexicon: Mesothelioma, Clinical Trial, Quality of Life.

Related studies

• Joint modeling of multiple longitudinal patient-reported outcomes and survival — L.A. Hatfield, M.E. Boye, et al. (2011)
• A radiotherapy community data-driven approach to determine which complexity metrics best predict the impact of atypical TPS beam modeling on clinical dose calculation accuracy — F.M.D. Brooks, M.C. Glenn, et al. (2024)
• Outcome of asbestos exposure (lung fibrosis and antinuclear antibodies) with respect to skin reactivity: an 8-year longitudinal study — A. Lange, D. Garncarek, et al. (1986)
• Mathematical modeling for Phase I cancer trials: A study of metronomic vinorelbine for advanced non-small cell lung cancer (NSCLC) and mesothelioma patients — F. Barlesi, D.C. Imbs, et al. (2017)
• Modeling Epidermal Growth Factor Inhibitor-mediated Enhancement of Photodynamic Therapy Efficacy Using 3D Mesothelioma Cell Culture — G. Cramer, M. Shin, et al. (2019)
• Prediction modeling using routine clinical parameters to stratify survival in malignant pleural mesothelioma patients complicated with malignant pleural effusion — S. Zhang, Y. Zhang, et al. (2021)

Browse return to medical research.