As for the best way to analyse longitudinal data, there is no straight answer. The most important thing to keep in mind is that the analysis techniques should reflect your objective. This depends on the objective of the study, its design, the amount of missing data, sample size, available software, etc.

The most common endpoints are as follows:

  • Analysis at each time point separately
  • Change from baseline
  • Time until event (predefined in/decrease)
  • Responder
  • Area under the curve or overall average over time

Longitudinal modelling is recommended over repeated univariate analysis.
This approach will try to model the data, including the time dependence. Estimates of time and group parameters allow you to test the significance of comparisons and/or construct confidence intervals for predictions. Many different types of models are possible according to the assumptions one is willing to make. Most commonly used are general linear models (GLM) and general linear mixed models (GLMM).

Some good references are:

  • Fairclough DL. Design and Analysis of Quality of Life Studies in Clinical Trials. Chapman and Hall/CRC Press, Boca Raton, Florida, 2002.
  • Diggle P et al. Analysis of Longitudinal Data, Oxford University Press, 1994.
  • Fahrmeir, Tutz. Multivariate Statistical Modelling using Generalized Linear Models Springer Statistics, New York, 1994.
  • Verbeke G & Molenberghs G. Linear Mixed Models for Longitudinal Data, New York: Springer – Verlag, 2000.
  • Coens C et al. International standards for the analysis of quality-of-life and patient-reported outcome endpoints in cancer randomised controlled trials: recommendations of the SISAQOL Consortium. Lancet Oncol. 2020 Feb;21(2):e83-e96.
  • Pe M et al. Statistical analysis of patient-reported outcome data in randomised controlled trials of locally advanced and metastatic breast cancer: a systematic review. Lancet Oncol. 2018 Sep;19(9):e459-e469.