Modeling survival data flexibly using multi-parameter regression : an application to lung cancer data / Kevin Burke.

Format:

Book

Author/Creator:

Burke, Kevin, active 2020, author.

Series:

SAGE Research Methods Cases: Medicine and Health.

SAGE Research Methods Cases: Medicine and Health

Language:

English

Subjects (All):

Lungs--Cancer--Data processing--Case studies.

Lungs.

Lungs--Cancer--Statistical methods--Case studies.

Physical Description:

1 online resource : illustrations.

Place of Publication:

London : SAGE Publications Ltd, 2020.

Summary:

In this case study, I describe my development of multi-parameter regression (MPR) as a flexible way of modeling survival data along with an application to lung cancer data (using the "mpr" package in the R programming language). I cover both the use of MPR as a data analysis research method and the thought process which led me to its development-the latter could be thought of as a method for developing a method! Therefore, the case study should appeal both to those interested in applying statistical methods and those interested in developing them (and everybody in between). If you have ever analyzed survival data, then you have almost certainly used Cox's proportional hazards (PH) model-but did you consider why this particular model should be used or what makes hazards proportional? More generally, critical assessment of the statistical models you use should never be overlooked; these models shape your research findings. In the context of this case study, the MPR method provides a vehicle for relaxing the PH assumption, but can be used in other contexts to create flexible models. Although these more flexible/complex models get closer to reality, you should always remember that all models are wrong. This is a key learning shared in this case study, which comes about naturally when you critically assess models (be they established or not) as standard practice.

Notes:

Includes bibliographical references and index.

Description based on XML content.

ISBN:

1-5297-1292-0

9781529712926

OCLC:

1142444140