Doctor: Arthur James
Title: Outcomes to assess health long-term after severe trauma
Supervisors: Philippe Ravaud and Viet-Thi Tran
Doctoral school: ED 393 Epidemiology and Biomedical Information Sciences, Université Paris Cité
Date of thesis defense: 21/11/2022
Jury: Julie Haesebaert, Pierre Bouzat, Anne-Claire Lukaszewicz, Bénédicte Sautenet, Viet-Thi Tran, Philippe Ravaud
A health assessment of severe trauma patients using only mortality leads to a dichotomous view of patient outcomes where all survivors are considered similar. For this reason, there is an ongoing challenge to identify new judgement criteria and we hypothesized that patients who survive severe trauma have a unique expertise that should be used. Therefore the objectives of this work were 1/ to provide a complete overview of domains that impair the daily lives of survivors and 2/ to evaluate the capacity of measurement tools used in research to encompass these domains. First, 340 severe trauma adult patients were recruited and semi-structured interviews were conducted between 6 months and 3 years after the trauma in order to identify problems encountered by patients in their daily lives. The number of patients required was determined by a mathematical model to estimate data saturation. Thematic analysis then allowed identifying 97 domains that patients considered useful in assessing their health. Most frequently reported domains were persistent pain (n=173, 51 % of respondents), difficulty to walk (n=115, 34 %) and an inability to work as before the accident (n=75, 22 %). Second, we determined how published and ongoing research takes is covering these 97 domains. For this purpose, a systematic review was conducted and included all studies that collected at least one outcome at least 180 days after severe trauma between 1 January 2014 and 30 September 2019. The systematic review was conducted using Medline, EMBASE and Clinicaltrials.gov databases. We included 116 reports representing 105,961 patients. A total of 154 different measurement tools were identified. Apart from mortality (20.7 %, n=24 reports), the three measurement tools most frequently used were the Glasgow Outcome Score (31.0 %, n=36), questions related to patients’ ability to return to work (20.7 %, n=24) and the EQ-5D (19.0 %, n=22). Of these measurement tools, 73.4 % (n=113) were used in only one single report, illustrating the heterogeneity of outcomes used. We then assessed the capacity of each measurement tools identified in the systematic review to address each domain identified by patients in the qualitative study. By convention, the 10 domains most frequently identified in the qualitative study were considered the most important to patients. We thus report the completeness of the collection of important domains with 10.3 % (n=12) of the included reports collecting no important domains, 69.8 % (n=81) of the reports collecting at least 3 important domains, 26.7 % (n=31) of the reports collecting at least 7 important domains, and only one report that was able to collect all 10 domains considered important. We also report the mismatch between the domains deemed important by patients and those prioritized by researchers with only five of the 10 domains most frequently mentioned by patients being among the domains most frequently captured in the systematic review. Through this work, we thus introduce the concepts of completeness and mismatch of the collection of important domains in research and strongly advocate for greater involvement of patients in the structuring of their care and related research.
Link to download the thesis (written in French)