Thèse, Nantes.| Entreprise/Organisme : | Université de Nantes | | Niveau d'études : | Master | | Date de début : | Premier trimestre 2020 | | Durée du contrat : | 4 ans | | Rémunération : | Salaire doctorant | | Description : | The project
Context
Many clinical studies attempt to measure important patients-centered characteristics1,2, such as Health-Related Quality of Life (HRQoL), using Patient reported outcomes measures (PROM) in which different items are usually grouped into several dimensions (e.g. physical, emotional, social…). PROM are directly reported by the patients without interpretation of their responses by a clinician or anyone else and pertains to the patient’s perceived health, HRQoL, or functional status associated with healthcare or treatment. Although HRQOL is recognized as a relevant outcome measure for patients admitted to the Intensive Care Unit (ICU), it is not routinely included in ICU practice or research3. This may be because measuring HRQOL is more time-consuming than mortality rates and is also more difficult to interpret. However, long-term outcomes of physical and psychological factors, functional status, and social interactions are becoming more and more important both for ICU physicians and nurses as well as for patients and their relatives. It is well-recognized that ICU survivors experience challenges during the recovery phase and that change in HRQoL, physical (e.g. neuropathy, reduced mobility), and psychological disorders (e.g. depression and anxiety) are essential to monitor after ICU discharge to evaluate the efficacy and efficiency of ICU interventions and treatments4
Changes in PROM can however be difficult to interpret in longitudinal settings as patients have to regularly adapt to their illness and as a consequence, patients might give different answers on PROM over time, not only because their health has changed, but also because their perception of what health or HRQoL means to them has changed. This phenomenon can be referred to as response shift5 (RS) and it has been evidenced in several health conditions6-8. In case of RS, it has been argued, on the one hand, that it might be impossible to disentangle, without appropriate methodology, perceived PROM changes from confounding RS effects (often triggered by adaptation), which is problematic for the interpretation of change and treatment effects because of biased estimations and poor power. On the other hand, the therapeutic importance of patients’ adaptation and of a better understanding of how patients adjust to their illness and life circumstances after ICU discharge has been highlighted4. Thus, adaptation, closely linked to RS, may also be one of the goals of therapy in helping patients to cope with their illness and to live with it. Whatever the adopted viewpoint, it is essential to be able to assess the changes experienced by patients after ICU discharge and to take into account RS, if appropriate in a reliable and unbiased manner for assessing the suitability of interventions.
To date, most statistical approaches used for RS analyses assume that the majority of patients experiences RS in the same way regardless of individual characteristics. However, it is indeed very likely that the propensity for RS depends on socio-demographic (gender, age…), clinical (treatments…), or psychological characteristics of the patients which might also interfere with changes in PRO over time. A method called the RespOnse Shift ALgorithm at Item-level (ROSALI)9 integrating Item Response Theory (IRT) and Rasch Measurement Theory (RMT) models was developed allowing for item-level RS analysis and estimations of PRO changes while adjusting for RS. The good performance of ROSALI based on RMT models for RS detection, estimation and adjustment for PROM change were recently evidenced in a simulation study10. The algorithm of ROSALI was recently deeply modified to integrate a single binary covariate enabling the investigation of RS at subgroup level, between two groups (manuscript submitted). This first step needs to be extended to allow for investigating RS in different subgroups defined by more than one covariate and to adjust the analyses on potential confounders which implies further developments and subsequent validation of the ROSALI algorithm.
Objectives
Methodological: Development and validation of ROSALI based on RMT to analyze RS at subgroup level defined by several covariates. Assessment of the performance and robustness of ROSALI using simulation studies.
Clinical: the HAP2 European project aims at comparing rHu-IL-12 and IFN-versus placebo regarding PROM changes and RS after ICU discharge in two clinical trials for prevention and treatment of hospital-acquired pneumonia (HAP). The validated ROSALI method will be used to analyze and compare changes in HRQoL, anxiety and depression according to treatment arms also inlcuding other covariates of interest (gender, country….).
The team
You will join a multidisciplinary, creative and inspiring environment full of expertise and curiosity within the INSERM SPHERE "methodS in Patient-centered outcomes & HEalth ResEarch" research unit (http://www.sphere-nantes.fr/) at the University of Nantes. The SPHERE unit aims to promote patient-centered methodological research and to reinforce the contribution of PROM data (health-related quality of life, fatigue, well-being…) in decision-making and assessment of healthcare. It consists of about 60 members including faculty, researchers, doctoral students, postdoctoral and visiting fellows. In the year 2019 alone, its research activities yielded over 80 publications in scientific journals in various areas, spanning from Biostatistics to Clinical Epidemiology, Medicine covering many specialties and topics (Innovative designs and methods in: Dermatology, Psychiatry, General Practice, Nephrology…).
Funding of the project: European Commission, H2020, 4-year full-time position (Engineer) from first quarter of 2020, possibility of doing a PhD.
The candidate will be welcomed at the unit's premises in Nantes, France.
Tasks (Under the responsibility of the principal investigator and a Research Engineer of the unit)
• Development and programming of the ROSALI algorithm to integrate several covariates
• Validation of the ROSALI algorithm:
• Design and programming of simulation studies
• Analysis and organization of the results for their interpretation
• Writing of the associated scientific publication(s)
• Application of the ROSALI algorithm to analyze the two clinical trials for prevention and treatment of HAP.
• Writing of the associated scientific publication(s)
Qualifications
• Master’s degree in Biostatistics
• Experience in longitudinal modeling using mixed effects models is mandatory; experience in longitudinal Structural Equation Modeling, Rasch or IRT models will be appreciated
• Knowledge of statistical software and computer programming skills are expected
• An experience in PROM analysis, Stata software and/or simulation studies will be appreciated
• Experience from research in epidemiology, medicine, or other substantive disciplines is valuable
• Highly organized, and able to work both as part of a team but also under their own initiative
• Willing to share knowledge and expertise with the other members of our research group
• Excellent teamwork and communication skills, both verbal and written
• Fluent in English (written, reading, verbal), French (not mandatory)
Application process
Applications must contain the following documents in English or French
• A CV with a complete list of internships completed, publications and conference presentations, if any
• A motivation letter
Contact Information
Please send your CV and a cover letter to: Véronique Sébille, veronique.sebille@univ-nantes.fr | | En savoir plus : | http://sphere-nantes.fr/
Research position in Biostatistics_Eng_PhD.pdf | | Contact : | veronique.sebille@univ-nantes.fr |
|
