From the barrier, I have read many of the topics and I would like to say how much I appreciate your discussion. Then, I am planing a new study and I would like to share my details and hopefully to ask you for some suggestions during the analysis and planing the protocol and study design.
- In my hospital, we receive around 120 patients/year with hip fracture. Many of them have strong comorbidities that affect the prognosis in terms of mortality, morbidity and recovery. There are some modifiable factors to avoid in order to improve the prognosis and potential functional recovery.
- One the main covariates investigated is the time-until-surgery. From the emergency consultation until the surgery for fixation of the hip fracture.
- In my country and place, there are not any study about the prognosis (functional or mortality) about these common patients but from literature we know that mortality is around 15% (30 days) after surgery.
- The main outcomes in which we are interested are mortality 30 days and 3 months and Quality of Life 30 days and 3 months.
- How should I plan the sample size. By funding and logistics we are able to collect all patients during a year (120) but I am afraid about potential post hoc modelling with that number of patients. We expect around 15% mortality 30 days and If we based our sample only in estimation of the mortality could be enough.
- Quality of life is very important as well, so, could we use that outcome for sample planning in terms of estimation ?
- We are interested in testing the effect of time-until-surgery in mortality or Quality of life. So, if mortality presets low number of events, should we change for quality of life ? Or just to focus in estimation and not in testing about potential associations.
- We are going to collect some covariates, so, please could you suggest me which type of analysis comes to your mind for the first outcomes (mortality and quality of life): QoL will be measured using EuroQoL 5D.
Thank you so much for you advice.
PD. In my place, there are no single statistician with high skills in biomedical research.