Part B – Data analysis General Instructions You are required to produce a poster which describes the study, the methods used to analyse the data, and the results of the analysis. Guidance on general rules for production of assignments is given in the MPH assignment submission guidelines (see the MPH Handbook, or Canvas), which you must follow. In particular you are reminded of the Department’s guidelines on Plagiarism, which can be found in the Handbook. Check to ensure that your submission meets the regulations. You should present your work as a POSTER with tables and figures (as appropriate) to illustrate your investigations. The poster should be created as A0 size (1189mm x 841mm) and it can be produced using commonly available software e.g. Microsoft PowerPoint. You can choose whether to present the poster in portrait style or landscape style. A template can be found on CANVAS, in the Practical Epidemiology and Statistics module page (in the assignment section). Like other assessments, you should not put your name on your poster, but only your student ID number. Please put your student ID in the top right hand corner. The minimum font size is 10. References and abstract are optional but not essential. There is no need for pictures or diagrams, and only include figures and tables that support your results. You should attempt all stages of the analyses that are described below. Assessment submission A PowerPoint and pdf version of the poster should be submitted via Canvas by 12thJanuary 2021. Task: Using the data from the cohort study provided, investigate which patient characteristics are associated with mortality for individuals who have suffered a cardiac arrest. Background Approximately 1,500 infants or children suffer a cardiac arrest in the United Kingdom and Ireland each year. Children resuscitated after a cardiac arrest remain at risk of dying. Currently, paediatric clinicians cannot accurately predict mortality in children post cardiac arrest. If factors predictive of mortality are known, interventions could be implemented with the aim of reducing the likelihood of mortality. In this assignment, you will work towards identifying patient characteristics that are associated with mortality. The study The data file you have been given contains simulated data on 500 patients. Patients included in the study were aged under 18, and experienced at least one diagnosed cardiac arrest during the study period. Patients were identified through Paediatric
Intensive Care Unit (PICU) electronic records. Patient characteristics are recorded when patients are admitted to PICU and include: age, ethnicity, gender, systolic blood pressure, amongst other characteristics. The primary objective of the study is to identify which patient characteristics are associated with mortality (at 3 months from admission). There is additional clinical interest in whether the patient characteristics that are associated with mortality differ based on the location of the cardiac arrest (in-hospital or out-of-hospital) The task Using simulated data based on the cohort study, your task is to investigate which factors are associated with mortality. The dataset you have been given includes the outcome (mortality at 3 months), demographic information about the patients in the study and other patient characteristics. The variables are summarised in the table below. The primary objective is to identify which patient characteristics are associated with mortality at 3 months. You also need to investigate whether the patient characteristics associated with mortality are different based on the location of the cardiac arrest by performing an analysis separately for in-hospital cardiac arrests and out-of-hospital cardiac arrests. The stages of your analysis should be:
1. Create a table of descriptive summary statistics to describe all the characteristics of the study population, stratifying as appropriate. You should briefly discuss how the characteristics compare across the stratification variable.
2. Perform univariable and multivariable analyses for the outcome (mortality).
Examine and describe the univariable and multivariable contribution that each of the patient characteristics makes to the outcome. Place particular emphasis on the point estimate and the confidence interval.
a. Discuss the characteristics that seem to be the most important predictors of the outcome
b. Discuss whether you think you have identified any characteristics that are not independent predictors of the outcome. 3
. Perform a multivariable analysis separately for in-hospital cardiac arrests and out-of-hospital cardiac arrests.
Discuss whether you think there is any evidence of differential factors. 4. Summarise the limitations of the study and what, if any, the clinical value might be from your findings