BASIC STATISTICS FOR MEDICAL SCIENCES IEHC0046
DATA ANALYSIS EXERCISE 2018-19
Investigators have been interested in high blood pressure of middle aged men and women. They were particularly interested in the association between education and systolic blood pressure. They decided to use continuous measure of systolic blood pressure and binary measure of hypertension (expressed as SBP>=140mmHg) as two study outcomes.
The data available for this exercise were drawn from a large cross-sectional screening study conducted among men and women aged 50-65 years.
The dataset assessment2018.dta
——————————————————————————-
age age in years
sex 1=Male, 2=Female
mi diagnosed or hospitalised for myocardial infarction
SBP systolic blood pressure
alco annual alcohol intake; 0=no,1<500g ethanol/year, 2<2000g/year, 3<6000g/year, 4=6000g and more
binge binge drinking >60g eth at least 1x a week; 0=no,1=yes
material material conditions;1=worst,4=best
ht hypertension;SBP>=140; yes=1 no=0
educ 1=primary or less, 2=vocational,3=secondary,4=university
marital 1=married,2=single,3=divorced,4=widowed
sports sport activities in a week:0=no,1=yes up to 4h/week,2=up to 7h/week,3=more
smoke smoking; 1=current smoker,2=past smoker, 3=never smoker
econact economic activity:1=working,2=houesewife,3=pensioner still employed,4=pensioner not employed, 5=unemployed
bmi body mass index
The aim and hypothesis: The investigators aimed to study the relationship between education and systolic blood pressure. It has been hypothesized that less educated people have higher systolic blood pressure and higher prevalence of hypertension.
Q1. Summarise the study sample
Before looking at relationships between the exposure of interest and study outcomes it is important to know the characteristics of the study sample. Describe the sample in terms of available variables. Use no more than two tables for this part.
Q2. Evaluate the association between the binary measure of hypertension and possible risk factors other than education.
Q3. Next, evaluate the association between binary measure of hypertension and education. Use classical MH analysis to identify possible confounders and effect modifiers (if any). Moreover, would you confirm the original hypothesis saying that less educated people have higher prevalence of hypertension than those with higher education?
Q4. Evaluate the association between education and systolic blood pressure (SBP; on a continuous scale) using classical (non-regression) analysis.
Q5. Next, look at the association between education and systolic blood pressure using linear regression. Identify possible confounders and effect modifiers (if any). Give an interpretation of these results.
Q6. Please discuss whether results from two types of analysis (using binary outcome in Q3 and using continuous outcome in Q5) agree or disagree. Make brief conclusions in response to the aim and hypothesis of the study.
For you assessment:
a) Introduce the study (in maximum 5 sentences)
b) Describe methodological aspects of your analysis (focus on description of the data, statistical methods used, assumptions)
c) Carry out the necessary analyses to enable you to draw conclusions about whether any variables do confound or modify the relationship between education and blood pressure/hypertension.
d) Present tables summarising your results in an appropriate way (please do not use tables directly copy-pasted from STATA; create your own tables directly in MS Word)
e) Briefly state your conclusions about the relationship between education and blood pressure/hypertension.
Your report:
The maximum length is 2000 words (excluding all the tables). Use maximum of 6 tables. Tables should be included after the main text. All tables should have their own titles and be clearly linked to the main text of your report.
Course administrator will check your word count – please do submit your report as a single document in MS Word format (not in Adobe Acrobat!)
Please add list of commands used in the analysis as the appendix in your document (your “do” file) – this appendix will not be included in your word count. Please do not include output from STATA (do not include your log file), just commands! Include all commands necessary to get all the results reported in your submission and do not include commands with errors in the syntax.
Your work will be marked between 0 and 100 points.
For work that exceed a specified word limit by 1-199 words the mark will be reduced by 5 percentage marks but the penalised mark will not be reduced below the pass mark assuming the work merited a Pass. For work that exceed a specified word limit by 200 words or more the mark will be reduced by 10 percentage marks but the penalised mark will not be reduced below the pass mark assuming the work merited a Pass. See https://www.ucl.ac.uk/academic-manual/chapters/chapter-4-assessment-framework-taught-programmes/section-3-module-assessment#3.13 for further regulations.
The deadline is on 14 January 2019 at 09:00am. You need to submit electronically to Moodle (there is a submission link via Turnitin on Moodle). Submission after the deadline will be penalised using standard UCL regulations for late submissions of modules at level 7 (https://www.ucl.ac.uk/academic-manual/chapters/chapter-4-assessment-framework-taught-programmes/section-3-module-assessment#3.12).
Use examination code provided by the College.