Question 2: Are males more likely to be diagnosed with pneumonia?

Question 2: Are males more likely to be diagnosed with pneumonia?

1.0 Analytical Plan:

1.1 Study Design:

Copy over + fix up blurb from Q1. Take out stuff about repeated measures.

Can we find slide/something online to explain different study types?

The Mexican government conducted a cross-sectional study . observe the outbreak of COVID-19 through both clinical and epidemiological data. This study followed 15000 patients and recorded data including care type, days with symptoms and intubation they experienced with COVID-19.

1.2 Variables:

Workbook one

The independent variable being assessed in this analysis is gender. Gender is classified as a dichotomous categorical variable ie. either male or female. This variable is represented as female (1) and male (2).In regard to missing variables, this will be marked as 9999 and not be included in analysis. The data included three cases of data marked as 10 which wont be included in analysis.

The dependent variable is whether or not patients had a diagnosis of pneumonia , this is a categorical and dichotomous variable. This variable is represented as yes (1) and no (2). In regard to missing variables, this will be marked as 9999 and not be included in analysis.

1.3 Hypotheses:

H0 There is no difference in the proportion of females compared to males with a diagnosis of pneumonia.

H1 There is a difference in the proportion of females compared to males with a diagnosis of pneumonia.

1.4 Univariate Analysis Plan

Independent Variable: Gender

Numerical summary = proportion (or %) of sample who are female compared to male which will be numerically displayed in a frequency distribution table.

Graphical summary = will be a bar graph representing the proportion of female compared to male.

Dependent Variable: Pneumonia

Numerical summary = proportion (or %) of sample who were diagnosed with pneumonia compared to those who werent which will be numerically displayed in a frequency distribution table.

Graphical summary = will be a bar graph. Those diagnosed with pneumonia will be represented yes or no.

1.5 Bivariate analysis Plan

Numerical summary:

2 x 2 contingency table

Graphical summary:

Side by side

1.6 Statistical test and assumptions

Dee Module 3 - PART 1, 2, 3

See workbook 3 - Week 6 - family of chi squared tests (comparing categorical data)

Which specific chi-squared test looks like an option?

Fishers exact test assumptions or Continuity Test

Which ones can you rule out as not appropriate for this situation?

1.7 Significance level Plan

P<0.05 will be used to indicate statistical significance.

2.0 Analysis

2.1 Univariate Analysis

Of the 14997 people analysed for this variable, 7123 (47.5%) were females, while 7874 (52.5%) were males.

Of the 14991 people analysed for this variable, 2774 (18.5%) were diagnosed with pneumonia, while 12225 (81.5%) were not.

2.2 Bivariate Analysis

From the Two way tables and the graphs above it looks as though a higher proportion of males compared to females were diagnosed with pneumonia (20.8% compared to 16.0%). This difference will be formally tested to check whether the difference is large enough to not be due to chance alone.

Look at module 1 workbook about rounding the numbers (two decimal places)

The odds ratio 0.725 means that females were 0.725 times as likely as men to be diagnosed with pneumonia. To find out the odds ratio for men compared to women, can divide 1 by this number.

Odds ratio is 1/0.725 = 1.38

The odds of being diagnosed with pneumonia was 1.38 times higher for males compared to females.

Risk ratio is 1/0.769 = 1.30

Males are 1.30 times more likely to be diagnosed with pneumonia than females.

The odds ratio is 1.38 (95% CI 1.26 - 1.49)

The risk ration is 1.30 (95% CI 1.22 - 1.39)

2.3 Statistical test and assumptions

See workbook 3 - Week 6 - family of chi squared tests (comparing categorical data)

As p<0.05, we reject H0 and conclude there is a statistically significant difference in the proportion of females with pneumonia compared to males with pneumonia (16.0% compared to 20.8%). Fishers exact test (1 = 57.2 ), p=0.001.

There is a statistically significant difference

The difference between the 2 groups is statistically significant

Module 3 Part 1 - slide 12/15

Module

3.0 Summary [200 to 500 words]