This page provides a step-by-step correction for building a simple MS Word report using the Quarto notebook exercise3.qmd as a starting point.

Important

To get the most out of your learning experience, attempt to solve the exercise on your own before looking at this correction. Remember, trials and errors are an essential part of the learning process, strengthening your understanding and helping you build confidence. Take your time, experiment, and learn actively!

1 Setup basic elements of the Quarto notebook

All the basic elements you need to setup are defined in the YAML header at the beginning of your Quarto notebook.

---
1title: My outbreak report
2author: Helene Langet
3date: 2023-12-31

4format: docx

5execute:
6  echo: false
7  warning: false
---

1: Update the title option in the Quarto notebook. Quotation marks are optional.
2: Set the author option to your name. Quotation marks are optional.
3: Add the date option and set it to 2023-12-31. Quotation marks are optional.
4: Set the format option to docx to generate a MS Word document.
5: Add an execute block to customize the execution behaviour for your Quarto notebook. This block allows you to control how code, warnings, and other outputs are displayed in the rendered document.
6: Add the echo option and set it to false to hide code in the rendered document.
7: Add the warning option and set it to false to hide warnings in the rendered document.

YAML indentation

Remember that YAML is a whitespace-sensitive language where indentation determines the structure; tabs are not recognised for indentation. The recommended practice is therefore to use two spaces per indentation level to ensure consistency and avoid errors.

2 Create publication-ready summary statistics tables

Creating publication-ready tables in R often involves leveraging specialized packages. Popular options include gtsummary, gt and flextable. In what follows, we will use gtsummary which is an excellent package for creating summary statistics tables. To install it, run the command install.packages("gtsummary") in your R console.

Table 1

```{r}
10#| tbl-cap: Population characteristics

#Table 1
#Generate a summary table displaying population characteristics
subdf |>                               
8  dplyr::select(age,
                sex,
                bmi,
                confirmed,
                death) |>
9  gtsummary::tbl_summary()
```

8: Use the dplyr::select() function to select specific columns (age, sex, bmi, confirmed, and death) from the subdf dataset.
9: Use the gtsummary::tbl_summary() function to create a table that summarizes the descriptive statistics of the selected variables in the whole population.
10: Add a caption to the table by using the tbl-cap chunk option.

Population characteristics
Characteristic	N = 65,669¹
age	50 (35, 65)
sex
1	33,114 (50%)
2	32,555 (50%)
bmi	29 (21, 38)
confirmed
0	13,235 (20%)
1	52,434 (80%)
death
0	64,455 (98%)
1	1,214 (1.8%)
¹ Median (Q1, Q3); n (%)

Table 2

```{r}
12#| tbl-cap: Demographic characteristics of deceased vs. alive

#Table 2
#Generate a summary table comparing the demographic characteristics
#of individuals who died versus those who are still alive
subdf |>                                                      
  dplyr::select(sex,
                age,
                bmi,
                death) |>
10  gtsummary::tbl_summary(by = death) |>
11  gtsummary::add_overall()
```

10: Use the gtsummary::tbl_summary() function to create a summary table grouped by the death column, showing separate statistics for individuals who have died vs. those who are still alive.
11: Use the gtsummary::add_overall() function to add a row to the table that shows overall summary statistics for the whole population, regardless of their death status.
12: Add a caption to the table by using the tbl-cap chunk option.

Demographic characteristics of deceased vs. alive
Characteristic	Overall N = 65,669¹	0 N = 64,455¹	1 N = 1,214¹
sex
1	33,114 (50%)	32,504 (50%)	610 (50%)
2	32,555 (50%)	31,951 (50%)	604 (50%)
age	50 (35, 65)	50 (35, 65)	52 (37, 67)
bmi	29 (21, 38)	29 (21, 38)	34 (28, 41)
¹ n (%); Median (Q1, Q3)

3 Create publication-ready figures

Figure 1

```{r}
13#| fig-cap: Weekly count of all cases, confirmed cases and deaths

#Figure 1
# Aggregate the data to get the weekly count of all cases, confirmed cases and deaths
weekly_data <- subdf |>
  dplyr::group_by(week) |>
  dplyr::summarise(count = dplyr::n(),
                   confirmed_count = sum(confirmed == "1"),
                   death_count = sum(death == "1"))

# Plot the weekly cases, confirmed cases and deaths
ggplot2::ggplot(weekly_data, ggplot2::aes(x = week)) +                        
  ggplot2::geom_line(ggplot2::aes(y = count,
                                  color = "All cases"),
                     size = 1) +
  ggplot2::geom_line(ggplot2::aes(y = confirmed_count,
                                  color = "Confirmed cases"),
                     size = 1) +
  ggplot2::geom_line(ggplot2::aes(y = death_count,
                                  color = "Confirmed deaths"),
                     size = 1) +
  ggplot2::labs(x = "Week",
                y = "Count",
                color = "Legend") +
14  ggplot2::scale_color_manual(values = c("All cases" = "#440e54",
                                         "Confirmed cases" = "#f8766d",
                                         "Confirmed deaths" = "#128984")) +
15  ggplot2::theme_minimal() +
16  ggplot2::theme(panel.grid.major.y = ggplot2::element_line(linewidth = 0.5,
                                                            linetype = "dashed",
                                                            color = "grey"), 
                 panel.grid.minor.y = ggplot2::element_line(linewidth = 0.5,
                                                            linetype = "dashed",
                                                            color = "grey"),
17                 panel.grid.major.x = ggplot2::element_blank(),
                 panel.grid.minor.x = ggplot2::element_blank())
```

13: Add a caption to the table by using the fig-cap chunk option. Remove the title parameter from ggplot2::labs to avoid the repetition.
14: Assigns specific colors to each line for better distinction.
15: Apply the ggplot2::theme_minimal() function to set a clean and simple aesthetic for the plot. This theme removes unnecessary elements such as background shading and focuses attention on the data by displaying only grid lines and axis labels.
16: Add dashed lines for horizontal (major and minor) grid lines.
17: Omits vertical (major and minor) grid lines for a cleaner look.

Weekly count of all cases, confirmed cases and deaths

4 Present statistical models and results

```{r}
18#| echo: true

#Logistic regression model
19model <- glm(death ~ bmi + age,
21             subdf |> dplyr::filter(confirmed == "1"),
20             family = binomial)
```

18: The echo: true option ensures that the code chunk used to generate the logistic regression model is displayed in the rendered output, which is helpful for transparency and reproducibility.
19: The glm() function fits a generalised linear model to predict death (binary outcome) from two predictors (bmi and age). The function returns a model object containing various components, such as coefficients (effects of predictors), residuals (errors), and fitted values (predicted probabilities).
20: Set the family argument to binomial to fit a logistic regression model. This is appropriate for binary outcomes like death, as it models the probability of the outcome occurring using a logit (log-odds) function (i.e. coefficients represent the effect size of each predictor on the log-odds of the binary outcome).
21: Filter data using the dplyr::filter() function to subset the dataset so that the model is estimated from confirmed cases only.

Table 3

```{r}
24#| tbl-cap: Formatted regression table

#Table 3
22gtsummary::tbl_regression(model,
23                          exponentiate = TRUE)
```

22: Use gtsummary::tbl_regression() to create a table summary of the logistic regression results.
23: Set the exponentiate argument to TRUE to convert the coefficients (log-odds) into odds ratios, which are more interpretable and useful for understanding the strength of associations.
24: Add a caption to the table by using the tbl-cap chunk option.

Formatted regression table
Characteristic	OR	95% CI	p-value
bmi	1.04	1.03, 1.04	<0.001
age	1.00	1.00, 1.01	0.003
Abbreviations: CI = Confidence Interval, OR = Odds Ratio

Reuse

CC-BY