index.qmd

# Preface {.unnumbered}

```{r setup, include=FALSE}
knitr::opts_knit$set(root.dir = rprojroot::find_rstudio_root_file())
here::i_am("index.qmd")
library(here)
library(gt)
library(dplyr)
```


::: callout-tip
### Goal
An **in-development** cookbook & code collection for sample size calculation in medical research using R
:::


### Who is book for 

This book this for someone who has 

-   Some basic `R` knowledge.

-   Some experience conducting statistical test (e.g., t-test). 

But you haven't done any sample size calculation before (at least in R) and you want a quick tutorial to do that in `R` with some examples.

### Resources

Many of the resources and code are from these sources:

- **Sample size calculation in clinical trial using R** [@sample_size_calc_clinical_R]

- **Reproducible Medical Research with R** [@rmrwr]

- [**{pwr} vignette**](https://cran.r-project.org/web/packages/pwr/vignettes/pwr-vignette.html)

### Functions

Here is the summary of functions that help you calculate sample size or power based on statistical test

```{r include=FALSE, eval=FALSE}
cat.prop1
cat.prop2
cat.fisher
cat.chi
cat.mcnemar
```




```{r ss_param_tbl, include=FALSE}
# Parametric Test
ss_param_tbl <- tibble(
  id = c("num_p", "num.cat2_p", "num.cat2_p.paired", "num.cat3_p"),
  name = c(
    "One-sample t test", "Two-sample t test",
    "Paired t test", "One-way ANOVA"
  ),
  group_no = c("1", "2", "2", "≥3"),
  pkg = rep("pwr", 4),
  fn = c("`pwr.t.test(type = 'one.sample')`",
         "`pwr.t.test(type = 'two.sample')`",
         "`pwr.t.test(type = 'paired')`",
         "`pwr.anova.test()`")
)
```


```{r ss_non.param_tbl, include=FALSE}
# Non-parametric Test
ss_non.param_tbl <- tibble(
  id = c("num_np", "num.cat2_np", "num.cat2_np.paired", "num.cat3_np"),
  name = c(
    "One-sample Wilcoxon test", "Mann-Whitney U test",
    "Paired Wilcoxon test", "Kruskal-Wallis test"
  ),
  group_no = c("1", "2", "2", "≥3"),
  pkg = rep("pwr", 4),
  fn = NA # TODO
)
```

```{r ss_tbl, include=FALSE}
# Combine
ss_tbl <- bind_rows(
  list(
    "Continuous ~ Categorical / parametric" = ss_param_tbl,
    "Continuous ~ Categorical / nonparametric" = ss_non.param_tbl
  ),
  .id = "type"
) |>
  mutate(
    pkg = case_when(
      pkg == "pwr" ~ "[pwr](https://cran.r-project.org/package=pwr)")
    )
ss_tbl 
```



```{r echo=FALSE}
ss_tbl |> 
  select(-id) |> 
  gt(groupname_col = "type", rowname_col = "name") |> 
  fmt_markdown(columns = c(pkg, fn)) |> 
  cols_label(
    type = md("**Type**"),
    name = md("**Name**"),
    group_no = md("**# Groups**"),
    pkg = md("**Package**"),
    fn = md("**Function**")
    ) |> 
  tab_style(
    style = list(
      cell_fill(color = "#faedcd"),
      cell_text(weight = "bold")
      ),
    locations = cells_row_groups()
  ) |> 
  cols_width(
    c(group_no, pkg) ~ px(60)
  )
```

**TODO:** make table above from here

![R Packages and functions for sample size calculation, from  [@sample_size_calc_clinical_R]](/pic/sample-size-pkg.png)

### R Package Used

- **pwr** [@pkg_pwr]

- **WebPower** [@pkg_WebPower]