Tutorial: Computing WHODAS 2.0 Scores Using R

Introduction

This tutorial demonstrates how to compute the WHODAS 2.0 (World Health Organization Disability Assessment Schedule 2.0) scores using R. The WHODAS 2.0 is a tool designed to measure disability across six domains of functioning. Although a manual exists detailing how to compute the overall score using SPSS (see Measuring Health and Disability. Manual for WHO Disability Assessment Schedule WHPDAS 2.0, a similar code for R has not been available until now. In this tutorial, we will go through a custom R function to calculate the WHODAS 2.0 scores.

Generating Example Data

First, let’s create some example data to work with. We’ll generate a synthetic dataset of WHODAS responses, simulating responses from 10 individuals.

Code

# Defining the response levels
whodas_responses <- fct(c("None", "Mild", "Moderate", "Severe", "Extreme or cannot do", NA))

# Defining the column names corresponding to the WHODAS items
column_names <- c(
  "D1_1", "D1_2", "D1_3", "D1_4", "D1_5", "D1_6",
  "D2_1", "D2_2", "D2_3", "D2_4", "D2_5",
  "D3_1", "D3_2", "D3_3", "D3_4",
  "D4_1", "D4_2", "D4_3", "D4_4", "D4_5",
  "D5_2", "D5_3", "D5_4", "D5_5", "D5_8", "D5_9", "D5_10", "D5_11",
  "D6_1", "D6_2", "D6_3", "D6_4", "D6_5", "D6_6", "D6_7", "D6_8"
)

# Generating the data frame
df_whodas <- data.frame(matrix(ncol = 1 + length(column_names), nrow = 10))
colnames(df_whodas) <- c("id", column_names)

# Populating the data frame with random responses
df_whodas <- df_whodas %>%
  mutate(
    id = row_number(),
    across(
      contains("D", ignore.case = F), ~ sample(
        x = whodas_responses,
        size = 10,
        replace = T)
      )
    )

The sample data is presented in Table Table 1.

Code

df_whodas

Table 1:

Sample data for the WHODAS tutorial

Recoding and Computing WHODAS Scores

Now that we have our data, we’ll define a function to recode the responses and compute the WHODAS scores. This function takes in the WHODAS data and an optional argument to specify whether to include remunerated work items in the computation.

This function takes two arguments:

1. dat: The dataset containing the WHODAS 2.0 responses.
2. remunarated_work_items: A flag (either TRUE or FALSE) indicating whether to include remunerated work items in the computation. The default is FALSE.

The f_recode_whodas function performs the following tasks:

• Checks for the presence of required columns in the dataset.
• Recodes the responses based on specified rules.
• Computes summary scores for each domain (e.g., Do1, Do2, etc.).
• Calculates the overall WHODAS summary score (st_s32 or st_s36), with or without the remunerated work items.
• Adjusts the domain and overall summary scores by applying weights to normalize the scores to a 100-point scale.
• Returns the original dataset with additional columns for the domain and overall summary scores.

Code

f_recode_whodas <- function(dat,
                            remunarated_work_items = F) {
  
  # List of original column names and corresponding new column names
  col_names <- c(
    "D1_1", "D1_2", "D1_3", "D1_4", "D1_5", "D1_6",
    "D2_1", "D2_2", "D2_3", "D2_4", "D2_5",
    "D3_1", "D3_2", "D3_3", "D3_4",
    "D4_1", "D4_2", "D4_3", "D4_4", "D4_5",
    "D5_2", "D5_3", "D5_4", "D5_5", 
    "D6_1", "D6_2", "D6_3", "D6_4", "D6_5", "D6_6", "D6_7", "D6_8"
  )
  
  # Include additional columns for remunerated work items if specified
  if (remunarated_work_items) {
    col_names <- c(col_names, "D5_8", "D5_9", "D5_10", "D5_11")
  }
  
  
  # Check for the presence of required columns in the input dataset
  if(length(setdiff(col_names, names(dat))) > 0) stop(cat(
    "All WHO-DAS D columns need to be present in dat. The following are missing:",
    setdiff(col_names, names(dat))))
  
  # Generate new column names by removing underscores
  new_col_names <- gsub("_", "", col_names)
  
  # Define recode rules for general and specific columns
  recode_rules_general <- c(0, 1, 2, 3, 4)
  recode_rules_specific <- c(0, 1, 1, 2, 2)
  
  # Iterate over each column to apply recode rules
  for (i in seq_along(col_names)) {
    original_col_name <- col_names[i]
    new_col_name <- new_col_names[i]
    
    # Apply specific recode rules for certain columns
    if (original_col_name %in% c("D1_5", "D1_6", "D2_2", "D2_3", "D3_1", "D3_3", "D3_4", 
                                 "D4_1", "D4_2", "D4_3", "D4_5", "D5_2", "D5_3", "D5_5", 
                                 "D5_8", "D6_1", "D6_3", "D6_6", "D6_8")) {
      dat[[new_col_name]] <- recode_rules_specific[dat[[original_col_name]]]
    } else {
      dat[[new_col_name]] <- recode_rules_general[dat[[original_col_name]]]
    }
  }
  
  # Summary scores of domains (do), where domain 1 is abbreviated as Do1, etc.
  dat$Do1 <- (dat$D11 + dat$D12 + dat$D13 + dat$D14 + dat$D15 + dat$D16)
  dat$Do2 <- (dat$D21 + dat$D22 + dat$D23 + dat$D24 + dat$D25)
  dat$Do3 <- (dat$D31 + dat$D32 + dat$D33 + dat$D34)
  dat$Do4 <- (dat$D41 + dat$D42 + dat$D43 + dat$D44 + dat$D45)
  dat$Do51 <- (dat$D52 + dat$D53 + dat$D54 + dat$D55)
  dat$Do6 <- (dat$D61 + dat$D62 + dat$D63 + dat$D64 + dat$D65 + dat$D66 + 
                dat$D67 + dat$D68)
  
  # WHODAS summary score without remunerated work items
  dat$st_s32 <- dat$Do1 + dat$Do2 + dat$Do3 + dat$Do4 + dat$Do51 + dat$Do6 
  
   # Include remunerated work items if specified
  if(remunarated_work_items){
    
    # Domain summary score for remunerated
    dat$Do52 <- (dat$D58 + dat$D59 + dat$D510 + dat$D511)
    
    # WHODAS summary score without remunerated work items
    dat$st_s36 <- dat$st_s32 + dat$Do52
  }
   
  # Apply weights to normalize domain and overall summary scores to a 100-point scale
  dat$Do1 <- dat$Do1 * 100/20
  dat$Do2 <- dat$Do2 * 100/16
  dat$Do3 <- dat$Do3 * 100/10
  dat$Do4 <- dat$Do4 * 100/12
  dat$Do51 <-dat$Do51 * 100/10
  dat$Do6 <- dat$Do6 * 100/24
  dat$st_s32 <-  dat$st_s32 * 100/92
  
  if(remunarated_work_items){
    dat$Do52 <- dat$Do52 * 100/14
      dat$st_s36 <- dat$st_s36 * 100/106
  }
  
  # Return the dataset with additional summary score columns
  dat
}

Running the function

To execute the function, we can pass the dataset as the dat argument. If the dataset includes remunerated work items, set remunarated_work_items to TRUE. The function will return a new dataset with the computed domain and overall summary scores.

The result is shown in Table Table 2.

Code

f_recode_whodas(dat = df_whodas, remunarated_work_items = T)

Table 2:

Results table after appyling function f_recode_whodas

Conclusion

With the f_recode_whodas function, calculating WHODAS 2.0 scores in R is now straightforward. We can easily integrate this function into the data analysis workflow to compute disability scores.