Skip to contents

Case Study: Hospital Discharge Analysis (DEIS Chile)

Rodolfo Tasso Suazo

2026-05-25

Source: vignettes/case-study-discharges.Rmd
case-study-discharges.Rmd

Introduction

This case study describes the technical workflow required to transform administrative health databases in Chile into a standardized analytical format. We use the Hospital Discharge databases published by the Department of Health Statistics and Information (DEIS) as a reference.

Health administrative records frequently present structural inconsistencies in ICD-10 coding. The two most common variations in the Chilean context are:

  1. Compact formats: Codes without a decimal point (e.g., J189 instead of J18.9).
  2. Filler suffixes: Use of the letter X to complete the field length in 3-digit categories (e.g., I10X for Essential hypertension).

These variations hinder interoperability and cross-referencing with international standards. The ciecl package automates the correction of these inconsistencies in a vectorized and efficient manner.

1. Original Data Structure

A synthetic dataset is generated below that replicates the structure and typical anomalies found in DEIS .csv files. The DIAG1 column represents the primary diagnosis with the aforementioned informal coding formats.

set.seed(42)

# Simulation of 200 records with typical DEIS Chile formats
discharges <- data.frame(
  DISCHARGE_ID = 1:200,
  PATIENT_ID   = sample(1:50, 200, replace = TRUE),
  YEAR         = sample(2018:2022, 200, replace = TRUE),
  DIAG1        = sample(
    c(
      "J189", "O800", "Z380", "K359", "N390",
      "I10X", "J449", "E119", "O829", "J069",
      "K922", "N185", "I509", "C509", "A099",
      "N40X", "K800", "I259", "J180", "E149"
    ),
    size    = 200,
    replace = TRUE
  ),
  stringsAsFactors = FALSE
)

head(discharges)
#>   DISCHARGE_ID PATIENT_ID YEAR DIAG1
#> 1            1         49 2018  J189
#> 2            2         37 2022  E119
#> 3            3          1 2021  E149
#> 4            4         25 2018  N390
#> 5            5         10 2022  I10X
#> 6            6         36 2021  C509

2. Code Normalization with cie_norm()

Normalization is the critical first step to ensure analysis integrity. The cie_norm() function processes codes by applying official MINSAL coding rules:

  • Suffix removal: Identifies and removes the trailing X.
  • Punctuation formatting: Inserts the decimal point in the standard position according to the ICD-10 hierarchy.
  • String cleaning: Removes whitespace and non-printable characters.
# Cleaning and standardization of diagnoses
discharges <- discharges %>%
  mutate(
    DIAG1_NORM = cie_norm(codes = DIAG1)
  )

# Comparison between original and normalized formats
discharges %>% 
  select(DIAG1, DIAG1_NORM) %>% 
  distinct() %>% 
  head(5)
#>   DIAG1 DIAG1_NORM
#> 1  J189      J18.9
#> 2  E119      E11.9
#> 3  E149      E14.9
#> 4  N390      N39.0
#> 5  I10X        I10

3. Semantic Enrichment with cie_describe()

After normalization, standardized clinical descriptions are assigned to facilitate result interpretation. The ciecl package provides the vectorized cie_describe() function, specifically designed to integrate into dplyr workflows efficiently.

Unlike a traditional left_join, cie_describe() directly returns a character vector, avoiding the creation of additional join columns and keeping the code cleaner.

# Direct integration of descriptions into the main dataframe
discharges_full <- discharges %>%
  mutate(
    description = cie_describe(DIAG1_NORM)
  )

head(discharges_full %>% select(DISCHARGE_ID, DIAG1, description))
#>   DISCHARGE_ID DIAG1
#> 1            1  J189
#> 2            2  E119
#> 3            3  E149
#> 4            4  N390
#> 5            5  I10X
#> 6            6  C509
#>                                                       description
#> 1                                       Neumonía, no especificada
#> 2                     Diabetes mellitus tipo 2 sin complicaciones
#> 3 Diabetes mellitus, no especificada, sin mención de complicación
#> 4              Infección de vías urinarias, sitio no especificado
#> 5                                Hipertensión esencial (primaria)
#> 6                 Tumor maligno de la mama, parte no especificada

For cases where full metadata is required (chapter, category, group), cie_lookup() can still be used in conjunction with a table join:

# Extracting full metadata via lookup + join
metadata <- cie_lookup(
  code = unique(discharges$DIAG1_NORM),
  full_description = TRUE
)

discharges_metadata <- discharges %>%
  left_join(metadata, by = c("DIAG1_NORM" = "codigo"))

In exploratory phases, where the exact code is unknown or transcription errors in the original clinical descriptions are suspected, cie_search() allows text searches using string similarity (fuzzy matching).

# Example of search with intentional typo ("diabetis")
# The function returns the most likely matches ordered by score
cie_search(text = "diabetis", threshold = 0.7)
#> # A tibble: 50 × 4
#>    codigo descripcion                                            score categoria
#>    <chr>  <chr>                                                  <dbl> <chr>    
#>  1 E10    Diabetes mellitus insulinodependiente                  0.917 E10 DIAB…
#>  2 E10.0  Diabetes mellitus tipo 1 con coma                      0.917 E10 DIAB…
#>  3 E10.1  Diabetes mellitus tipo 1 con cetoacidosis              0.917 E10 DIAB…
#>  4 E10.2  Diabetes mellitus tipo 1 con complicaciones renales    0.917 E10 DIAB…
#>  5 E10.3  Diabetes mellitus tipo 1 con complicaciones oftálmicas 0.917 E10 DIAB…
#>  6 E10.4  Diabetes mellitus tipo 1 con complicaciones neurológi… 0.917 E10 DIAB…
#>  7 E10.5  Diabetes mellitus tipo 1 con complicaciones  circulat… 0.917 E10 DIAB…
#>  8 E10.6  Diabetes mellitus tipo 1 con otras complicaciones esp… 0.917 E10 DIAB…
#>  9 E10.7  Diabetes mellitus tipo 1 con complicaciones múltiples  0.917 E10 DIAB…
#> 10 E10.8  Diabetes mellitus tipo 1 con complicaciones no especi… 0.917 E10 DIAB…
#> # ℹ 40 more rows

5. Comorbidity Index Calculation with cie_comorbid()

An advanced application of ciecl is population risk stratification through comorbidity indices. The cie_comorbid() function maps normalized diagnoses to Charlson or Elixhauser categories, adapted to the reality of Chilean data.

# Requires the 'comorbidity' package to be installed
# Calculation of the Charlson Index by patient identifier
comorbidities <- cie_comorbid(
  data = discharges,
  id   = "PATIENT_ID",
  code = "DIAG1",
  map  = "charlson"
)

# The result allows immediate use in statistical models
head(comorbidities, 10)

Process Summary

The ciecl workflow enables a reproducible transition from raw administrative data to an analytical dataset in four stages:

  1. Standardization: Format correction using cie_norm().
  2. Contextualization: Assignment of official glosses with cie_lookup().
  3. Validation: Term discovery and checking with cie_search().
  4. Aggregation: Generation of complex clinical indicators with cie_comorbid().

Data source: This tool uses the ICD-10 catalog standardized by the Centro FIC of the DEIS, Ministry of Health of Chile. For more information, visit deis.minsal.cl.