Skip to contents

Adjust Service Data for Coverage Analysis

Source: R/1c_adjust_service_data.R
adjust_service_data.Rd

The adjust_service_data function processes DHIS-2 health service data, correcting for incomplete reporting, applying k-factor adjustments for accurate scaling, and managing outliers for consistency in data analysis. This standardization supports reliable comparisons in immunization coverage and health service utilization studies.

Usage

adjust_service_data(
  .data,
  adjustment = c("default", "custom", "none"),
  k_factors = NULL
)

Arguments

.data

A cd_data dataframe, typically containing health facility data from DHIS-2 with monthly service counts by district.

adjustment

A character string specifying the type of adjustment to apply:

  • "default": Applies a default k-factor of 0.25 across all indicator groups.

  • "custom": Uses user-defined k_factors for different indicator groups, with values between 0 and 1.

  • "none": Returns the data without any adjustments.

k_factors

A named numeric vector of custom k-factor values between 0 and 1 for each indicator group (e.g., c(anc = 0.3, idelv = 0.2, ...)). Used only if adjustment = "custom".

Value

A cd_adjusted_data object containing adjusted service data, where outliers are flagged and managed, and missing values are imputed.

Details

This function prepares service data through a series of steps to ensure data quality and consistency:

  1. Validation: Checks the structure of .data and ensures that the adjustment argument is valid. For custom adjustments, k_factors must be specified and contain valid values.

  2. k-Factor Defaults: Default k-factor values are set to 0.25 for each indicator group, unless overridden by user-provided values in k_factors.

  3. Reporting Completeness: Flags any district-year reporting rates below 75% and imputes missing data using district-level medians to account for reporting inconsistencies.

  4. k-Factor Scaling: Adjusts service counts based on the k-factor and reporting rate using the following scaling formula: $$AdjustedValue = Value \times \left(1 + \left(\frac{1}{ReportingRate/100} - 1\right) \times k\right)$$ where ReportingRate is the reporting rate in percent, and k is the k-factor for the indicator group.

  5. Outlier Detection: Identifies and flags extreme outliers using Hampel's X84 method, marking values that exceed 5 Median Absolute Deviations (MAD) from the median.

  6. Data Imputation: Replaces remaining missing values with district-level medians to ensure data completeness.

See also

new_countdown() for creating cd_data objects and generate_adjustment_values() for generating adjustment summaries.

Examples

if (FALSE) { # \dontrun{
# Default adjustment
adjusted_data <- adjust_service_data(data, adjustment = "default")

# Custom adjustment with specific k-factors
custom_k <- c(anc = 0.3, idelv = 0.2, pnc = 0.35, vacc = 0.4,
              opd = 0.3, ipd = 0.25)
adjusted_data_custom <- adjust_service_data(data, adjustment = "custom",
                                            k_factors = custom_k)

# No adjustment
unadjusted_data <- adjust_service_data(data, adjustment = "none")
} # }