mt_adjust: Adjustment for multiple testing

In brittanyblouin/ANCRTAdjust: Adjusting HIV prevalence from ANC-RT data for common biases

Description

Adjusts the variables n_status_c (the number of women with an ascertained HIV status) and totpos_c (the total number of women with HIV) when multiple testing is suspected based on the presence of HIV testing coverage values greater than 100%.

Usage

mt_adjust(data, adjust_option)

Arguments

data

The ANC-RT dataset. The functions check_data and data_clean should have been run on the data to properly prepare the data for use here. The data set must have the following variables: n_status_c: Cleaned n_status (generated using the data_clean function) n_status_c_impute: n_status_c adjusted using the impute adjustment option (generated using the data_clean function) n_status_c_remove: n_status_c adjusted using the remove adjustment option (generated using the data_clean function) n_status_c_setmax: n_status_c adjusted using the set to maximum adjustment option (generated using the data_clean function) totpos_c: Cleaned totpos (generated using the data_clean function) totpos_c_impute: totpos_c adjusted using the impute adjustment option (generated using the data_clean function) totpos_c_remove: totpos_c adjusted using the remove adjustment option (generated using the data_clean function) totpos_c_setmax: totpos_c adjusted using the set to maximum adjustment option (generated using the data_clean function) n_clients: The number of women who attended the facility during the time period

adjust_option

The adjustment option chosen. Possible options include: impute: When n_status_c > n_clients, replace n_status_c with its facility's mean coverage (only including valid coverage observations to calculate the mean) multiplied by n_clients. remove: When n_status_c > n_clients, replace n_status_c as missing. setmax: When n_status_c > n_clients, replace n_status_c with n_clients. none: When n_status_c > n_clients, no adjustment is made.

Details

This function was designed to adjust the variables n_status_c and totpos_c, using one of the three adjustment options, when multiple testing is suspected (i.e. coverage values greater than 100% are present). The three options include: 1) impute; 2) remove; and, 3) set to maximum.

Value

A data set with the variables n_status_c and totpos_c, adjusted for multiple testing using the selected adjustment option, is returned. The adjustment option variables (e.g. n_status_c_impute, n_status_c_remove, totpos_c_impute, etc.) are removed from the dataset. Coverage and prevalence (for each facility-time period) are calculated using the newly adjusted variables.

Author(s)

Mathieu Maheu-Giroux

Brittany Blouin

brittanyblouin/ANCRTAdjust documentation built on Oct. 28, 2019, 4:53 a.m.