In Thaliehln/timci: Data management and visualisation for the TIMCI project

write(" o Facility checks", stderr())

Facility information can be extracted in five different manners:

• fid_ra: facility selected by the research assistant from a drop-down list - not implemented in Senegal;
• fid: facility ID (Fxxxx) that is extracted from the first 5 characters of the child ID (X-Fxxxx-Pxxxx), e.g., X-F0001-Pxxxx corresponds to facility F0001;
• fid_from_device: derived calculation from the facility ID (fid) that is the most used on a specific day with a specific device, assuming 1 device is used to screen children in 1 unique facility each day;
• fid_from_main_device: derived calculation from the facility ID (fid) that is the most used with a specific device (in Senegal, 1 unique device has been used to screen children in each facility);
• latitude and longitude from the tablet GPS coordinates - not implemented in Kenya.

facility_data <- facility_data %>% 
  dplyr::select(facility_cols)

Data from different facilities collected on the same device on the same day

This check is implemented for Tanzania only.

write("\to Data from two different facilities collected on the same device on the same day (Tanzania only)", stderr())

# Parameters for manual corrections
mc_description <- paste0('Manual correction of device IDs based on operational information from the field.')
to_correct_df <- facility_data
correction_type <- "correct_device_ids"

cat(knitr::knit_child('database_export_sub_corrections.Rmd',
                      envir = environment(),
                      quiet = TRUE))

n_deviceid_edit_records <- n_mc
facility_data <- corrected_df

Non-valid device IDs [Context check r qc_screening_w_nonvalid_deviceid]

write("\to Non-valid device IDs", stderr())

qc_description <- "Devices different from the ones whose IDs are listed in Section 2.3 of this report are considered as *non-valid* and may have been used for instance for training purposes. Records submitted by non-valid devices should not be considered as part of the main database."
qc_rule <- "Records submitted by devices different from the ones listed in Section 2.3 of this report are deleted."
qc_type <- "nonvalid_deviceids"
df <- facility_data
qc_text <- "non-valid device IDs"
qc_idx <- qc_screening_w_nonvalid_deviceid
qc_export_label <- "nonvalid_deviceids"
qc_export_description <- "the device ID is not valid"

cat(knitr::knit_child('database_export_sub_quality_check.Rmd',
                      envir = environment(),
                      quiet = TRUE))

n_nonvalid_deviceid_records <- n_detected

facility_data <- cleaned_df

Facility IDs for the CDSA pilot [Context check r qc_screening_other_fids]

write("\to Facility IDs for the CDSA pilot (India only)", stderr())

r if ( !is_india ) { paste0('This check is only implemented in TIMCI India.') }

qc_description <- "Records submitted by facilities which are part of a different TIMCI study and stored in the same ODK Central project should not be considered as part of the main database."
qc_rule <- "Records submitted by non-study facilities listed in Section 2.3 of this report are deleted."
qc_type <- "other_study_fids"
df <- facility_data
qc_text <- "facility IDs corresponding to another TIMCI study"
qc_idx <- qc_screening_other_fids
qc_export_label <- "screening_other_fids"
qc_export_description <- "the facility ID correspond to another TIMCI study"

cat(knitr::knit_child('database_export_sub_quality_check.Rmd',
                      envir = environment(),
                      quiet = TRUE))

other_timci_study_df <- qc_reuse_df
n_other_fid_records <- n_detected

facility_data <- cleaned_df

Facility start date [Context check r qc_screening_before_facility_startdate]

write(" o Facility start date context", stderr())

This check is implemented for Senegal only.

qc_description <- "Screening data are considered valid only from the specific facility start date, when a facility started recruiting after the study start date. Data may have been entered before this date for training purposes."
qc_rule <- paste0("Records entered before the specific facility start date are deleted")
qc_type <- "anterior_to_facility_startdate"
df <- facility_data
qc_text <- "an entry date anterior to the specific facility start date"
qc_idx <- qc_screening_before_facility_startdate
qc_export_label <- "anterior_to_facility_startdate"
qc_export_description <- "the entry date is anterior to the specific facility start date"
cat(knitr::knit_child('database_export_sub_quality_check.Rmd',
                      envir = environment(),
                      quiet = TRUE))

facility_data_before_start <- facility_data_before_start %>%
  dplyr::bind_rows(qc_df %>%
                     dplyr::select(date_visit, child_id))
n_before_facility_startdate_records <- n_detected

facility_data <- cleaned_df

Non-valid enrolling facility IDs [Compliance check r qc_nonvalid_fid]

write(" o Non-valid enrolling facility IDs", stderr())

r if ( is_kenya | is_tanzania ) { paste0('::: {custom-style="redparagraph"}\n\nInitial check ', qc_nonvalid_fid, 'a\n\n:::')}

qc_description <- "The facility ID **fid** should correspond to the ID of one of the research facility listed in Section 2.2 of this report."
qc_rule <- "Records corresponding to facility IDs not listed in Section 2.2 of this report are detected and manually corrected where possible."
qc_type <- "nonvalid_facility_ids"
df <- facility_data
idcol1 <- "fid"
refdf <- research_facilities
idcol2 <- "facility_id"
qc_text <- "facility ID not valid"
qc_idx <- paste0(qc_nonvalid_fid, 'a')
qc_export_label <- "nonvalid_fids"
qc_export_description <- "the facility ID is not valid"

cat(knitr::knit_child('database_export_sub_quality_check.Rmd',
                      envir = environment(),
                      quiet = TRUE))

n_nonvalid_fid_records <- n_detected

# Parameters for manual corrections
mc_description <- "Correction of the facility ID"
to_correct_df <- facility_data
correction_type <- "correct_non_valid_facilities"

qc_description <- "The facility ID **fid** should correspond to the ID of one of the research facility listed in Section 2.2 of this report."
qc_rule <- "Records corresponding to training data are deleted."
qc_type <- "nonvalid_ids"
df <- facility_data
idcol1 <- "fid"
refdf <- research_facilities
idcol2 <- "facility_id"
qc_text <- "facility ID not valid"
qc_idx <- paste0(qc_nonvalid_fid, 'b')
qc_export_label <- "nonvalid_fids"
qc_export_description <- "the facility ID is not valid"

cat(knitr::knit_child('database_export_sub_corrections.Rmd',
                      envir = environment(),
                      quiet = TRUE))

n_edit_nonvalid_fid_records <- n_mc

facility_data <- cleaned_df

Discrepancy in facility information 1 [Logic/Context check r qc_inconsistent_facility_info1]

write(" o Discrepancy in facility IDs (vs. device)", stderr())

::: Initial check r qc_inconsistent_facility_info1a :::

qc_description <- "Check the facility ID from the first 5 characters of the child ID (**fid**) versus from the device(s) which has(ve) been used in this facility on that day (**fid_from_device**)."
qc_rule <- action_alert_no_modification
qc_type <- "inconsistent_facility_info1"
df <- facility_data
qc_text <- "inconsistent facility info"
qc_idx <- paste0(qc_inconsistent_facility_info1, "a")
qc_export_label <- "inconsistent_facility_info1"
qc_export_description <- "facility information between the child ID and the device is inconsistent"

cat(knitr::knit_child('database_export_sub_quality_check.Rmd',
                      envir = environment(),
                      quiet = TRUE))

n_inconsistent_facility_info1 <- n_detected

mc_description <- ""
to_correct_df <- facility_data
correction_type <- "edit_day0_child_ids_to_correct_facilities_from_check1"

# Parameters for the quality check following manual corrections
qc_idx <- paste0(qc_inconsistent_facility_info1, "b")
qc_export_label <- "inconsistent_facility_info1"

cat(knitr::knit_child('database_export_sub_corrections.Rmd',
                      envir = environment(),
                      quiet = TRUE))

n_edit_inconsistent_fid1a <- n_mc
facility_data <- corrected_df

mc_description <- ""
to_correct_df <- facility_data
correction_type <- "correct_day0_inconsistent_facilities1"

# Parameters for the quality check following manual corrections
qc_idx <- paste0(qc_inconsistent_facility_info1, "c")
qc_export_label <- "inconsistent_facility_info1"

cat(knitr::knit_child('database_export_sub_corrections.Rmd',
                      envir = environment(),
                      quiet = TRUE))

n_edit_inconsistent_fid1b <- n_mc
facility_data <- corrected_df

Discrepancy in facility information 2 [Logic/Context check r qc_inconsistent_facility_info2]

write(" o Discrepancy in facility IDs (vs. manual entry)", stderr())

This check is not relevant and not implemented for Senegal.

r if ( is_rct | is_kenya ) { paste0('::: {custom-style="redparagraph"}\n\nInitial check ', qc_inconsistent_facility_info2, 'a\n\n:::')}

qc_description <- "Check the facility ID from the first 5 characters of the child ID (**fid**) versus from the research assistant's manual entry (**fid_ra**)."
qc_rule <- action_alert_no_modification
qc_type <- "inconsistent_facility_info2"
df <- facility_data
qc_text <- "inconsistent facility info"
qc_idx <- qc_inconsistent_facility_info2
qc_export_label <- "inconsistent_facility_info2"
qc_export_description <- "facility information between the child ID and the research assistant's entry is inconsistent"

cat(knitr::knit_child('database_export_sub_quality_check.Rmd',
                      envir = environment(),
                      quiet = TRUE))

n_inconsistent_facility_info2 <- n_detected

mc_description <- ""
to_correct_df <- facility_data
correction_type <- "edit_day0_child_ids_to_correct_facilities_from_check2"

# Parameters for the quality check following manual corrections
qc_idx <- paste0(qc_inconsistent_facility_info2, "b")
qc_export_label <- "inconsistent_facility_info2"

cat(knitr::knit_child('database_export_sub_corrections.Rmd',
                      envir = environment(),
                      quiet = TRUE))

n_edit_inconsistent_fid2 <- n_mc
facility_data <- corrected_df

mc_description <- ""
to_correct_df <- facility_data
correction_type <- "correct_day0_inconsistent_facilities2"

# Parameters for the quality check following manual corrections
qc_idx <- paste0(qc_inconsistent_facility_info2, "c")
qc_export_label <- "inconsistent_facility_info2"

cat(knitr::knit_child('database_export_sub_corrections.Rmd',
                      envir = environment(),
                      quiet = TRUE))

facility_data <- corrected_df
n_inconsistent_facility_info2 <- n_detected

Discrepancy in facility information 3 [Logic/Context check r qc_inconsistent_facility_info3]

This check is relevant and implemented only for Senegal.

write(" o Discrepancy in facility IDs (vs. main device) - Senegal only", stderr())

r if ( is_senegal ) { paste0('::: {custom-style="redparagraph"}\n\nInitial check ', qc_inconsistent_facility_info3, 'a\n\n:::')}

qc_description <- "Check the facility ID from the first 5 characters of the child ID (**fid**) versus from the device which is normally used in this facility **fid_from_main_device**."
qc_rule <- action_alert_no_modification
qc_type <- "inconsistent_facility_info3"
df <- facility_data
qc_text <- "inconsistent facility information"
qc_idx <- paste0(qc_inconsistent_facility_info3,"a")
qc_export_label <- "inconsistent_facility_info3"
qc_export_description <- "facility information between the child ID and fid_from_main_device is inconsistent"

cat(knitr::knit_child('database_export_sub_quality_check.Rmd',
                      envir = environment(),
                      quiet = TRUE))

n_inconsistent_facility_info3 <- n_detected

mc_description <- ""
to_correct_df <- facility_data
correction_type <- "correct_day0_inconsistent_facilities3"

# Parameters for the quality check following manual corrections
qc_idx <- paste0(qc_inconsistent_facility_info3, "b")
qc_export_label <- "inconsistent_facility_info3"

cat(knitr::knit_child('database_export_sub_corrections.Rmd',
                      envir = environment(),
                      quiet = TRUE))

n_edit_inconsistent_fid3 <- n_mc
facility_data <- corrected_df

n_edit_inconsistent_fid_records <- n_edit_inconsistent_fid1a + n_edit_inconsistent_fid1b + n_edit_inconsistent_fid2 + n_edit_inconsistent_fid3
n_inconsistent_fid_records <- n_inconsistent_facility_info1 + n_inconsistent_facility_info2 + n_inconsistent_facility_info3

write(" o Recalculate fid from device", stderr())

facility_data <- facility_data %>% 
  dplyr::select(facility_cols)

facility_data <- facility_data %>% 
  timci::allocate_screening_facility(research_facilities) %>%
  dplyr::select(c(facility_cols,
                  "fid_from_device",
                  "fid_from_main_device"))

Thaliehln/timci documentation built on April 8, 2024, 3:38 p.m.