Skip to contents

Environmental surveillance performance by district

Source: R/es_performance.R
calc_es_district_performance.Rd

Computes ES performance for each district over the reporting window.

Usage

calc_es_district_performance(
  samples,
  as_of = Sys.Date(),
  from = NULL,
  to = NULL,
  admin_units = NULL,
  guid_lookup = NULL,
  pv_pattern = "^WPV|^cVDPV|^VDPV|^aVDPV|^iVDPV",
  ev_threshold = 50,
  min_samples = 5L,
  measured_label = "Measured",
  thin_label = "Too few samples",
  absent_label = "No ES site",
  window_months = 12L,
  resolved_min = 0.8,
  settle_months = 3L,
  digits = NULL,
  cols = list()
)

Arguments

samples

Cleaned environmental samples, one row per sample, as returned by clean_es().

as_of

The reporting date, as a Date or a string that as.Date() reads. Defaults to Sys.Date(). Samples collected after it, or with no collection date, are left out.

from, to

First and last collection date of the reporting window. When both are NULL (the default) the window is computed from samples and as_of with es_reporting_window(), using window_months, resolved_min and settle_months.

admin_units

Optional table of every district to report, with the columns named by cols$samples$adm1, cols$samples$adm2 and cols$samples$adm2_guid. A district with no sample in the window is kept with zero counts. NULL (default) reports only districts with a sample in the window.

guid_lookup

Optional table with columns from and to that maps a retired district key to its current key. Keys not listed are kept.

pv_pattern

Regular expression a virus type must match to count as a poliovirus. Default "^WPV|^cVDPV|^VDPV|^aVDPV|^iVDPV", wild and vaccine-derived poliovirus, not Sabin or NPEV.

ev_threshold

Enterovirus share, in per cent, at or above which a unit meets the standard. Default 50.

min_samples

Fewest resolved samples for a percentage to be measured. Default 5.

measured_label, thin_label, absent_label

Labels for the three states. Defaults "Measured", "Too few samples" and "No ES site".

window_months

Number of collection months in the window. Default 12.

resolved_min

Share of a month's samples that must be resolved for the month to count as complete. Default 0.80.

settle_months

Number of months after which a month still below resolved_min is taken as final. Default 3.

digits

Decimal places to round percentages to, or NULL (default for districts) for no rounding.

cols

Named list overriding the default column names, keyed by table, in the same shape as in calc_polio_indicators(), for example list(samples = list(vtype = "virus_type")). The samples defaults are the polished ES columns, with site_name = "site_name", plus vtype = "vtype", site_guid = "site_guid", longitude = "site_x_coordinate", latitude = "site_y_coordinate" and district_key = "adm2_guid", the column samples are counted by district on.

Value

A tibble with one row per district: level ("adm2"), guid, name, the province name adm1, sites, samples, resolved, pending, ev_n, pv_n, ev_pct, pv_pct, ev_state and ev_meets_standard. ev_pct and pv_pct are on a scale of 0 to 100.

Details

  • Enterovirus share (ev_pct): samples with any virus type recorded (ev_n, a non-polio enterovirus or any poliovirus) over resolved samples (resolved), times 100. The GPEI standard is 50 per cent or more (ev_threshold). This differs from the ev_rate indicator in calc_polio_indicators(), which divides by all samples, and from sites_with_entero_pct, which counts a site as meeting the standard above 49 per cent. Here samples still at the laboratory are left out of the denominator, and the standard is met at 50 per cent or more.

  • Poliovirus share (pv_pct): samples whose virus type matches pv_pattern (pv_n) over resolved samples, times 100. No threshold.

  • Minimum-sample rule (ev_state): a district with fewer than min_samples resolved samples is marked as having too few, and one with none as absent. See es_sample_state().

A percentage is NA when the district has no resolved samples. Samples are counted by district on district_key, so a retired district GUID can be repaired before the call (supply the repaired key as a column and point cols$samples$district_key at it) or through guid_lookup.

Examples

es <- tibble::tibble(
  adm1 = "P1", adm2 = "D1", adm2_guid = "G1", site_guid = "S1",
  collection_date = as.Date("2024-01-01") + c(0, 31, 60, 91, 121, 152),
  classification_all = "DONE",
  vtype = c("NPEV", NA, "NPEV", "cVDPV2", NA, "NPEV")
)
calc_es_district_performance(es, as_of = as.Date("2024-07-15"), window_months = 6)
#> # A tibble: 1 × 14
#>   level guid  name  adm1  sites samples resolved pending  ev_n  pv_n ev_pct
#>   <chr> <chr> <chr> <chr> <int>   <int>    <int>   <int> <int> <int>  <dbl>
#> 1 adm2  G1    D1    P1        1       6        6       0     4     1   66.7
#> # ℹ 3 more variables: pv_pct <dbl>, ev_state <fct>, ev_meets_standard <lgl>