Results marked ‘non-urgent discussion’ - rate of notification without a result notification procedure

library(magrittr)
library(dplyr)
library(tidyr)
library(highcharter)

Executive Summary

During four measurement periods in 2015/2017, the Kensington site had approximately 100 results per week marked for ‘non-urgent discussion’. Although there was no mechanism to recall those patients if the result was not yet notified, about 55% of those results were notified within four weeks.

Introduction

Without any procedure to ‘recall’ patients who have a result marked ‘non-urgent discussion’, how many results are notified over the following weeks?

Results could be notified, even in the absence of a follow-up procedure because some, perhaps many, patients intentionally make an effort to find out the results after, for example, they have made the effort to have a test done.

Potentially time consuming and anxiety provoking efforts to ‘recall’ patients who have a result marked ‘non-urgent’ could consider

the number of results marked ‘non-urgent’
whether patients with a result marked ‘non-urgent’ already have an appropriate review appointment
the likelihood of a patient being notified in a timely fashion of a ‘non-urgent’ result, even in the absence of a deliberate ‘recall’.

Data

Collected with Medical Director at Kensington site of coHealth during months in 2015 and 2017.

At the time, Kensington site did not have an active procedure to ‘recall’ patients who had a result marked ‘non-urgent discussion’.

d <- data.frame(
  Age = c("0-1", "1-2", "2-3", "3-4", "4-5", "5-6", "6-7", "7-8", "8-9", "9-10", "10-11", "11-12"), # age in weeks
  Jul2015 = I(list(list(Not.Notified = 88, Notified = 16), # notified, not notified
                   list(Not.Notified = 79, Notified = 35),
                   list(Not.Notified = 81, Notified = 52),
                   list(Not.Notified = 30, Notified = 45),
                   list(Not.Notified = 25, Notified = 43),
                   list(Not.Notified = 36, Notified = 51),
                   list(Not.Notified = 14, Notified = 53),
                   list(Not.Notified = 22, Notified = 96),
                   list(Not.Notified = 31, Notified = 71),
                   list(Not.Notified = 21, Notified = 56),
                   list(Not.Notified = 31, Notified = 94),
                   list(Not.Notified = 26, Notified = 72))),
  Mar2017 = I(list(list(Not.Notified = 86, Notified = 15), # notified, not notified
                   list(Not.Notified = 41, Notified = 20),
                   list(Not.Notified = 26, Notified = 37),
                   list(Not.Notified = 75, Notified = 44),
                   list(Not.Notified = 42, Notified = 51),
                   list(Not.Notified = 23, Notified = 64),
                   list(Not.Notified = 43, Notified = 73),
                   list(Not.Notified = 38, Notified = 40),
                   list(Not.Notified = 30, Notified = 69),
                   list(Not.Notified = 47, Notified = 91),
                   list(Not.Notified = 47, Notified = 52),
                   list(Not.Notified = 11, Notified = 10))),
  Jun2017 = I(list(list(Not.Notified = 63, Notified = 4), # notified, not notified
                   list(Not.Notified = 49, Notified = 33),
                   list(Not.Notified = 47, Notified = 36),
                   list(Not.Notified = 41, Notified = 44),
                   list(Not.Notified = 38, Notified = 74),
                   list(Not.Notified = 22, Notified = 42),
                   list(Not.Notified = 45, Notified = 77),
                   list(Not.Notified = 18, Notified = 23),
                   list(Not.Notified = 26, Notified = 42),
                   list(Not.Notified = 38, Notified = 39),
                   list(Not.Notified = 34, Notified = 64),
                   list(Not.Notified = 26, Notified = 39))),
  Sep2017 = I(list(list(Not.Notified = 96, Notified = 32), # notified, not notified
                   list(Not.Notified = 68, Notified = 41),
                   list(Not.Notified = 96, Notified = 51),
                   list(Not.Notified = 85, Notified = 49),
                   list(Not.Notified = 91, Notified = 52),
                   list(Not.Notified = 91, Notified = 47),
                   list(Not.Notified = 78, Notified = 65),
                   list(Not.Notified = 82, Notified = 85),
                   list(Not.Notified = 55, Notified = 108),
                   list(Not.Notified = 62, Notified = 88),
                   list(Not.Notified = 69, Notified = 99),
                   list(Not.Notified = 45, Notified = 76))),
  row.names = "Age"
)

# create PercentageNotified
for (i in names(d)) {
  d[[i]] <- lapply(
    d[[i]],
    function(x) {
      list(Notified = x$Notified,
           Not.Notified = x$Not.Notified,
           Percentage.Not.Notified = round(x$Not.Notified / (x$Notified + x$Not.Notified) * 100))
      
    })
}

# create means
d$Means <- lapply(
  seq(to = nrow(d)),
  function(x) {as.list(rowMeans(sapply(slice(d, x), unlist)))}
)

# change to long format
d_long <- data.frame()
for (i in names(d)) {
  d_long <- rbind(
    d_long,
    do.call(
      rbind,
      mapply(
        function(x, y, z) {
          data.frame(Not.Notified = x$Not.Notified,
                     Notified = x$Notified,
                     Percentage.Not.Notified = x$Percentage,
                     Month = y,
                     AgeWeeks = z)
        },
        d[[i]], i, row.names(d[i]),
        SIMPLIFY = FALSE)
    )
  )
}

d_long <- d_long %>%
  gather(key = Group, value = n, "Not.Notified", "Notified", "Percentage.Not.Notified") %>%
  mutate(Series = paste(Month, Group))

Charts

Proportion of results notified

highchart() %>%
  hc_add_series(
    d_long %>% filter(!grepl("Means", Series)),
    "line",
    hcaes(x = AgeWeeks,
          y = n,
          group = Series),
    visible = FALSE
  ) %>%
  hc_title(text = "Results marked 'Discuss' as of 7th July 2015, 24th March 2017, 5th June 2017,19th September 2017") %>%
  hc_subtitle(text = "Results notified/not-yet-notified") %>%
  hc_xAxis(title = list(text = "Weeks")) %>%
  hc_yAxis(title = list(text = "n")) %>%
  hc_size(height = 600) %>%
  hc_add_series(
    d_long %>% filter(grepl("Means", Series) & !grepl("Percentage", Series)),
    "line",
    hcaes(x = AgeWeeks,
          y = n,
          group = Series),
    visible = FALSE
  ) %>%
  hc_add_series(
    d_long %>% filter(grepl("Means", Series) & grepl("Percentage", Series)),
    "line",
    hcaes(x = AgeWeeks,
          y = n,
          group = Series),
    visible = TRUE # just show this series initially
  )

In the average (mean) of four series, 55% results marked ‘non-urgent discussion’ are discussed within four weeks of the result. After four weeks, relatively few results marked ‘non-urgent’ are likely to be marked ‘discussed’ in the subsequent six weeks.

Number of results marked ‘non-urgent discussion’

highchart() %>%
  hc_add_series(
    d_long %>% filter(!grepl("Means", Series)),
    "line",
    hcaes(x = AgeWeeks,
          y = n,
          group = Series),
    visible = FALSE
  ) %>%
  hc_title(text = "Results marked 'Discuss' as of 7th July 2015, 24th March 2017, 5th June 2017,19th September 2017") %>%
  hc_subtitle(text = "Results notified/not-yet-notified") %>%
  hc_xAxis(title = list(text = "Weeks")) %>%
  hc_yAxis(title = list(text = "n")) %>%
  hc_size(height = 600) %>%
  hc_add_series(
    d_long %>% filter(grepl("Means", Series) & !grepl("Percentage", Series)),
    "line",
    hcaes(x = AgeWeeks,
          y = n,
          group = Series),
    visible = TRUE
  ) %>%
  hc_add_series(
    d_long %>% filter(grepl("Means", Series) & grepl("Percentage", Series)),
    "line",
    hcaes(x = AgeWeeks,
          y = n,
          group = Series),
    visible = FALSE
  )

In the months sampled, approximately 100 results are marked ‘non-urgent discussion’ every week.