Estimate Measures of Injury Epidemiology

2023-02-01

library(injurytools)
library(dplyr)
library(knitr)
library(kableExtra)

Whenever one collects and prepares the data, the next natural step is to summarize and explore these data. In this case, from a sports-applied point of view, one wants to know e.g.: how many and what type of injuries have occurred, how often they have occurred or what the load has been.

This document shows convenient functions from injurytools to describe sports injury data, in terms of measures used in sports injury epidemiology (see Bahr and Holme (2003) and Waldén et al. (2023)). Below, these measures are explained and then, injsummary() and injprev() functions are illustrated.

Measures of occurrence

Rates

As Hodgson Phillips (2000) state,

“Sports injuries occur when athletes are exposed to their given sport and they occur under specific conditions, at a known time and place.”

Thus, when attempting to describe the distribution of injuries it is necessary to relate this to the population at risk over a specified time period. This is why the fundamental unit of measurement is rate.

A rate is a measure that consists of a denominator and a numerator over a period of time. Denominator data can be a number of different things (e.g. number of minutes trained/played, number of matches played). As such, it reflects the speed at which new “injury-related” events occur.

Injury incidence rate

Injury incidence rate is the number of new injury cases (\(I\)) per unit of player-exposure time, i.e.

\[ I_{r} = \frac{I}{\Delta T}\] where \(\Delta T\) is the total time under risk of the study population.

Injury burden rate

Injury burden rate is the number of days lost (\(n_d\)) per unit of player-exposure time, i.e.

\[I_{br} = \frac{n_d}{\Delta T}\] where \(\Delta T\) is the total time under risk of the study population.

Prevalence

Prevalence, period prevalence, is a proportion that refers to the number of players that has reported the injury of interest (\(X\)) divided by the total player-population at risk at any time during the specified period of time (\(\Delta T\) time window). This includes players who already had the condition at the start of the study period as well as those who acquired it during that period.

\[P = \frac{X}{N}\] where \(X\) is the number of injury cases and \(N\) the total number of players in the study at any point in the window of time \(\Delta T\).

In practice

Again, as in the prepare-data vignette, we use the data sets available from the injurytools package: data from Liverpool Football Club male’s first team players over two consecutive seasons, 2017-2018 and 2018-2019, scrapped from https://www.transfermarkt.com/ website1.

- injsummary()

Prepare data 
df_exposures <- prepare_exp(raw_df_exposures, player = "player_name",
                            date = "year", time_expo = "minutes_played")
df_injuries  <- prepare_inj(raw_df_injuries, player = "player_name",
                            date_injured = "from", date_recovered = "until")
injd         <- prepare_all(data_exposures = df_exposures,
                            data_injuries  = df_injuries,
                            exp_unit = "matches_minutes")

Now, the preprocessed data is passed to injsummary() to calculate injury summary statistics:

injds <- injsummary(injd)
#> Warning in injsummary_unit(unit, injds, quiet): 
#>   Exposure time unit is matches_minutes
#>   So... Injury incidence and injury burden are calculated per 100 player-matches of exposure (90 minutes times 100)
#> Warning in injsummary_unit(unit, injds_overall, quiet): 
#>   Exposure time unit is matches_minutes
#>   So... Injury incidence and injury burden are calculated per 100 player-matches of exposure (90 minutes times 100)

We notice that it throws some warning messages (unless quiet = TRUE). They are thrown to make it clear what the exposure time unit is2.

What injsummary() returns as its output is a list of two elements, i.e. the output stored (here) in the injds object consists of,

str(injds, 1)
#> List of 2
#>  $ playerwise: tibble [28 × 9] (S3: tbl_df/tbl/data.frame)
#>  $ overall   : tibble [1 × 14] (S3: tbl_df/tbl/data.frame)
#>  - attr(*, "class")= chr [1:2] "injds" "list"
#>  - attr(*, "unit_exposure")= chr "matches_minutes"
#>  - attr(*, "unit_timerisk")= chr "100 player-match"

two data frames (two tables), which can be accessed by typing injds[[1]] (or injds[["playerwise"]]) and injds[[2]] (injds[["overal"]]).

To present the results in a more tidier and comprehensible way (instead of R code styled output) the following can be done:

# format the 'playerwise' data frame for output as a table
injds[[1]] %>% 
  arrange(desc(injincidence)) %>% # sort by decreasing order of injincidence
  head(10) %>%
  kable(digits = 2, col.names = c("Player", "N injuries", "N days lost", 
                                  "Mean days lost", "Median days lost", "IQR days lost",
                                  "Total exposure", "Incidence", "Burden"))
Player N injuries N days lost Mean days lost Median days lost IQR days lost Total exposure Incidence Burden
adam-lallana 6 302 43.14 43.0 18.5-52.5 700 77.14 3882.86
daniel-sturridge 3 122 30.50 33.5 12-52 927 29.13 1184.47
divock-origi 1 5 2.50 2.5 1.25-3.75 366 24.59 122.95
philippe-coutinho 3 62 15.50 18.5 9-25 1117 24.17 499.55
naby-keita 3 89 22.25 19.0 13.5-27.75 1393 19.38 575.02
dejan-lovren 6 160 22.86 13.0 9-28.5 3109 17.37 463.17
jordan-henderson 8 91 10.11 7.0 4-11 4154 17.33 197.16
xherdan-shaqiri 2 67 33.50 33.5 23.25-43.75 1057 17.03 570.48
fabinho 3 22 5.50 5.5 1.5-9.5 2013 13.41 98.36
james-milner 5 48 8.00 9.0 6.25-11.75 3548 12.68 121.76

We used RMarkdown, in particular knitr::kable() function, to report these tables in this way.

# format the table of total incidence and burden (main columns)
injds[[2]] %>% 
  select(1:8) %>% 
  data.frame(row.names = "TOTAL") %>% 
  kable(digits = 2,
        col.names = c("N injuries", "N days lost", "Mean days lost",
                      "Median days lost", "IQR days lost",
                      "Total exposure", "Incidence", "Burden"),
        row.names = TRUE) %>% 
  kable_styling(full_width = FALSE)
N injuries N days lost Mean days lost Median days lost IQR days lost Total exposure Incidence Burden
TOTAL 82 2049 18.97 7.5 1-20.25 74690 9.88 246.9

Note that to provide numbers that are easy to interpret and to avoid small decimals, injury incidence and injury burden are reported ‘per 100 player-match exposure’. As in this example exposure time is minutes played in matches, we multiply the rates by 90*100 (i.e. 90 minutes lasts a football match). The reported incidence rate is estimated by \(\hat{I}_r = \frac{82}{74690}\times90\times100\).

# format the table of total incidence and burden (point + ci estimates)
injds_tot_cis <- injds[[2]] %>% 
  select(7:last_col()) %>% 
  data.frame(row.names = "TOTAL")
injds_tot_cis$ci_injincidence <- paste0("[", round(injds_tot_cis$injincidence_lower, 1),
                                    ", ",round(injds_tot_cis$injincidence_upper, 1), "]")
injds_tot_cis$ci_injburden <- paste0("[", round(injds_tot_cis$injburden_lower, 1),
                                    ", ",round(injds_tot_cis$injburden_upper, 1), "]")

conf_level <- attr(injds, "conf_level")*100
injds_tot_cis %>% 
  select(1, 9, 2, 10) %>% 
  kable(digits = 2,
        col.names = c("Incidence",  paste0("CI", conf_level, "% for \\(I_r\\)"), 
                      "Burden", paste0("CI", conf_level, "% for \\(I_{br}\\)")))
Incidence CI% for \(I_r\) Burden CI% for \(I_{br}\)
TOTAL 9.88 [8.1, 11.7] 246.9 [237.9, 255.9]

Players with the highest injury incidence rate (all type of injuries) were Adam Lallana and Daniel Sturridge with 77.1 and 29.1 injuries per 100 player-matches respectively. The teams overall injury incidence was of 9.9 injuries per 100 player-matches and the injury burden of 246.9 days lost per 100 player-matches.

These summaries can be done by type of injury:

injstats_pertype <- injsummary(injd, var_type_injury = "injury_type", quiet = T)

These are the results of the team regarding injury incidence and injury burden by type of injury:

injstats_pertype[["overall"]] %>% 
  select(1:5, 7:11) %>% 
  mutate(ninjuries2 = paste0(ninjuries, " (", percent_ninjuries, ")"),
         ndayslost2 = paste0(ndayslost, " (", percent_dayslost, ")"),
         median_dayslost2 = paste0(median_dayslost, " (", iqr_dayslost, ")")) %>% 
  select(1, 11:13, 8:10) %>% 
  arrange(desc(injburden)) %>% 
  kable(digits = 2,
        col.names = c("Type of injury", "N injuries (%)", "N days lost (%)",
                      "Median days lost (IQR)",
                      "Total exposure", "Incidence", "Burden"),
        row.names = TRUE) %>% 
  kable_styling(full_width = FALSE)
Type of injury N injuries (%) N days lost (%) Median days lost (IQR) Total exposure Incidence Burden
1 Muscle 25 (30.49) 735 (35.87) 21 (12-36) 67266 3.34 98.34
2 Ligament 9 (10.98) 596 (29.09) 28 (7-54) 67266 1.20 79.74
3 Unknown 21 (25.61) 332 (16.2) 7 (4-18) 67266 2.81 44.42
4 Concussion 16 (19.51) 213 (10.4) 10.5 (5.75-14.5) 67266 2.14 28.50
5 Bone 11 (13.41) 173 (8.44) 9 (4.5-16.5) 67266 1.47 23.15

- injprev()

Prepare data 
df_exposures <- prepare_exp(raw_df_exposures, player = "player_name",
                            date = "year", time_expo = "minutes_played")
df_injuries  <- prepare_inj(raw_df_injuries, player = "player_name",
                            date_injured = "from", date_recovered = "until")
injd         <- prepare_all(data_exposures = df_exposures,
                            data_injuries  = df_injuries,
                            exp_unit = "matches_minutes")

We calculate the injury prevalence and the proportions of injury-free players on a season basis:

prev_table1 <- injprev(injd, by = "season")
prev_table1
#> # A tibble: 4 × 5
#>   season           type_injury     n n_player  prop
#>   <fct>            <fct>       <int>    <int> <dbl>
#> 1 season 2017/2018 Available       7       23  30.4
#> 2 season 2017/2018 Injured        16       23  69.6
#> 3 season 2018/2019 Available       2       19  10.5
#> 4 season 2018/2019 Injured        17       19  89.5

Making use of knitr::kable():

kable(prev_table1,
      col.names = c("Season", "Status", "N", "Total", "%"))
Season Status N Total %
season 2017/2018 Available 7 23 30.4
season 2017/2018 Injured 16 23 69.6
season 2018/2019 Available 2 19 10.5
season 2018/2019 Injured 17 19 89.5

Overall, there were more injured players in the 18-19 season than in the previous season. Let’s calculate it monthly:

prev_table2 <- injprev(injd, by = "monthly")

## compare two seasons July and August
prev_table2 %>%
  group_by(season) %>% 
  slice(1:4)
#> # A tibble: 8 × 6
#> # Groups:   season [2]
#>   season           month type_injury     n n_player  prop
#>   <fct>            <fct> <fct>       <int>    <int> <dbl>
#> 1 season 2017/2018 Jul   Available      21       23  91.3
#> 2 season 2017/2018 Jul   Injured         2       23   8.7
#> 3 season 2017/2018 Aug   Available      18       23  78.3
#> 4 season 2017/2018 Aug   Injured         5       23  21.7
#> 5 season 2018/2019 Jul   Available      16       19  84.2
#> 6 season 2018/2019 Jul   Injured         3       19  15.8
#> 7 season 2018/2019 Aug   Available      15       19  78.9
#> 8 season 2018/2019 Aug   Injured         4       19  21.1


## compare two seasons January and February
prev_table2 %>%
  group_by(season) %>% 
  slice(13:16)
#> # A tibble: 8 × 6
#> # Groups:   season [2]
#>   season           month type_injury     n n_player  prop
#>   <fct>            <fct> <fct>       <int>    <int> <dbl>
#> 1 season 2017/2018 Jan   Available      18       23  78.3
#> 2 season 2017/2018 Jan   Injured         5       23  21.7
#> 3 season 2017/2018 Feb   Available      21       23  91.3
#> 4 season 2017/2018 Feb   Injured         2       23   8.7
#> 5 season 2018/2019 Jan   Available       9       19  47.4
#> 6 season 2018/2019 Jan   Injured        10       19  52.6
#> 7 season 2018/2019 Feb   Available      12       19  63.2
#> 8 season 2018/2019 Feb   Injured         7       19  36.8

Looking at monthly basis, there were more differences w.r.t. player availability in Liverpool FC 1st male team, during the winter January/February months. More injured players in the 18-19 season.

prev_table3 <- injprev(injd, by = "monthly", var_type_injury = "injury_type")

Tidy up:

## season 1
prev_table3 %>% 
  filter(season == "season 2017/2018", month == "Jan") %>% 
  kable(col.names = c("Season", "Month", "Status", "N", "Total", "%"),
        caption = "Season 2017/2018") %>% 
  kable_styling(full_width = FALSE, position = "float_left")
## season 2
prev_table3 %>% 
  filter(season == "season 2018/2019", month == "Jan") %>% 
  kable(col.names = c("Season", "Month", "Status", "N", "Total", "%"),
        caption = "Season 2018/2019") %>% 
  kable_styling(full_width = FALSE, position = "left")
Season 2017/2018
Season Month Status N Total %
season 2017/2018 Jan Available 18 23 78.3
season 2017/2018 Jan Ligament 1 23 4.3
season 2017/2018 Jan Muscle 3 23 13.0
season 2017/2018 Jan Unknown 1 23 4.3
Season 2018/2019
Season Month Status N Total %
season 2018/2019 Jan Available 9 19 47.4
season 2018/2019 Jan Bone 1 19 5.3
season 2018/2019 Jan Concussion 2 19 10.5
season 2018/2019 Jan Ligament 1 19 5.3
season 2018/2019 Jan Muscle 3 19 15.8
season 2018/2019 Jan Unknown 4 19 21.1

Further implementation

In the near future there will be available the negative binomial (method = "negbin" argument), zero-inflated poisson (“zinfpois”) and zero-inflated negative binomial ("zinfnb") methods in injsummary() function.

Finally, this document shows how to perform descriptive analyses for injury epidemiology, but naturally, following these analyses further statistical inferences or multivariate regression analyses may be chosen to infer about the player’s/athletes population properties (e.g. to test whether there are differences between the injury incidence rates of two cohorts) or to evaluate the influence of independent factors (e.g. previous injuries, workload) on the injuries occurred.

References

Bahr, R, B Clarsen, and J Ekstrand. 2018. “Why We Should Focus on the Burden of Injuries and Illnesses, Not Just Their Incidence.” British Journal of Sports Medicine 52 (16): 1018–21. https://doi.org/10.1136/bjsports-2017-098160.
Bahr, R, and I Holme. 2003. “Risk Factors for Sports Injuries—a Methodological Approach.” British Journal of Sports Medicine 37 (5): 384–92. https://doi.org/10.1136/bjsm.37.5.384.
Hodgson Phillips, L. 2000. “Sports Injury Incidence.” British Journal of Sports Medicine 34 (2): 133–36. https://doi.org/10.1136/bjsm.34.2.133.
Waldén, Markus, Margo Mountjoy, Alan McCall, Andreas Serner, Andrew Massey, Johannes L Tol, Roald Bahr, et al. 2023. “Football-Specific Extension of the IOC Consensus Statement: Methods for Recording and Reporting of Epidemiological Data on Injury and Illness in Sport 2020.” British Journal of Sports Medicine. https://doi.org/10.1136/bjsports-2022-106405.

  1. These data sets are provided for illustrative purposes. We warn that they might not be accurate, there might be a mismatch and non-completeness with what actually occurred.↩︎

  2. Something that it is important to be aware of.↩︎