Data sources

Overview

This vignette outlines the data sources for the packaged datasets. It also summarises the cleaning process used for each dataset - although it is recommended that these details are checked against the coded implementations.

Bovine tuberculosis (TB) incidence estimates

Humans

Raw data

• Source from this systematic review by Muller et al.
• Data represents a list of studies estimating the proportion of overall TB cases that have zoonotic TB. Data includes zoonotic TB cases, and sample size (TB cases).
• Data is restricted to 61 countries.
• Data quality is extremely low with multiple issues.

Data cleaning

Initial data cleaning

• Added a logical variable capturing if studies are multi-year.
• Standardized missing data to be listed as NA.
• Extract country name by assuming it comes before the first comma in the area string.
• Extract the study end date by assuming (with some manual post processing) that the last 4 digits in the study period are the final year of the study.
• Replace NA with 0 zoonotic TB cases for studies that have no TB cases.

Exclusions

• Replace missing sample size data with TB case data.
• If the number of zoonotic TB cases is unknown attempt to back calculate using the sample size and proportion of cases that have zoonotic TB.
• Drop studies with an unknown study period.
• Drop studies from the EU as a whole.
• Drop all studies from the US due to low quality data.
• Drop all studies that look at subset of ages.
• Drop all studies that look at a specific gender.
• Exclude studies that did not look at the general population, identified TB cases, hospital populations, or identified TB cases in hospital populations.
• Drop studies that are not country wide if that country has country wide data available. (This is a major assumption that could be re-evaluated).
• If a study reports data on TB cases then drop duplicated data on the general/hospital population.
• Drop all studies that don’t list a sample size (This may be to restrictive as some studies may use the overall population size as the sample size).
• Split data-set into country-wide and non-country wide studies for further cleaning.

Country-wide data

• Excluded studies in sub-populations or high risk populations via manual inspection.
• Dropped data from New Zealand as a study is duplicated with differing sample sizes.
• Excluded duplicate entries for France and Ireland.
• Final data-set has data from 36 countries across 18 years (1990 - 2007)

Non-country wide data

• No exclusions based on sub-populations as all studies are effectively sampling sub-populations (risk, location, etc). This needs to be appropriately adjusted for at the analysis stage.
• Excluded duplicate entries from Sierra Leone and Argentina
• Final data-set has data from 14 countries across 15 years (1991 - 2008).

Cleaned data

• Country-wide and non-country wide data-sets joined
• Manually adjusted country names to map to those found in other TB incidence datasets. See EstZoonoticTB::link_data for error checking of this.
• Proportion of TB cases that were zoonotic TB estimated for each study
• Upper and lower confidence intervals (95%) also included.
• Standard error also included for later use in modelling/simulation.

EstZoonoticTB::zoonotic_tb_humans
#> # A tibble: 201 x 16
#>       id study_id dirty_id country geo_coverage study_pop sampling_strat
#>    <int>    <dbl>    <dbl> <fct>   <fct>        <fct>     <fct>         
#>  1     1        2      279 Austria Country-wid… Country-… All notified …
#>  2     2        2      280 Belgium Country-wid… Country-… All notified …
#>  3     3        2      281 Bulgar… Country-wid… Country-… All notified …
#>  4     4        2      282 Cyprus  Country-wid… Country-… All notified …
#>  5     5        2      283 Czechia Country-wid… Country-… All notified …
#>  6     6        2      284 Denmark Country-wid… Country-… All notified …
#>  7     7        2      285 Estonia Country-wid… Country-… All notified …
#>  8     8        2      286 France  Country-wid… Country-… All notified …
#>  9     9        2      287 Finland Country-wid… Country-… All notified …
#> 10    10        2      288 Germany Country-wid… Country-… All notified …
#> # … with 191 more rows, and 9 more variables: study_period <fct>,
#> #   study_end <dbl>, multi_year_study <fct>, cases <dbl>,
#> #   sample_size <dbl>, tb_z_prop <dbl>, tb_z_prop_lo <dbl>,
#> #   tb_z_prop_hi <dbl>, tb_z_prop_se <dbl>

Limitations

• Studies are often not representative of overall TB and zoonotic TB incidence.
• Zoonotic TB incidence is highly heterogeneous, being more likely to occur in rural areas and in populations that routinely eat raw dairy.

Animals

• Sourced from the OIE via personal communication
• Data cleaning details can be found in /data-raw/zoonotic_tb_animals.R
• Data available for 2018 split into half years.
• Data is available for domesticated and wild animals - stratified into present, limited, suspected, limited + suspected, not present and missing.
• Incidence or incidence rates are not available in this data set.
• Data quality is low.
• Last recorded status is also available but this field has currently been excluded due to its poor quality.
• Data for half years has been combined into data summarised across a year.

EstZoonoticTB::zoonotic_tb_animals
#> # A tibble: 187 x 5
#>    country                    country_code  year dom           wild        
#>    <fct>                      <fct>        <dbl> <fct>         <fct>       
#>  1 Afghanistan                AFG           2018 present       <NA>        
#>  2 Albania, Republic of       ALB           2018 present       <NA>        
#>  3 Algeria, People's Democra… DZA           2018 present       <NA>        
#>  4 Andorra, Principality of   AND           2018 not present   not present 
#>  5 Angola, Republic of        AGO           2018 not present   <NA>        
#>  6 Argentina, Argentine Repu… ARG           2018 limited       suspected +…
#>  7 Armenia                    ARM           2018 suspected + … not present 
#>  8 Australia, Commonwealth of AUS           2018 not present   not present 
#>  9 Austria, Republic of       AUT           2018 present       not present 
#> 10 Azerbaijan, Republic of    AZE           2018 not present   not present 
#> # … with 177 more rows

Tuberculosis (TB) incidence

• Sourced from the WHO using {getTBinR}.
• Extracted TB incidence, TB incidence rates (+ CI’s), proportion with extra-pulmonary TB, and proportion HIV positive (+ CI’s).
• Data prior to 2000 is dropped due to the high proportion of missing data.
• Data sourced using EstZoonoticTB::tb_data(). See ?EstZoonoticTB::tb_data() for documentation.

EstZoonoticTB::tb_data()
#> Downloading data from: https://extranet.who.int/tme/generateCSV.asp?ds=estimates
#> Saving data to: /tmp/RtmpTZSQOf/tb_burden.rds
#> Downloading data from: https://extranet.who.int/tme/generateCSV.asp?ds=mdr_rr_estimates
#> Saving data to: /tmp/RtmpTZSQOf/mdr_tb.rds
#> Joining TB burden data and MDR TB data.
#> Getting additional dataset: Latent TB infection
#> Downloading data from: https://extranet.who.int/tme/generateCSV.asp?ds=ltbi_estimates
#> Saving data to: /tmp/RtmpTZSQOf/latent_tb_infection.rds
#> Getting additional dataset: Notification
#> Downloading data from: https://extranet.who.int/tme/generateCSV.asp?ds=notifications
#> Saving data to: /tmp/RtmpTZSQOf/notification.rds
#> Getting additional dataset: Drug resistance surveillance
#> Downloading data from: https://extranet.who.int/tme/generateCSV.asp?ds=dr_surveillance
#> Saving data to: /tmp/RtmpTZSQOf/drug_resistance_surveillance.rds
#> Getting additional dataset: Non-routine HIV surveillance
#> Downloading data from: https://extranet.who.int/tme/generateCSV.asp?ds=tbhivnonroutinesurv
#> Saving data to: /tmp/RtmpTZSQOf/non-routine_hiv_surveillance.rds
#> Getting additional dataset: Outcomes
#> Downloading data from: https://extranet.who.int/tme/generateCSV.asp?ds=outcomes
#> Saving data to: /tmp/RtmpTZSQOf/outcomes.rds
#> Getting additional dataset: Budget
#> Downloading data from: https://extranet.who.int/tme/generateCSV.asp?ds=budget
#> Saving data to: /tmp/RtmpTZSQOf/budget.rds
#> Getting additional dataset: Expenditure and utilisation
#> Downloading data from: https://extranet.who.int/tme/generateCSV.asp?ds=expenditure_utilisation
#> Saving data to: /tmp/RtmpTZSQOf/expenditure_and_utilisation.rds
#> Getting additional dataset: Policies and services
#> Downloading data from: https://extranet.who.int/tme/generateCSV.asp?ds=policies
#> Saving data to: /tmp/RtmpTZSQOf/policies_and_services.rds
#> Getting additional dataset: Community engagement
#> Downloading data from: https://extranet.who.int/tme/generateCSV.asp?ds=community
#> Saving data to: /tmp/RtmpTZSQOf/community_engagement.rds
#> Getting additional dataset: Laboratories
#> Downloading data from: https://extranet.who.int/tme/generateCSV.asp?ds=labs
#> Saving data to: /tmp/RtmpTZSQOf/laboratories.rds
#> Joining TB burden data and additional datasets.
#> # A tibble: 4,474 x 12
#>    country iso3  g_whoregion  year tb_cases tb_inc tb_inc_lo tb_inc_hi
#>    <chr>   <chr> <chr>       <int>    <int>  <dbl>     <dbl>     <dbl>
#>  1 Afghan… AFG   Eastern Me…  2000    39000    190       123       271
#>  2 Afghan… AFG   Eastern Me…  2001    41000    189       123       271
#>  3 Afghan… AFG   Eastern Me…  2002    43000    189       122       270
#>  4 Afghan… AFG   Eastern Me…  2003    45000    189       122       270
#>  5 Afghan… AFG   Eastern Me…  2004    47000    189       122       270
#>  6 Afghan… AFG   Eastern Me…  2005    48000    189       122       270
#>  7 Afghan… AFG   Eastern Me…  2006    50000    189       122       270
#>  8 Afghan… AFG   Eastern Me…  2007    51000    189       122       270
#>  9 Afghan… AFG   Eastern Me…  2008    52000    189       122       270
#> 10 Afghan… AFG   Eastern Me…  2009    54000    189       123       270
#> # … with 4,464 more rows, and 4 more variables: prop_tb_ep <dbl>,
#> #   prop_hiv <dbl>, prop_hiv_lo <dbl>, prop_hiv_hi <dbl>

Demographics

• Sourced from the FAO.
• Data cleaning can be found in /data-raw/demographics.R.
• Population estimates are available by country from 1950 until 2018.
• The proportion of each country that is rural is also available.

EstZoonoticTB::demographics
#> # A tibble: 14,729 x 5
#>    country     country_code  year population prop_rural
#>    <fct>              <int> <int>      <dbl>      <dbl>
#>  1 Afghanistan            2  1950    7752118      0.940
#>  2 Afghanistan            2  1951    7839510      0.938
#>  3 Afghanistan            2  1952    7934980      0.936
#>  4 Afghanistan            2  1953    8038596      0.934
#>  5 Afghanistan            2  1954    8150447      0.931
#>  6 Afghanistan            2  1955    8270581      0.929
#>  7 Afghanistan            2  1956    8399030      0.927
#>  8 Afghanistan            2  1957    8535807      0.924
#>  9 Afghanistan            2  1958    8680946      0.921
#> 10 Afghanistan            2  1959    8834445      0.919
#> # … with 14,719 more rows

Animal Demographics

Domesticated

• Sourced from the FAO.
• Data cleaning can be found in /data-raw/animal_demographics.R.
• Animals included:
  • Cattle
• Additional animals that are vectors of Zoonotic TB could also be included as needed/thought appropriate

EstZoonoticTB::animal_demographics
#> # A tibble: 11,224 x 4
#>    country     country_code  year  cattle
#>    <fct>              <int> <int>   <int>
#>  1 Afghanistan            2  1961 2900000
#>  2 Afghanistan            2  1962 3200000
#>  3 Afghanistan            2  1963 3300000
#>  4 Afghanistan            2  1964 3350000
#>  5 Afghanistan            2  1965 3400000
#>  6 Afghanistan            2  1966 3600000
#>  7 Afghanistan            2  1967 3600000
#>  8 Afghanistan            2  1968 3633000
#>  9 Afghanistan            2  1969 3600000
#> 10 Afghanistan            2  1970 3700000
#> # … with 11,214 more rows