The health_system_za_hospitals_v1.csv is the first version of a bigger collaborative file (contributed by about 4 people) with data ranging from clusters in which hospitals belong to, to the number of beds per hospital amongst other hospital related attributes.
This issue gives some context and background on how we worked together in collecting, contributing and combining the data.
Since there are various sources in various formats used; I did not programmatically populate the datasets into that one file. Instead, I manually made changes per hospital row using data given about it. The changes.txt file has a record of what changes have I made from the original file when populating data that is in the file with bed numbers then turned it to this Hospitals v1 file. The main reason for manually populating data is because some of the hospital names have changed over time (different sources have different names) and for the hospitals whose names have not been changed, the naming convention differs from one source file to the other. Eg in this file where I started to work from, most names are shortened while on this file - full names are given. So it’s not easy to immediately identify a correct hospital match from one data source (where it could be shortened) to another different data sources (where it could be written in full name).
The table below shows what sources were used and the year when sources were last updated.
| Source | Contributor | Use of the source | Last Updated |
|---|---|---|---|
| Data from annual reports of some public hospitals | Herkulaas | Identification of Hospitals per province,areas of locations etc. | 2019 |
| Data from the Data library ** from DOH | Nompumelelo | Geo coordinates of public and private hospitals | 2018 |
| Data from various research outputs | Anelda | Hospital Bed numbers, number of surgeons in the public and private health sector. And more | From 2016 |
| Data from Gauteng's Cluster Policy | Vukosi | Public Hospital Number of beds per cluster in Gauteng | 2019 |
** To download health facility data from the data dictionary: Choose Data Dictionary > Data File > NIDS integrated > Groups > All Groups > Download > Org Unit Level > Down to Level 5 and then click submit.
Other sources used to validate minor data uncertainties like classification of a hospital include : this Gem file , Medpages and Google search engine. For number of beds in some public hospitals, this National Health Act doc was used. Statistics of South Africa URL - For Estimates of population per distriction from 2002 -2021. Raw Data here URL
Since this work is work in progress, it is incomplete. There could be inconsistencies with actual names used for hospitals or districts from which the hospitals are located to. These inconsistencies are mainly due to different sources of data from different times and mostly have hospital naming conventions that are not the same across various files used. The classification of hospitals is also a grey area.Some hospitals could classified as Regional hospital in one source and District in another. So at this point, there is no one complete perfect and reliable way of classifying hospitals. There are still about 353 hospitals without the bed numbers data. And there is also a small number of hospitals with hospital bed data but no geo coordinates. These limitations are to make the potential users of this data aware of the level of accurancy (or lack thereof) of this data at this point and that it might not be exactly the actual current real information about hospitals.
As it is apparent how incomplete this first version of hospital data is; from here we will work with the recent hospital data contributions made by other github users.This link shows what types of information is still missing.
In the hospital branch , there should be a couple of files related to hospitals (including the ones listed below). They contain portions of this main file (and a bit more in some files). They used for various purposes and they contributed to this main v1 of hospital data file.
1.health_system_za_hospital_id- This file contains all hospitals with their assigned ID's.
2.health_system_za_private_hospitals - This file has a list of private hospitals, not classified into what types of hospitals becuase access to private hospital data is limited at this point.
3.health_system_za_public_hospitals - This file has a list of different types of public hospitals.
4.health_system_za_public_hospitals_contacts - This file has contact details of some hospitals.
5.health_system_za_public_hospitals_extended_details - This file has a thorough description of hospitals and other details relating to minimum and maximum capacities in hospitals.
The sources below are indirectly used or to be used. For more healthcare data and research purposes, these links below are worth looking at:
1.Afrimapr work
2.South African Doctors
3.South Africa administrative levels
4.South African Medical Journal
5.Dell, Angela June's Thesis and raw data
6.Anelda van der Walt: merging open hospital datasets