Productivity Commission admits to $3.6b error

Questions were raised after this year’s report released historical total expenditure figures dating back to 2011–12 that were, on average, $465 million lower than those reported in 2018.
 
Despite the unprecedented nature of the wholesale reductions, the Productivity Commission  initially dismissed the changes as a result of ‘adjustments’ made to historical expenditure data and the application of a ‘deflator’ to present the figures in 2017–18 dollars.
 
However, upon further questioning from newsGP, Assistant Commissioner Catherine Andersson, Head of Secretariat for the Steering Committee for the Review of Government Service Provision, conceded that errors had been made.
 
‘Please note in further reviewing attachment table 10A.2 [Australian Government total expenditure on GPs], staff have identified that Primary Health Network [PHN] expenditure and some GP incentive program expenditure were omitted from the table,’ Ms Andersson told newsGP.
 
‘The omission of PHN/Medicare Local reporting was the result of a coding error, meaning the originally published table presented Medicare Benefits Schedule [MBS] claims data only. This has been addressed for future reports.’
 
Overall, the Productivity Commission was forced to add more than $3.66 billion to its initially reported figures, including $619 million to 2017–18 funding, which increased last year’s total expenditure to $9.46 billion.
 
The revision did not impact 2011–12 historical figures, which remain $47 million lower in the 2019 Report on Government Services (RoGS) than the prior year.
 
GP numbers also affected
Expenditure data was not the only anomaly in this year’s report, and questions remain over prior revisions made to RoGS – particularly in relation to GP numbers and how they are calculated.
 
The 2019 RoGS saw historical Full Service Equivalent (FSE) GP numbers from 2013–14 to 2016–17 – supplied by the Department of Health (DoH) – dramatically cut.
 
Ms Andersson initially explained these changes as a result of ‘refinements DoH has made to its calculation methods’.
 
However, when it was pointed out that the most recent major change in methodology had occurred ahead of the 2016 RoGS – in which GP numbers again dropped substantially compared to prior years – the Productivity Commission revised its answer.
 
‘It is standard practice in the production of the RoGS to allow provision for data providers to revise historical data ... the DoH has advised that historical revisions usually reflect refinements to the Derived Major Specialty [DMS] classification,’ Ms Andersson said.
 
‘A more significant, and atypical, revision was made to published GP workforce data in early 2018 due to a coding error resulting in some double counting.’

The coding error apparently inflated FSE numbers for 2016–17 by nearly 2000, to the point where it was higher than the 25,149 recorded for the past 12 months.
 
However, the 2019 RoGS revised the total number of FSE GPs for 2016–17 to 23,911, allowing the Government to say it had increased numbers.

Breakdown of historical GP numbers as published in annual Reports on Government Services

 
Ms Andersson also pointed to a statement that was published on the DoH website at the time of the revision, but subsequently taken down.
 
‘A code error that affected the FSE figures released in 2015–16 and 2016–17 has been corrected. The error resulted in double counting of a group of fast growing items. Consequently, the published FSE figures were inflated,’ the statement read.
 
‘Headcount figures were minimally affected. The revised figures differ from figures previously published as they also reflect [DMS] classification changes since 2015–16.’
 
This statement appears to contradict some aspects of the 2019 RoGS.
 
Namely, FSE GP numbers for 2013–14 and 2014–15, apparently not affected by the coding error, still dropped substantially between the 2018 and 2019 RoGS. Additionally, 2015–16 figures purportedly corrupted by the coding error in 2017 are very close to the final reported figures in 2019.
 
Additionally, it is not clear how the change to DMS classification ahead of the 2016 RoGS resulted in substantial reductions to both historical total GP headcount numbers and FSE figures from 2006–07 to 2013–14.
 
newsGP sought clarification on these issues from the DoH, as well as more information on how the change to DMS classification and subsequent ‘refinements’ result in historical figures that change year-on-year, but it declined to respond before publication.
 
The DoH’s notes on workforce statistics describe DMS as a means for deriving the most appropriate speciality in each period for reporting purposes.
 
Prior to the change to DMS classification, historical total headcount statistics and FSE figures remained constant year-on-year.

Department of Health primary healthcare Productivity Commission