*2.3. Data Management and Analysis*

All dollar values were adjusted for inflation using the Reserve Bank of Australia's online inflation calculator and reported in second quarter of 2020 Australian dollars [16]. The main outcome variables were number of services per quarter and total cost of services per quarter. The number of providers was used as the denominator to calculate the following secondary outcome variables: number of services per provider per quarter and total cost per provider per quarter. The number of individuals with general treatment cover was used as the denominator to calculate the following secondary outcome variables: number of services per 100,000 insured population per quarter and total cost of services per 100,000 insured population per quarter.

Time series forecasting involved fitting seasonal autoregressive integrated moving average (ARIMA) models of service utilization data from 2015 to 2019 using the methods described by Hyndman and Athanasopoulos [17]. The seasonal ARIMA models provided estimates that account for seasonality and trends over time. Point forecast estimates with 95% prediction intervals for 2020 Q1 and Q2 were calculated from the seasonal ARIMA models and compared against observed values. The resulting mean errors and mean percentage errors were used as measures of absolute and relative impact. All statistical analyses were conducted using R, version 3.5.1 (R Foundation for Statistical Computing, Vienna, Austria) using the forecast and hts packages.
