Figure 3: Actual and forecasted national estimated dementia diagnosis rate for patients aged 65+, including upper and lower confidence limits, December 2019 - March 2020, forecast from April 2020 – November 2021
In Figure 3, an Autoregressive Integrated Moving Average Model (ARIMA model) has been used to investigate and forecast any trend in the national diagnosis rate, using the most recent 40 months of data (December 2016-March 2020). This has been done with a view of determining the likelihood that the national diagnosis rate will remain above the national ambition of 66.7% over the next year.
As can be seen in Figure 3, the national estimated diagnosis rate forecast remains stable and above the national ambition for the foreseeable future. The lower limits forecast by the model predict that the national rate dropping below the national ambition is a possibility in 2021, however as the lower limit it is unlikely to happen. This graph does not give any indication of whether the national diagnosis rate is likely to become statistically significantly different to the set ambition.
The ARIMA model used data to 31st March 2020; since the forecast was calculated, using data to 31st March 2020, data for the 30th April 2020 was collected and released with a national diagnosis rate for patients aged 65+ of 65.4%. When the diagnosis rate recorded in April 2020 is plotted against the ARIMA forecast for April 2020, the model shows the published diagnosis rate to be an outlier, falling far outside the prediction confidence intervals. This suggests the decrease observed was not due to chance, and was not a likely outcome given the trends seen in the previous 40 months of data and therefore a result of the coronavirus (COVID-19) pandemic. As the forecast of national diagnosis rate was not trained using COVID-19 affected data, the rates forecast by this ARIMA model could be compared to the actual national diagnosis rate from April onwards in order to estimate the impact of COVID-19.