Skip to main content
. 2023 Jan 31;9:e34982. doi: 10.2196/34982

Table 1.

PCCa and RMSEb values obtained for the entire prediction period (May 2014 to March 2018) at the national and regional levels, with all the combinations of data sources.

Levels and data sources Real time 1-week forecast 2-week forecast 3-week forecast

PCC RMSE PCC RMSE PCC RMSE PCC RMSE
National

AR(52)c 0.946 d 16.16 0.910 22.69 0.898 26.95 0.884 30.69

Google 0.830 42.75 0.803 44.99 0.801 41.27 0.770 38.96

EHRe 0.477 48.35 0.512 45.59 0.489 47.37 0.519 44.65

AR(52) and Google 0.941 18.10 0.896 24.17 0.871 26.98 0.847 28.24

AR(52) and EHR 0.932 16.41 0.880 21.58 0.820 26.15 0.823 25.93

Google and EHR 0.836 36.09 0.846 34.48 0.779 34.23 0.795 32.32

AR(52), Google, and EHR 0.936 21.26 0.903 20.94 0.856 24.16 0.845 25.33
Regional

AR(52) 0.725 40.75 0.705 44.18 0.670 47.65 0.681 49.12

Google 0.652 65.84 0.603 64.79 0.594 60.33 0.596 61.67

EHR 0.462 59.83 0.538 55.62 0.546 55.87 0.582 52.90

AR(52) and Google 0.738 42.07 0.665 46.44 0.616 47.82 0.619 47.74

AR(52) and EHR 0.697 40.99 0.685 42.38 0.637 46.48 0.634 46.31

Google and EHR 0.608 60.70 0.610 60.97 0.615 57.50 0.628 59.72

AR(52), Google, and EHR 0.724 42.12 0.689 45.24 0.646 47.37 0.620 52.19

aPCC: Pearson correlation coefficient.

bRMSE: root mean squared error.

cAR(52): autoregressive model of order 52.

dItalicization highlights the 2 highest correlations and lowest errors obtained with the models for real time and 1-week, 2-week, and 3-week forecasts.

eEHR: electronic health record.