Results for the interventional autoregressive integrated moving average model summarizing the association of the COVID-19 pandemic and monthly age-standardized incidence rates for new pediatric cancer diagnoses per million population
Parameter | ARIMA (p d q)(P D Q)12* | Level shift intervention† | Trend change intervention‡ | ||
---|---|---|---|---|---|
Estimate of β (95% CI) | p value | Estimate of β (95% CI) | p value | ||
Cancer type | |||||
All cancers combined | (1 0 0)(0 0 0)¶** | 4.98 (−15.1 to 25.04) | 0.6 | −2.09 (−5.63 to 1.44) | 0.25 |
Leukemia and lymphoma | (1 0 0)(0 0 0) | 10.27 (−0.72 to 23.25) | 0.1 | −1.00 (−3.27 to 1.27) | 0.4 |
CNS tumour | (0 0 0)(0 1 1)¶** | −5.11 (−11.7 to 1.45) | 0.1 | NA‡ | – |
Extracranial solid tumour | (0 0 0)(0 1 1)¶** | −1.33 (−11.4 to 8.73) | 0.8 | NA | – |
Geographic region | |||||
Atlantic§ | (1 0 1)(0 0 0) | 11.98 (0.38.3 to 62.22) | 0.6 | NA | – |
Quebec | (1 0 1)(0 0 0) | −16.0 (−51.1 to 19.12) | 0.4 | NA | – |
Ontario | (2 0 0)(0 0 0) | −7.09 (−32.6 to 18.42) | 0.6 | NA | 0.7 |
Prairies¶ | (0 1 1)(0 1 1) | −38.7 (−126.0 to 48.41) | 0.4 | NA | – |
British Columbia | (0 0 0)(0 0 0) | 25.22 (−3.01 to 53.45) | 0.09 | NA | – |
Note: ARIMA = autoregressive integrated moving average, ASIR = age-standardized incidence rate, CI = confidence interval, CNS = central nervous system, NA = not applicable.
↵* For methodological considerations regarding ARIMA models, see Appendix 1, Supplement 2, available at www.cmaj.ca/lookup/doi/10.1503/cmaj.210659/tab-related-content.
↵† The level change measures an immediate change after the onset of the pandemic.
↵‡ The trend (slope) change measures a progressive increase or decrease in ASIRs in the pandemic period compared with the hypothetical continuations of the trends from the baseline period of March 2016 to February 2020.
↵§ Includes Nova Scotia, Prince Edward Island, Newfoundland and Labrador, and New Brunswick.
↵¶ Includes Manitoba, Alberta and Saskatchewan.
↵** Indicates that the trend change variable was not selected in the final model.