Maximum likelihood estimation for ARIMA testing the short-term effects of the COVID-19 pandemic on emergency department encounters* for suicide attempt and intentional self-harm, UMHS 2015–2020†
| Parameter | β‡ (SE) | P value |
| Step effect§ | ||
| Intercept | 1.963 (0.052) | <0.001 |
| AR lag 1 | 1.434 (0.079) | <0.001 |
| AR lag 2 | −0.444 (0.073) | <0.001 |
| MA lag 1 | −1.077 (0.084) | <0.001 |
| MA lag 2 | 0.110 (0.070) | 0.12 |
| COVID-19 pandemic | −0.509 (0.127) | <0.001 |
| Ramp effect¶ | ||
| Intercept | 1.954 (0.055) | <0.001 |
| AR lag 1 | 1.425 (0.074) | <0.001 |
| AR lag 2 | −0.434 (0.070) | <0.001 |
| MA lag 1 | −1.066 (0.080) | <0.001 |
| MA lag 2 | 0.103 (0.069) | 0.13 |
| Time from onset of COVID-19 pandemic | −0.003 (0.001) | <0.001 |
*Natural log of total encounters per day.
†1 October 2015 to 31 October 2020; observation days n=1858.
‡Can be converted to percentage change using the equation (eβ−1)*100.
§ARIMA(2,1,2); AIC=3015.257.
¶ARIMA(2,1,2); AIC=3347.07.
AIC, Akaike information criterion; AR, autoregressive; ARIMA, autoregressive integrated moving average; MA, moving average; UMHS, University of Michigan Health System.