Using predicted length of stay to define treatment and model costs in hospitalized adults with serious illness: an evaluation of palliative care

Table 5 Main results for Aim 1: model diagnostics and treatment effect estimation for models differentiated by the use of estimated LOS in defining treatment

	Model diagnostics								Treatment effect estimation
	R²	RMSE is	MAPE is	RMSE os	MAPE os	MPE os	AIC	BIC	ATET	95%CI	CI width	t-test
*Model (i)*	0.14	47,261	27,501	47,459	27,614	49	60,450	60,515	− 11302**	−14,289 to −8314	5975	t = 0.83
*Model (ii)*	0.14	47,352	27,585	47,562	27,700	−9	60,482	60,547	− 9140**	−13,260 to − 5020	8240

**p < 0.01; *p < 0.05. For explanation of models, see Table 1. RMSE root-mean-square error, MAPE mean absolute percentage error, MPE mean percentage error, is in-sample, os out-of-sample, AIC Akaike Information Criterion, BIC Bayesian Information Criterion, ATET average treatment effect on the treated, CI Confidence interval

ISSN: 2191-1991