Comparison of Rx-defined morbidity groups and diagnosis- based risk adjusters for predicting healthcare costs in Taiwan

Table 3 Predictive models for total cost

Morbidity Index	Source of morbidity	Predictors	Model performance - prediction of total cost
			Concurrent (year 2006)			Prospective (year 2007)
			R²	MAPE	MAPE*(%)	R²	MAPE	MAPE*(%)
(none)		Age + gender	0.089	505.9	99.5	0.092	582.9	102.9
Deyo's CCI	Diagnosis	Age + gender + CCIs	0.345	409.0	80.5	0.273	496.4	87.6
Elixhauser's Index	Diagnosis	Age + gender + E. Index	0.373	390.8	76.9	0.294	480.3	84.8
ADG	Diagnosis	Age + gender + ADGs	0.411	360.0	70.8	0.252	486.1	85.8
Rx-MG	Medication	Age + gender + Rx-MGs	0.618	297.5	58.5	0.360	448.5	79.1
ADG + Rx-MG	Diagnosis & medication	Age + gender + ADGs + Rx-MGs	0.650	277.4	54.6	0.382	430.4	75.9
ADG + Rx-MG	Diagnosis & medication	Age + gender + ADGs + Rx-MGs + prior total cost				0.465	386.1	68.1
ADG + Rx-MG	Diagnosis & medication	Age + gender + ADGs + Rx-MGs + prior total cost + (prior total cost)²				0.465	389.2	68.7

ISSN: 1472-6963