BMC Medical Informatics and Decision Making

Table 3 Results of original Charlson models ¹ and enhanced Charlson ² models

From: The predictability of claim-data-based comorbidity-adjusted models could be improved by using medication data

MRDx³	c statistics		Hosmer-Lemeshow test
MRDx³	Charlson⁴	Enhanced⁵	Charlson⁴	Enhanced⁵
	(95%CI⁶)	(95%CI⁶)	chi-square (P)	chi-square (P)
ICH⁷	0.655	0.654	6.2 (0.63)	10.0 (0.27)
ICH⁷	(0.642-0.669)	(0.641-0.667)	6.2 (0.63)	10.0 (0.27)
Pneumonia	0.882	0.884	28.9 (<0.01)	26.7 (0.00)
Pneumonia	(0.876-0.888)	(0.878-0.890)	28.9 (<0.01)	26.7 (0.00)
Ischemic infarct	0.715	0.716	10.5 (0.23)	12.7 (0.12)
Ischemic infarct	(0.698-0.750)	(0.699-0.733)	10.5 (0.23)	12.7 (0.12)
AMI⁸	0.766	0.770	9.2 (0.32)	10.9 (0.21)
AMI⁸	(0.750-0.782)	(0.754-0.786)	9.2 (0.32)	10.9 (0.21)
Non-alcoholic liver disease	0.740	0.750	30.3 (<0.01)	32.2 (<0.01)
Non-alcoholic liver disease	(0.724-0.756)	(0.734-0.766)	30.3 (<0.01)	32.2 (<0.01)
Intracranial injury	0.724	0.724	18.9 (0.02)	16.3 (0.04)
Intracranial injury	(0.705-0.743)	(0.705-0.747)	18.9 (0.02)	16.3 (0.04)
CRF⁹	0.752	0.756	5.6 (0.69)	5.0 (0.76)
CRF⁹	(0.733-0.771)	(0.737-0.775)	5.6 (0.69)	5.0 (0.76)
COPD¹⁰	0.719	0.726	3.6 (0.89)	2.8 (0.95)
COPD¹⁰	(0.696-0.742)	(0.704-0.748)	3.6 (0.89)	2.8 (0.95)
Alcoholic liver disease	0.696	0.708	30.5 (<0.01)	27.1 (<0.01)
Alcoholic liver disease	(0.673-0.719)	(0.685-0.731)	30.5 (<0.01)	27.1 (<0.01)
Aspiration pneumonia	0.658	0.658	12.4 (0.13)	8.8 (0.36)
Aspiration pneumonia	(0.631-0.685)	(0.631-0.685)	12.4 (0.13)	8.8 (0.36)
CHF¹¹	0.633	0.641	6.2 (0.62)	4.7 (0.80)
CHF¹¹	(0.604-0.662)	(0.613-0.669)	6.2 (0.62)	4.7 (0.80)
Coronary atherosclerosis	0.847	0.861	16.3 (0.04)	21.0 (0.01)
	(0.827-0.867)	(0.842-0.880)

1 & 2 Multiple logistic regression models for predicting in-hospital mortalities composed of age + sex + status of health insurance + admission category (emergent or not) + operation (yes or no) + Charlson index score, before (Charlson models) and after (enhanced Charlson models) adding comorbidities inferred by drug prescription information, 3 Most responsive diagnoses, 4 Charlson models, 5 Enhanced Charlson models, 6 95% confidence interval calculated by bootstrapping, 7 Intracranial hemorrhage, 8 Acute myocardial infarction, 9 Chronic renal failure, 10 Chronic obstructive pulmonary disease, 11 Congestive heart failure.

Back to article page

ISSN: 1472-6947

Contact us

General enquiries: journalsubmissions@springernature.com