Primary care practices’ ability to predict future risk of expenditures and hospitalization using risk stratification and segmentation

Table 3 Difference between practices using adjudication (N = 4, patients scored 2,088), clinical intuition on key metrics (N = 1, patients scored 359), or computer algorithm alone (N = 1, patients scored 934)

	Algorithm plus adjudication			Clinical intuition			Computer algorithm
	Practice	HCC	Δ	Practice	HCC	Δ	Practice	HCC	Δ
ED ≥ 2 (12%)
Sensitivity	0.40	0.37	0.03	0.15	0.33	− 0.18*	0.14	0.17	− 0.03
PPV	0.30	0.28	0.02	0.15	0.32	− 0.17*	0.19	0.23	− 0.04
Accuracy	0.72	0.70	0.02	0.78	0.82	− 0.04	0.87	0.87	0.00
Hospitalization ≥ 1 (14%)
Sensitivity	0.36	0.44	− 0.09*	0.13	0.30	− 0.17*	0.10	0.16	− 0.06*
PPV	0.30	0.37	− 0.07*	0.15	0.34	− 0.19*	0.16	0.27	− 0.12*
Accuracy	0.70	0.73	− 0.03	0.76	0.81	− 0.05	0.84	0.86	− 0.02
Expenditures ≥ 30k (9%)
Sensitivity	0.37	0.46	− 0.09*	0.14	0.41	− 0.27*	0.07	0.22	− 0.15**
PPV	0.23	0.27	− 0.04	0.06	0.19	− 0.13*	0.06	0.20	− 0.14**
Accuracy	0.71	0.74	− 0.03	0.82	0.86	− 0.04	0.88	0.90	− 0.02

ED = Emergency Department use; PPV = Positive Predictive Value; 30k = 30,000; HCC = Hierarchical Condition Category score
^*p value < .05; **p value < .01; ***p value < .001

ISSN: 1472-6947