The statistical approach in trial-based economic evaluations matters: get your statistics together!

Table 5 Cost–effectiveness probabilities at different willingness to pay thresholds

	REALISE study			HYPOAWARE study
Analysis	0 €/QALY	10,000 €/QALY	23,300 €/QALY	0 €/QALY	10,000 €/QALY	23,300 €/QALY
1.- Two t-tests (1 for cost; 1 for effects) complete-case	0.67	0.57	0.55	0.59	0.53	0.51
2.- Two bootstrapped t-tests; complete-case	0.66	0.65	0.62	0.59	0.56	0.54
3.- Two regressions with correction for confounders; complete-case	0.54	0.51	0.50	0.52	0.49	0.47
4.- Two bootstrapped regressions with correction for confounders; complete-case	0.54	0.52	0.50	0.52	0.46	0.41
5.- Two t-tests; mean imputation	0.74	0.59	0.57	0.59	0.54	0.54
6.- Two bootstrapped t-tests; mean imputation	0.74	0.73	0.71	0.59	0.62	0.62
7.- Two regressions with correction for confounders; mean imputation	0.71	0.58	0.56	0.58	0.54	0.55
8.- Two bootstrapped regressions with correction for confounders; mean imputation	0.72	0.71	0.69	0.58	0.64	0.68
9.- Two t-tests; MI	0.55	0.53	0.53	0.28	0.42	0.45
10.- Two bootstrapped t-tests; MI	0.55	0.57	0.59	0.26	0.31	0.38
11.- Two regressions with correction for confounders; MI	0.53	0.52	0.53	0.26	0.42	0.46
12.- Two bootstrapped regression with correction for confounders; MI	0.53	0.56	0.58	0.26	0.32	0.40
13.- SUR with correction for confounders; MI	0.58	0.60	0.61	0.28	0.33	0.39
14.- SUR with correction for confounders; MI	0.54	0.57	0.59	0.27	0.33	0.40

ISSN: 1472-6963