Table 3.

Effect of statin use versus non-use following myocardial infarction on outcomes among frail older adults after propensity score matching

	Percent with Outcome		HR (95% CLs)	Risk Difference (95% CLs)*,†	NNT / NNH (95% CLs)‡	Difference in RMST (95% CLs)§
Outcome	Statin	No Statin
Mortality	34.3	40.5	0.80 (0.73, 0.87)	−6.2 (−8.7, −3.6)	NNT 17 (12, 28)	16 (10, 22)
Rehospitalization	52.9	50.5	1.06 (0.98, 1.14)	2.4 (−0.3, 5.0)	NNH 43 (NNT 313 to ∞ to NNH 20)	−4 (−12, 3)
Functional Decline	18.6	17.8	1.00 (0.88, 1.14)	0.9 (−1.2, 2.9)	NNH 118 (NNT 86 to ∞ to NNH 35)	−1 (−6, 5)

Abbreviations: PY, person-years; HR, hazard ratio; CLs, confidence limits; NNT, number needed to treat; NNH, number needed to harm; RMST, restricted mean survival time.

^*Reported as a percent rather than a proportion.

^†Confidence intervals estimated using bootstrapping with 10,000 replicates.

^‡Confidence intervals for non-significant NNT/NNH expressed in the format recommended by “Altman DG. Confidence intervals for the number needed to treat. BMJ. 1998 Nov 7;317(7168):1309–12”.

^§Restricted mean survival time is interpretable as the average gain or loss in event-free days due to statin use versus non-use during a 1-year follow-up period; for example, residents who initiated statins would increase the time they survived after myocardial infarction by 16 days on average over a 1-year follow-up period.

Results before propensity score matching in Table A4.