. 2017 Dec 6;12(12):e0178918. doi: 10.1371/journal.pone.0178918

Table 5. Sensitivity analyses (QALYs).

Analysis	Acupuncture vs. usual care				Alexander lessons vs. usual care
Analysis	Sample size	Incremental cost (£)	QALYs gained	ICER (£)	Incremental cost (£)	QALYs gained	ICER (£)
Base case	293	451.32^a	0.032^b	18,767/QALY	667.24^a	0.025^b	25,101/QALY
Base case	293	(285.29 to 634.83)	(0.001 to 0.062)	(4,426 to 74,562)^c	(472.28 to 896.42)	(-0.007 to 0.058)	(-150,208 to 248,697)^c
Inclusion of societal costs	293	509.44^d	0.032^b	20,151/QALY	861.70 ^d	0.025^b	35,552/QALY
Inclusion of societal costs	293	(252.11 to 775.93)	(0.001 to 0.062)	(3,659 to 86,635)^e	(491.52 to 1286.13)	(-0.007 to 0.058)	(-172,253 to 329,091)^e
Including only healthcare costs relating to neck pain	293	375.46^a	0.032^b	15,364/QALY	576.81^a	0.025^b	20,065/QALY
Including only healthcare costs relating to neck pain	293	(328.58 to 425.78)	(0.001 to 0.062)	(4,156 to 56,763)^c	(522.85 to 627.61)	(-0.007 to 0.058)	(-112,735 to 241,192)^c
Imputation of QALYs and cost	507^f	690.02^a	0.019^b	43,838/QALY	884.41^a	0.010^b	121,269/QALY
Imputation of QALYs and cost	507^f	(516.39 to 894.26)	(-0.005 to 0.044)	(-216,427 to 395,047)^c	(727.87 to 1059.61)	(-0.014 to 0.034)	(-854,671 to 1,014,592)^c

^aAdjusted for baseline NHS healthcare costs and practice size. 95% non-parametric bias-corrected confidence intervals based on 1,000 bootstrap replications.

^b Adjusted for baseline EQ-5D index score and practice size. 95% non-parametric bias-corrected confidence intervals based on 1,000 bootstrap replications.

^cBased on 1000 bootstrap cost-effect pairs. Adjusted for baseline EQ-5D, baseline healthcare costs and practice size

^dAdjusted for baseline societal healthcare costs and practice size. 95% non-parametric bias-corrected confidence intervals based on 1,000 bootstrap replications.

^e Based on 1000 bootstrap cost-effect pairs. Adjusted for baseline EQ-5D, baseline societal costs and practice size

^fExcluding 2 participants who died