Methods for network meta-analysis of continuous outcomes using individual patient data: a case study in acupuncture for chronic pain

Table 4 IPD NMA ANCOVA synthesis model (model 2) results with adjustments, EQ-5D preference endpoint

IPD NMA results: EQ-5D preference scores endpoint^a			Model 2, ANCOVA, with adjustment for baseline score, age and treatment-by-age interactions, median MCMC posterior sample (95 % CrI)
Relative treatment effects	Osteoarthritis of the knee	SHAM vs UC	0.040 (-0.006, 0.084)
		ACU vs UC	0.066 (0.025, 0.105)
		ACU vs SHAM	0.026 (-0.012, 0.066)
	Headache	SHAM vs UC	0.056 (0.012, 0.098)
		ACU vs UC	0.060 (0.023, 0.095)
		ACU vs SHAM	0.004 (-0.036, 0.043)
	Musculoskeletal	SHAM vs UC	0.045 (-0.001, 0.094)
		ACU vs UC	0.074 (0.038, 0.109)
		ACU vs SHAM	0.029 (-0.009, 0.067)
Main effects	Age		-0.002 (-0.002, -0.001)
Main effects	Age ²		0.000 (0.000, 0.000)
Age common interactions	Age		0.000 (0.000, 0.001)
Age common interactions	Age ²		0.000 (0.000, 0.000)
	Between-study variance		0.001 (0.000,0.003)
	Total residual deviance ^b		15,590 (15,210; 15,970)
	Deviance information criterion ^c		-6,462.0

^aUC usual care, SHAM sham acupuncture, ACU acupuncture, Headache group headache, migraine and TTH, Musculoskeletal group neck, shoulder and low back pain
^bCompare to approx. 14, 800 observations
^cDeviance information criterion (DIC) is a statistical measure of model fit and model comparison. Models should be preferred with smaller DIC

ISSN: 1471-2288