PHARMACOLOGICAL INTERVENTIONS FOR FAMILIAL HYPERCHOLESTEROLAEMIA IN CHILDREN AND ADOLESCENTS- AN EXPLORATORY EVALUATION USING ADVANCED HIERARCHICAL NETWORK META-ANALYSIS TECHNIQUES
Author(s)
Langford BE, Wickstead RM, Murphy D, Marsh W, Beale RC, Evans JS, Kusel J, Griffiths M
Costello Medical Consulting Ltd, Cambridge, UK
Presentation Documents
OBJECTIVES: To evaluate the use of advanced hierarchical network meta-analysis (NMA) techniques for the comparison of pharmacological interventions for familial hypercholesterolaemia (FH) in children and adolescents. METHODS: A systematic literature review (SLR) was conducted in Epub-Ahead-of–Print, MEDLINE, MEDLINE In-Process, Embase and The Cochrane Library on 19/05/2016 to identify randomised controlled trial (RCT) data on pharmacological interventions for FH patients aged ≤18 years. Six RCTS identified through the SLR were ultimately included in this NMA. Outcomes of interest included percentage change from baseline in total cholesterol (TC) and low-density lipoprotein (LDL-C). Three networks of increasing complexity were generated. The first basic NMA modelled each treatment dose as a separate node. The second, a hierarchical NMA, created a class-based relationship between different treatments; this assumed the treatment effect of differing doses was equivalent. The third more complex hierarchical NMA incorporated dose constraints, assuming a dose-efficacy relationship within each class. RESULTS: In the basic NMA, atorvastatin 10/20 mg was the most effective treatment for reduction of LDL-C. In the hierarchical NMA without dose constraints, higher doses of the same drug did not demonstrate a statistically significant increase in efficacy. The dose-efficacy relationship was, however, apparent in the dose constraints hierarchical model. CONCLUSIONS: To our knowledge, this is the first NMA to test hierarchical techniques for the comparison of pharmacological interventions for FH in children and adolescents. This evaluation demonstrates that the assumptions made in complex hierarchical models do not compromise the robustness and validity of the results. However, not applying dose constraints where appropriate can undervalue the treatment effect of higher doses. Hierarchical modelling could be useful in more complex disease areas, both for treatments with similar efficacy grouped into classes and for treatments that exert a dose-efficacy relationship.
Conference/Value in Health Info
2016-10, ISPOR Europe 2016, Vienna, Austria
Value in Health, Vol. 19, No. 7 (November 2016)
Code
PRM128
Topic
Methodological & Statistical Research
Topic Subcategory
Confounding, Selection Bias Correction, Causal Inference, Modeling and simulation
Disease
Cardiovascular Disorders