SIMULATED TREATMENT COMPARISONS – AN ALTERNATIVE APPROACH TO INDIRECT COMPARISON WHEN STANDARD METHODS ARE NOT FEASIBLE OR APPROPRIATE
Author(s)
Ishak KJ*1, Proskorovsky I1, Benedict A2, Chen C3 1Evidera, Dorval, QC, Canada, 2Evidera, Budapest, Hungary, 3Pfizer Global Pharmaceuticals, New York, NY, USA
Health technology assessments (HTAs) rely on comparative evidence about new treatments and competing therapies, which are typically derived using indirect or mixed treatment comparisons (ITC/MTCs). These are not always feasible or appropriate, particularly in rapidly evolving therapeutic areas, like oncology. For instance, some comparisons may not be possible due to incomplete evidence networks; or, heterogeneity between studies due to differences in design or population may make an MTC inappropriate. There is, therefore, a need for alternative techniques, such as Simulated Treatment Comparisons (STCs). This technique is designed to derive comparisons between treatments after adjustment for differences between the populations of the two studies. This targeted comparison requires individual patient-level data (IPD) for at least one of the treatments (the index), and are appropriate when the trials used for the comparison are sufficiently comparable in design and methods, but differ in the profiles of their population in measured risk factors. The differences can be adjusted analytically using IPD via regression equations. This produces endpoint estimates for the index treatment that reflect the profile of the comparator population. These can then be contrasted with published results for the comparator to obtain a measure of difference between treatments. Since only measured risk factors can be included in the adjustment, the potential for residual confounding remains. Another potential bias is a possible “study effect” whereby other differences between studies distort the comparisons. This can be assessed using the reference groups of the trials, if these received the same treatment. STCs have been used in HTA submissions, and it is likely that its use and that of other alternative techniques will increase particularly in areas with rapid drug development. In the presence of heterogeneity or incomplete evidence networks, STCs can provide comparative evidence where these may be otherwise deemed unavailable due to limitations of ITCs/MTCs.
Conference/Value in Health Info
2013-11, ISPOR Europe 2013, The Convention Centre Dublin
Value in Health, Vol. 16, No. 7 (November 2013)
Code
PRM228
Topic
Methodological & Statistical Research
Topic Subcategory
Confounding, Selection Bias Correction, Causal Inference
Disease
Multiple Diseases