Skip to content

The Case for Network Meta-Interpolation to Handle Effect Modifiers in Indirect Treatment Comparisons

When performing indirect treatment comparisons, effect modification can create complexities in the event of high study-to-study heterogeneity. A number of methods can be used to handle these complexities, but according to Dr. Ofir Harari, Senior Research Principal at Cytel, Network Meta-Interpolation overcomes the shortcomings of a number of other techniques.

Understanding Complications Caused by Effect Modification

Effect modification occurs when the effect of a treatment on an outcome varies along with the values that (one or more) other variables take. For example, if a treatment has less of an effect when someone is younger or of a particular sex, then age and sex are said to be effect modifiers. A robust approach would need to capture the relationship between the treatment effect and such effect modifiers in order to accurately establish the relationship between treatment and outcome.

Commonly used approaches assume for simplicity that the effect modifiers impact all treatments in the evidence network in the exact same way. This is known as the shared effect modification (SEM) assumption and can be handled with network meta-regression (NMR), a form of network meta-analysis that utilizes an aggregate data model. As Dr. Harari notes, the assumption of shared effect modification is “extremely limiting” and “rarely reasonable to make.”

What other options are there?

Multi-Level Network Meta-Aggression (ML-NMA)

Assuming the availability of at least one individual patient-level data (IPD) per treatment class, this method can accommodate models that depart from the shared effect modification assumption – but this is an extremely hopeful assumption to make!

Matching-Adjusted Indirect Comparison (MAIC)

This method relies on weighting the IPD data and can – in its crude form – be used when there is one IPD study and one aggregate-level data study. Indirect treatment comparison when using this method is only valid to a population that follows the effect modifier distribution identical to that of the aggregate-level data study.

Simulated Treatment Comparison (STC)

STC, which relies on regression equations to balance the aggregate and individual patient-level datasets on effect modifiers, is limited by the same constraints as MAIC.

Network Meta-Interpolation (NMI)

NMI, a new method developed by Cytel, can bypass many of the restrictions of the methods mentioned above. Unlike MAIC and STC, NMI can model more than two studies and is not limited by the effect modification conditions of any of the studies. It takes advantage of reported subgroup analysis results – which are often available but have not been used before in ITCs – to overcome non-shared modification data generating mechanisms.

In a recent Cytel webinar, Dr. Harari provides an example of how to utilize NMI and gives an account of its various technical benefits. Click below to watch the webinar.

Watch the Webinar

 

Read more about this webinar series:

5 Steps to Adjust for Effect Modifiers

New Directions thumbnail


About the Author of the Blog: 

Untitled_self

Dr. Esha Senchaudhuri is a research and communications specialist, committed to helping scholars and scientists translate their research findings to public and private sector executives. At Cytel Esha leads content strategy and content production across the company's five business units. She received a doctorate from the London School of Economics in philosophy, and is a former early-career policy fellow of the American Academy of Arts and Sciences. She has taught medical ethics at the Harvard School of Public Health (TH Chan School), and sits on the Steering Committee of the Society for Women in Philosophy's Eastern Division, which is responsible for awarding the Distinguished Woman in Philosophy Award.