Biosimilars – Current Treatment – Biosimilar Target Product Profiles (Oncology)

For each of the six attributes included in the conjoint analysis, we defined three to five levels of performance in either quantitative or qualitative terms. The range of performance levels for each attribute was set to allow for reasonable trade-offs in the performance of an emerging biosimilar relative to competing biosimilars or to the reference brand that physicians would be willing to make when considering which biosimilar to prescribe.

For the price attribute, we defined a range of biosimilar prices extending from a list price that is the same as that of a reference brand to a list price that is 50% lower than that of a reference brand.

Surveyed U.S. and European medical oncologists were presented with a series of TPPs constructed by varying the combination of performance levels across the attributes included in the study (excluding any prohibited combinations). In the screening part of the ACBC module, surveyed medical oncologists were asked to indicate whether they would prescribe each of the four TPPs presented to them, as well as identify any levels that are completely unacceptable or absolutely necessary. Next, medical oncologists were presented with sets of three, semi-randomly generated TPPs (based on their answers in the screening section) and were asked to indicate which profile they would most likely prescribe. Finally, medical oncologists were presented with a series of TPPs and asked to rate their likelihood to prescribe each profile (from “definitely would not” to “definitely would” prescribe).

Surveyed medical oncologists’ responses in the ACBC survey module were analyzed to calculate, at an individual respondent level, the part-worth utility of each attribute level. Within a given attribute, part-worth utilities are measures of the relative attractiveness of each attribute level to surveyed medical oncologists in relation to the other attribute levels included in the conjoint ACBC. Conjoint part-worth utilities also allow for determining the relative importance of each of the attributes in surveyed medical oncologists’ preferences when making biosimilar prescribing decisions.

Conjoint part-worth utilities were also used to build the Target Product Profile Simulator. The Target Product Profile Simulator allows for the creation of user-defined TPPs with varying performance across the attributes. Using conjoint part-worth utilities derived from surveyed medical oncologists’ responses, the Target Product Profile Simulator models how medical oncologists’ preference and likelihood to prescribe vary across the set of user-defined TPPs. With it, users can conduct simulations of customized sets of TPPs and explore the preference share and likelihood to prescribe each TPP in the scenarios. Results allow users to gain an understanding of the trade-offs across biosimilar attributes that physicians are willing to make when considering which biosimilar to prescribe, as well as to gauge how physician preference is affected by changes to one or more attributes of a particular TPP. Notably, findings are presented for the full sample and by region (United States and Europe).

1. Share of preference: the probability that a given TPP in each simulation would be selected and represents the relative desirability of each product in a given scenario. Share of preference values sum to 100% across the collection of drugs simulated.

2. Prescribing likelihood: the estimated prescribing likelihood for TPPs specified in the simulator, where each product is considered independently. The likelihood of prescribing projection for each drug is given on a 0-100% scale.