4100_DemandEstimation_Blog-1Emily Mandell co-wrote this blog post with Yasasvi Popuri.

What clinical profile should our pipeline drug pursue to compete in a crowded market? For which patients will our drug be preferred and why? How will the adoption of our in-market drug change with upcoming biosimilars, competitive launches and changing guidelines? These multibillion-dollar questions are very familiar to commercial and medical teams of pharmaceutical and biotech companies, but just how do they feel about answering them? Simply put: uncertain.

To be sure, the task of predicting anything is difficult. That difficulty only multiplies in the pharmaceutical and biotech context for a variety of reasons. First, the decision-making process behind drug selection is complex. While physicians are the primary decision makers, patients are increasingly important influencers, and treatment guidelines and payers also play a role. Second, the emergence of paradigm-shifting drugs with novel mechanisms of actions makes the task of predicting future treatment behavior challenging. Third, the competitive landscape is evolving with the emergence of precision medicine that targets patients with specific genetic markers and the carving out of patient niches. Finally, the steady rise of biosimilars is presenting interesting challenges in terms of predicting drug potential.

Admittedly, medical and commercial teams must live with a healthy dose of uncertainty in drug potential prediction. However, they can and should mitigate the extent of uncertainty through simple and logical market research. Traditional quantitative survey instruments do not adequately capture how physicians or patients consume, internalize and act on information from their peers, other patients or stakeholders such as payers. More importantly, traditional research techniques rely on the assumption that respondents are willing and able to process nuanced information such as multiple drug profile scenarios easily and provide estimates of their intended drug use. Imagine the fatigue a physician might experience when asked to predict the proportion of patients who would be prescribed an upcoming gene therapy, with four other competitors in the market, for three different patient sub-types, and across three possible variations of the gene therapy itself. Even setting aside the relative novelty and paradigm-shifting nature of gene therapies, that is a total of 60 different numbers that the respondent must estimate in a 30 to 45-minute online survey!

How then might commercial teams fine tune market research to deliver reliable insights that aid decision making in a constantly changing world? The answer lies in keeping research simple, real and interesting.

  1. Keep it simple. When multiple product profile and competitive scenarios are at play, breaking the problem into digestible chunks reduces respondent fatigue and allows for better information processing. This can be achieved through multi-day reflection studies where respondents review a new scenario or stimulus each day and answer forward-looking questions about their reactions, drug use, patient selection or competitive positioning. Responses can also be aggregated between days and given to respondents to show how peer perspectives might influence their own.

  2. Keep it real. When physicians or payers are faced with treatment decision-making in a complex landscape, particularly in therapeutic areas with paradigm-shifting medications, decisions are rarely made in a vacuum. Instead, these decision makers draw from the collective wisdom of opinion leaders and fellow practitioners. The best way to mimic this dynamic in a research setting is through prediction panels. This is achieved through either an in-person or digital focus group where a mix of opinion leaders and practitioners discuss emerging treatment trends and debate their role in the treatment algorithm.

    A variation of the prediction panels is consensus, which is an approach based on peer-to-peer conversations to get to a common view using blockchain. Respondents participate in multiple rounds of conversation with one another, sharing their perspectives on a forward-looking question such as the potential of a new drug. Over multiple one-on-one gossips with different partners, respondents have the option of refining their estimates. Obtained this way, the estimates reflect information exchange in the real world and tend to be a more reliable representation of steady state. Prediction panels can be conducted across different physician or patient archetypes or across geographies to capture variability of responses. Because the numerical estimation is closely tied to the justification behind it, such estimates tend to be robust even with relatively smaller sample sizes.

  3. Keep it interesting. No matter how simple or real, the task of predicting the future of a pipeline or in-market drug can be dull if not daunting. Prediction research can be made more engaging through gamification. This could take the form of respondents betting on the most likely future outcomes such as which product is likely to lead in a certain patient sub-type. Over multiple respondents or groups of respondents, commercial teams can discern the likelihood of different market scenarios. This is particularly helpful when respondents have multiple complicated scenarios to consider for predicting their anticipated behavior. Another approach could be mock patient visits, where a smart device app is used to present patient profiles and mimic a consultation. The respondent is asked to review patient history and make a treatment choice against a backdrop of different market and competitive scenarios. Subsequently, the respondent is asked to present their reasoning through a voice memo to a mock tumor board or pharmacy and therapeutics (P&T) committee. Role play exercises such as these can be done through mobile phone surveys or in an in-person interview setting and create better potential estimates and rationale.

Not all of these options are necessary or even useful in every drug potential prediction context. It is critical, however, to think beyond the standard quantitative survey for forward-looking drug perception and share estimation studies. While it’s hard to claim a magic bullet for predicting the performance of pipeline or in-market drugs in a rapidly evolving landscape, these options can help generate reliable estimates that inform decision-making and strategy.

Topics: go-to-market strategy, predictive analytics, market research, predictions, commercialization