Article
5 questions on exploring healthcare decision-maker sentiments
-
Darlena Le
Discover insight from AMCP Annual 2025
Q&A with Darlena Le
In anticipation of the 2025 AMCP meeting, we had the opportunity to speak with Darlena Le, PharmD, Research Fellow, Health Outcomes and Market Access about her poster “Understanding healthcare decision-maker sentiments: Using natural language processing to predict formulary coverage outcomes.” This work utilizes AI to analyze qualitative data, offering fresh insights into the predictive nature of healthcare decision-maker perceptions. Hannah Bischel, PharmD; Claire Gorey, PhD; Christopher Klein; Jason Lynch, MBA; Joseph Washington, PharmD, MS; Zade Hikmat, PharmD, MS; Andrew Gaiser, PharmD, MBA, MS; and Melissa McCart, PharmD, MS are the poster’s co-authors.
*Available in English only
Here, Darlena shares the inspiration behind her research, key takeaways, unexpected findings, and next steps
What inspired this research?
Darlena Le: Artificial intelligence (AI) has been a hot topic, and I was eager to research within this field. While AI is known for creating efficiencies in various processes, I wanted to focus on analyzing qualitative data, which traditionally requires significant time. We wanted to focus on sentiment analysis, a specific natural language processing (NLP) technique that evaluates whether textual data convey a positive, negative, neutral, or mixed sentiment.
While NLP has been used extensively in e-commerce and other areas, limited research has explored whether healthcare decision-maker sentiments about a product can predict formulary coverage outcomes. Given our access to FormularyDecisions, which allows healthcare decision-makers to provide qualitative and quantitative feedback through product surveys, we were inspired to investigate how AI could classify these responses and see how certain sentiments help predict formulary coverage outcomes.
While NLP has been used extensively in e-commerce and other areas, limited research has explored whether healthcare decision-maker sentiments about a product can predict formulary coverage outcomes. Given our access to FormularyDecisions, which allows healthcare decision-makers to provide qualitative and quantitative feedback through product surveys, we were inspired to investigate how AI could classify these responses and see how certain sentiments help predict formulary coverage outcomes.
Was there a hypothesis that was confirmed through the research?
Darlena Le: Initially, we hypothesized that positive sentiments would lead to more favorable formulary coverage outcomes. Instead, we discovered that mixed sentiments, relative to neutral, were associated with more favorable formulary coverage outcomes. This was surprising and shed new light on our understanding of the data.
What are the key takeaways from your research?
Darlena Le: We obtained open-ended FormularyDecisions survey responses on FDA-approved products from healthcare decision-makers from June 16, 2022 to June 1, 2024.
We examined three product attributes: clinical efficacy, patient value, and economic value. The AI Builder feature from the Microsoft Power Platform was used to carry out sentiment analysis to classify responses on these attributes as positive, negative, neutral, or mixed sentiment.
Our findings revealed that mixed sentiments, relative to neutral, were the strongest predictor of more favorable formulary priority levels and formulary preference. In other words, mixed sentiment type was associated with higher priority rating in managing coverage for a product for formulary review and a greater likelihood of listing or planning to list the product at preferred status.
We examined three product attributes: clinical efficacy, patient value, and economic value. The AI Builder feature from the Microsoft Power Platform was used to carry out sentiment analysis to classify responses on these attributes as positive, negative, neutral, or mixed sentiment.
Our findings revealed that mixed sentiments, relative to neutral, were the strongest predictor of more favorable formulary priority levels and formulary preference. In other words, mixed sentiment type was associated with higher priority rating in managing coverage for a product for formulary review and a greater likelihood of listing or planning to list the product at preferred status.
Was there anything in the research that was surprising, that you didn't expect, that you found out?
Darlena Le: The most surprising finding was that mixed sentiments, rather than positive ones, were associated with more favorable formulary coverage outcomes. Our research suggested that mixed sentiments might reflect a balanced assessment of a product. This idea aligns with a previous study that found physician reviews containing both positive and negative inputs were perceived as more objective, facilitating more informed decision-making. However, this explanation requires further investigation.
What are the next steps from this research?
Darlena Le: Additional research can help determine whether mixed sentiments indeed reflect a balanced assessment of a product and serve as a potential indicator for formulary coverage. Other factors could be explored that may influence formulary coverage outcomes, beyond the three product attributes we examined. Comparative analysis of competitor products could also provide further insights, as the surveys did include questions about competitor products, but our current focus was solely on the product itself.
Citations relevant to the content described herein are provided in the article linked here. Readers should review all available information related to the topics mentioned herein and rely on their own experience and expertise in making decisions related thereto.
Connect with our team
Our team of leading value experts is dedicated to transforming evidence, policy insights, and market intelligence into effective global market access strategies. Let us help you navigate today’s complex healthcare landscape with confidence. Reach out to discover how we can support your goals.
