Pricing pharmaceuticals: The transition from prospective qualitative research to predictive data analytics

By Preeti Patel 

Finding the right price in a cost-constrained global healthcare environment

Needless to reiterate that costs of healthcare continue to increase. Some of the main drivers behind this trend are the development costs of novel and efficacious medicines which have the potential to extend life expectancy and overall patient health in previously hard to treat conditions. On average, pharmaceutical spending accounts for more than a sixth (17%) of all health expenditure across OECD countries.1 In addition, aging populations create new challenges by increasing the number of chronic and age-related conditions that contribute to overall healthcare spending.

The high price of medicines has been identified as a significant contributor to overall healthcare spending, to a point where tensions between manufactures and financing providers or payers have been rising dramatically. 2 In the US, the Congress hearings of the last few days are amongst the latest occasions where the complexities of medicine price setting have made their way into the public domain.

High prices of pharmaceuticals are a result of extensive development programmes, lengthy development times and uncertain research portfolio success rates. Manufacturers need to find an optimal launch price for their medicines to at least off-set some research and development (R&D) costs and to provide adequate return on investment (ROI). Once it comes to launch, there are additional pricing considerations driven by target patient number and expected sales volume, as well as the intricate margin and rebate waterfalls across the value chain that US executives have highlighted to Congress over the last week.

With price being the key driver of revenue and ultimately profit, pricing activities, strategy and planning hold a key place in biopharma commercialisation.3 Setting the price too low can have detrimental revenue implication with non-linear impact of price decreases on a global scale. In case of too high a launch price, market access can be jeopardised, either through restrictions to product use or by excluding reimbursement entirely. The latter case would not only have a revenue impact but political implications, disappointing patients and restricting the availability of new treatment options and therapeutic advances. Thus, it is important to find an optimal launch price that falls between the maximum that “the market could bear” and the minimum that a manufacturer could accept in terms of ROI.4

Traditional pricing research based on prospective data

Finding the right or viable price range for a new medicine is amongst the most important challenges for decision makers in the biopharmaceutical industry. To tackle and overcome this challenge, companies usually formulate pricing strategies at an early stage of product development, aiming to answer the following questions:

  • What is the optimal price for my product (within the competitive and market environment)?
  • What is the likely sales impact of an increasing or decreasing in my price?
  • How could market share evolve in case of price change?
  • What is the maximum and minimum viable price for my product?
  • How does this vary across markets?
  • What are the products attributes that help to support my desired price?
  • Should I price my product in line with competitors or consider a different strategy to achieve maximum revenue?

To answer these questions, companies have been utilising a variety of market research tools including qualitative payer and industry expert research (Table 1.). 5-6 Some of the best known and most frequently utilised pricing research methods are direct and indirect elicitation methods, as well as more sophisticated product/price mix models. All aim to determine the product price range that “costumers” could afford or be willing to pay. These survey-based qualitative prospective data collection methods require the respondent to answer specific questions based on given settings or scenarios. Some of these methods are continuous and less complicated, involving simple interviews with qualitative questions. Others are more complex; for instance, for more sophisticated conjoint models, the design of the complete research involves multiple stages: First, respondents preselect specific product attributes that serve as the basis of a second more complex research phase where all pricing scenarios for these specific product attributes or target product profile can be tested.

Table 1. Overview of traditional prospective pricing research methods

Source: Stan Lipovetsky, Shon Magnan, Andrea Zanetti Polzi. Pricing Models in Marketing Research. Intelligent Information Management, 2011, 3, 167-174 doi:10.4236/iim.2011.35020 Published Online September 2011 (


Each methodology has its own advantages and drawbacks, some are more susceptible to bias, and others are more sophisticated in avoiding it. 7 In addition, at times it is hard to collect adequate and reliable data due to respondents’ personal preconceptions, opinions and inconsistencies, or simply by not having enough respondents to formulate any clear conclusion.

For this type of research, time is a crucial limiting factor as collection of accurate survey data, in a prospective manner, takes a considerable amount of time. As mentioned above, more reliable methods even involve multiple research phases and are therefore highly time-consuming. Where research takes longer time to completion, findings, once available, may not provide up to date insights. As a result, research conclusions can become irrelevant or require rounds of updating.

Accuracy and completeness of data in primary research can be compromised if respondents do not have enough time to provide good information. Where there are gaps in survey answers or any uncertainties arise at time of analysis, initial rounds of investigation may even have to be supplemented with follow-ups. In some cases, respondents may give one answers during the interview phase and another during the follow-up, or even provide input that is at odds with publicly available pricing and access documents. For good quality research, trained and experienced personnel are required to successfully prepare and conduct primary research whether it is in the form of logical surveys or structured interviews. In the absence of an experienced team, research efforts could increase in costs and suffer from unreliability and inefficiency.

Considering all these concerns, companies have started to explore methods beyond survey-based pricing research to find more appropriate options to help inform pricing strategy.

Leap forward to advanced predictive data analytics

Data analytics, put simply, is the process of examining data sets in order to draw conclusions about the information contained. This can be through statistical or modelling methods, but more often now through enhanced software or machine learning support.8 The process is not new but becoming increasingly prevalent across industries as many businesses use analytics to understand market trends and learn from past events.

Data generated within the healthcare industry is also growing rapidly and is used more frequently with advanced analytical methods to help decision making processes.9-11 Nowadays, over 65% of pharmaceutical manufacturers continue to invest in data analytics.12 Within the pricing space, new data-centric approaches have started to become of interest to inform better pricing strategies, to recognize pricing opportunities or to achieve more effective pricing for pharmaceuticals. Considering that an increase in pricing can translate directly into revenue and profit improvements, the potential of new analytics is substantial. Where on average, a one percent increase in product price could translate into 8.7% increase in operational profit (in the context of no loss of volume) companies think it worth to take a closer look at the wealth of data available today.13

To utilise large or complex data sets, the process of advanced data analytics provides a feasible solution for integration and analysis. Data analytics in pricing research involves, as a first step, raw data collection, aggregation and systematic filtering to collect reliable data from multiple sources (see Chart 1.). As a second step, based on quality of data, an appropriate data analytic framework is required to conceptualise all elements which impact the price of the product. These elements in the model framework can be further tested with network analytics to understand associations and linkage between the analysed variables. Once casual inference between variables has been established, a model then generates results on the basis of historical data. In this way analytical models can extrapolate and predict future outcomes. In other words, “the systematic use of data and related business insights developed through applied analytical disciplines (e.g. statistical, contextual, quantitative, predictive, cognitive, other models) can help to drive fact-based decision making for planning, management, measurement and learning”.14

Chart 1: Data analytics workflow in the context of pricing research

Source: GPI, 2019

One of the advantages of using data analytics that is often overlooked, is the ability to go beyond what has happened in the past, given that predictive analytics can take a higher level and forward-looking view. While retrospective methods provide a look at what has already happened in the market, helping to understand why those events happened, past events can formulate forward thinking model structures to predict future scenarios. Using hard data instead of biased interview responses provides more accurate business decision in a time-efficient way. To avoid costly mistakes or missed opportunities, predictive data analytics in pharma pricing also allow to test multiple advanced scenarios and integrate pricing, access and revenue outcomes (e.g. potential impact on profitability or demand in case of price change).

Despite the advantages of using data analytics, there are some potential barriers to overcome where such framework is chosen.15 Managing and integrating large datasets is challenging, especially with view to maintaining data quality and consistency. Fragmentation in silos with no linkage between individual data points across categories also imposes significant challenges. Moreover, the absence of standard methodologies may lead to doubts around the accuracy of results. In the medium term, the validation of new methods can shed light on inappropriate aspects of new modelling frameworks. Lastly, it remains a challenge to select the right data analytic tool as not all will fit all purposes. One framework can test multiple scenarios based on pre-set conditions that might not be applicable to answer other business questions (see Table 2.). While no model and no analysis will ever provide a precise prediction of the future, analytical frameworks can reduce the bias from human respondents in qualitative research and further provide another starting point for more accurate up to date recommendations for decision makers.

Table 2. Advantages and limitations of qualitative & data analytic research methods

Summary and outlook

  • Increasing healthcare spending across the globe puts high pressure on manufacturers to establish viable prices that still provide enough ROI
  • Traditional pricing research methods are subject to biases of respondents and can be time-consuming in case more complex methods are used
  • Data availability and accuracy remain a significant issue in traditional pricing research, and absence of timely results, effective surveying and experienced researchers can compromise research quality
  • Despite the complexity of large structured databases, companies have started to successfully realise the potential of data analytics tools
  • Analytical and data-centric research can take advantage of the wealth of historic pharmaceutical market data and react in real time to pricing changes in highly competitive markets
  • Advanced data analytics consider internal and external factors affecting the product’s viable price that could provide more meaningful results than traditional pricing research
  • More importantly, reducing human bias in pricing research enables generation of more accurate and consistent results
  • Reduced research costs through use of new innovative methods allow to optimize internal resources leading more pharmaceutical companies to invest and engage in these analytical activities

Global Pricing Innovations (GPI) is a market leader in business intelligence, analytics and innovative solutions for pharmaceutical pricing and access. See how we can support you on price decision-making through a combination of robust data-driven insights and bespoke analytical solutions, as well as our specialist consultant team’s knowledge of P&MA dynamics across Europe.

Alexandrosz Czira is Project Manager in GPI’s Pricing and Access Consultancy team with experience across P&MA global strategy, value development & evidence generation and conceptualisation of complex model frameworks.

Henrike Granzow is Director of Pricing and Access at GPI with experience across P&MA and commercialisation strategy, forecasting and modelling.




  1. Health at Glance 2013. OECD report. Available at:
  2. Editorial. The great medicines scandal. BMJ 2006; 332:1345-1346, doi:10.1136/bmj.38868.651736.47 Available at:
  3. The art of pricing by By Michael V. Marn, Eric V. Roegner, and Craig C. Zawada, February 2003 | Article, Available at:
  4. Pricing Medicines: Theory And Practice, Challenges And Opportunities by Nigel Gregson, Keiron Sparrowhawk , Josephine Mauskopf and John Paul. Nature review 2005. Volume 4. Available at:
  5. Pricing Models in Marketing Research by Stan Lipovetsky, Shon Magnan, Andrea Zanetti Polzi. Intelligent Information Management, 2011, 3, 167-174 doi:10.4236/iim.2011.35020 Published Online September 2011 Available at:
  6. Ed Schoonveld. The price of Global Health: Drug Pricing Strategies to Balance Patient Access and the Funding of Innovation. Chapter 15 – Payer and Pricing Research. Second Edition. Gower Publishing Limited 2015
  7. Limitation of marketing research by Hari Krishna. 2014. Available at:
  8. Definition of Data analytics. Search data managment [online] Available at:
  9. Big Data and Analytics in Healthcare by Tan et al. Methods Inf Med.2015;54(6):546-7. doi: 10.3414/ME15-06-1001. Epub 2015 Nov 18. Available at:
  10. Use of big data for drug development and for public and personal health and care by Leyens et al. Genet Epidemiol.2017 Jan;41(1):51-60. doi: 10.1002/gepi.22012. Epub 2016 Nov 21. Available at:
  11. The role of big data and advanced analytics in drug discovery, development, and commercialization by Szlezak et al. Clin Pharmacol Ther.2014 May;95(5):492-5. doi: 10.1038/clpt.2014.29. Epub 2014 Feb 20. Available at:
  12. Five Critical Success Factors to Achieve Global Pricing Excellence in Life Sciences. 2015 Annual Survey – Global Pricing in Life Sciences Available at:
  13. Using big data to make better pricing decision by By Walter Baker, Dieter Kiewell, and Georg Winkler. Article June 2014. Available at:
  14. Cortada JW, Gordon D, Lenihan B. The value of analytics in healthcare: From insights to outcomes. IBM Global Business Services, Life Sciences and Healthcare, Executive Report, Jan 2012
  15. SOCR data dashboard: an integrated big data archive mashing medicare, labor, census and econometric information by Husain et al. Journal of Big Data (2015) 2:13 DOI 10.1186/s40537-015-0018-z. Available at: