Current and Future Use of HEOR Data in Decision Making

The growing power of payers in healthcare decision-making has put an increased emphasis on health economics and outcomes research (HEOR), as the data collected through that research is key to demonstrating the value of your drug. As HEOR teams have grown in importance, their roles have expanded to be included in nearly every facet of drug development to ensure the right information is being collected. Furthermore, a report published by PharmaForce International earlier this year revealed that pharma companies increased their field-based HEOR personnel by 53% in 2016. So basically, any pharma company that hasn’t upped their HEOR game yet should strongly consider it. To help you, PM360 asked several experts in this field:

  • What datasets should HEOR teams look at to better understand a drug’s value and how to best present that value to payers? Are there any new datasets that companies should be particularly focused on now or in the near future?
  • How else can HEOR data benefit pharma companies outside of market access strategies? In what ways can it be used to impact sales strategies, marketing practices, physician engagement, etc.? How can companies ensure they are getting the most use out of the data?
  • Given the increase of data now available, what are the best methods for collecting and analyzing data for HEOR in a timely matter? What sources of information are the best for companies to focus on if they have limited resources? Are there any new sources that should be of particular interest?
  • How do you measure the success of your HEOR efforts to ensure HEOR teams are delivering ROI or are able to adjust to changing marketplace factors or new payer needs?

Mike Eaddy, PharmD, PhD

The complexities of therapies—and the healthcare system—require manufacturers and HEOR teams to evaluate more than safety and efficacy data to prove the economic viability of a therapy. Value stories should be informed by public and private sources, and structured and unstructured aspects of these sources.

Unstructured data, particularly from electronic health records (EHRs), has been a focus for HEOR teams recently. However, more than 100 million people share health information on social media and patient support websites. One website, for example, has 31 million data points covering 2,500 diseases. Much of this data goes beyond the structured patient reported outcomes stakeholders have traditionally relied upon or unstructured data in EHRs. Social media will present new opportunities for data sharing as patients turn to online communities for support. Unstructured and structured data can be integrated and operationalized so that product value from the economic perspective can be better determined.

Maximizing Your Data

Data is critical in supporting the shift to value-based care. We already see the industry preparing for this, as evidenced most recently by UnitedHealth and Merck’s deal to link payments to drug performance. This shift will require manufacturers to mine, organize, and analyze data like never before. Manufacturers can ensure they are getting the most use out of their data by partnering with HEOR experts who can use innovative and transparent techniques to help communicate product value.

Manufacturers can look to outside partners with a deep understanding of payer, provider, and patient perspectives and their decision-making processes to supplement existing resources. Through mindful data capture and analysis, they can help transform evidence and market intelligence into effective communications and tools that demonstrate the relevant value to each of these stakeholders. HEOR experts’ ongoing analyses of product performance can also support the shift to value-based care.

Richard Gliklich, MD

In order to better understand drug value and to present that information to payers, HEOR teams need to better demonstrate the true value of a drug—its patient-centric outcomes divided by its costs. Until recently, analyses in large claims datasets have been viewed as sufficient for presenting to payers. However, that is now changing. Many studies have demonstrated that claims data are very limited or inaccurate in assessing clinical outcomes and, increasingly, payers are focusing on clinical outcomes. This means that HEOR teams need access to more clinical data, ideally combined with claims data, to provide a more complete picture of the patient journey and the impact of their product on that journey.

Understanding Value at the Payer Level

Unfortunately, clinical data, largely from EMRs and other sources are poorly standardized and often unstructured. However, by applying new and advanced big data technologies, such as machine learning and natural language processing, to data sources that include deep clinical, claims, and other data, a much richer resource can be generated. With longitudinal patient information, which can include extensive outcomes data across large populations, companies can now understand value at a level much more relevant to a payer.

These datasets and outcomes are where companies should turn their focus as they can be used to directly demonstrate and monitor drug value for payers. The same data can also be used in other ways, including understanding the impact of a multitude of factors ranging from patient selection to physician behaviors to narrow networks to changes in reimbursement models to payer policies. Understanding all of the patient, provider, and payer factors that impact patient outcomes will be increasingly critical for pharma to ensure that their life-saving and life-improving medicines are given to the right patients at the right time.

Julie C. Locklear, PharmD, MBA

When it comes to what kind of data HEOR teams should look at, it starts with the real-world data (RWD) that is easily accessible to pharma companies, including medical, pharmacy, medical charts, EMR data, and prospective observational data. Additionally, over the years, we are becoming more familiar with the value of patient registries, patient surveys, and the Internet of Things (social media, trolling or listening for signals, mobile technology for capturing data, etc.). And the latter is likely to increase in the next few years.

Quality Over Quantity

The goal is not the amount of RWD, however, it is the quality of the data. We always need to keep two things in mind: What is the question and who needs these data? For instance, considering that the key stakeholders utilize data for decision-making, the first step is understanding the needs of those key stakeholders—early and often throughout a product’s lifecycle. One way to achieve this is by mapping out the patient journey. This is critical to understand the path from symptoms to diagnosis, treatment, and outcomes. Additionally, in the U.S., it is also important to understand the reimbursement system and present data to payers that is relevant to their patient population.

David Rees

HEOR teams recognize a growing need to cultivate more meaningful data to support their products’ value propositions. This is especially true in Europe, where countries like Germany require biopharmaceutical companies to submit data on products’ incremental benefits over specified comparative treatments as well as patient-relevant outcomes.

Here and abroad, it also is becoming increasingly critical that marketers share real-world evidence to demonstrate their product offers meaningful clinical or cost benefits for specific patient populations. Today, most providers’ electronic health records are designed to capture the metrics that can reinforce a product’s real-world value, such as impact on complications or hospital readmissions. For this reason, many forward-thinking biopharmaceutical companies have partnered with their more “wired” organized customers, such as integrated delivery networks and accountable care organizations, to gain access to longitudinal patient data in HIPAA-compliant ways.

Preparing for Value-based Payment Models

For example, providers can use these patient datasets to determine the rate of cardiovascular events—and the associated cost of care—in patients with a body mass index (BMI) greater than 35 who have diabetes and modestly elevated lipids. Using the same data, a partnered pharmaceutical company could predict the real-world savings associated with the use of a hypoglycemic drug that reduces cardiovascular events in addition to lowering HbA1c in this specific patient population. As value-based payment models proliferate, providers will leverage such big data approaches to determine the patient groups for whom specific treatments are most cost-effective.

Similarly, as payers look to restrict the use of high-cost drugs to the specific patient segments that will realize the greatest benefit from these therapies, biopharmaceutical companies should cultivate the datasets that will help them identify both who these patients are and how to improve their care. This requires designing clinical trials that demonstrate their products’ incremental clinical and economic benefits—in other words, their value—to specific patient populations.

Richard J. Willke, PhD

Three main sources of HEOR data can be used to support a product’s value proposition:

Clinical Trial Data

HEOR data can be collected and analyzed as part of clinical trials. This approach can be the timeliest source of cost-effectiveness and patient-reported outcomes evidence for new products when conducted in Phase III. Costs vary depending on the amount of data collected, but expenses typically are relatively small compared to clinical trial costs. The limitation of this approach is that as with all trial data, while results are very applicable to the clinical trial population, they may or may not be generalizable to other populations.

Real-World Data

Real-world data can come from a variety of sources (e.g., administrative claims, electronic medical records, telephone or Internet surveys, patient registries, etc.). Prior to launch, real-world data can help document the unmet need in a disease state by providing cost-of-illness and disease burden data. A year or two post-launch, real-world data can provide evidence on the actual effectiveness and economic impact of a new product. Careful analysis and interpretation of real-world data is critical due to the lack of randomization.

Health Economic Models

Finally, health economic models can be useful for estimating cost-effectiveness and budget impact evidence. They can be produced at launch based on trial efficacy data and other parameters obtained from trials, real-world data, or the literature and can be modified to be applicable to specific populations or alternative treatment-related assumptions.

ISPOR has produced many Good Practices for Outcomes Research Task Force Reports to provide guidance in these areas that are made freely available as open-access articles at as part of the Society’s mission.

David Melvin

To optimize resources when collecting HEOR data, target high-cost categories. Retrospective studies can identify costs, components of care, and established baseline assumptions. Next, follow-up with prospective studies that can then demonstrate cost-effectiveness of care, either as a replacement for existing therapy or for niche uses within the category. It is critical to have hard endpoints (this is where oncology has an advantage, as the endpoint is typically mortality or progression-free survival). Lab results (e.g., HbA1C, cholesterol components) or events (ER visits, hospitalizations, fractures) that are clearly defined are most widely acceptable.

It is also critical to partner with reputable organizations, such as large integrated delivery networks or university medical centers to minimize the impact of anti-pharma bias. Best practices in HEOR (see ISPOR good practices) should be followed to ensure that the data are acceptable for peer-reviewed journals.

However, data sources remain challenging:

  • Electronic Medical Records are not yet fully integrated: Even a closed system using a single vendor/system has challenges in data collection. Most organizations have a mix of old/new systems and custom integration strategies. As a result, data collection nearly always requires customized efforts.
  • Administrative claims/billing is limited by what is paid for. (For example, several times in my career, I was asked to look at indwelling catheters for blood clots or blood infections and was stymied by the fact that these events are not consistently recorded since there is minimal reimbursement in the hospital environment.)
  • Hospital/event ER visits and hospitalizations remain the preferred source of data, providing clarity in admission and discharge diagnoses.
  • Surveys are limited by completions, respondent self-selection, and data accuracy.
  • Longitudinal cohorts while desirable, require a sufficient time frame to demonstrate economic impact. This may be effective for high-mortality and high-cost categories but may not be appropriate for chronic illnesses.

J. Cameron Tew

The January FDA guidance on payer communications brings outcomes data center stage. Many companies view the guidance as establishing strong guardrails for how outcomes data can be used, especially how to speak to payers regarding outcomes information around investigational drugs.

It also means that many pharma companies will need their health outcomes groups to work more closely with medical affairs to share real-world data as it relates to duration of treatment, disease burden, validated surrogate endpoints, clinical outcome assessments, and quite a number of other health outcome measures. This will require even greater collaboration between the two groups.

The FDA draft guidance appears to be a positive step toward allowing companies to share routine health economic information analyses—short of communications about unapproved uses claims—with a key group.

Determining the Best Data Points

In research that Best Practices conducted regarding real-world data, we found that only 5 of 36 types of transactional, online, scientific, and machine-generated data were rated highly valuable by most of our benchmark partners. What those companies rated highly were: Claims data, EHR, health outcomes (provider/payer reported), real-world studies, and registries.

Companies with limited resources should consider focus on these data subsets, but I’d also encourage them to look at partnerships with payer groups. A majority of study participants said that partnerships with health plans/payers and data aggregators have a high impact on big data projects and programs. Respondents also said that about 20% of data is sourced from partnerships.

Yi Han, PhD, MBA

Given the legal restrictions about how HEOR results can be communicated to healthcare providers, it is easy to overlook the value of providing physicians with relevant HEOR evidence that can impact their practice and their patients. This includes HEOR analyses and models generated from both clinical trial and real-world data.

For example, systematic literature reviews coupled with network meta-analysis, can help physicians stay abreast of the latest developments for managing diseases. In the absence of head-to-head trials, these types of reviews and analyses can help summarize all available evidence for healthcare providers, enabling them to put it into context and make more informed clinical decisions.

Real World vs. Clinical Trials

It is well-known that real-world performance of many drugs and devices may differ from that observed in clinical trials due to patient diversity, care delivery mechanisms, and chronic drug use. Much of the real-world evidence HEOR analyses for safety and efficacy can help to bridge the information gap between clinical outcomes and real-world performance. Effective communication with providers about real-world comparative studies can inspire development of more cost-effective clinical guidelines based on these insights, and improve efficiency and outcomes overall.

Keeping physicians up-to-date with patient-centered HEOR studies is critically important in today’s evolving healthcare landscape. Patient preference studies and adherence analyses can help providers optimize disease intervention strategies, improve treatment protocols, and ultimately enhance patient care.

While pivotal clinical trial data demonstrates efficacy and safety, HEOR evidence provides a rationale for use in the context of unmet medical need and incremental impact on utilization and cost. HEOR data helps to complete the scientific story, and further strengthens the evidence-based value platform that is foundational to scientific, marketing, and sales initiatives. You should develop a thoughtful HEOR plan as to what specific evidence is needed early in the development and commercialization process.

Joshua Ransom, PhD

Increasingly, all healthcare stakeholders are focused on treatment value, which is the very definition of HEOR’s remit in many organizations. To capitalize on this, it is imperative to transparently help physicians and payers to understand how a given decision will impact both outcomes and cost. Transparency into the assumptions that determine value both drives and requires stakeholder engagement resulting in a virtuous feedback loop.

Companies should democratize access to the HEOR data and empower non-programmers. This will allow them to leverage the curiosity of the crowd and to glean the most insights from HEOR data. Doing so means companies must effectively leverage technological advancements such as common data models, analytic software with guided workflows for non-programmers, etc.

Best Information Sources for Companies with Limited Resources

The specific resource limitations that an organization faces can determine which sources of information will deliver the most value. For example, if the limiting resource is cash flow, then focusing on either 1) a slice of a commercial claims data source for a key indication; or 2) a partnership with a physician society to conduct research on their registries with their principal investigators could each yield strategic value. If on the other hand, the limiting resources is a lack of people with the right skill sets, then pursuing pan-therapeutic area access to multiple data sources and leveraging common data models to speed up the ability to conduct comparative analytics is likely to provide a large productivity boost and result in a reduction in outsourcing.


[sam_pro id=0_12 codes="true"]