Diving for Data
Finding the right data to underpin outcome-based contracts is not as easy as it looks
Over the past couple of years, there has been a sea-change in the attitudes of healthcare payers towards how they assess the value of new medicines. Momentum is building around the use of outcomes-based contracting as payers seek to better understand how individual treatments impact total cost of care.
Real-world data form the foundations of many such deals, yet, as the number of pioneering deals struck with pharma grows, clear tensions are building around who has access to what data, the quality of that evidence and whether it is trustworthy.
The ideal situation in many cases would be to secure data direct from the payer, yet this does not always happen, says Aaron Mitchell, Managing Principal and Global Lead for R&D Excellence Practice at ZS. Reasons include the proprietary nature of data, concerns around privacy, and the potential for patient re-identification.
In addition, while payers have claims data related to payments made for medical care, they may not have access to medical data, like medical records and lab results, to evaluate the efficacy of this care.
“It is not always possible to get comprehensive data for every patient, so it will still be necessary to supplement patient data with other data. Yet, in some cases, payers may not be comfortable sharing their data. They can benefit more from running the analysis themselves during negotiations,” he says.
Claims data have a good history of providing useful evidence, says Joseph Dye, Head of HEOR, Neurology US at UCB, but using them is not straightforward. “It gets more challenging when the data we need isn’t easy to get our hands on, say for conditions that don’t have clear metrics that can be delivered in claims. In these cases, it can take some creativity and diligence to develop, test and validate new solutions using the data in hand or it will be necessary to do some prospective data collection to answer these questions.”
Sometimes, payers simply don’t have the capability to provide useable data. They may not be able to provide information in a fit state for computerized analytics. While drugs are easy to track because of their NDC number, there are often concerns around whether the right diagnosis codes are provided or how other data have been entered on the host records.
“There are loads of caveats that go with data, and not all payers have an analytics group to extract the data and organize it so that it can be shared with collaborative partners,” says Dye. “Some of the big payers have developed their own real-world evidence divisions to support that kind of work.”
The usual path taken by companies is to purchase syndicated data, and use it to try to develop patient cohort statistics similar to the payer population, says Mitchell. “The payer might also ask pharma companies to do a subset analysis that is more reflective of their patient population. This could involve leveraging data with propensity score matching to show that the result should be consistent with their patient population.”
Generally, however, payers are happy with this approach, he adds. “They recognize the reality of the situation if you go the extra mile to show that it is reflective of their patient population and you can show them it will work.”
Closer matching of third-party claims data to specific patient populations is increasingly being required, says Dye. “A lot of payers are saying they want data that looks like it comes from patients like theirs.” He cites Medicaid or Medicare. “Medicare patients may face very different real-world circumstances, for example, less mobility, fewer resources and more severe conditions,” he says. “If you take a commercial study done around a particular disease or therapeutic area to a Medicare decision maker, they will be skeptical because they know their patients are not like commercial patients.”
New options are emerging to create more useful data sets, which may help to answer some of payers’ concerns. The new breed of providers, such as FIGmd and HealthVerity, that have emerged alongside the established vendors of large aggregated, de-identified data sets are creating a more sophisticated data marketplace of readily available databases at reasonable costs, says Michelle Hoiseth, Corporate VP for Biopharmaceutical Services at PAREXEL International.
“We are seeing an incredible number of small companies emerging as a consequence of where they are plugging into the healthcare data flow to create interesting new opportunities,” she says.
Whether data are derived from third-party sources or direct from payers, another way to improve the quality of the analysis – and the insights gained – is closer collaboration between pharma and payers, says Dye.
“Ideally, you would do this together in a collaborative way to inform the decision-making process for both parties, but whether the payer is prepared to share data outright, share results only or anything in between, needs to be worked out,” he says.
“I would love to do a mutual analysis with the payer, using their data to demonstrate the value in our products,” says Sean McElligott, Director, Global Market Access Lead Dermatology, at Janssen. “Every payer in the US will set different levels of evidence [that they require]. If there is a question about whether a product delivers value, we are more than happy to provide the analyses to demonstrate the value,” he says. “It’s hard to know what is going to be important for a payer and I would like a more open dialogue with them.”
Closer collaboration between payer and pharma can only help all parties gain a deeper understanding of the impact of a treatment on the total cost of healthcare, says Hoiseth, citing the progress in infectious liver disease therapies as an example of the savings on offer to payers and the superior returns available to pharma.
“Established therapies such as interferon were not completely effective and were still costly. Patients had side effects like chronic fatigue and were often on disability if they couldn’t work. Sustained virologic response occurred in approximately 50% of patients, leaving healthcare providers grappling with the cost of cirrhosis, liver failure and liver cancer in the non-responsive patients. When newer alternative curative treatments launched at a high price, about $85,000 for a 12-week course of treatment, payers balked at the cost. But, in fact, $85,000 is a drop in the bucket when you quantify the total cost of the healthcare consumption of patients with advancing liver disease, which now can be prevented.”
Sophisticated data analysis will be needed to capture these various important healthcare dimensions. Drivers of the ways in which healthcare is paid for and of list prices for drugs are complex and at odds, so all parties should expect change to be slow, says Hoiseth. “Discussion around this is definitely part of the public discourse now but this is going to turn slowly like an ocean liner as incentives become more aligned.
“The reimbursement of volume-based agreements is such a complex calculation to unwind, and people acknowledge that it is sensible but a hard thing to do. It takes a lot to move people off tried-and-true, low-risk ways of measuring the performance of a drug.
Since you're here...
... and value our content, you should sign-up to our newsletter. Sign up here