by Dan Tarjan 
figures by Krissy Lyon

The EpiPen, the antiparasitic drug Daraprim, the blood pressure medication Nitropress. These life saving drugs have recently been in the news because their prices spiked by over 100% year-to-year without any apparent reason except increasing profits. And they’re not alone. Across the US healthcare industry, specialty drug prices are rising. These brand name products marked a 16.2% increase from 2014 to 2015 alone. This trend leads to the question: Are we getting anything in return for these drug prices? An international report by the OECD predicts that, without reform, current healthcare trends among developed nations will become financially unsustainable by mid-century. The United States is leading the way on rising healthcare costs by spending almost twice what the runner up France spent in 2013. That’s not a race the US wants to win.

To justify higher prices and save on costs, many patient advocates support using outcomes based healthcare, sometimes known as value based care. The idea is simple: medical providers and drug companies should be paid based on how much they improve patients’ lives. The implementation of this idea, however, is much harder. Outcomes based healthcare aims to incentivize medical treatments that measurably improve patients’ health. To make this model work several challenges need to be addressed: What marker of a patient’s health does a physician measure? What target is set for that marker to demonstrate improved health? Who is responsible for collecting and processing this data?

Figure 1: In an outcomes-based healthcare framework how well patients respond to a given treatment informs the size of the payment for that treatment. Over time this flow of information can be used to select the best treatments for a given condition.
Figure 1: In an outcomes-based healthcare framework how well patients respond to a given treatment informs the size of the payment for that treatment. Over time this flow of information can be used to select the best treatments for a given condition.

Heartbeats or headaches: what should physicians track?

Clinical and basic scientists have worked for a long time to find the best way to measure or identify a disease state. Such indicators of a disease state are known as “biomarkers.” In addition, physicians need to be able to assess a biomarker reliably. These measurements must work from day-to-day, and from hospital-to-hospital. Without validated biomarkers to track the treatment of a disease, it is difficult to construct an outcomes based framework to compare different treatments and the value they bring patients.

Some diseases like diabetes and cardiovascular disease have well-established biomarkers including blood sugar and blood cholesterol levels, respectively. As the name implies, these biomarkers reside in the blood and are therefore easy for a physician to access and measure. On the other hand, it has proven difficult to identify biomarkers for diseases such as Alzheimer’s. This disease starts as the aggregation of proteins called beta-amyloid in the brain that over the course of years kill nerve cells and damages the brain. Eventually those afflicted suffer from changes in behavior, memory loss, and decline in general cognitive ability. Early indicators of the presence of beta-amyloid aggregates could be one potential biomarker that could help catch the disease before brain function is affected and also track the efficacy of any potential treatments. However, it is much harder to measure what is happening in the brain than in the blood. The Alzheimer’s Association is hopeful that measuring the levels of key proteins in the cerebrospinal fluid or performing brain imaging may yield a useful biomarker for the disease. Sampling a patient’s cerebrospinal fluid is invasive, costly, and carries its own set of risks for the patient’s health that need to be balanced with the clinical usefulness of the information derived from this procedure. For many diseases, such expensive approaches may nevertheless be the only way to assess how patients are doing.

Figure 2: Biomarkers provide quantitative ways to describe a patient's health or disease. Well established ranges for healthy vs unhealthy blood sugar and blood cholesterol help doctors decide which patients need treatment. Beta-amyloid detected by a PET brain scan can track an Alzheimer's patient's disease.
Figure 2: Biomarkers provide quantitative ways to describe a patient’s health or disease. Well established ranges for healthy vs unhealthy blood sugar and blood cholesterol help doctors decide which patients need treatment. Beta-amyloid detected by a PET brain scan can track an Alzheimer’s patient’s disease.

Depression, chronic pain and attention deficit disorder represent another class of diseases that are proving difficult to find biomarkers for. This class of ailments frequently relies on subjective patient-reported information. Pain assessment, for example, often works by asking patients to rate their pain on a ten-point scale. Many physicians see such scales as unreliable as patients compare only to their own experience. Reliably tracking outcomes for treating pain and other diseases without true biomarkers will remain a challenge for health systems for the foreseeable future.

Doctor meet computer; computer meet frustrated doctor

A healthcare system that operates on outcomes not only requires reliable measures of disease, but also a system to track these measures once patients are treated. While doctors and hospitals keep archives of notes and test results for their patients to manage their care, these data systems are not designed to track patient or population health outcomes. Several efforts are currently under way to better capture data about patients from automatic interpretation of doctors’ written notes with natural language processing to standardizing information formats for medical records. Genomic information is also increasingly being used in making treatment decisions. Despite these advances no medical data system can currently track patients specifically for determining how well a drug or treatment worked for them. This reality poses a challenge for determining whether treatments helped individuals or groups of patients.

New data systems will need to be created that can capture biomarker and disease relevant patient data at the source and incorporate it into a bigger picture of how well patients are responding to a given treatment. Importantly, physicians must adopt these systems into their daily practice. Doctors are already struggling with adopting electronic medical records and a recent survey indicates they question their utility for patients. If these data fed into a system tracking patients’ health outcomes and providing feedback about which treatments are working it could go a long way towards convincing doctors that the data entry burden is worth their time.

Data analysis in healthcare faces the challenge of strict privacy regulation under the Health Insurance Portability and Accountability Act (HIPAA). These regulations protect patient privacy but also act as a barrier to determining medical treatment outcomes. Medical records may one day be stored in a shared format yet will still reside with individual healthcare providers and hospitals. Billing and therefore cost information resides with a variety of insurers. Collecting all these data under one roof for a comprehensive analysis must overcome the regulatory burden of complying with HIPAA and get many different organizations to agree to share. Health care providers are reluctant to share patient data for fear of violating their privacy as well as fear of losing business to competitors. However without such sharing large-scale analysis of healthcare outcomes is impossible.

A preview of an outcomes-based future for US healthcare

The US government and private companies are already implementing early examples of what outcomes-based healthcare could look like. The Centers for Medicare & Medicaid Services (CMS), perhaps the most influential organization in US healthcare, is currently introducing value-based programs for reimbursing physicians based on metrics of their performance. By 2018 all physicians that take Medicare patients will be subject to these programs. While CMS can unilaterally implement such programs, their regulations currently hinder efforts by private companies to do the same. Insurer Anthem and pharmaceutical firm Eli Lilly recently lobbied for changes to regulations that would allow private companies to implement value-based contracting while staying compatible with the Medicare marketplace.

Nonetheless drug companies like Novartis have recently announced deals including an outcomes-based component. Novartis recently agreed with insurers Aetna and Cigna to take payment for their new heart failure drug Entresto based on how many patients the drug keeps out of the hospital. The insurers already collect this information because they get billed for hospital visits. This deal side steps the need for a relevant biomarker and shows the limitations of what outcomes it is possible to track today. It’s a great start, yet hospitalizations are a crude measure of patient health and importantly, that information comes too late to help patients.

Several challenges remain towards bringing healthcare outcomes into the equation of US healthcare. Basic research needs to establish reliable disease biomarkers where it’s possible, and find the best possible metrics where it’s not. The stakeholders involved in collecting, sharing, and interpreting data about medical treatments and patients’ health need to reach agreement to enable population wide outcomes analysis. Finally the regulatory landscape will need updating to allow for innovative models in both the public and private sectors. These challenges are great, but solvable. If we overcome them we might finally have an answer to the question: are these drug prices worth it?

Dan Tarjan is a PhD Candidate in Biological and Biomedical Sciences at Harvard Medical School. He is passionate about the next generation of medical treatments and the business models that will deliver them.

This article is part of our Special Edition: Dear Madam/Mister President.

For more information:

  1. Express Scripts. (2016, March). 2015 Drug Trend Report.
  2. OECD (2015), Fiscal Sustainability of Health Systems: Bridging Health and Finance Perspectives, OECD Publishing, Paris.
  3. Value-Based Health Care Is Inevitable and That’s Good,” Harvard Business Review, Sept 24, 2013.
  4. Drug Goes From $13.50 a Tablet to $750, Overnight,” NY Times, Sept. 20, 2015.

Cover image from here. Licensed by CC.


One thought on “Should we pay for drugs or cures? How tracking drug effectiveness could improve US healthcare spending

  1. Prescription drugs constitute only 10% of overall healthcare costs, so why do they get so much attention?

    It’s phenomenally hard, risky and expensive to develop a new drug, and drug companies need reasonable profits in order to survive. There may be a couple of egregious examples of overpricing, but overall I think we get good value for money.

    If we truly want to control healthcare costs, we need to focus more on prevention, but the current U.S. insurance system has an extremely short-term business model: only some kind of State or national control of healthcare priorities would cause a shift in prevention versus treatment.

Leave a Reply

Your email address will not be published. Required fields are marked *