Grand Rounds Blog

Last month, I had the pleasure of speaking on a panel at Health Datapalooza alongside industry influencers, including Charles Ornstein of ProPublica, Niall Brennan of Centers for Medicare & Medicaid Services (CMS), Robert Krughoff of Consumers’ Checkbook, Fred Trotter of CareSet, and David Vivero of Amino.

As a data scientist, I was especially excited to be at a conference dedicated to what my team and I strive to do every day: finding innovative ways to use health data to improve patient outcomes and eliminate wasteful health care spending. And with the availability of new data sources, our jobs are even more exciting and important — we have much better insight into trends around quality and costs to help make the health care system more efficient.

Here are the top three themes we discussed, and what Grand Rounds is doing to move the needle around data-driven, patient-centric care.

Physician Quality Is About More Than Just Procedural Skill

How we think about physician quality

One of the major efforts on my team is assessing physician quality, which we then use to decide which physicians to send our patients to. When it comes to physician quality, it’s most common to think about procedural skill (often measured by volume of experience). And patients often base their ratings on how well a physician communicates with them. While both of these factors are important, at Grand Rounds, we like to focus on a physician’s clinical judgment, which refers to whether they are:

  • Correctly diagnosing a problem;
  • Acting in the most appropriate way to treat the patient; and
  • Over- or under-using specific medications, diagnostics or procedures.

We dive into our data to try to find trends where aspects of clinical judgment might separate physicians based on quality. And luckily, we now have access to various datasets that allow us to approach this problem in more nuanced ways than ever before. Examples of such datasets include:

  • Public datasets, including those from CMS;
  • Commercial datasets that we’ve bought; and
  • Proprietary datasets acquired from Grand Rounds operations

While all this data is immensely helpful to our analyses, there are flaws and limitations to all datasets. For example, some are only summary-level and others don’t have complete patient or provider coverage. That is why we try to look at everything that is available and combine data sources and types to get a better sense of real trends around physician quality.

Data Findings Offer a Glimpse Into Interesting Health Care Trends

Data reveal interesting trends

Through our ongoing analyses, we’ve discovered many interesting trends that would otherwise not be measurable. For example, in 2013 shortly after actress Angelina Jolie publicly announced that she had the BRCA1 gene, we found that BRCA screenings increased by 50%. We also saw an even larger spike in screenings during Breast Cancer Awareness Month later that year.

We’ve also found some interesting trends around drug prescriptions. For example, most prescriptions are filled on Mondays or equally across the days of the week, whereas ED drugs (such as Viagra) are filled most commonly on Fridays. Additionally, in looking at claims data along with open payments data, we’ve found that doctors who regularly interact with pharmaceutical reps are more likely to prescribe drugs from that particular manufacturer, even when they aren’t necessary. Even more, physicians at the lower end of our quality algorithm appear to be more likely to do this than physicians of higher quality. This is an example of a clinical judgment trend — some physicians are influenced by such interactions, rather than exercising their best medical judgment when making decisions.

Physician-patient Matching: The Right Care From the Beginning

As my fellow panelist Fred Trotter mentioned, “A doctor doesn’t have to be good at everything, but has to be good at what he/she is doing for you.”

At Grand Rounds, we focus on physician quality to ultimately guide our patients to the doctor who is most qualified for their specific needs, right from the start. After all, health care is what happens between a doctor and a patient, and it is our mission to ensure that experience is as best as it can be.

As I noted earlier, our goal is to use clinical datasets to identify features that accurately reflect clinical judgment. Even more, we look at their impact on physician behavior and accumulate these findings to better understand physician quality so we can better serve our patients’ needs.

Addressing the information void

Since patients don’t have access to this type of data, we aim to close this information gap and empower them to make informed decisions about their care. All this robust data also allows physicians to have a holistic picture of a patient’s health so they can treat them properly from the beginning. This leads to better outcomes and less wasted resources. A win-win.

We’ve found that Grand Rounds-endorsed expert physicians help patients get to the correct diagnosis faster and help prevent unnecessary — and often times costly — testing and procedures. In doing so, higher quality care reduces, on average, thousands of dollars of health care spend per case. This is moving us closer to eliminating the more than $1 trillion of wasted health care spend in the U.S. each year, and more importantly, improving patient outcomes.

Other things you might be interested in.
Employee Benefits That Actually Mean Something
Measuring Quality, Part 2: How Physician Quality Should Be Measured
Clinical Expertise Matters: Grand Rounds Works With the Nation's Preeminent Physicians and Institutions