Rapidly advancing technology is enabling the production and consumption of healthcare data in a way not previously possible. With so much at stake in medicine, can Big Data help us decide on the “right” doctor to meet our medical needs?
What defines the “right” doctor depends on a number of variables:
- The complexity, rarity, severity and urgency of the illness;
- Various preferences that the patient may have for the provider, such as availability, proximity, gender and / or languages spoken; and
- The provider’s clinical expertise and outcomes.
For a common and urgent diagnosis, such as strep throat, we assume that many physicians have the baseline competency to treat this. As such, shortest time to availability or proximity to one’s home or office is likely to define the “right” doctor. But just how does one go about identifying the “right” doctor for a less common and/or more complex diagnosis? And, how do these available data help us understand those situations more clearly?
In recent years there have been numerous studies linking physicians who perform a large volume of procedures with better outcomes. Studies of carotid endarterectomies, aortic dissection repair and hip and knee arthroplasty have all suggested that physicians and/or teams performing more cases had better outcomes than their less prolific peers. This observation is consistent with reports from other industries as well. In Outliers, bestselling author Malcolm Gladwell repeatedly mentions the “10,000-Hour-Rule” – where the key to success, regardless of field, is merely a matter of practicing more and achieving a greater volume of repetition. Can we assume then that high volumes are a proxy for better outcomes in healthcare? And, if so, should volume be prioritized in the search for the “right” doctor?
Unfortunately, the answer may be more complex than a simple “yes” or “no”. For one, the association between volume and outcomes has not been demonstrated uniformly. The 2012 New York State Department of Health report on cardiac bypass surgery showed that in one hospital, providers with the highest volumes actually had worse risk-adjusted outcomes. Perhaps exceedingly high volumes can lead to less attention to detail, a willingness to take more risks, or result in provider fatigue. Might there be a sweet spot or threshold after which additional case volume may not be beneficial?
Secondly, while we may be able to assess the relationship between procedure volume and major outcomes, such as mortality, demonstrating correlations between more ambiguous outcomes and for non-procedural specialties, such as oncology and endocrinology, may be more difficult. For non-procedural specialties, the ability to obtain a thorough medical history and knowing what tests to order might be more relevant than volume.
Given the limitations of using volume as a proxy for quality, Kyruus has increasingly started to think about volume as a proxy for “clinical fluency” instead. Consider the metaphor of learning a foreign language. After some initial lessons in Spanish, you may become conversational – that is, you have the ability to talk with someone and carry out basic interactions. This may be enough if you want to go to a café in Barcelona and order a cup of coffee. Like diagnosing and treating strep throat, this interaction requires basic and widely available competency.
By contrast, if you wish to engage in a business transaction in Barcelona, it would require a greater degree of skill or what is known in the field of language fluency as a “professional working proficiency.” By spending a year in Spain, you may gradually attain this level of proficiency. “Professional working proficiency” could be thought of as the equivalent of the volume thresholds above which a physician is able to perform consistently well.
Now, what if you wanted to teach at a university in Spain? This would require what is known as “native or bilingual proficiency.” Perhaps by spending five years in Spain, immersed in the culture and in social interactions, you could attain this level of mastery. This might be the equivalent of having expertise in treating complex medical conditions.
Does the foreign language metaphor suggest that more volume (or in this case, time spent in Spain) equals better outcomes? Yes, but not always. What if you spent a lot of time in Spain but only spoke with other non-native speakers? Or, alternatively, what if you were able to only learn “street” vernacular? You might learn the language, but may not attain fluency or mastery. In other words, repeating the same task is only good if you have solid fundamentals to begin with.
Alternatively, what if you had natural talent for learning languages and already spoke Italian and French? Would it take you five years to become fluent or would three months in Spain be enough? This might be the equivalent of evaluating a physician who has only performed 3 right hip replacements but has performed 500 left hip replacements. Because of his/her existing mastery on the left, gaining mastery on the right hip may not require high volumes.
What the language metaphor suggests is that additional time spent practicing a language, after one has developed some baseline competence, increases one’s likelihood of fluency. Similarly, having high provider volumes of diagnosing or treating a specific medical condition, after reaching a threshold for baseline competency, increases one’s likelihood of mastery. But other factors matter as well. Just like learning a language from someone with good grammar skills is important for attaining fluency, training in a high volume and/or high quality organization likely impacts the potential and speed with which one can attain clinical fluency. In addition, continued medical education and relationships with colleagues may play a role in maintaining and enhancing the provider’s fluency over time.
Big Data holds big promises for healthcare. While the answer will likely not come in the form of a single data point, we believe that the concept of clinical fluency can help us better understand how to identify and connect patients with the “right” doctor for their condition. Over time, by leveraging the expertise of our talented engineers, our team hopes to further understand the intersection of healthcare and data and, ultimately, how to identify what elements are associated with better outcomes. We look forward to sharing these learnings as we go and to ushering in a more information-driven healthcare ecosystem.