In healthcare's era of big data, primary care providers have instant access to years of data on each patient's history, risks, preferences, and test results, all right in the exam room. As a result, clinical decisions are more informed and more personalized. All that should be celebrated as a major advance in medical care.

So, what's the catch?

This data, more often than not, are being used not to personalize patient care, but for mass outreach to large populations of patients. Why not, the thinking goes, let efficient algorithms and automated emails prompt people to get much-needed screenings?

While automated prompts and reminders are valuable new tools, as providers we must ask: How do we know if a screening test saves lives? And which lives? For example, with mass mammogram screenings, are we identifying and screening the correct women for breast cancer? And how do we balance the cost of the repeated radiation from mammograms with the benefit of early detection and treatment of breast cancer?

There is an inherent tension between population health management and person-centered care – one that must be resolved if we are to realize all the benefits of big data in healthcare.

Big data shows us our limitations as well as our opportunities, and we need to honor those limitations as well. The fact is, we don't always know whether what we advise people to do works. Uncertainty is the nitrogen of primary care – it's in the air, but you can easily forget about it because the oxygen is dominant.

There may be a way out of this dilemma for both individual providers and health systems. Rather than using the number of mammograms a population has gotten as a surrogate for healthcare, we can focus on factors that drive mortality and are confirmed by data. Smoking (bad!), access to primary care (good!), vaccines (good!) and other airtight interventions for populations are much better places to concentrate — both to make us healthier and to bend the cost curve in healthcare.

Healthy People 2020 and 2030 are both expressions of these priorities by our government, but hardly any attention is paid to these initiatives by the healthcare sector. Of the United States Preventive Services Task Force's 50 recommendations, only 16 are designated 'Grade A,' meaning we know with certainty that they improve health.

And yet in primary care, we increasingly promote and judge based on Grade B, C, and D recommendations, ignoring not only their as-yet unproven efficacy but also that they frequently trample on a healthcare consumer's wishes and better judgment.

Person-centered care, which is what all of medicine has to potential to be, requires honest, transparent conversations. And those are infrequent when providers must rush through visits to maximize throughput or require patients to comply with “recommended" metrics. The phrase “if you really cared about your health…" has been used as a weapon for a long time. We are not honest in our clinical encounters if we do not admit our uncertainties and allow time for both people in the room to process their meaning.

Do I encourage my patients with diabetes to control their blood sugar? Absolutely. I have staked my career on designing and perfecting systems to maximize the use of teams and get the very metrics I am questioning as high as possible in a challenging environment.

Why? Because I discovered a while ago that patients experience the attention our team gives to their health metrics as love, not data. Our ER visits and re-hospitalization rates went down when we went after blood sugars. They felt that we cared when we called them and reminded them how to take care of themselves. People don't necessarily imbue that same emotion to a robo call from a computer reminding them to check their blood sugar. Relationships still matter.

In the era of big data, we deserve a much happier marriage between public health and healthcare. Rather than using it to drive overuse and misdirected screening tests, we have the opportunity to use these powerful tools to improve our health, not to rely on false reassurance from a negative mammogram that was never warranted. We can do better.

Kirsten Meisinger, M.D., medical staff president of Cambridge Health Alliance (CHA), special projects coordinator for the Family Medicine Department at Tufts-CHA, and a family physician at the Union Square Family Health Center, a CHA practice.