In his State of the Union address, President Obama announced a precision medicine initiative “to bring us closer to curing diseases like cancer and diabetes.” The goal of precision medicine is to more accurately identify diagnoses and treatments based on a patient’s genetic information. This information will hopefully lead to better screening, earlier diagnosis, and more personalized treatment.

But it’s hard to imagine that patients will get this kind of medical care if one of the most basic forms of precision medicine — finding the right doctor or team of doctors for each patient — remains one of the most rudimentary processes in health care.

When I was in medical school, I never realized that a critical part of my job as a primary care doctor would be finding the right specialists for my patients. Yet, it seems like every day I am asked not only by patients but also by friends and family to recommend doctors.  Patients trust that I know who the “best” doctors are, and if a referral goes well or doesn’t go well, it is often a reflection on my own skills and judgment.

What most people don’t realize is how challenging this process is and how doctors use very unscientific methods for these decisions.

Consider a patient I recently saw in my office who needed to see a shoulder specialist for a rotator cuff tear.  When I offered the names of a handful of orthopedic surgeons to whom I regularly refer, she replied that the hospital did not take her insurance. When I suggested we look for doctors at another hospital, she said that she had a bad experience in that hospital and did not want to return. I began to research surgeons at a third hospital. As I scrolled through the list on the hospital’s website, it occurred to me that I know very little about the people to whom I refer and, in this particular case, I felt like I was almost picking a name blindly from a hat.

I am not alone. Surveys of doctors show that we choose other doctors based on word of mouth from other doctors, experiences fed back to us from our own patients and personally knowing another doctor.  Not only are these criteria relatively subjective and narrow, but they do not help us if we are trying to find a doctor with whom we have no personal experience.  This may be particular problematic in areas where there are physician shortages, for patients with narrow insurance networks, for low-income patients with public health insurance, or for patients with rare medical conditions. In those situations, physicians may feel as I did — like we are picking a doctor’s name blindly from a hat.

In an era of big data and easily accessible information, it’s alarming that doctors use such haphazard and unscientific methods to find other doctors for their patients. If a key aspect of precision medicine is making sure patients get to the right doctors, we must find a way to improve that process.

In an ideal world, doctors would have easy access to both basic and higher-level information about other doctors.  By basic, I mean information on whether a physician takes new patients, how long it takes for a new patient to get an appointment, and what types of insurance they accept.  By higher-level, I mean information on a physician’s quality of care — defined not only by clinical metrics but also by metrics of care coordination and communication with other doctors.

Unfortunately, most of this information is not readily available to doctors either because the data are not available or because available data are difficult to access. For example, a 2010 study found a scarcity of quality measures for specialty doctors (even though there are a plethora of metrics for primary care doctors). In addition, metrics of how well a specialist communicates with other doctors and coordinates care with primary care doctors barely exist.

Even when important information is available, it may not be easy to access.  Several states, including New York, publish bypass surgery mortality rates for cardiac surgeons. I would think that most patients and doctors want this information when recommending a surgeon. But when I recently tried to access the New York State cardiac surgeon report, I had to click through several links and eventually found a 64-page document with lists of surgeons organized just by hospital and name. I can’t imagine any doctor sifting through this document during a 20-minute visit with a patient.

Some doctors might say that the current process works fine — after all it’s been used for centuries. But I would argue that we live in a different world where on-line physician rating companies and increased transparency are empowering patients to learn more about their doctors.  Although current online rating sites are geared to patients rather than other doctors, they are signs that these sites might affect referral choices. In my own practice, a patient recently questioned one of my specialist recommendations because she had read negative reviews online. I had never imagined that in addition to my own judgment, I would also have to review the on-line rating of doctors to whom I refer.

We need to realize that finding the right doctor for a patient is a critical part of doctoring and should be an essential component of precision medicine both now and in the future.  The status quo of how doctors find other doctors for their patients is rudimentary and unscientific, and there’s a lot of potential for improvement.

Tara F. Bishop is an internal medicine physician who blogs at her self-titled site, Tara Bishop MD.