Skip to main content

SDoH: Poised to Care for the Complete Patient

By September 26, 2019November 25th, 2019No Comments
SDOH: Will population health and data science deliver the tools to beat disease?

I must confess that I chose to go to medical school because I had a little bit of a hero complex that drove me into the field. I felt like medicine was the way I could best help others solve all their problems and thereby be honored and respected as a hero. The self-delusional nature of this dream quickly faded during my third year of medical school as I spent many nights and weekends seeing patients at the local homeless clinic.

One patient, in particular, stands out in my memory as a significantly large piece of kryptonite that began to unravel my superman complex. I wish I could say that I remembered his name, but sadly I don’t. I do remember his bright blue eyes, though shining out from behind silvery, salt-and-pepper hair that still resembled the shape of the dirty trucker hat he held in his hands. His dark tanned skin looked more like worn leather than the typical pale-skinned faces I was used to seeing on the monochromatic streets of Salt Lake City.

I remember how proud I was that I had used my super sleuth medical student skills to deduce that his chief complaint of vision changes, combined with decreased sensation in his feet, and increased urination pointed to a diagnosis of diabetes. I remember walking back to fill in my attending, reciting in my head my assessment and plan: “We’ll get some labs done to confirm, then start him on medications, teach him about proper diet, and show him how to monitor his blood sugar levels; we’ll have him follow up next month to see how he’s doing and see if he has any questions.”

I remember too well the look my attending gave me that screamed, “I don’t know who’s smarter: You or my shoe.” He then gently explained that my patient had no place to sleep, let alone a refrigerator or a place to store his medications, and that due to the migratory nature of his profession, he likely wouldn’t be around for follow up visits. I remember driving home that night with a touch of hopelessness, realizing that even though I knew the diagnosis and treatment, there were still times when the ability to help my patient was beyond my reach.

Looking back at that moment and comparing it with the current state of health care, I can’t help but feel optimistic about the changes that have been introduced. Since the ACA began moving the industry from a fee-based compensation to a payment method based on improving patient outcomes, analyses of how to solve the underlying factors that prevent or hinder access to treatment has gained new emphasis. This new emphasis and accompanying payment models have been seen as both a blessing and a curse for clinicians.

For some providers, the ACA has added another layer of issues outside the purview of medical care that must be addressed in care plan development, and thereby placing one more figurative piece of straw on an already wobbly camel’s back. In an environment where physician job satisfaction is low, many complain of total burnout.

On the other hand, the ability to consider all the social determinants of health in a patient’s life has provided the physician the opportunity to care for the complete patient, and get to the underlying causes contributing to the health status of their patients. This is where population health tools and analyses can have a significant impact. With the right tools and care teams in place to address the social determinants of health surrounding a patient, these changes can not only be an opportunity to improve care, but also rejuvenate physician job satisfaction while increasing revenue for clinics.

The social determinants of health (SDoH) are the conditions in which people are born, grow, work, live, and age, and the wider set of forces and systems shaping the conditions of daily life. Some examples include: availability of resources to meet daily needs, access to quality education and training, job opportunities, transportation, social support, local socioeconomic conditions, etc. Though this list is long, and many are difficult to measure, there seems to be a way to divide them into two overlapping groups that can help us understand their interaction with population health.

Actionable Issues

The first group are those issues that are actionable. These include issues where, through either increased community interaction or direct involvement with a patient, a care team has the ability to effect change in the patient’s life to improve their situation, hopefully thereby improving their overall health. Within this group are issues like lack of transportation, housing difficulties, lack of food, health insurance, unemployment, etc.

Admittedly, unless a clinic is willing to allow patients to work for food or to receive medical care for “helping out” (as seen in the 1984 motion picture, Patch Adams, starring Robin Williams), a physician can’t remedy these circumstances on their own. They must rely on the members of the community that have expertise in finding solutions to those problems, and have a mechanism to reach out and consult on patient’s needs, just like they would with a peer in another specialty.

Many physicians nationally have done just that by partnering with their local food banks, low-income housing authorities, or public transport agencies in order to find resources for their patients to find assistance. The growth of UberHealth and the growing number of insurance companies providing housing for high-risk patient populations demonstrate the incredible need in this area. There have also been several new start-up companies whose platforms focus on connecting physicians with various social organizations, which give them the ability to alert the organization of the patient’s need along with contact information. With this infrastructure in place, determining those needs then becomes the focus.

Non-actionable Issues

As we try and determine which patients could potentially benefit from some sort of social intervention, the role of population health becomes much more apparent. This is where the second group of social determinants plays a key role. There are a set of factors that a physician has no power to impact but still play a significant role in a patient’s health; examples include zip code, education, occupation, race, gender, culture, socioeconomic status, and more. While a caregiver may not have the ability to alter any of these factors it is essential to understand how they influence an individual’s health and contribute to the risk of disease.

As a medical community we have begun to lean on population health to help us understand that relationship and even predict the progression of disease states based off a combination of social determinants, clinical information, and family history. Though clinical information and family history can be gathered from EMRs and claims data, social data must often be inferred from demographic information, census data, and well-designed population studies. By combining these various direct and inferred data sets, analysts can create profiles for different populations and stratify them by risk of hospital admission, disease type, or risk of adverse event.

Predicting Outcomes

As these algorithms become more sophisticated, our ability to predict outcomes at the level of the individual improves. This leads to the ability to develop alerts triggered when a patient’s risk of a significant adverse event crosses a predetermined threshold. This alert can be sent to care managers who can contact patients, remotely assess their current health status, and, if indicated, schedule an in-office visit with the primary care physician.

Alerts like these have been successful in decreasing emergency room visits, unplanned hospitalizations and readmissions. Because of their demonstrable success, data scientists have begun looking for other sources of information that can improve the ability to generate algorithms and better predict health issues that previously were unknown. Some of these new data sources include change in address databases, which can not only indicate a significant life stressor of moving, but if the move results in relocation to a higher-risk zip code, it can indicate a change in socioeconomic status with all the accompanying risk factors.

Another data set that has just begun being explored is credit score. Sudden negative changes in credit score can be an indication of job loss or other social issues that have had an impact on the patient’s social equilibrium, and possibly a negative impact on the patient’s health.

Looking back to that first experience with social issues in medical school, I can’t help but wonder how my view of medicine and my role as a physician would be different if I had access to the tools that physicians do today. If I were a third-year medical student today, working in the homeless clinic caring for an undiagnosed diabetic migrant worker, would I be able to provide the referrals necessary to help my patient make the social connections he so desperately needed to care for himself and manage his disease?

After all the billions of dollars that have been invested over the last decade to improve health care, with all my heart I would love to unreservedly reply “yes!”

However, I fear that desire comes from the stubborn optimistic child within who still believes that everyone can be saved. It is that optimistic belief, though, that keeps our work fresh and pushes out discouragement.

As population health and data science has advanced and begun to provide the tools to get ahead of disease, it is much easier to look forward to the day when physicians will have all the tools they need to appropriately and efficiently care for the complete patient. It is the vision of that day that makes population health such an exciting and rewarding field today. There are still hurdles to overcome, but we have the technology to develop the tools necessary to remove those hurdles, improve the health of our patients and, hopefully, heal the medical system.

Leave a Reply