≡ Content Category ≡ Main Menu

Patient-Centered Outcomes Research: Passing the “Grandma Test”

As I flip through the short stack of medical journals I’ve accumulated each week, I often ask myself whether the research studies I am reading would pass the “grandma test.” Could I explain these studies to my grandmother in a way she could understand? Would she think they were important? Physicians, patients­—all of us—are inundated with information in today’s fast-paced, technocentric world.

African medical nurse talking to a senior patientHow do we decide that a particular research question is useful or important? As a cardiologist and researcher, I recently seized the opportunity to participate in the newly created Patient-Centered Outcomes Research (PCOR) training program at Albert Einstein College of Medicine and Montefiore Health System. PCOR is a nascent research field aimed at answering questions that matter to patients, their families and other stakeholders and involving them throughout the research process. 

Using Data to Understand Patient Behavior
As the PCOR training program has an emphasis on using “real-world” data to learn how to conduct this type of research accurately, I have spent the past few months working at Montefiore Health System’s Care Management Organization (CMO) under the supervision of Urvashi Patel, Ph.D. M.P.H., who leads their evaluation and outcomes research unit. The CMO contracts with commercial insurance companies, Medicare and Medicaid to coordinate healthcare for 350,000 patients. The goal of the CMO is to achieve the triple aim advocated by the Institute for Healthcare Improvement: improving health, enhancing the patient experience and reducing the cost of care. 

The CMO receives information about medical claims paid for these patients by their health insurers. It also conducts a detailed survey of its patient members that seeks to identify reasons why they may become sick and to provide interventions to keep them well. Part of keeping people well is avoiding unnecessary hospital readmissions.

The number of avoidable hospital readmissions has become a key quality metric by which hospitals are judged as readmissions are thought to reflect fragmented or poorly coordinated care.

Patient Dread Fuels Passion for Research
Patients also uniformly seem to dread the idea of being readmitted to the hospital. As part of the PCOR program, I was given wide latitude to choose an area of research that addressed an unmet need. Given the national focus on hospital readmissions, it seemed to be the perfect area for me to experiment with the CMO’s “real-world” data.

Specifically, I have been asking if researchers and clinicians could find characteristics beyond basic social, demographic factors and illnesses that predict whether someone will be readmitted to the hospital.

Here are some of the questions we’re trying to answer:

  • Are patients who take many medications or have trouble paying for medications or food more likely to be readmitted to the hospital within 30 days of discharge?
  • Are people who live alone or in certain neighborhoods more likely to wind up back in the hospital?

Isolating the effect of any of these can be difficult because one must take into consideration the effects of other social determinants of health, such as race, education and income. However, these data can be missing for patients in large “real world” datasets. For example, determining someone’s neighborhood of residence can be challenging if even a small part of the address is missing.

Despite these limitations, we hope our research will get us closer to understanding why some patients in our health system are readmitted soon after they leave the hospital, and lessen this burden on both the patient and the healthcare system.

Like what you’ve read? Subscribe to The Doctor’s Tablet!

Comments on this entry are closed.

  • Dr. Elyas Fraenkel Isaacs July 16, 2015, 12:21 PM


    Patient centered outcomes are what we should all ascribe and aspire to whether for optimal and understandable ER care and service or from trying to understand what exactly human outlying data to the 10th standard deviation level can mean for better or for worse.

    I think the Bell Normal curve for IQ data is a very good example. What is the difference between the top and bottom 1% or 2%’s? And, what is their relevance to the vast body of “normal” data, individuals/people, who lie within the + and – two standard deviations, that is, those in the “middle”?

    Your comments, opinions, and if possible answers are welcomed here. Thank you.

    Always yours,
    Dr. Elyas Fraenkel Isaacs; PHD, PHD, PHD, DPH, DDiv., PHD, MD[CC/PH]
    from dba “The FRANCES YORK(R)” Institute in the City of New York

  • Paige July 21, 2015, 5:41 PM

    I think this is something great to research. No one likes to be readmitted to the hospital, so if you can figure out who is likely to be readmitted, you can help prevent it from happening. Thanks for sharing with us!