Visualizing the Complexities of Health Care

Jacqueline Merrill, PhD, is associate professor of nursing in biomedical informatics at Columbia Nursing. She is a public health nurse and health services researcher, directing the Laboratory for Informatics, Complexity and Organizational Study at the School of Nursing. Her public health experience includes six years with NYC Department of Health and Mental Hygiene. She has been at Columbia since 1996.

You have had a non-typical nursing career path. Please tell us about it.

My mother was a public health nurse in Rhode Island and this had a strong effect on me.  Inspired by her, I obtained a diploma in nursing from Massachusetts General Hospital School of Nursing. After eight rigorous years in critical care, I wanted a change so moved into the world of design. I had my own women’s clothing design company, and then was a photo stylist for Bloomingdale’s.

I enjoyed design but found that I missed nursing. I went back and within 10 years obtained my BA, MPH, and PhD , all from Columbia.  This included participating in Suzanne Bakken’s doctoral research training program here at Columbia Nursing, which aims to promote health in underserved populations by applying informatics.  I drew on my experience in design to investigate how computers can generate information that helps public health managers run their organizations—in other words, I study what is known as ‘socio-technical systems’.

How does your design career affect your nursing research?

The whole experience of having to succeed in the business environment is central to what I do now. I saw that public health managers were experts in caring for populations but, because public resources nearly always are constrained, they could benefit by leveraging the organization itself as an asset. For example, technology could help them understand how their organization was working, and provide information they could use to make business-oriented decisions about how to use resources.  Making every dollar go further allows services to reach more people in the community.

In my work as a designer I learned that design and visualization go hand in hand. The same is true for health care organizations. When trying to design a better performing organization, visualizations are a great way to get insight into the complexities of how the people and the work interrelate. You can see it the same way a clinician would see an X-Ray.

This led me to apply the socio-technical approach to model how patients move through health care systems, including clinics and facilities here at Columbia.  Where do they go first? Where do they go next? Is this the right order for doing things? What gaps have to be filled?  These kinds of questions can’t be answered by guessing. You need evidence. Network analysis is crucial here. This entails collecting data to break down a complex system into its component parts and plotting interdependencies and interrelationships. Once you determine what is actually happening, you can then develop a simulation model, using the computer, which allows you to deploy “what if” case scenarios to see if doing things differently would  lead to more effective or efficient care.

Can you give us an example of your own research on these types of questions?

In one study we used network analysis to examine care delivery in patients with congestive heart failure, a chronic condition with 30-day hospital readmission rates as high as 27 percent.

My colleagues and I looked at clinical and administrative data for more than 4,800 congestive heart failure patients throughout the Columbia University Medical Center system. Using a software algorithm, we assessed and plotted the total volume of visits these patients made between services, such as inpatient and outpatient internal medicine, and inpatient and outpatient cardiology, regardless of the order in which they occurred.

At the macro level we found that these patients used a broad range of services, with no one of them dominant. Nearly every possible pair of services was used by at least one patient. But more refined analysis helped us see that individual patients frequented a small, somewhat predictable number of services in comparatively clear sequence, often according to their admission status.

We then classified patients as never-admitted, admitted-once, single-readmission, and multiple-readmissions. The admitted-once group stood out from the others. They had many more transitions and a 50-percent increase in links between services (what we call greater “density” of care) compared to the never-admitted group. And it wasn’t just specialty care related to heart failure. They received higher concentrations of all kinds of care, probably because newly admitted patients have multiple events happening which need to be identified and addressed, perhaps for the first time. On the other hand, in the multiple readmission group, care density drops, suggesting a more focused care pattern, likely in response to the acuity that resulted in readmission.

We also explored complex, non-linear relationships between the services used by inpatients and outpatients using variables as critical care unit admission, cardiac surgery, heart transplant, use of cardiology diagnostics and social services, and emergency department visits.  This system-oriented approach is producing evidence that clinicians and administrators can understand and use to design and evaluate how care is coordinated.  The goal is to improve patients’ experience of care and to ensure that their outcomes are the best possible.

The bottom line is that network analysis can help us measure real-world health care utilization and allow us to plan and use resources better. We don’t have to limit our categories to admission status. We can look at how patients transition through service visits according to their age, gender, medication usage, and so on. Network analysis raises the realistic possibility of assessing which sequences of care lead to better results for patients with congestive heart failure patients and other chronic conditions.

You were recently appointed to the American Association of Colleges of Nursing's Health Policy Advisory Council. Can you tell us about the council and its mission?

In 2015 the American Association of Colleges of Nursing established its Health Policy Advisory Council to guide the association in assessing and taking formal positions on federal public policy proposals on health, health care, and higher education topics, particularly those that require insight from academic nursing expertise. This includes legislation, rules and regulations, notices, and other requests for comment.

We also will be identifying academic nursing experts to represent the association on public policy issues as well as highlight current and emerging policy issue areas that can affect academic nursing.

Can you tell us about your teaching at Columbia Nursing?

I am teaching a core course for masters’ students called Health and Social Policy in the Context of Practice.  I aim to convey the importance of policy as the underpinning of clinical endeavors, and to familiarize students with the policy environment that will define their professional lives. 

The nursing profession always has had a stake in policy development but since the 2010 Institute of Medicine report, The Future of Nursing, individual nurses are increasingly participating in the policy process. To prepare our students for this growing role, I assign a class project involving real-world policy development. Students work in small groups to research and develop health policies at the state level, for example, by picking an issue such as smoking cessation in that state and developing policy at six levels of intervention: individual, community, health care practitioners, legislation, organizations, and coalitions and networks (stakeholder groups) – the whole “spectrum of prevention.” The objective is to generate an understanding of how to create policy by doing it.  The policy documents they create will live online.

Things move fast in health policy these days so I update the readings in my syllabus throughout the semester to reflect current events.  For example, in 2014 and 2015 major Supreme Court decisions on the Affordable Care Act were issued during the summer health policy course!

In the course I try to convey that nurses are innately interdisciplinary and holistic and that nursing is therefore an ideal profession to drive systemic reform in health care.