May/June 2017 Issue

Population Health Management — The Social Work Connection
By Susan A. Knight
Social Work Today
Vol. 17 No. 3 P. 10

Many U.S. health systems have seen the value that social workers can bring to population health management initiatives, particularly around care coordination.

Throughout the health care industry, population health management is on everyone's radar. Effective care coordination among individuals with chronic and complex health conditions is a key priority, as is the need to promote patient engagement and effective self-management. The evolution of population health management has largely been enabled by the advancement of modern technologies that support health data analytics. Health service providers are increasingly using analytics to gain greater insight into the populations they serve. Along with providing new opportunities for informed clinical decision making, analytics are enhancing the delivery of preventive care and facilitating the integration of behavioral and primary care services. And social workers have an important role to play as everything goes forward.

"Health data analytics isn't new," says W. "RP" Raghupathi, MBA, PhD, LLM, a professor of information systems, program director of the MS in business analytics, and director of Fordham University's Center for Digital Transformation. Raghupathi also runs the university's Gabelli School of Business Design Lab. "Descriptive health analytics have been in use for quite some time," he says. "Health service organizations would use reporting tools and technologies to analyze data and view health trends across the population."

What has changed, he explains, is the amount of data now available and the increased capacity for these data to contribute to improved health outcomes and a higher quality of care. "It [health data analytics] has taken on increased prominence in recent years. Health care entities are discovering that they have a lot of data, and they can use these data to inform decision making and provide better care to patients."

And just how much data do these heath care entities have at their disposal? The volume of data available for analysis, referred to as "big data," is staggeringly large. So much so that it's difficult to comprehend the amounts being referred to. Most people refer to gigabytes when discussing routine computer data storage. Analysts and data scientists, on the other hand, refer to exabytes, zettabytes, and yottabytes to describe existing and anticipated data stores within the U.S. health care system.

Aggregating Data From Multiple Sources
The availability of modern technologies capable of analyzing huge amounts of data is promising, but it is not without its challenges. Raghupathi explains that one of the first steps in health data analytics is data preparation, i.e. the need to get the data into a suitable format so they can be analyzed. "Data preparation presents a significant challenge for most health care entities," he says. With health service providers using different software solutions from different vendors, the data sources are often incompatible.

"You need to reconcile all those different solutions and platforms," Raghupathi explains. He notes that different entities and institutions may be governed by different laws, based on geographical location, and that also needs to be taken into consideration as part of the process.

"The need for system interoperability is an ongoing problem in informatics," says Timothy Patrick, PhD, an associate professor and chair in the department of health informatics and administration in the College of Health Sciences at the University of Wisconsin-Milwaukee. He adds that data compatibility issues aren't always necessarily technical in nature. "It's not just about linking data from different sources," he says. "There may be differences in how the data are expressed and what they mean. All of this needs to be resolved in order for the data to be aggregated and analyzed in a meaningful way."

Some of these nontechnical issues are immediately apparent when one considers the task of aggregating primary care data with data from other community sources. "Combining data between primary health care services and social services is critical," Patrick says, "but one of the issues is handling differences in terminology and coding systems. There are a variety of terminologies used across the spectrum of social services and health care. Linking these terminologies semantically is very important."

According to David S. Pieczkiewicz, PhD, director of graduate studies and a clinical assistant professor in the Institute for Health Informatics at the University of Minnesota, the challenges around preparing and aggregating health data are further exacerbated by the ever increasing volume of data available to be processed. "Making sense of all the various and disparate streams of data we have available to us has always been a challenge, but even more so now, as the variety, volume, and velocity of those data increase."

Tools for Managing Big Data
Whenever technical challenges arise, corresponding innovations are never far behind. The rapid rate at which technology has advanced has enabled the development of powerful tools and systems that are increasingly capable of addressing the challenges around data preparation and aggregation. Pieczkiewicz explains that the creation of systems that are equipped to intelligently handle data and sift through them has led to "massive strides" in our ability to manage and use big data.

Sophisticated data visualization software is one such tool. These tools allow for data to be presented as charts, graphs, and even animations. This type of presentation can be extremely useful in helping people gain a better understanding of the data they're working with, and there is widespread agreement that data visualization will continue to be an important aspect of health analytics going forward.

"Our ability to visually process information, especially when it is presented graphically, is still a tremendous asset," Pieczkiewicz states. "More recently, data visualization has become sexy, and I've been glad to see so many people in different disciplines, including outside of the sciences, pay more attention to how we can visualize information, reason with it, and communicate our findings to others."

Pieczkiewicz also points to the value of data visualization in providing a method to support inductive, rather than deductive, reasoning. He explains that beyond analyzing the data for the purpose of testing hypotheses and answering preformed questions, data visualization can be valuable in helping us to look at the data in new ways, which in turn can lead to the formation of new questions and hypotheses.
Patrick highlights machine learning as another area where there have been tremendous advancements in the analytical tools available. Algorithms can be developed with the capacity to identify relationships in existing data; those algorithms can then be used to identify similar relationships in new data.

Data mining is also a tool that is making it possible to do more with all of the data available. "There's a lot of interest in data mining," Patrick says, "where we look for patterns within the data and use these patterns as a basis for predictive modeling." Data mining has been used in other industries for decades, but its use within the health care industry is relatively recent. The ability to identify patterns that would normally be hidden makes it possible to gain greater insight into the population and its care needs. For instance, high-risk patients can be identified and targeted interventions can be developed to better meet the needs of those individuals.

Social Work Roles
It's easy to see how social workers might feel that computer applications, algorithms, and data mining are all far outside of the typical social work domain, but Susan C. Westgate, MBA, MSW, LCSW-C, a clinical instructor at the University of Maryland School of Social Work, challenges this notion. She believes social workers have a valuable role to play not just in aiding in the design and implementation of population health management models but also in the administration of all the related tools and applications.

Westgate states, "I think that it is critical for social workers to be engaged in population health management and to be engaged in the more complex big data applications that are proliferating throughout the health care industry. I would also offer that social workers are needed to ensure the quality and integrity of the data that are collected and subsequently applied." A growing number of social workers do appear to be delving into the technical administration aspects that Westgate refers to, either by contributing to projects in a consulting capacity or by working directly for population health management solution providers, where they are helping to shape the development of applications and tools.

Raghupathi also views social workers as having a very important role to play in population health management initiatives, and sees opportunities for engagement at different levels and in a variety of capacities. "This is an exciting space for social workers right now," he says, "being able to take steps at both the micro and macro level." Along with face-to-face client work, experienced social workers are also assuming senior leadership roles in health care institutions and health technology companies.

Predictive Analytics
As previously mentioned, the use of data mining and predictive analytics is now well underway within the health care industry. Whereas descriptive analytics examines the data with a focus on what has happened up until this point, predictive analytics examines the data with a focus on what to expect going forward. More and more health services providers and institutions are utilizing predictive analytics because of its immense value in supporting the development of targeted population health management initiatives.

Pieczkiewicz describes the shift to predictive analytics as significant and one that we're only just starting to take advantage of. He contrasts the current use of health data to earlier days when clinical data were typically collected and used solely for direct patient care, whereas data for research were often collected in a purpose-built manner. He states, "If we wanted to know the answer to a clinical, health services, or population health question, we often had to make a special effort to collect data specifically for our study." He explains that each type of data—clinical and research—was often walled off from wider use.

Pieczkiewicz continues, "We're routinely collecting more and more varied clinical data in the course of patient care, to the point at which some research studies can be carried out using these data, rather than having to expend time, money, and energy collecting 'bespoke' data for a particular study." He describes a blurring of boundaries between the storage/sharing of clinical vs. research information as a key factor enabling the shift to predictive analytics, as it makes it so much easier to access the required data.

With its capacity to support the design and delivery of preventive care services, predictive analytics goes hand in hand with population health management. "Predictive analytics is certainly helpful in terms of conceptualizing patient risk," Westgate says. "Our data tools can increasingly tell us both who is at risk and what sorts of interventions would benefit them and ideally improve their outcomes."

Taking things a step further, Westgate suggests that social workers should be part of the build phase of these models. "It is important to remember again that the sum total of the value of any sort of technological infrastructure is derived from baseline population assumptions, as well as from the types of questions that we ask." She advises that social workers should be involved in both creating risk assessments and evaluating programs, in order to lend their expertise. "It is important to be able to see what is and what is not working, and why," she adds.

Integration and Care Coordination
The successful implementation of population health management initiatives requires successful integration and care coordination across multiple agencies and sectors. "It's very important to break down barriers," Patrick says. "Integration across all the different agencies and sectors for various populations is incredibly important." This includes social services, justice, and education along with health service providers. Social workers have a critical role to play, given their active involvement in these agencies and across these sectors.

Beyond the data collection and analytics that occurs independently, Patrick identifies the need for information sharing in order to deliver the most effective care possible. "If all those agencies aren't sharing information in an integrated way, across the various agencies and sectors, they're going to miss the full story of what's happening with the client." He explains that without that "full story" on a client, service providers are unable to determine with accuracy exactly what is taking place or what the appropriate responses and interventions should be.

Patrick cites numerous examples, both in the United States and abroad, of cases where children fell through the cracks, sometimes with devastating consequences. In many cases, a wealth of information was readily available across multiple care settings that pointed to risks and/or unmet needs. However, due to a lack of effective information sharing and care coordination, none of the service providers involved had a complete picture of what was taking place, making it difficult to select the most appropriate intervention strategies.

Patrick refers to birth to three years programs as a positive example of successful integration, in which early childhood education, family services, and medical services all come together with a focus on the entire family. "This kind of approach is truly needed," he says, "but it still remains an underutilized approach for improving health outcomes across the population."

Westgate also sees gaps in service integration and coordination, and points out that this is a longstanding issue within the health care industry. "Health care systems have long struggled with trying to make fragmented care and clinical pathways more efficient," she says. She believes that social workers' knowledge can and should be leveraged to help resolve these issues. "Social workers have significant knowledge about the structural and process barriers that discourage efficiency and consequently raise costs. So again, social workers being part of the planning, execution, and evaluation of these programs is critical."

Social Work and Population Health Management
Westgate believes there should be a "natural bridge" between social work and population health management, as the latter offers "a potential strategy to aid in operationalizing interventions." But she notes that this "proverbial bridge is not always so obvious to social workers indoctrinated in traditional clinical and health models." Organizational structures can make it difficult for social workers to play an active role. For instance, senior executives in any given health system will likely be able to appreciate the ways in which population health management is deeply intertwined with statistical analyses and various models of business intelligence. However, "The social worker on the ground may not necessarily be aware of and consequently included in system-based and corporate strategies to improve outcomes."

Many health systems across the United States have already seen the value that social workers can bring to population health management initiatives, particularly around care coordination. Their ability to view patients holistically, with an appreciation for how psychosocial needs and environmental issues can impact a patient's health choices, is viewed as an asset. Not surprisingly, social workers are increasingly being integrated into health care service delivery teams, where they play a critical role in identifying and addressing barriers to effective patient care.

As Westgate summarizes, "Social workers have valuable community-based and anecdotal knowledge about the populations that they serve. The overall challenge is to effectively integrate the strategy, the technology, the intelligence, and the social work workforce."

— Susan A. Knight works with organizations in the social services sector to help them get the most out of their client management software.