For health and human services, data is the key to prevention

By | 5 minute read | April 19, 2018

An apple a day... Image by Holly Mindrup via Unsplash

According to the U.S. Department of Health and Human Services, 1 in 25 adults in the U.S. had a serious mental illness in 2016. In 2017, more than 550,000 Americans were homeless on any given single night. And in 2018, more than 115 Americans die every day after an opioid overdose.

Some or all of these issues may be related—but agencies often have limited data to help solve them. To learn how to increase prevention, we spoke with Mary-Sara Jones, IBM Health & Human Services Industry Lead for US State & Local Government.

What are the biggest issues health and human services faces today? And how can we address them?

Addressing individuals with complex needs—particularly people who receive services across multiple programs or agencies. They’re between 5-20 percent of individuals receiving services, but account for between 50-80 percent of the total cost.

It’s critical that agencies identify complex clients, and more proactively address the root causes of their needs. One of the ways agencies can better improve outcomes for this population is by addressing the social determinants of health, because housing, food vulnerability, physical activity, education, and income can account for over 50 percent of a person’s overall wellness.

Baltimore, Maryland citizens, for example, can have a 20-year difference in life expectancy, although individuals live just six blocks apart. So your ZIP code may be more important than your healthcare.

What will it take for HHS to shift towards increased prevention?

Agencies are primarily mission-focused, which creates a lack of visibility. So the housing agency is not aware of somebody at risk until they become homeless. Prevention requires visibility to the individuals who will be homeless tomorrow, but are not today.

Prevention also requires an upfront investment. But agencies can’t just shift dollars without negatively impacting outcomes. They need additional funds to invest in prevention. Complicating funding is the “wrong pocket problem,” which means that the agency that spends money on preventive actions is not the agency that saves money from those actions.

There are opportunities to leverage models similar to accountable care organizations (ACOs), where the healthcare provider receives money from Medicaid for achieving certain healthcare targets. It’s win‑win because the cost is less than what Medicaid saves overall.

There’s also the issue of data sharing. Data today is siloed within health and human services, and many federal, state, and agency policies restrict the way data can be shared. But technology has finally advanced to where we can more safely and widely share data, and remain compliant with all the relevant privacy restrictions. This would enable multi‑disciplinary teams to more effectively collaborate to achieve improved outcomes for complex clients.

What’s possible through insights into that data?

Most of the valuable, rich data within health and human services is unstructured and in different formats, like case notes, hotline audio files, and images.

Tapping into and analyzing that unstructured data greatly improves an agency’s ability to provide person‑centered care, and helps agencies address underlying issues as well as the symptoms.

Homelessness, for example, tends to be a symptom of unemployment, mental health, or substance abuse. Providing housing alone would probably not be a sustainable solution without also addressing the underlying issues.

Having comprehensive data helps agencies connect people to the appropriate services, which improves access and outcomes. And ultimately it reduces costs: earlier intervention yields better results, higher client satisfaction, and lower costs.

How can AI help?

Although caseworkers bring great expertise about their clients, they don’t necessarily understand all the issues impacting each client.

AI will be increasingly important because of the vast amounts of data being created. AI can digest all of the data and  discover new  patterns to understand how different correlations of factors result in certain risk levels.

With the opioid crisis, for example, agencies know some factors common to individuals with substance use disorder, like mental health issues, lack of a strong social network or income level.

If agencies understood all the patterns, they could build a more accurate risk indicator to proactively address the driving factors and change the client’s trajectory.

What does “single view” mean for HHS?

Today, client data lives in many different organizations. Single view means pulling all the data together in one place so agencies can better deliver care and optimize outcomes.

Say Ted has type 2 diabetes. With a more holistic view, an agency can see that Ted recently went through a divorce, lost his job, and is being treated for  severe depression—all of which exacerbated his physical health. Effectively addressing Ted’s diabetes requires understanding and assisting Ted with his underlying issues.

How can agencies evolve to meet citizens’ expectations?

By emphasizing digital government and user‑focused design, agencies can create an integrated experience across processes, solutions, and devices.

Many health and human services clients may not have computer access, but may have a smartphone—so can access mobile-ready services.

Agencies are also starting to overcome data sharing challenges by using a Memorandum of Understanding to create trust agreements between agencies.

Some municipalities are enacting new legislation that focuses on permitting data sharing for multidisciplinary teams, which bring together individuals with different expertise to work together to help a single client.

What will health and human services look like in 2025?

I’m very optimistic that agencies will move to new business models and develop the data management capabilities necessary to deliver services in a more individual-focused manner.

As costs become more constrained, community‑based organizations will become increasingly important in the service delivery chain. This will increase the need for integration and data sharing to send service delivery data back to agencies so they make better decisions and more informed policies.

What agencies already effectively share data?

Sonoma County is just starting a project to provide a shared data platform to address the health, housing, and social services needs of homeless families.

LA County has introduced legislation that allows data sharing across multidisciplinary teams based on specific use cases, such as child welfare investigations. For the first 30 days of an investigation, child welfare investigators have access to an extensive amount of data, such as criminal history and mental health. When they complete the investigation, they have the information needed to determine if a child is safe in an environment.