Build or Buy? Boston’s experience implementing SDOH technology

Takeaways from the HelpSteps story


A discussion paper from the National Academy of Medicine shared how the City of Boston is addressing the social determinants of health (SDOH) through the development, maintenance, and evolution of a public resource directory. It offers a useful look at how a basic information product can develop into community-wide resource if given the right amounts of leadership, strategy, and time. We also found however, that it doubles as a commentary on the challenges of a homegrown approach to building the technical infrastructure to support SDOH management across a city.

Background, Evolution, and Impact

The tool is called HelpSteps — it was developed as a resource directory in 2004 and now lives in the Boston Children’s Hospital. Over its 14 year lifespan, HelpSteps evolved from a search engine into a screening tool and ultimately a referral system used by healthcare workers and community-based social service agencies and organizations. Ultimately, the city’s public health commission incorporated the tool into a program called the Mayor’s Health Line (MHL), which provides additional public health navigation services from insurance enrollment to eldercare programs and more.

Along its journey, HelpSteps underwent a number of studies as a screening tool, all of which illustrated the level of value that health systems can add by capturing social determinants data during clinical visits.

Helpsteps.com — Screenshot from original discussion paper (via National Academy of Medicine)

One interesting tidbit: A recurring challenge in the emerging literature base around SDOH implementation seems to be the thorny issue of asking patients difficult and potentially awkward questions about their behavior. 84 percent of the respondents in a HelpSteps survey, however, reported they thought it was acceptable and appropriate to screen for these issues during a medical visit. The caveat? The 401 patients surveyed were 15 to 25 years old, illustrating both a potential cultural divide along agelines, as well as the potential opportunities for more robust engagement of millennial populations with SDOH screening.

The overall impact of Helpsteps is also notable, particularly the amplification achieved by partnering up with the mayor’s office and the public health commission: Access grew fourfold from 2011–2015 (1,104 to 4,167 users), who are now selecting over 12k services annually using the tool.

Challenges and Implications

The paper offers a few considerations for health systems and city or county health departments interested in implementing a new resource directory, screening process, and referral system to address SDOH in their communities.

ROI? YMMV - A tool is not a solution
The first takeaway is a cautionary tale as we move into an era of excitement with new approaches, tools, and funds for addressing SDOH. One of the cardinal rules of health IT: “If you build it, they will come” is just not true.

“Sustained implementation of screening and referral for social needs is difficult to achieve. Despite users’ positive feedback, case managers frequently resort to either institutional knowledge or other established mechanisms for referring families to social services, rather than using HelpSteps. Part of this reluctance includes limited access to computers and the challenge of changing old habits.”

When planning an SDOH program, workflow planning involves figuring out “when and who.” In this case, HelpSteps seemed to hit the common speedbump of behavior change. If we’ve learned anything from studies on doctors referral patterns, it’s that people will default towards what they’ve always done. Training programs can help, but the adoption curve is steep and can just take time to climb. Expecting social workers to use new tools or look up new resources when they’ve been referring patients to the same shelter, the same agency contact, or the same rehab service provider for the last decade is unrealistic.

Helpsteps’ ongoing collaboration with the Mayor’s office may help drive a more systemic change over time. This could also be an opportunity to introduce design features that make responsive, dynamic resource matching more intuitive or automated, e.g. geoselection based on a patient’s zip code. In the short term, the HelpSteps is developing mobile-friendly access through smartphones and tablets, a tactic that’s led to new use cases and dramatic improvements to efficiency at places like Thedacare in Wisconsin.

Resources required, and then some more
The second point is the obvious one: Creating an aggregate resource is a lot of work.

“Compiling available resources involved hundreds of hours of accessing a variety of listing including binders of printed information, Rolodexes, Excel files, static websites, private databases, as well as interviewing social workers, resource specialists, and other professionals. In doing so, it became clear that there was limited collaboration among community agencies, many of which duplicated efforts by maintaining independent resource lists that were often incomplete and quickly out of date.”

Moreover, beyond the manual effort of developing the tool, there’s the long-tail issue of maintaining it. The quality of information directories is so bad in regular provider directories, for example, that CMS started a penalty program for health plans whose provider lists are out of date, incomplete, or otherwise inaccurate. Philosophically this represents ‘the last mile’ of engagement with which healthcare so often fails to cross the finish line — pointing the horse to water, only for the horse to trot over and discover the water evaporated out of network.
If the maintenance issue is bad among the regulated, well-defined clinical provider networks of healthcare, it must be even tricker for the hodgepodge of community based organizations, social workers and other agencies that comprise the social services resource directories. As the paper notes:

“A major part of HelpSteps’ value is in the accuracy, quality, and quantity of information in the tool’s database. The annual updating process is labor intensive: each of the 1700 agencies is called, and it can take 5 to 10 minutes to update the agency’s information; new agencies take 10 to 15 minutes to be entered into the system.”

The study authors maintain HelpSteps for free public use, but like any good software managers also sell a freemium model, a “fee-for-service tool that allows agencies to track client information and view both individual and aggregate data on clients.” With this in mind, their description of the tool’s worth doubles as a subtle sales pitch, one that raises the timely questions of value and ROI as SDOH continues to emerge as a business strategy, not just a community service.

Buy Or Build?

The paper doesn’t take a clear side on the issue of whether it’s better for SDOH programs to be developed at home, or licensed from a third party. Patchwise Labs spoke with lead author Dr. Eric Fleegler by phone, who stressed that while a lot of attention has shifted towards SDOH, we are at the early stages of figuring out how to manage them.

“Most hospitals, until recently would give lip service to social determinants, at best,” he told us. “The ACA is certainly pushing these things forward. But I think we’re still developing an understanding from the molecular level to the societal level how social things affect health. I think people are hoping for a magic bullet of, ‘ Oh if we just do things like connecting people to rideshare, or a food bank, we’ll lower ED utilization.’ That’s unrealistic.”

So where does that leave us?

Looking back at the last 10 years, a homegrown approach has made a lot of sense for several reasons. As Dr. Fleegler points out, the ROI question is still a fuzzy one, meaning there have been no budgets to buy technology or other support for these sorts of efforts; in turn, that’s amounted to a lack of dedicated commercial solutions outside of a few companies. Moreover, the hodgepodge of CBOs and agencies makes addressing SDOH an intrinsically hyperlocal problem that outside companies weren’t very well-suited to solve, as well as a fragmented market that’s difficult to enter, and scale from a business perspective.

Today however, we can see that things have changed. Looking ahead at the next 10 years, it’s clear they’ll continue evolving. Beyond the shifting sands of value based payment and the drumbeat, zeitgeist, and hashtagging of SDOH, partnerships between agencies and health plans is on the uptick, one part of the broader trends in Medicaid managed care. We’ll see more data connectivity between agencies and central health systems as EHR vendors slowly expand their tentacles outward.

And perhaps most importantly, a slew of specialized vendors, many of them non-profits, are forging long-term relationships with communities around the country. Just like HelpSteps, their investments and specialization will be tough to replicate for an overstretched public health department, let alone distracted health systems.

At this early stage of market’s development, we may not be able to quantify to a clear-cut business case for SDOH technology. This explains why several vendors are using freemium model as a middle path, offering a free public-facing resource for communities and health workers, as well as a slew of business-enhancing upgrades available. Popular features include closing the loop on referrals to track outcomes, analytics tools to enable more precise risk adjustment, or dashboards and reporting tools for compliance with value-based care programs’ reporting requirements, and more.

We anticipate this space will continue evolve rapidly in 2018 for both sellers and buyers (and maybe even a few builders) of SDOH technology - We’ll be following progress closely as it unfolds in communities around the country.

Are you interested in SDOH-related innovation? Let us know! Share any thoughts, ideas, examples, or other resources with us in the comments, or get at us on Twitter! @patchwiselabs @onboardhealth


Naveen Rao is Managing Partner of Patchwise Labs.