Making Sense of Social Determinants of Health
By Susan Chapman, MA, MFA, PGYT
For The Record
Vol 23 No. 6 P. 14
These data, which provide invaluable insights to patient care, are difficult to collect, protect, and quantify. Nevertheless, the industry must develop trustworthy processes that enable widespread use.
In July, AHIMA released “Social Determinants of Health: Improving Capture and Use by Applying Data Governance Strategies,” a whitepaper that addresses the impact of “the conditions in which people are born, grow, live, work, and age,” or the social determinants of health (SDoH). The paper focuses on “identified opportunities and challenges associated with the collection and use of social determinants as well as how facilities may improve their governance of social determinants of health data. … To improve health outcomes, public health and health inequities, AHIMA believes that SDOH [sic] data must be collected, accessed, managed, and shared in a culturally respectful way to support whole person healthcare [sic].”
An Impetus for Analysis
Margaret Foley, PhD, RHIA, CCS, an associate professor at Temple University and one of the paper’s coauthors, says that the effort to understand SDoH and its data governance has stemmed, in part, from “a call from health insurers, including CMS [Centers for Medicare & Medicaid Service], to increase the screening for social determinants of health and intervening as needed. The insurers are particularly interested in this data for their various value-based purchasing programs. With value-based purchasing, the insurers are trying to share the risk with the health providers. They’re telling providers, ‘We will share the profits and the losses.’”
Insurer payment models differ, Foley explains. For instance, sometimes a payer offers a bonus for good outcomes but exacts a penalty for bad results. “Or, an insurer might say to a health system, ‘Here is the $20 million. Manage it.’ These payment models are shifting the financial risk for managing the care from the insurer to the providers. With bonuses and penalties for good and bad outcomes, respectively, addressing SDoH is critical because they have a huge impact on outcomes. So, there is an uptick in providers’ screening for and reporting SDoH data in their records,” Foley says.
Foley notes that SDoH had been a concern prior to COVID-19’s arrival, but that the pandemic has shone a brighter spotlight on the topic, a point reflected in the AHIMA report:
“The Coronavirus pandemic 2019 (COVID-19) … further illuminated disparities and inequities and the disproportionate effect on people who identify as Black, Indigenous, and People of Color (BIPOC). The pandemic also emphasized the need for health and social care organizations to identify tools for addressing needs and accelerated the call for documentation of a patient’s social circumstances.”
Sanjay Seth, MD, an executive vice president at HealthEC, believes SDoH data for the general population are not available. Sethsays, “No one has been collecting that kind of data. There are specific populations for whom that data are available: the Medicaid population and anyone on programs like WIC [The Special Supplemental Nutrition Program for Women, Infants, and Children]. These data are available specifically for those programs. The data for women and children are only available to the people who get that service. But, as a general population, we don’t know that information.”
Emily Mortimer, senior director for health care strategy at LexisNexis Risk Solutions, believes otherwise. “These data elements have been collected and aggregated outside of the health care system and clinically validated as correlated with health outcomes. While many surveys are collected at the point of care and vulnerable to subjectivity or incompleteness, SDoH indicators based on objective measures offer a robust alternative to survey data. These data can be used in combination with survey responses or as an alternative, allowing health care organizations to focus on care management and programming to remove SDoH barriers and improve health outcomes, rather than data collection.”
Challenges to Data Collection
However, many experts agree that collecting SDoH data has proven to be a daunting task. The AHIMA report highlights a number of these challenges, including the following:
• hospitals’ and providers’ constrained resources;
• ensuring patient privacy;
• a lack of workforce education and development in order to integrate SDoH data into the care continuum;
• a lack of standardization as to which data should be prioritized, which impacts coding productivity; and
• a lack of organization policy around data collection.
HIM professionals are eager to overcome these hurdles but documentation issues are proving tricky. For example, when Judy Littrell, RHIT, CPHQ, presented at a rural health clinic on SDoH, the audience immediately focused on solving the issues she described. “Afterward, all they [the audience] could talk about was how to get patients food, shelter, and transportation to health care services. They were so overwhelmed with those immediate needs, the idea of documenting and collecting useful data couldn’t penetrate their immediate concerns,” she says. “In my experience with at least two nearby hospitals, coders have been sent basic coding information on SDoH but without emphasis, direction, or in-service on its importance. If no one is engaged in documenting SDoH, then there is nothing for a coder to code.”
Corroborating Littrell’s observations, Jennifer Mueller, MBA, RHIA, FACHE, FAHIMA, vice president and privacy officer for the Wisconsin Hospital Association, says that SDoH data collection requires administrative support. “Otherwise, it’s difficult to get the rest of the chain of events that need to occur to happen. Once you have that buy-in, your organization has to meet to discuss what areas to focus on because, with SDoH, there are a lot of areas to focus on,” she says. “For an HIM person, for example, we want SDoH collected in the medical record to assign a Z code [a code that reflects factors that influence health], of which there are 88 in ICD-10. Doing so would capture the environment and situation that these patients live in. We want to do mapping of our communities. A lot of health care organizations are starting to ask these questions and then make a formal community resource referral. We want to be able to use the data to improve the health of our communities.”
Mueller believes that health care organizations must identify who is going to collect the information. “Normally in the medical record, the only documentation that gets a Z code would have to be done by a physician. But with SDoH, anyone on the care team can do it. So the organization or provider has to say who is going to collect it and how,” she explains.
Overall, Mueller says that there must be a structured plan and a build-out process that states who is going to ask the questions and document the record. “And it all has to be in the EHR,” she says.
Kristi Kempe, vice president for Partnerships at Aunt Bertha, a public benefit corporation that connects people with social care providers, elaborates on Mueller’s latter point. “Many EHRs have developed ways to make SDoH data both viewable and actionable across care teams within their systems in the form of patient summary reports and/or alerts that let a provider know someone requires social care,” she says.
One way that SDoH data are collected is through patient surveys, which can be tailored to the audience and type of provider that is requesting the information.
“There are many surveys available. Hospitals will choose based on their patient population,” Foley says. “Some survey tools screen for several social risks such as housing insecurity and transportation needs. If a patient has a specific problem, like literacy, they may provide surveys written in different languages, for example. Another distinction is literacy vs health literacy. You may have a patient who has a high school diploma and reads the newspaper every day but may not be able to understand information about health care.”
Mueller notes that the location of where the questions are asked varies depending on the provider. For example, one hospital may ask in the emergency department, while another may pose survey questions during an inpatient stay. “For a lot of organizations, this information is being asked of patients during wellness visits at their physicians’ offices. There are intake forms that people can fill out in their health history with these types of questions included,” Mueller says.
“The survey has to be customized, based on some objectives,” Seth says. “The way the survey tools are currently designed is inadequate. ‘Are you concerned about your safety?’ ‘Yes.’ But we don’t know what that means. Does it mean physical harm, violence, or do they sleep on the street? We don’t know. We have to redesign these questions to get specific answers. Questions are being asked differently, and there is no guidance to ask the question to elicit a specific response. And patients are often choosing not to answer, which doesn’t help at all.”
Mueller agrees that the personal nature of the questions can pose a problem, and, as Seth points out, patients may elect not to answer certain questions. They also may not answer the questions honestly. “People don’t like being asked if they’re homeless or if they have safety issues at home. The people who do this have to ask carefully,” Foley says. “The other issue is that sometimes people will do better if they fill the questionnaire out privately, rather than if they’re asked directly about these areas of their lives.”
To confirm the data’s validity, Trent Haywood, MD, chief medical officer for Zing Health, believes that the survey methodology must rigorously adhere to standards. “Organizations must follow well-established methodological protocol,” he says. “If organizations lack such expertise, then they should collaborate with experienced community-based research organizations to assist them.”
Ensuring Patient Privacy
In whatever manner SDoH data are collected, ensuring patient privacy can be difficult, given that patients are often being referred to professionals and services outside of the health care setting. “The privacy piece is complicated,” Foley says. “We in health care have a pretty good sense of privacy. Now, we’re bringing in these social agencies and going beyond the traditional health care provider, which is where it gets gray. If I’m a primary care physician referring you to a cardiologist, that is one thing. But the rules haven’t been as well defined with social agencies. That process, how to refer to outside agencies, can be at risk of violating privacy.”
Kempe concurs that SDoH information should be made available to those who need it to ensure excellent care coordination but adds that an individual should know how their information is shared and be able to give consent as to who can access it. “The relationship between a social worker, care coordinator, or other helping profession and an individual is built on trust that develops over time,” Kempe says. “Our approach is guided by the following principles: People should give consent to organizations before their information is shared, know who has access to their information, and be able to remove organizations that have access to their information.”
“Health care organizations should coordinate with their information security team when considering methods of data collection and storage for SDoH data,” Mortimer says. “Acting as responsible stewards of patient information, whether that information is considered PHI [protected health information] or PII [personally identifiable information], increases patient and public trust in organizations.”
Incorporating SDoH Data Into the Care Continuum
Once health care providers are confident that they have accurate SDoH information, the focus can shift to integrating this information into the care continuum. However, most experts agree with Seth that sufficient data are not yet available to effectively accomplish this goal.
“Right now, in Wisconsin, 3% of our hospital inpatient claims have a Z code, and only 0.5% of our emergency department patients have a Z code,” Mueller says. “In my opinion, that is a very small percentage. There are a lot of social issues out there that can, and should, be documented in the medical record and Z code captured.”
Mueller says that instead of using Z codes, she and her colleagues use data points from the United States Census Bureau’s American Community Survey (ACS). “There are some general points on commute time, income poverty, etc. These can relate to the SDoH in some ways. But we don’t have enough Z codes to do a proper analysis. In the future, the goal is that hospitals and health systems can start to collect this information in a standardized way, with correct Z codes applied. Ultimately, organizations and providers can be reimbursed, but right now Z codes are not reimbursable. We have a community health assessments mapping tool; we use discharge claims for the clinical conditions of our community and the community characteristics from the ACS—the points that most correlate to health. We can see a prevalence of diabetes, for instance, in different parts of the community. We can map down to the census-tract level where there is poverty, where commute times are high. Once we have Z codes, we can get even more specific. If there is a transportation issue, then we can do something about a bus stop or food deserts that would need farmer’s markets or other food options. We need to get granular to put in the resources,” she explains.
Haywood believes that when health care organizations and providers utilize established surveys that incorporate SDoH information, they can see opportunities to intervene before clinical, social, or behavioral conditions worsen for their patients. “In many settings, the SDoH data are part of routine screenings at the point of interaction,” he says. “In other settings, the SDoH data determine new programs or services that should be offered. SDoH data illustrate that the care continuum extends beyond the clinical walks into the community.”
But Foley points out that data are lacking on how well some of these new programs work. “What we know is that more than 50% of all hospital readmissions are because of a social risk issue,” she says. “First, we have to determine what is the cause for each patient. It could be a literacy issue, someone maybe can’t read their prescription directions, or a money issue for the prescription itself. It could also be transportation, and the patient couldn’t get to the pharmacy to pick up the prescription. As a way to address some of these social risks, health care systems are launching pilot programs. For instance, a provider is working with Uber to get patients to their health appointments. We know that people who are at risk socially have poor health outcomes, and we know that the improvement of health outcomes reduces cost. But there is a dearth of data to analyze what initiatives are working well. Additionally, the question becomes, if we roll these pilot programs out to a larger population what will continue to work? We have to play our role in gathering and analyzing this data.”
Ultimately, for health care organizations to ensure the availability, integrity, security, and usability of SDoH data, the concept has to become a priority. “There has to be intentionality,” Foley says. “Health care organizations have to recognize the role SDoH play and direct efforts at improving the gathering, coding, and analysis of SDoH data.”
Seth notes that the only way for SDoH data to play a role in the care continuum is if there is an economic impact. “Altruistic solutions do not solve the problem,” he believes. “The government and insurance companies need economic bases for solutions. Otherwise, nothing gets solved.”
— Susan Chapman, MA, MFA, PGYT, is a Los Angeles–based freelance writer and editor.