Task Force Report

Background

Collection of patient race and ethnicity data is important for health research and patient health outcomes. It is also critical to quality improvement activities that aim to improve equity, as measurement of sociodemographic data and stratification of quality metrics is needed to: 1) identify areas of focus for equity-focused quality improvement goals and activities; and 2) measure the effectiveness of quality improvement activities in reducing inequities and improving outcomes. 

For example, in a classic Plan-Do-Study-Act cycle of quality improvement, stratified quality measures may reveal an inequity in surgical outcomes by race/ethnicity. A plan is then made for a quality improvement intervention to address root causes of the inequity in outcomes, such as specific components of pre-, intra-, or post-operative care. The quality improvement team then conducts the planned intervention (the ‘do’) and studies its effectiveness by comparing stratified quality measures afterward. Following this, the team reflects on the intervention and refines quality improvement activities to act on further improvements in equitable health care delivery.

Among Michigan hospitals, there is currently variability in race/ethnicity data collection practices, which can lead to difficulty identifying and addressing disparities across hospitals. This includes variation in question wording and response categories, as well as data coding and submission for health information exchange (HIE). While there is a requirement to collect these data, there is no standard guidance to hospitals on how to collect and submit race/ethnicity data in the BCBSM P4P program or otherwise.

From 1997 until early 2024, the federal government required five minimum race categories: White, Black or African American, American Indian or Alaska Native, Asian, and Native Hawaiian or Other Pacific Islander. This race question followed the yes/no ethnicity question: Are you Hispanic or Latino? The Task Force recommends that hospitals submit race/ethnicity data in statewide HIE per the revised federal guidance, which took effect on March 28, 2024. The Michigan Department of Health and Human Services (MDHHS) has proposed a similar revision and is currently working on finalizing their guidelines for implementation at the state level.

Recommendations and Rationale

Goal 1: Emphasize self-reported race/ethnicity data collection. Currently many health systems make assumptions about patients’ race/ethnicity, whether based on patients’ appearance or based on other factors available in the electronic health record (e.g., surname, address). However, self-reported data collection that allows patients to self-identify is the gold standard for this type of data collection. MSHIELD recommends this as the primary/sole method for race/ethnicity data collection. Hospital staff should not make assumptions about patient identification to satisfy administrative requirements.

Goal 2: Ask race/ethnicity in a single question and update the categories based on revised federal guidance. While currently race and ethnicity are asked in separate questions, the updated federal recommendation is to move to a single question asking patients to report on all of their identities (“Select all that apply” approach). Hospitals should collect data on race/ethnicity using a single question and, at a minimum, include the following categories: American Indian or Alaska Native; Asian; Black or African American; Hispanic or Latino; Middle Eastern or North African; Native Hawaiian or Pacific Islander; White. See example federal data collection form in Appendix B.

Of note, the updated list of categories includes the addition of Middle Eastern and North African (MENA), as separate and distinct from the “White” category. This is particularly important in Michigan, where we have the second highest Middle Eastern or North African (MENA) population in the US. Many individuals in the MENA community do not share the same health care experiences as White people with European ancestry, do not identify as White, and are not perceived as White by others.

While the federal guidance does not require additional categories for “unknown” or “other”, MSHIELD recommends including “prefer to self-identify” (with write-in) and “prefer not to answer” as best practices in patient questionnaires and MDHHS provides similar guidance. MSHIELD does not recommend including “unknown” or “other” options in a patient-facing questionnaire, but hospital staff could enter “Unknown/Race or Ethnicity Not Asked” when they were unable to ask the patient to self-identify or the information is otherwise unavailable.

Goal 3. Hospitals should submit race/ethnicity data using the minimum categories above, but may collect more granular data on race/ethnicity (see Appendix B for different options in the updated federal forms). Hospitals may continue to use different data collection practices and questionnaires when asking patients about race/ethnicity, including asking more detailed categories than required for data coding and submission. While data collection practices may vary, MSHIELD recommends coding and submitting the data under the same broad categories for the combined race/ethnicity question. This will facilitate a uniform approach to race/ethnicity in the HIE and allow future stratification of quality measures by race/ethnicity across hospitals.

For coding and submission of data to the statewide HIE, hospitals should submit race/ethnicity data under the broad categories listed above but may encounter missing data fields. MSHIELD recommends an additional “Unknown” category for data coding/submission that can include “Prefer Not to Answer” responses to the questionnaire, as well as any other reason for missing data. Because the updated data collection guidance allows for multiple selection, storage systems for the data need to include multiple fields—at least one for each of the above race/ethnicity categories and one for “Unknown”.

Background

The social drivers of health (SDOH) are the conditions in which people live, learn, work, and play, which includes food, housing, transportation, education, income, and safety. SDOH are responsible for major differences in health outcomes across populations, such as morbidity, mortality, and life expectancy. While SDOH refer to the broader systems that shape socioeconomic conditions, social needs are the manifestations of the impact of SDOH on patients. Interventions to address social needs can lead to reduced readmissions, better management of chronic diseases, and improved adherence to medications, all of which result in reduced costs to the healthcare system, improved health outcomes for patients, and better experiences for patients and providers.

Historically, the responsibility of addressing social needs has been in the arena of community-based organizations. However, clinicians and hospitals are increasingly being charged with screening and addressing social needs to improve the health outcomes of their patients. For example, in January 2024, CMS started requiring hospitals to screen all inpatients for 5 specific social needs–food insecurity, housing instability, transportation needs, utilities security, and interpersonal safety–and report data on rates of social needs screening and of needs identified. In addition, as of July 2023, the Joint Commission started requiring hospitals to assess social needs and stratify quality measures by these data, but hospitals may decide what types of SDOH data to use for such purposes. Many hospitals are already screening patients for social needs and linking them to resources. Having a consistent way of coding social needs data can help hospitals and CQIs better understand and address the needs of patients, as well as potential gaps in services.

Currently the most common ways to code SDOH data have been through either claims or EHR coding. In claims, International Classification of Diseases-10 (ICD-10) Z codes have been developed to code for social needs or SDOH, including housing, food insecurity, education, and employment. Currently these codes are applied to less than 1-2% of health care encounters nationally and frequently represent broad categories of SDOH rather than specific social needs. In addition, EHR-based codes called Logical Observation Identifier Names and Codes (LOINC) have been developed to code for specific social needs, including specifying the screening questions used to identify social needs and patients’ responses to these questions. LOINC uses a common language (set of identifiers, names, and codes) for identifying health measurements, observations, and documents, as well for a comprehensive set of social needs. LOINC is currently an optional component of meaningful use incentives and several federal and state programs, including in data being collected by the Michigan Health Information Network Shared Services (MiHIN). Currently some payers favor Z codes while others favor LOINC for value-based incentive programs.

Recommendations and Rationale

Goal 4: Hospitals should submit SDOH LOINC data for 5 social needs domains (food, housing, transportation, utilities, interpersonal safety) to statewide HIE. These 5 social needs domains were selected, based on the current CMS requirement to screen inpatients for these needs, value-based incentive programs emphasizing social needs screening for incentive payments, and given the effectiveness for addressing such needs in prior social needs interventions nationally. Screening and documentation of social needs will assist hospitals and health care providers with meeting current federal requirements and with obtaining financial incentives from current and future value-based programs by CMS and other payers.

The Task Force debated between emphasizing the use of LOINC versus Z codes, as current federal requirements emphasize the need for screening but not necessarily the form of documentation. Ultimately the Task Force favored LOINC for structured SDOH data collection and submission for several reasons. (1) Our assessment of evolving policies in this area suggest that more specific documentation of social needs will be required in the future. Since LOINC provides a high degree of clarity in identifying specific social needs compared with the broad domains outlined in Z codes, they are more likely to be actionable for quality improvement activities. (2) LOINC is derived from the elements of the EHR where social needs screening is commonly collected and reviewed by health care teams, whereas Z codes require a health care provider to spend time entering additional diagnosis codes in encounters. Minimizing additional burdens for front-line clinicians was thus an important factor in the decision to recommend LOINC rather than Z codes for HIE. (3) Concern was raised in the Task Force’s discussion about use of Z codes, which label patients with “diagnoses” that may seem permanent and hard to remove from the medical record, while LOINC describes an observation of a social need at a single point in time and implicitly acknowledges that social circumstances may change.

We also considered the current practices and use of structured SDOH data by health systems and community-based organizations in the state and conducted an informal environmental scan of larger hospital systems, rural health providers, federally qualified health centers, and community hubs. We learned that each modality (LOINC and Z codes) is currently utilized by some health systems and organizations in the state, but there was a preference for LOINC by several organizations. Finally, in February 2024, MiHIN released a new version of their SDOH use case, which focuses on the submission of LOINC for the statewide HIE; currently some health systems are already voluntarily submitting LOINC data to MiHIN.

Background

Conducting the work of incorporating equity into quality improvement activities requires the knowledge, skills, and commitment of health care teams in the area of health equity. MSHIELD has already been providing training, resources, and support to staff members in each CQI coordinating center designated as “Health Equity Champions”. As the Joint Commission also rolled out new requirements in 2023 to have designated hospital leaders to address health care disparities as a quality and safety priority, the Task Force discussed approaches to building on the foundation of CQI Health Equity Champions to expand capacity building.

Recommendations and Rationale

Goal 5: CQIs should partner with MSHIELD to offer at least one equity-focused training at a collaborative-wide meeting in 2025. Participation reflects the commitment that each CQI will be making to promoting health equity and should be recorded by coordinating centers as part of general attendance. MSHIELD will not be tracking attendance at these meetings; rather, these sessions will be incorporated into individual CQI collaborative wide meetings already taking place during the calendar year. While trainings will be required by 2025, CQIs may request to schedule these trainings in 2024.

The Task Force considered several possible avenues of capacity building trainings, including trainings or small group meetings for hospital leaders. Presentations for the identified options (below) will be developed by the MSHIELD team, in preparation for the launch in FY25.

• Let’s Get Uncomfortable (Level-setting workshop focused on building comfort to engage in difficult conversations about equity and anti-racism)

• Data & Anti-Racism (Best practices for talking about, collecting, and analyzing demographic data)

• Increasing Awareness & Moving Towards Action (2-part series focused on systemic racism in health care, interventions, and how to develop an action plan); Aligns with The Joint Commission’s health equity action plan requirements for hospitals

• Health Equity & Social Drivers of Health (Drawing connections between health equity, SDOH, and how they impact specific areas of care – tailored to each CQI’s population of focus)

The intention of this recommendation is to heighten awareness and focus on anti-racism and the connection to health equity, build skills/capacity for anti-racism and equity-focused quality improvement, and facilitate on-going action plans in the quality improvement work of CQIs and hospitals. This goal is complementary to Goals 1-4, as skilled health care leaders can then make use of race/ethnicity- and SDOH-stratified data to generate equity-focused quality improvement goals and initiatives. Further detail on implementation of this CQI-focused recommendation can be found in Appendix C.

The initial trainings at collaborative-wide meetings can help CQI members from across the state learn and share knowledge, and also eventually be incorporated into CQIs’ core quality improvement activities. Thus, individual CQIs should work with MSHIELD to tailor selection and content of trainings to their patient populations, needs, and priorities.