Simple Questions, Better Care
Asking patients their race, ethnicity, and language preferences can feel awkward for busy hospital registration staff, but Cincinnati-area hospitals have found that asking these questions reveals important information that can help them to more effectively identify and address potential disparities in the care they deliver.
Today more than one-third of Americans are racial or ethnic minorities, and according to the U.S. Census Bureau, the proportion will increase to more than half of the U.S. population by 2043. This shift in national demographics will have a tremendous impact on health care providers and the patients they serve, and collecting information about race, ethnicity, and language preference (R/E/L) will be a powerful and increasingly important tool to ensure all patients are receiving high-quality care.
Historically, registration staff or providers often relied on "eyeballing" patients’ race and ethnicity; patients were rarely, if ever, asked to report this information themselves—and assumptions and mistakes were made. Guessing patients' race and ethnicity came under federal government scrutiny, and, in 2009, the Institute of Medicine (IOM) of the National Academies—a nonprofit often referenced by lawmakers seeking information to frame health policy—released its report Race, Ethnicity, and Language Data: Standardization for Health Care Quality Improvement. The report found that disparities in the quality of care racial and ethnic minorities receive have persisted over time, and real progress toward eliminating these disparities will not be accomplished without access to accurate, reliable information on patients’ racial and ethnic backgrounds.
Going forward, R/E/L data will become more critical than ever and serve as a cornerstone to reducing health care disparities and improving clinical outcomes for patients. This was exactly what the Greater Cincinnati Health Council discovered when it recently introduced a new R/E/L data collection initiative across 19 member hospitals in the regions. The endeavor has led to more complete and more accurate data, giving hospitals and providers an opportunity to correct course if necessary, address problems more quickly and efficiently, and elevate the standards of care for all patients regardless of race or ethnicity.
Inspiring a National Model
Greater Cincinnati Health Council has served the Cincinnati region since 1957 and today consists of 33 urban, suburban, and rural hospitals serving 14 counties that stretch across southwest Ohio, northern Kentucky, and southeastern Indiana. A long-standing partner with the Cincinnati Aligning Forces for Quality initiative, led by The Health Collaborative, the Greater Cincinnati Health Council has experienced tremendous success with implementing R/E/L data collection—so much so that a group of leading health care organizations, including the American Hospital Association, has recognized this work as a national model in their call to action to eliminate health care disparities.
The Health Council began working with member hospitals to standardize R/E/L data collection in 2010, expanding to primary care practices over the next two years. More than 1,200 registration staff members at area hospitals have been trained to obtain these self-reported data, and now R/E/L training has been integrated into new staff orientation and yearly training programs. This has led to improved decision-making supported by re-engineered information technology systems that better capture standardized, accurate, and self-reported R/E/L data. Having these data can inform ways to improve care delivery in inpatient and outpatient settings.
"The race, ethnicity, and language data serve us best when we had a specific focus and looked at smaller data sets instead of large, bigger picture situations," explained Nancy Strassel, senior vice president, external affairs and communication of Greater Cincinnati Health Council and co-Project Director for Cincinnati Aligning Forces for Quality. "A few years ago, such questions weren't top of mind for hospital staff, but Cincinnati is growing and becoming more culturally diverse, and this data intervention has allowed us to change how we frame questions and how we approach care for all patients. For example, hospitals can now review data by service line to see if certain populations need preventive services or screenings. We can ask 'Ok, why are more African American men going to the ER? Do they wait for their conditions to worsen before seeking care?' The data continue to give us 'a-ha' moments and help us identify what works and what needs to change."
Lost in Translation
Consistently documenting a patient's language preference is an important part of Cincinnati's R/E/L improvement work. As Cincinnati-area hospital staff quickly learned, a patient with limited-English proficiency (LEP) can encounter issues that affect every aspect of care. One of the Council's member hospitals, Good Samaritan of TriHealth, quickly experienced its own “ah-ha” moment when review of R/E/L data showed inconsistent use of interpreters in the emergency department.
TriHealth, a member of the Greater Cincinnati Health Council, is an integrated health care system composed of two hospitals and several outpatient facilities in the Cincinnati area that serves a culturally and linguistically diverse patient population. For many patients, emergency departments are often a first point of entry for care, and this is especially true for many LEP patients, who are more likely to lack a primary care physician in the region, explained Lisette Martinez-Davis, director of diversity services for TriHealth.
When TriHealth staff encounter a patient with limited English proficiency, the protocol is to notify a hospital operator to contact the interpreter agency, and the agency will either send an interpreter to the facility or have the interpreter dial in by conference call. However, R/E/L data showed emergency department staff were not always following this protocol.
"Initially, staff members believed that a family member that spoke English and the patient’s language did not need an interpreter assigned," said Rhondalyn Prince, director of inpatient quality improvement for the Greater Cincinnati Health Council. "An interpreter should always be offered. The Joint Commission strongly discourages the use of family and friends for interpretation due to patient confidentiality and the Health Insurance Portability and Accountability Act (HIPAA). Medical interpreters are bound by HIPAA and are skilled, literate, and educated in medical interpretation. "
Hospital staff can't ensure that any family member is proficient in both English as well as the patient's own language. Hospital staff also need to recognize that trained interpreters with unique qualifications can benefit them as much as they benefit patients by helping boost trust and treatment compliance.
The problem turned out to involve different components: Not all staff were aware of hospital policies regarding interpreter services availability and protocols, and patients and families with limited English may not have known to ask about such services or may have opted to rely on a family member.
Further, understanding a physician's recommendations can be a challenge in any language; there are drugs with unusual names, follow-up tests and procedures, and treatment regimens to adhere to. Interpreters, available onsite at the hospital and by phone, are trained to translate medical terminology, but they weren’t being used consistently.
"This was a surprise to us since we have existing hospital policies encouraging [the use of] interpreter services," said Martinez-Davis. "There are many patients who do not speak English as a first language who come to us via the interstate or the airport. We need to be sure we can appropriately accommodate them."
TriHealth continues to study this problem to address both the issue of consistently documenting whether a patient receives an interpreter and also why he or she might refuse one. It sees value in finding out more in order to continue to fine-tune its interventions across its hospitals, noting that patients with limited English proficiency are more likely to schedule and keep follow-up appointments if they have interpreter services available.
Some interventions in the emergency departments have included placing dual handsets at the front desk to make it easier for staff and patient to listen in together when an interpreter is called. Speaker phones are also used so family members can listen in on physician instructions. Additionally, the hospital is making sure electronic medical record systems such as EPIC capture any patient refusal for interpreter services. Hospitals also need to document that any waiver of interpreter services has been properly explained, and patients need to sign that hospital staff discussed with them the risks of using friends or family members as interpreters.
While reasons why patients might refuse an interpreter need more exploration, the hospitals suspect that a lack of trust or comfort with being in an unfamiliar hospital setting may be a factor in making patients turn instead to a family member they know. "What it boils down to," said Strassel, "is, 'Do we know who our patients are?' R/E/L data give us a clearer picture of what is really going on." Identifying how many limited-English patients are being seen, determining how many receive interpreter services, figuring out why patients might refuse services, and finding out whether any of this is documented in the electronic medical record are critical steps. “The ability to look at patient data by R/E/L doesn’t always provide immediate answers, but it helps providers to ask important questions that can ultimately lead to better care,” Strassel said.
In addition to an absence of documented medical interpreter requests, TriHealth R/E/L data showed a growing number of patients speaking Russian and French. Preliminary data suggest there are language gaps TriHealth will need to fill by making Russian and French interpreters available to patients. "It was unexpected to see so many people mark Russian as their first language," said Prince. "We also see more French-speaking patients who are from Senegal and Ghana, which is a recent development. We need to be sure we have enough Russian interpreters as well as French interpreters for our patients."
Discovering High Compliance in the Data
R/E/L data can also be used to evaluate different aspects of a health care delivery system and sometimes serve as a compass to point leadership in the direction of what holds promise. Good Samaritan Hospital of TriHealth learned, for instance, that R/E/L data showed a 100 percent adherence rate in a special prenatal and perinatal program for Hispanic women, “Spanish Language Social Workers in the Community.” While the program had been in place before R/E/L data collection began, the recent rollout of R/E/L enabled analysis to show the program was effective. The obstetrics program included two bilingual social workers who attended community events to identify pregnant women in need of care, followed up with mothers-to-be throughout the course of their pregnancies by calling them at home, and maintained weekly communication with these women to ensure they kept their prenatal appointments, ate healthfully, and were supported throughout labor, delivery, and postnatal care. Hispanic women participating in the program also received assistance in addressing any paperwork or payment problems. The R/E/L data showed a strong adherence rate among expectant Hispanic mothers receiving care at Good Samaritan Hospital, and the hospital points out that higher adherence is associated with better clinical outcomes. "Hispanic women traditionally have faced barriers to good health care," said Martinez-Davis. "Having bilingual case workers speaking one-on-one in Spanish with women who were expecting proved to be very effective."
While the program was not set up to have a comparison group, the data have pointed hospital leadership to a department program that might work for other obstetric patients or perhaps be replicated for other service lines.
Learning to Ask Questions
Behavior change at every level is essential to successful R/E/L data collection, and TriHealth hospital staff are still adapting a few years after launch. Initially, staff reported feeling uncomfortable about asking patients their race. "This was a big change, and there was concern about patients pushing back," said Martinez-Davis. "But this change had to happen because staff can't assume or volunteer race or ethnicity information. Sometimes a patient would say, 'Well, I'm part of the human race, what does it matter?' We trained staff to be prepared for these types of reactions so patients would quickly be put at ease, and staff could explain why asking about race, ethnicity, and language preference was important. Most patients were fine about being asked."
TriHealth prepared a script that encouraged self-reported information from patients. Hospital staff role-played scenarios to develop familiarity with the script before taking it to the registration desks. Questions were asked in this order:
• What language do you feel most comfortable speaking with your doctor or nurse?
• For statistical purposes, please tell me your marriage status and religion?
• Are you Hispanic or Latino?
• What category best describes your race? (Staff presents patients with a list of categories to choose from)
If patients became uncomfortable with the questions or verbally pushed back, staff were trained to assure them such information would be used to improve their care. If R/E/L data were not fully captured during the registration process, with paperwork marked with "Unknowns" or "Unavailable," staff were trained to validate this information by asking the patient during their hospital stay or course of their care when it was appropriate to do so. This might occur, for example, when an ambulance brings in an unconscious patient or a patient is unable to communicate after an accident.
Ensuring R/E/L data accuracy doesn't hinge on a single visit. Staff also were trained to verify these data every three months or at follow-up patient visits, updating this information in the patients' electronic medical records.
Building on Success
Implementing a process and training for collecting R/E/L data was just the beginning, said Prince. "We want to explore the process of creating a computer dashboard for the patient population, in our community, and perhaps we include scores to more clearly signal what's working, who needs help and what needs to change." Examples could include readmission rates or the incidence of emergency department visits stratified by R/E/L, or perhaps some data on specific conditions that are more prevalent in a certain population, such as kidney disease or hypertension. Knowing if there are differences could provide important information and become increasingly important as Cincinnati's demographics continue to shift, with immigrant populations from Russia and African countries presenting with their own unique health challenges.
R/E/L data also could be used to enhance existing community outreach efforts. "If the data show that African American men are not getting prostate screening, do we partner with neighborhood barber shops or grocery stores or churches to get the word out?" asked Prince. "Given our success with Hispanic women who are expecting, could those connections help foster changes in getting more women to undergo cervical or breast cancer screening? How can we use R/E/L data to develop trust with different communities? We really have only just started to tap into the variety of possibilities."