Blog from October, 2021

Hannah Nelson


Standardized EHR documentation practices for SDOH such as race and ethnicity could help prevent data quality issues that can lead to bias.


Social determinant of health (SDOH) data quality issues signal the need for standardized SDOH EHR documentation practices to avoid bias and promote health equity, according to a study published in JAMIA.

Researchers conducted a review of 76 studies related to SDOH data quality.

The majority of articles that discussed race, ethnicity, or country-of-origin data (65 percent) examined data plausibility, which refers to data accuracy.

Accurate race/ethnicity data is key for clinical research, especially as the industry continues its focus on SDOH data and health equity.

However, researchers noted misclassification bias as a problem or a potential problem in more than half of the articles about race/ethnicity data plausibility. What’s more, several studies reported that implausible data and misclassification errors were more likely for certain groups.

Notably, 14 studies reported that Hispanic patients were more likely to be misclassified in terms of their ethnicity. Patient misclassification includes missing ethnicity information or misclassification into the “Other” category.

Several studies speculated that this may be due to the fluid nature of the definitions of race and ethnicity.

“The fluidity of these definitions leads patients to respond inconsistently to questions about their race/ethnicity, thus causing problems with data reliability,” the literature review noted. “Further, the fact that these categories are so broad and poorly defined leads to difficulties with data validity.”

Misclassification can have profound impacts on clinical research, the study authors noted.

“When patients from one racial or ethnic group are lost in another group or mistakenly categorized as ‘Other,’ subsequent analysis can cause those groups to be under-represented in research results,” they explained. “Misidentification of the race or ethnicity of groups of patients can inadvertently lead to the erasure of those groups from clinical research.”

Several studies speculated that variations in how healthcare organizations collect and record race/ethnicity information have impaired data quality.

“Consistently applied standards for SDOH data collection in the EHR would result in improved data quality, which in turn would lead to more robust research, care coordination, and population health management,” they noted.

The review authors noted that the integration of patient-facing health IT could also help mitigate race/ethnicity misclassification.

“It is possible that the increasing use of dynamic patient-facing data entry tools may allow people to inform and correct their own demographic information, thus helping to improve the quality of race, ethnicity, and country-of-origin data in the future,” the study authors suggested.

The quality of data elements is key in supporting interoperability for large-scale research, data analytics, and care coordination, the review authors emphasized.

“Privacy-preserving record linkage (PPRL) methods identify when records from different sources belong to the same entity while minimizing the exposure of sensitive personal information,” the review authors wrote. “These techniques often rely heavily on patient address along with name and date of birth. When there are errors or missing address data, linkage quality suffers.”

The review authors added that healthcare organizations originally collected demographic data such as race, ethnicity, insurance status, and address, for administrative purposes. Repurposing administrative data for secondary, retrospective research can result in poor data quality.

“Given the increasing importance of social determinants in health equity research and intervention, it is crucial that healthcare institutions work to improve the quality and availability of these data,” the authors noted.

The review revealed several evidence-based solutions to mitigate issues associated with data quality problems. The researchers grouped these recommendations into five main suggestions: avoid complete case analysis, impute data, rely on multiple sources, use validated software tools, and select addresses thoughtfully.





Hannah Nelson


Using machine learning to extract SDOH data from EHR clinical notes could aid in the development of clinical decision support systems, a study says.


Machine learning offers significant potential to extract social determinant of health (SDOH) data from EHR clinical notes, which may aid in the development of clinical decision support systems, according to a study published in JAMIA.

The researchers conducted a literature review of 82 publications focused on the extraction of SDOH data from EHR clinical notes.

Despite increased interest in capturing SDOH in EHRs, data is typically locked in unstructured clinical notes, the study authors explained.

In general, the researchers observed two major steps associated with SDOH extraction systems from the literature.

“The first step is gathering SDOH-related keywords to create lexicons for each SDOH category, and the second step is developing rule-based or supervised systems to locate clinical notes associated with SDOH categories or extract SDOH concepts,” they wrote.

Rule-based approaches require manual chart review while supervised systems leverage machine learning approaches, the researchers explained.

In total, 22 out of 82 publications used rule-based methods to identify SDOH in clinical notes. Health IT researchers leveraged rule-based systems more frequently for housing, transport, and social isolation.

On the other hand, researchers used machine learning techniques more frequently for smoking, alcohol, and substance use data extraction efforts.

Insufficient volumes of structured data for social support and homelessness may explain why rule-based systems were more common for these variables, the researchers suggested.

The study authors pointed out that integrating automated SDOH data extraction systems into EHRs may aid in clinician burden.

“In clinical settings, providers report spending less time on patient care and more time on administrative burdens that are byproducts of data management in the EHR,” they wrote. “Manual screening of SDOH could potentially further complicate and delay the process for healthcare staff.”

“We believe that the NLP-based SDOH identification and the developed outcome analysis tools may offer an optimal solution that may minimize impact on current documentation routines while guiding providers to make better, informed and holistic clinical decisions, they explained.

The researchers also noted that as SDOH categories grow in number and complexity with the industry’s focus on health equity, storing SDOH in a structured framework could become inefficient and require frequent maintenance.

“With increasing recognition of nonclinical factors that define patients’ health risks, needs, and outcomes, it becomes equally imperative that social and behavioral concepts are captured in order to be leveraged during clinical decision-making related to diagnosis and therapy planning,” they wrote.

“Devising novel ways in which such data can be extracted and leveraged with as little impact on current documentation routines of providers is an ideal solution,” the authors continued. “With the valuable knowledge of the relatively new literature in this area, researchers can leverage such reviews to steer their study in innovative ways.”

The researchers noted several opportunities for future NLP research that extracts less-studied SDOH such as child and sexual abuse, financial issues, transportation, neighborhood, social isolation, family problems, employment, education, food insecurity, and access to healthcare.

“Another interesting study would be to compare different aspects of NLP algorithms, such as system performance, amount of annotated data, type of NLP systems, and so forth with the difficulty of SDoH extraction,” the researchers added.





Hannah Nelson


A primary care EHR integration helped boost referrals, especially in settings with high clinician engagement.


EHR integration in primary care settings can increase referrals, but use of that health IT usually varies based on clinician engagement, according to a recent study published in the Annals of Family Medicine.

The study, which focused specifically on EHR integrations that help refer patients with hearing loss to audiologists, underscores the importance of clinician engagement in health IT use.

Primary care physicians (PCPs) rarely screen for hearing loss due to time constraints, conflicting priorities, and limited experience managing hearing loss, the study authors explained.

The researchers implemented an EHR integration at 10 clinics across two health systems to see if EHR prompts would boost PCP assessment of hearing loss.

The EHR implementation set off an EHR prompt to screen for hearing loss among patients ages 55 years or older, asking a validated one-question screener.

Overall, audiology referrals increased from 2.2 percent to 11.5 percent.

“This prompt may have reduced some of the uncertainty clinicians feel around hearing loss, thereby increasing their screening and referral rates,” the study authors suggested.

Through clinic observations and semistructured interviews with 27 family medicine clinicians, the study authors found several main themes.

First, the researchers noted that clinicians found the prompt overwhelmingly simple to use and considered it an effective way to increase conversations with patients about hearing loss.

“Our results indicated that the EHR prompt was easy to use and incorporate into typical workflows, likely because it was designed through extensive feedback from PCPs,” the study authors explained.

Next, the study revealed that clinician engagement and buy-in played a vital role in implementation of the health IT tool. For example, clinicians who shared personal experiences with hearing loss described strong support for the EHR integration, the study authors explained.

“They often emphasized the difficulty of addressing their own family members’ hearing loss due to an unwillingness to admit hearing loss or wear hearing aids,” the researchers wrote. “They viewed the prompt as an opportunity to discuss the impact of hearing loss on quality of life with patients.”

The research also revealed that medical assistant (MA) involvement in the EHR hearing loss prompt workflow varied by health system, clinic, and clinician.

Some physicians reported unclear expectations about how staff and clinicians should interact with the prompt. For instance, at one health system, MAs were responsible for addressing prompts related to colorectal and breast cancer screenings. The authors suggested that systematically leveraging MAs could be an effective way to increase the utilization of screening prompts in primary care.

“Moreover, expanding MA responsibilities may lead to innovations for panel management, health coaching, or patient navigation that reduce patient-level barriers to hearing loss,” the researchers added.

The researchers also found that clinicians prioritized the prompt during annual wellness exams compared to acute visits when there were more pressing health needs and concerns.

Lastly, while the prompt resulted in more conversations about hearing loss and more referrals, clinicians reported uncertain impact on patient outcomes.

“Clinicians were unsure whether the prompt had increased use of hearing aids, though most expected to see an increase over time,” the study authors wrote. “Many reported that additional barriers, particularly cost, would limit the number of patients who obtained hearing aids.”

“Longitudinal research is needed to know whether a prompt leads to improved hearing loss screening and referral over time,” the researchers noted.