Why EHR Data Is Key to Gathering More Accurate Medical Assessments

Jake Hochberg, Executive Director, Customer Insights at Arcadia

Patient data in electronic health records (EHRs) holds a wealth of insights that can enable health systems to better serve their patient populations.  

However, data from EHRs is challenging because this data is messy due to the lack of standardization in how clinical data is collected, and often times valuable information is buried in unstructured fields such as notes. Additionally, EHR data is disconnected from EHR data at other sites of care resulting in an incomplete view of the patient.  

For example, there are dozens and dozens of ways for blood pressure to be recorded in an EHR. Aggregating data from many EHR sources need to have quality processes in place to ensure the standardization of clinical information like blood pressure is reliable.  

Having the ability to scrub and accurately aggregate data from multiple clinical systems is step 1 in the process of empowering health systems to gain a fuller picture of their patient population. Layering on top of this aggregated data set the ability to match patients across clinical sources and even tie it to claims data allows health systems to unlock the true power of their clinical data.  

EHR data in action: Insights into long COVID
The primary advantage of EHR data is the low latency compared to the 90-day lag of claims data. For example, in the early stage of the pandemic, clinical researchers in collaboration with the COVID-19 Patient Recovery Alliance needed to study patients with long COVID-19. EHR data proved particularly important during this time before there was an ICD code to identify COVID patients. Researchers were able to look at the free-text notes about symptoms and diagnoses for patients who had COVID within EHR data. The ability to analyze timely, early COVID-19 data enabled researchers to obtain insights that they would have had to wait months to extract from more structured forms of data, such as claims. 

The long-COVID analysis was based on COVID-19 infection and vaccination data gathered during a 15-month period from February 2020 to May 2021. The study included about 1.1 million patients infected with COVID-19, of which more than 240,000 had available a complete longitudinal history prior to the pandemic as well as at least one encounter with a healthcare provider at least eight weeks after their COVID-19 diagnosis. 

Key findings from the study include:

– Patients who received at least one dose of any of the three COVID vaccines (Pfizer-BioNTech, Moderna, and Johnson & Johnson) prior to their diagnosis with COVID-19 were up to ten times less likely to report two or more long COVID symptoms compared to unvaccinated patients. 

– Unvaccinated patients who received their first COVID-19 vaccination within four weeks after SARS-CoV-2 infection were up to six times less likely to report multiple long-COVID symptoms.

– Those who received their first dose 4-8 weeks after diagnosis were three times less likely to report multiple long-COVID symptoms compared to those who remained unvaccinated.

– Vaccination had a protective effect even when the first dose was administered up to 12 weeks after diagnosis.

Despite the many challenges posed by the COVID-19 pandemic, there is one key difference that enabled our 21st-century response compared with past public health challenges: timely EHR data. To fight such a devastating virus, the world did not have time to wait for years-long randomized controlled trials to reveal every single detail about vaccine response. We needed timely data to enable immediate action, and, ultimately, the aggregation and organization of patient records have given us the tools to create a better recovery.

About Jake Hochberg

Jake Hochberg is Executive Director of Customer Insights for Arcadia, a leading data analytics platform for healthcare and life sciences, transforming data from disparate sources into targeted insights, putting them in the decision-making workflow to improve lives and outcomes.