Discrete Data Are Key to Realizing Electronic Health Records’ Full Potential

Hamid Emamekhoo, MD

Following the passage of the Health Information Technology for Economic and Clinical Health Act (HITECH) of 2009, electronic health records (EHRs) were rapidly adopted by health systems across the United States.¹ Based on reports published in 2019, 89.9% of office-based physicians use EHR systems.^1,2 EHRs have the potential to improve quality of care and patient outcomes.^3,4 EHR adoption has clear benefits, such as facilitating access to past medical records, making order placement easier, and improving communication with other health care providers. However, EHR systems’ design, usability, and interoperability issues can hinder integration and use.

Paper to Electronic Transformation

As a practical first step in EHR integration, paper reports were converted to digital copies. This improved accessibility and eliminated the need for storing hard copies of these documents in conventional hospital archives.⁵ However, these files were stored as digital images, which, at best, could be converted to plain text. Following this path, providers started typing or using transcription services to generate progress notes and most other reports. Unfortunately, this approach drastically limits the potential advantages and applications of EHRs.

Despite improvements in order entry, most of the results, except for laboratory tests with discrete data, were entered into the EHR as plain text reports. Some of these reports (eg, pathology and genetic testing) still are generated from discrete data elements initially recorded in another database. This trend created a large number of reports that were better organized and more accessible but only comprehensible to human users because the computers running the EHR do not understand English.

One of the primary functions of EHR platforms is to facilitate human-machine interactions by translating our phrases into code (machines’ language). For example, malignant diagnosis and disease stage might be clearly documented in an oncology office visit note. However, this information is not accessible to the EHR unless it is converted to numbers, such as International Classification of Diseases codes, and stored in a prespecified location in the EHR, such as the list of current medical problems. Further, entering the diagnosis into the problem list requires additional steps to open that section of the EHR, find the accurate diagnosis with necessary relevant details, and enter cancer staging information.

Routine Clinical Workflow

In addition to interacting with the patient to gather information and perform a physical examination, generating a report (eg, consult notes or progress notes) might be the focus of the clinicians’ routine workflow. These reports are the primary method of communication with other care team members and are commonly used to support billing. Documentation styles and preferences are highly variable among clinicians. They might use dictation transcription services, voice recognition technologies, note templates and macros, or free text to generate reports. Although most providers acknowledge the value of discrete data availability in EHR, many consider the required additional steps for data entry at the point of care highly disruptive to their clinical workflow and a barrier to practicing medicine a way that is most effective.⁶

Let us return to our example of malignant diagnosis. Finding the accurate diagnosis and relevant details in an unstructured note can be challenging. Reviewing multiple prior reports might be necessary, which will increase the review time and the reader’s cognitive burden. Adding diagnosis to the problem list will make it easier for other care team members to retrieve this information. In addition, the discrete data can support secondary clinical use, such as integration of clinical decision support tools and pathways, clinical trial recommendations, and billing.

Challenges and Potential Solutions

Multiple factors, including EHR design, inadequate training, and insufficient appointment time augment the challenge of discrete data collection at the point of care. Artificial intelligence and natural language processing platforms may help identify and extract clinically relevant information from our notes or conversations with patients during a visit. In the interim, we should focus on improving our EHR design and developing user-friendly tools that are least disruptive to clinical workflows.

It is imperative to engage clinicians as early as possible in designing any new informatics tool. Discussing the need for change, the selected approach’s rationale, and the outcome’s potential effect on the practice can lead to easier product integration in the established workflows and improve user adherence. Applying these principles, the University of Wisconsin Carbone Cancer Center team implemented an automatic validation step when closing the encounter, improving the capture of discrete data on malignant diagnosis and staging.⁷ Learning from the routine clinical workflows, we have embedded data collection forms in our standardized note templates.

These forms integrate collection of discrete data related to disease and treatment into the process of writing a note. The collected data serve multiple purposes. They generate parts of the disease assessment and treatment response evaluation in the progress notes and support rule-based EHR functions such as clinical decision support tools and clinical trial recommendations. In addition, this information is displayed in a longitudinal reporting dashboard to facilitate a comprehensive review of the disease course at a glance. Discrete data are the key to realizing EHR’s full potential.

Hamid Emamekhoo, MD, is an assistant professor and faculty member in the Department of Medicine, Division of Hematology, Medical Oncology, and Palliative Care at the University of Wisconsin School of Medicine and Public Health in Madison. He is also a member of the University of Wisconsin Carbone Cancer Center.

References

1. 2022 Medicare Promoting Interoperability Program requirements. Centers for Medicare & Medicaid Services. Updated November 15, 2022. Accessed November 21, 2022. bit.ly/3Er0Uiq
2. Electronic medical records/electronic health records (EMRs/EHRs). Centers for Disease Control and Prevention. Updated September 6, 2022. Accessed November 21, 2022. bit.ly/3OrWoEF
3. Buntin MB, Burke MF, Hoaglin MC, Blumenthal D. The benefits of health information technology: a review of the recent literature shows predominantly positive results. Health Aff (Millwood). 2011;30(3):464-471. doi:10.1377/hlthaff.2011.0178
4. McCullough JS, Casey M, Moscovice I, Prasad S. The effect of health information technology on quality in U.S. hospitals. Health Aff (Millwood). 2010;29(4):647-654. doi:10.1377/hlthaff.2010.0155
5. Dujat C, Haux R, Schmucker P, Winter A. Digital optical archiving of medical records in hospital information systems--a practical approach towards the computer-based patient record? Methods Inf Med. 1995;34(5):489-497.
6. Diaz-Garelli F, Strowd R, Ahmed T, et al. What oncologists want: identifying challenges and preferences on diagnosis data entry to reduce EHR-induced burden and improve clinical data quality. JCO Clin Cancer Inform. 2021;5:527-540. doi:10.1200/CCI.20.00174
7. Emamekhoo H, Carroll CB, Stietz C, et al. Supporting structured data capture for patients with cancer: an initiative of the University of Wisconsin Carbone Cancer Center Survivorship Program to improve capture of malignant diagnosis and cancer staging data. JCO Clin Cancer Inform. 2022;6:e2200020. doi:10.1200/CCI.22.00020