
5 minute read
The Role of Electronic Health Records (EHRs) in Patient Care
Dr. Sophie Laurenson (BSc Hons., Ph.D. Cantab.)
Electronic Health Records (EHRs) contain heterogeneous data types from multiple sources, often recorded as a narrative in unstructured free text. Extracting structured data is necessary for analysis to enable improved clinical decision-making.
Introduction Electronic Health Records (EHRs) is a collective term used to describe clinical information systems that collect, store and present longitudinal patient data in a digital format. In recent years, healthcare providers have been encouraged to adopt EHRs to improve patient outcomes and safety, boost efficiencies and aid in clinical decision making. Attention is now diverting away from EHR adoption and implementation and moving toward realising the potential benefits of digital records.
The History and Development of EHRs The first examples of clinical information systems date back to the 1960s when discrete systems were implemented by several large healthcare providers. Reports published by the National Academy of Medicine in the 1990s promoted more widespread EHR adoption in the US healthcare system based on their potential to improve patient safety 1,5 . These reports also defined the core functionalities and barriers to adoption. The 1990s saw a global expansion in the adoption of EHRs as they were touted as an essential tool to increasing quality of care and improving patient outcomes. Today, most healthcare providers in high-income countries have implemented EHR systems and many organisations in low and middle-income countries have followed suit.
Core Components and Functions of EHRs EHRs include a range of data types from different sources including demographics, medical history, treatment regimen, allegies, immunization status, laboratory and pathology test results, medical images, vital signs and billing information. This data may be entered and stored in a variety of formats including both structured and unstructured data types. Information is entered by physicians during patient consultations and may be amended by annotations or codifications and assimilated with data from other clinical sources at various timepoints 6 . Electronic data storage eliminates the need to locate paper records and improves the accuracy and legibility of data. It may also reduce the risk of data replication, reinforces the regular updating of records and decreases the risk of lost paperwork.
One of the key benefits of EHRs is the timely delivery of patient data to the medical practitioner. Combining multiple types of clinical data can aid clinicians in decision making and patient risk stratification. Appropriate access to data can support healthcare providers to achieve a variety of goals including improved care coordination, disease prevention and management and patient monitoring or support outside traditional care settings. Increased visibility across healthcare agencies can also lead to improvements in cost effectiveness through reducing unnecessary procedures and allocating resources more efficiently.
Issues with Usability and interoperability of EHRs Since the widespread implementation of EHRs, several issues have surfaced surrounding the usability and interoperability of current commercial systems. A recent review of the available literature revealed that the quality and usability of EHR systems is commonly poor 7 . Several issues with EHR documentation have been identified. These include structural problems manifesting in convoluted workflows and documentation quality deficiencies. In particular, the use of free text fields to record clinical narratives are common in many areas of medical practise and have demonstrated to be prone to error. Another common structural problem is the lack of standardisation in EHR systems across different aspects of healthcare delivery leading to problems with interoperability 8 .
FUJITSU LABORATORIES’ CO-CREATION STRATEGY WITH LEADING HEALTHCARE PARTNERS IS HELPING TO UNLOCK VALUABLE PATIENT DATA AND EVOLVE EHR PLATFORMS.
Each of these issues has received significant attention in recent years and studies are underway to investigate interventions to improve EHR design and use 9 . One issue highlighted by physicians is the additional workload required to complete EHR documentation and reporting. The original purpose of EHR implementation was to improve billing processes and a significant proportion of the total time and financial cost associated with EHR-based encounters is related to billing 10 . Despite the exponential increase in the amount of data recorded in EHRs, physicians have reported a lack of actionable data that can be applied to patient care. A recent survey reported that 65 percent of providers do not have the ability to view and utilise all the patient data they need during a consultation. Furthermore, only 36 percent were satisfied with ability to integrate data from external sources such as laboratories or referrals. Systems that can extract and analyse clinical data are of significant value in realising the full benefits of digital patient data 11 . However, extracting data in a structured format that can be readily analysed is both time- and cost-intensive.
The Importance of Structured Data in EHRs Data recorded in EHRs is highly heterogeneous, containing various and often incompatible formats. The most common data formats include: • structured and terminology encoded data, • structured data with limited or no encoding, • unstructured machine-readable text data, • unstructured scanned text or images 12 . Many EHR data entries are recorded in narrative format, enabling providers to articulate opinions and impressions via free text fields or dictation. This allows physicians and clinicians to record a nuanced version of data, without negotiating a structured entry system. However, this freedom of expression creates pragmatic issues with accurately communicating information between providers. Recording patient data as a narrative history can lead to comprehension issues between providers or specialisms that arise due to differences in jargon or terminology. One approach is to prospectively encode the data in a standardised format that can be subsequently analysed. Artificial Intelligence (AI)-based methods such as Natural language processing (NLP) can be employed to extract structured data from free text narratives.
For EHRs to be useful in guiding clinical decision making, it is necessary to establish standards in data structure and display. Currently, many EHR vendors utilize the ICD-9/10 and CPT code standards to apply a structure to clinical data 13 . Further solutions could include common data elements, such as the North American Association of Central Cancer Registries and STandards for Oncology Registry Entry standards used by national cancer registries. Terminologies for semantic tagging and annotation, such as Systematized Nomenclature of Medicine − Clinical Terms, Logical Observation Identifiers Names and Codes, and RxNorm would also be beneficial for standardising data structures 14 . Applying new automated artificial intelligence algorithms to the process of encoding and annotating medical data would enable more patient data to be extracted and available for analysis.
Conclusion Many chronic diseases require a multidisciplinary approach to care, involving teams of providers from disparate clinical disciplines. The coordination of care activities is facilitated through clinical documentation in EHRs, which relies on a consistent communication method between providers and between provider and patient. Automated methods of retrospectively applying standardisation and structure to information through transcoding and extracting data are required to fully realise the benefits of digital patient data.