Quality and standards
High-quality data are essential for the safe, high-quality care of individual patients, as well as for secondary uses of the data such as contracting, audit and service planning. Data quality encompasses the accuracy, completeness and timeliness of clinical, demographic and administrative information recorded in both paper and electronic health records. Comprehensive and accurate recording of clinical data is time consuming, but will compromise care if not done properly. Clinical input into the design of electronic record systems is essential if this is to be made more efficient. At the present time, in most hospitals, secondary data are extracted from paper and electronic records, and coded (ICD-10 for diagnoses and OPCS-4 for procedures) by trained coders for central returns and aggregate analysis. In due course, clinical terms will be coded automatically using the international standard structured clinical vocabulary for use in an electronic health record (SNOMED-CT) at the time of entry into electronic records. However, while secondary coding is in place physicians should ensure that a process is established for the validation of the clinical content of central returns. The HIU has published recommendations for improving the quality of coding, to ensure that hospital activity is accurately recorded.
Clinical record standards
Standards for the structure and content of patient records have been developed using published evidence and consultation with doctors, patients, nurses and allied healthcare professionals. These cover hospital referral letters, inpatient clerking, handover communications, discharge summaries and outpatient letters. The standards were signed off as fit for purpose for the whole medical profession by the Academy of Medical Royal Colleges in April 2013. Implementing the standards in electronic patient records will improve the quality of record keeping. To achieve this, systems procurements should include a requirement for suppliers to use the AoMRC standards. During implementation, clinicians will need to be closely involved in the design of screens and outputs as the AoMRC standards provide a super-set of headings and local decisions will be needed about which are required for specific specialties and services, and the sequencing and layout of the information to best support the care process.
Clinical information modelling
Clinical information modelling is a way of specifying the data structures for storing clinical information in electronic health records and for communicating information between electronic health record systems, to ensure that data items with the same logical meaning are represented consistently. If all electronic health record systems use the same data models, patient data can be faithfully transferred between them according to clinical need. Clinical information models are formally represented as ‘archetypes’ in the openEHR specification for interoperable electronic health records. This is being used by NHS Digital as advocated in the NHS England Interoperability Handbook.
Physicians need to be involved in the specification of these data models to ensure that they are clinically meaningful and practical for routine clinical use. The data models should also align with audit and reporting datasets, so that information only needs to be collected once for multiple purposes.
Clinical information models have been developed, based on the AoMRC clinical record standards (above) for electronic transfer of care messages such as discharge summaries, outpatient letters etc. This enables structured and coded information on diagnoses, procedures, allergies, medications etc to be incorporated into the GP electronic record systems to improve continuity of care and reduce transcription errors (eg Clinical Knowledge Manager). System procurements should include a requirement for suppliers to use these technical standards to produce electronic discharge summaries.