Answers are grounded in this paper and cite their sources.
BackgroundThe demand on diagnostic services in England has risen year-on-year.The National Health Service (NHS) has identified the need for radical investment and reform of diagnostic services, both to prevent missed or delayed diagnoses, and to transform the process itself that can be duplicative and inefficient.New service delivery recommendations emphasize virtual consultations and community diagnostics to relieve pressure on acute services, an issue only exacerbated by the COVID-19 pandemic.There is considerable focus on the robust diagnosis of respiratory disease, with targets to reduce variation in spirometry testing nationally through improved training of primary care personnel.However, there is currently no widely accepted approach for evaluating diagnostic performance.Measuring diagnostic quality must consider more than whether the final diagnosis is correct, but also the efficiency, timeliness, and rigorousness of the process.There is opportunity for advanced analytic methods able to incorporate a spectrum of information and account for the dynamic nature of the diagnostic process.Such methods would require appropriate input data sources with clinically rich information and linkage across healthcare tiers. 1 Routinely collected health data holds huge potential for clinical research.Each patient interaction presents an opportunity to improve services and standard of care through data-driven analytics.However, the nature of routine data can limit its applications in research.The availability and quality of all variables required for analysis must first be assessed to evaluate the capacity of the data in answering the intended research question.Existing literature recognizes the challenges of using routine data for evaluating diagnostics due to incomplete and erroneous coding, with an emphasis on the critical role of physicians in influencing the quality of routine data for research purposes.2,3 This study aimed to extend the discussion of the suitability of routine data for measuring diagnostic quality by considering the problem in the context of a specific data source and disease area.We aimed to illustrate key barriers using electronic health records (EHRs) from the Morecambe Bay Community Data Warehouse (CDW) applied to chronic respiratory disease (CRD). MethodsThis report retrospectively analyzed primary and secondary care data from the Morecambe Bay CDW, an SQL Server owned by the University Hospitals of Morecambe Bay NHS Foundation Trust.The CDW uses pseudonymized NHS Numbers to individually link healthcare data across the Morecambe Bay, north-west England, covering 3 hospitals, 32 general practices (GPs), and a population exceeding 360,000.Patient demographic information was obtained from GP records.CRD diagnoses (asthma, chronic obstructive pulmonary disease [COPD], bronchiectasis, and interstitial lung disease [ILD]) were identified from GP records by the first recording of a relevant SNOMED diagnosis code between 01/07/18 and 30/06/23.Only patients aged ≥35 years at time of diagnosis and with ≥12 months of continuous medical records prior to diagnosis were considered.Patients with missing age or sex information were removed from analysis.If an individual had more than one CRD diagnosis, these were treated as distinct events.
No takes yet. Share an insight, caveat, or question.
Mountain et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: