Article Text

Download PDFPDF

13 Clinical informatics to direct community echocardiography: an electronic healthcare record pilot
  1. Kamatamu Amanda Mbonye1,
  2. Alireza Yazdi1,
  3. Shane Cashin2,
  4. Nikhil Ahluwalia2,
  5. Nabila Laskar2,
  6. Sanjeev Bhattacharyya2,
  7. Carl Hayward2,
  8. Guy Lloyd2,
  9. Abhishek Joshi2
  1. 1Royal London Hospital, Bart’s NHS Trust
  2. 2Bart’s Heart Centre, London, UK


Background Delayed diagnosis of valvular heart disease carries a poor prognosis, and early identification is desirable. We undertook a retrospective analysis of echocardiographic and electronic health care record data from the largest single cardiovascular service in the UK, to identify the burden of acute presentations with previously undiagnosed valvular heart disease and to determine the geographical and demographic distribution.

Methods and Results Automated text mining analysis was retrospectively applied to all echocardiographic examinations performed between 2015 and 2019 at Bart’s Health NHS trust, identifying 2043 reports containing text or numerical data indicating severe valvular lesions. Demographic and clinical data was integrated with the echocardiographic dataset, identifying the postcode and GP practices for with the highest proportion of patients with severe valvular disease that were diagnosed during acute inpatient admissions. 376 individuals had severe valvular lesions identified during acute admission, of which 269 (72%) had no previously documented echocardiogram. A cluster of 11 GP practices (9%, 11 of 117 practices) were identified as having a higher proportion of diagnoses of severe valvular disease on acute admissions [figure 1]. These 11 were plotted geographically, alongside correlating postcodes, to identify geographical hotspots [figure 2]. Analyses were undertaken using Matlab, R and ggplot2.

Conclusions A geographical cluster of GP practices, centred around a single hospital, had a higher proportion of patients diagnosed with severe valvular disease during acute admissions without a previous echocardiogram. Outreach echocardiography provision in these regions could potentially identify patients with valvular disease before acute decompensation. Further work should focus on improving methodology to identify cases and investigating risk factors that predispose to diagnosis of severe valvular disease in extremis.

Conflict of Interest none

  • Echocardiography
  • Clinical Informatics
  • Valves

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.