Improving event prediction using general practitioner clinical judgement in a digital risk stratification model: a pilot study
2024

Improving Event Prediction with GP Clinical Judgement

Sample size: 31392 publication 10 minutes Evidence: moderate

Author Information

Author(s): Emma Parry, Kamran Ahmed, Elizabeth Guest, Vijay Klaire, Abdool Koodaruth, Prasadika Labutale, Dawn Matthews, Jonathan Lampitt, Alan Nevill, Gillian Pickavance, Mona Sidhu, Kate Warren, Baldev M. Singh

Primary Institution: Keele University

Hypothesis

Can combining digital risk stratification with GP clinical judgement improve the prediction of urgent care and mortality events?

Conclusion

Adding GP clinical judgement significantly improved risk prediction for urgent care and mortality events.

Supporting Evidence

  • 3968 out of 31,392 patients were data stratified into the Escalated group.
  • The 30-day event rate was 19.0 in the whole population and 168.0 in the Concern group.
  • GP clinical judgement improved the prediction of 30-day and 90-day events.
  • The ROC c-statistic for GP clinical judgement was 0.603.
  • Adding GP judgement increased the negative predictive value significantly.

Takeaway

This study shows that doctors can make better predictions about patient needs when they use both data and their own judgement.

Methodology

A cohort study involving 6 GP practices in a UK city, using digital risk stratification and GP clinical judgement.

Potential Biases

GP assessors may not be representative, and missing data from patients not reviewed could affect results.

Limitations

The study was small-scale and may not represent all GP assessors; outcomes related to quality of care were not gathered.

Participant Demographics

Patients from 6 GP practices in a deprived, multi-ethnic UK city.

Statistical Information

P-Value

p<0.001

Confidence Interval

95% CI 0.19–0.33

Statistical Significance

p<0.001

Digital Object Identifier (DOI)

10.1186/s12911-024-02797-5

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication