Abstract
Introduction: Pneumonia in hospitals is common with 359,000 finished consultant episodes with a spend of £705 million (2013/14 NHS reference costs).
Activity is coded to diagnosis driven Healthcare Resource Groups (HRG's) of broncho-pneumonia (BP) or lobar pneumonia (LP) by hospital teams coding from discharge letters/casenotes. This does not reflect the clinical concepts of hospital acquired pneumonia and community acquired pneumonia. Moreover coding to either BP or LP may affect the hospital's mortality statistics.
To determine the accuracy of coding for patients who died of pneumonia we have reviewed the CXRs / discharge summaries of patients over a 6 month period at our 1000 bedded hospital serving a population of 500,000.
Methods: Letters and CXR were reviewed and diagnosis re-classified when possible. The tabled results for the first 3 months are as below.
Results:
Lobar Pneumonia changed to broncho-pneumonia | 69 | 48.3% |
No change indicated (correct diagnosis) | 34 | 23.8% |
Not legally possible to change | 35 | 24.8% |
Pulmonary infection but not Pneumonia | 5 |
Conclusions:
In a large teaching hospital pneumonia coding is inaccurate.
Patients are often coded as having LP but on review of the information 69 (48.3%), have BP. This is an important aspect as it improves the performance of the Trust as a greater mortality is permitted for patients with BP and may also change trust income.
35 (24.5%) of cases did not have pneumonia but because it was mentioned in the discharge letter, the coding could not be changed.
In only 34, (23.8%) did the coded discharge diagnosis agree with the retrospective review.
- Copyright ©ERS 2015