How accurately is hospital acquired pneumonia documented for the correct assignment of a hospital acquired complication (HAC)?

BACKGROUND

In 2016, the Australian Commission on Safety and Quality in Healthcare (ACSQHC) released a list of 16 categories of potentially preventable, high impact hospital-acquired complications (HAC) identified by using administrative coded data (ACD). An important category are hospital-acquired infections (HAI). Within this category, hospital-acquired pneumonia (HAP) is among the most frequent complications documented. There are no published studies concerning the current ACSQHC approach to HAI surveillance using ACD and no pneumonia-specific ACD studies reported from Australia. Published work indicates that ACD detection of HAP has low a sensitivity and positive predictive value (PPV). The current study was designed to examine whether coders correctly reflected the documentation of HAP that was present in the medical record and also evaluated the medical documentation that was present.

METHODS

One hundred patients with ACD encoded HAP were selected for review, drawn from admissions to 2 Hunter New England Health hospitals during 2017. Patient records and the eMR were reviewed by two medical officers to assess medical and radiological documentation of pneumonia. The district coding manager reviewed the accuracy of coding of a subset of 23 cases where medical review had not located documented evidence of HAP.

RESULTS

Of the 100 reviewed cases, the median patient age was 75 years (range 0-95 years) with 3% under 16 years of age. Twenty one were intensive care-associated of which 13 were associated with ventilation. In 23 cases the documentation was disputed and a secondary review took place - the coding manager confirmed coding changes in 14 of these 23 cases.

CONCLUSIONS

This study found that administrative coded data of HAP, utilizing the ACSQHC method reliably reflected the available documentation with a PPV of 86% (95% binomial exact confidence interval 77-92%), much higher than documented by previous ACD studies. The actual documentation of pneumonia by medical staff frequently used the non-specific term 'lower respiratory infection (LRTI)' which we recommend to be avoided. Radiological confirmation was absent in one third of cases. We recommend the adoption of a medical note template checklist for HAP to prompt clinicians with the accepted diagnostic criteria. We also recommend documenting a reason as to why any antibiotic has been commenced in a hospitalized patient in accord with the ACSQHC Antimicrobial Stewardship Clinical Care Standard.