Big Data Studies
Identifying potentially preventable emergency department (PPED) visits among patients with cancer in Ontario.
Background: Cancer patients (pts) often visit the emergency department (ED) when symptoms and side effects occur. Growing evidence suggests that some treatment-related toxicities can be managed proactively in outpatient clinics, improving patient experience and optimizing acute care utilization. Understanding PPED visits is crucial to developing and evaluating such improvement efforts. Our objectives were to quantify the extent of PPED visits in Ontario among cancer pts and identify the best measure of PPED for province-wide quality improvement. Methods: By linking Activity Level Reporting to the Discharge Abstract Database and the National Ambulatory Care Reporting System, we identified cancer pts who had ED visits or hospitalizations up to 30 days after receiving chemotherapy and/or radiotherapy from April 1, 2014 to March 31, 2015 in Ontario. Episodes were stratified as ED Only or Indirect Admission (ED visit leading to hospitalization). We mapped the presenting Canadian Emergency Department Information System (CEDIS) complaints against the PPED metric proposed by Panattoni et al (JOP, 2018) that combined the CMS preventable visits typology with the STAR PRO tool (Basch et al, JCO, 2016) which found 49.8% were considered potentially preventable (PP). Results: We identified 43,593 ED visits (67% ED Only& 33% Indirect Admissions) among 64,407 pts. The most common presenting CEDIS complaints were pain (20%), fever (13%) and shortness of breath (SoB, 7%) among chemotherapy pts, and pain (19%), SoB (11%) and general weakness (9%) among radiotherapy pts. By applying the CEDIS-based PPED definition, which includes 17 presenting complaints, 50% of ED Only and 68% of Indirect Admission visits were considered PPED. Conclusions: We were able to adapt the PPED algorithm for the Canadian context, which can aid cross-jurisdiction comparisons. We found a substantial proportion of PPED visits. While common presenting complaints had face validity for being PP and a similar proportion of visits were PP compared to the US-based definition, further validation of this approach against healthcare records and in other jurisdictions would be helpful as these metrics become increasingly used for quality improvement.