Rapid response team interventions for severe hyperkalemia: evaluation of a patient safety initiative

This study evaluates outcomes and process measures associated with a rapid response team (RRT) intervention for patients with severe hyperkalemia.

Inpatients on medical-surgical floors (excluding dialysis or comfort care patients) at a 1000-bed tertiary hospital from 2005 to 2009 with severe hyperkalemia (defined as potassium [K(+)] ≥ 6.3 mEq/L).

Retrospective administrative data and medical record review. Hyperkalemia incidence (based both on coding data and laboratory test results) was assessed, as was the association between hyperkalemia and mortality. Independent physician reviewers adjudicated selected cases for death directly attributable to hyperkalemia and potential for preventability with the RRT intervention. All 115 Baylor University Medical Center (Dallas, TX) cases receiving the RRT hyperkalemia intervention over a 12-month period (December 2006-December 2007) underwent in-depth process assessment.

Hyperkalemia occurred as a codable diagnosis in approximately 3.2% of all hospital discharges annually (5-year average of 42 000 discharges), and K(+) values ≥ 6.3 mEq/L were observed in 0.8% to 0.9% of all K(+) assays run by the laboratory in the months sampled. Deaths determined to be directly related to hyperkalemia and potentially preventable were rare, with a total of only 4 events during the study period (3 of these were in the pre-implementation phase), precluding statistical analysis on mortality related to the intervention. The RRT averaged 6 to 10 interventions for hyperkalemia monthly (representing 10% of all inpatient K(+) values ≥ 6.3 mEq/L). Mean initial K(+) level triggering the RRT cascade was 6.7 ± 0.3 mEq/L; average time from floor notification of critical K(+) level to bedside RRT arrival was 14.6 ± 12.1 minutes. Over 24 to 36 hours, K(+) declined 1.7 ± 1.1 mEq/L between patients' initial and final K(+) values (P < 0.001).

Hyperkalemia occurs frequently in inpatient settings. Rapid response team intervention for this condition facilitates timely correction of critical laboratory test results and consistent treatment through use of a standardized protocol. Benefit of the intervention on mortality could not be reliably demonstrated in this study due to event rarity and challenges with case ascertainment. Further research with a prospective, multi-site cluster design using electronic medical records and larger sample sizes could demonstrate which RRT hyperkalemia intervention components warrant widespread adoption.