In a recently release issue of Pediatrics, Dr. Kelly Bergmann and colleagues present a Quality Improvement (QI) initiative (10.1542/peds.2019-1104) to reduce hospitalization rates for youth and young adults <21 years old with Type 1 diabetes mellitus (T1DM) presenting in “low risk” diabetic ketoacidosis (DKA). When I first read the article title, I have to confess I had never considered the possibility that DKA could be considered “low risk” in any way. The authors have convinced me, however, with this thoughtful QI initiative, that “low risk” DKA is a defined entity that has potential for outpatient intensive management.
What is “low risk” DKA? The authors created a working definition that incorporates both biochemical and social factors. “Mild DKA” was previously defined in the International Society for Pediatric and Adolescent Diabetes Guidelines as a pH of 7.2-7.3, a bicarbonate level of 10-15 mmol/L, a glucose >200mg/dL and ketonemia or ketonuria. Since clearly factors beyond the child’s laboratory values must go into clinical decision making, the authors added the following key factors to “mild DKA” to further define “low risk DKA”: established diagnosis of T1DM, good knowledge of sick day management and ability to do home care, and no concerning social conditions or comorbidities that would impede discharge. This pragmatic definition identified a group of patients for the QI intervention.
This QI project was initiated in April 2016, and compared a pre-period (January 1, 2012 to project start) to a post-period ending December 31, 2018; 165 patients with low risk DKA (of a total of 3,132 T1DM Emergency Room visits) were included. The authors’ Key Driver Diagram maps out their SMART (Specific, Measurable, Applicable, Realistic, and Timely) Aim, their Primary Key Drivers (High Quality Evidence, Shared Awareness and Provider Compliance) and their Secondary Drivers or Change Strategies. The authors compared the baseline pre-implementation hospitalization rate for their patients of interest with quarterly post- implementation hospitalization rates. (Due to low numbers each quarter, the analysis actually looked at groups of 9 patients – this is explained and shown in a statistical process control chart.) With the primary outcome of hospitalization rate for low risk DKA, the authors selected return to an Emergency Room (ER) within 3 days as their balancing measure – this is an endpoint that serves as a “warning signal” to alert the QI team that the primary outcome could otherwise be impacting safety or care (which it did not here). Overall the project led to a reduction in the rate of hospitalization (74% [95% CI 64-82%] to 55% [95% CI 42%–67%]), but the fun is in how the QI work got there.
Perhaps the most impressive features of this QI initiative for me were the inclusiveness and persistence of the QI team in implementing their Secondary Drivers/Change Strategies. For example, the “Guideline Team” was multi-disciplinary, and included nurses, doctors from both Endocrine and Emergency specialties, pharmacists, a family representative from the Family Advisory Council, a statistician and health economist. The Team solicited and incorporated stakeholder feedback to create their intervention Guideline, and worked with Information Technology to create a new single “ED DKA” order set that included appropriate options for low, medium and high risk patients with DKA. Multiple nursing education sessions and quarterly physician training sessions and updates were conducted. The authors frankly discuss challenges they faced, including physician and nursing concerns about prolonged ER stays and potential inadequate resolution of acidosis, and generalized hesitancy.
This article is a beautiful example of QI work, and congratulations to the authors and team for a job well done.