Every year, approximately 1.3 million emergency department visits in the U.S. are related to adverse drug events (ADEs).
Patients on multiple medications for comorbidities such as heart disease, COPD and diabetes, who often have outdated medication lists in their EHRs, are at high risk for ADEs. While pharmacy-led interventions can reduce the risk of ADEs, they are also resource-intensive. Manually prioritizing the patients who would benefit most from these interventions is often challenging and time-consuming.
Parkland Health, often operating at or near full capacity, knew that many of their ADEs were preventable. To address the challenge, the hospital converted their manual prioritization process to target high-risk patients for pharmacy-led interventions to a real-time predictive model. The process incorporated risk factors such as medications, disease complexity, prior healthcare utilization, demographics and social determinants of health data.
Partnering with a data scientist from the Parkland Center for Clinical Innovation, an interdisciplinary working group of clinicians, IT professionals and pharmacy professionals conducted a retrospective analysis of hospital admissions and aligned the results with literature and industry knowledge to formulate a Patient at Risk for Adverse Drug Events (PARADE) predictive scoring model.
To support efficient use of the PARADE score, the team integrated it into the clinical workspace. End-user workflows include a dashboard, a detailed score report and a column to sort incoming patients by risk level. By embedding the score into their clinical workflow, the pharmacy team could prioritize interventions for patients most at risk for ADEs. To aid in reporting, and for ongoing model evaluation and fine-tuning, pharmacy users also document the outcomes of their detailed medication reconciliation by including the count of individual interventions and whether an ADE was prevented.
As a result of the PARADE project, Parkland was able to consistently screen incoming patients for their risk of an ADE in real time and manage their workloads to prioritize patients who were high-risk. After two years, the PARADE process has screened more than 87,000 patients, identifying more than 8,000 high-risk patients—16% of whom received pharmacy interventions, preventing more than 2,000 ADEs.
In an analysis done on 5,948 high-risk patients with high PARADE scores, the 954 patients who received a pharmacy-led intervention were deemed less likely to be readmitted, with a 12% readmission rate, versus the 16% readmission rate of the nearly 5,000 patients who didn’t receive a pharmacy consult. These results translate to a potential savings of over $17 million through reduced readmissions and eliminated ADEs.
HIMSS is pleased to recognize Parkland Health for their HIMSS Adoption Model for Analytics Maturity (AMAM) Stage 7 validation. “Parkland Health is a model story of building analytics from the ground up, creating a trusting and supportive environment that embraced their ‘analytics first’ mantra,” said James E. Gaston, MBA FHIMSS, senior director, analytics, HIMSS. “Projects start with users asking, ‘What data do we need?’ Everyone from nurses to the chief MD appreciate the way transformed data informs their work, helping them understand their patient‘s needs and optimize care.”
“The use of data we demonstrated during the validation is a direct reflection of the way everyone at Parkland cares for patients,” said Fred Cerise, MD, MPH, president and chief executive officer, Parkland Health & Hospital System.
“While we are honored to receive this designation, the true representation of this recognition lies in the many innovative ways that Parkland’s use of data improves the health of our patient population. The HIMSS AMAM Stage 7 validation is not an end, but rather a launching point to advance our dedication to the health and well-being of the individuals and communities entrusted to our care.”
Discover how other healthcare organizations are innovating for improvement in more Stage 7 success stories: