Image by Shapelined

Quality Improvement Tool Kit For Congenital Cardiologists

Be Ready For The Transition To A Value-Based Healthcare Environment

Michael A. Rebolledo, MD, MBA, MPH Gina-Lynne C. Guasco, MT (ASCP) H. Jane Hanafin, MHA

The field of Quality Improvement (QI) involves devising and tracking the impact of targeted interventions designed to improve healthcare services.[1] Occasionally, the term performance improvement is used interchangeably with QI; however, this term is used more frequently in managerial or administrative systems.[2] Over the past two to three decades, the QI field has evolved through four major stages.[3]

FIGURE 1 Ishikawa Cause-and-Effect Diagram[12]

In the first stage, passive diffusion, there was an assumption that clinicians would take actionable information directly from the latest clinical research. In the second stage, there was the publication of guidelines and systematic reviews to effect behavior change among clinicians. In reality, it has been demonstrated that adults receive only about half the amount of recommended care.[4] There are likely several barriers to implementation, but because medicine is still perceived as an art (vs. a science), many clinicians continue to practice with limited reference to guidelines. In the third stage, there was the introduction of a more proactive-style total quality management from well-established industries. Several common QI methodologies were popularized during this stage including plan-do-study-act (PDSA) cycles, Lean and Six-Sigma.[5] Stage four focuses on systems re-engineering to design safer and more effective healthcare delivery systems, e.g., electronic health records or computerized physician order entry. There continues to be further QI evolution, which has taken lessons from other industries to develop high-reliability organizations, e.g., aviation.[6] Clinical decision support modules fall in the realm of systems re-engineering. The debut of the Watson Supercomputer by IBM, Inc., which processes structured and unstructured data fields using natural language processing, is an example of a large-scale artificial intelligence clinical decision support.[7]

Quality in healthcare is often described using Donabedian’s conceptual model: structure, process, and outcome.[8] Structure refers to the attributes of the e.g., Beta-blocker use post-myocardial infarction. However, process measures must be reliable, valid, lack systematic bias, and most importantly, linked to outcomes. Validating process measures is a lengthy systematic exercise that is very resource intensive. Outcomes, i.e., what happened to the patient, represent clinical or patient-reported outcomes. Outcomes analysis requires robust risk-adjustment because of potential con-founders, e.g., commonly used 30-day readmission rate. The context may affect all three components of this conceptual model.

To read the full article, please go to the February 2020 Issue of CCT, where it was originally published.

1 view0 comments