A Systems View of Peer Review

A Framework for Understanding Clinical Peer Review

Since 2005, I've been looking for opportunities to improve this fundamental professional activity through the lens of a systems framework.  In helping clients navigate multiple PDCA cycles, I've been refining a new model for clinical peer review process.  I call it the QI Model to distinguish it from the prevailing QA Model, which has for so long limited the utility of peer review (see QA Model History).

The QI Model seeks to conduct peer review as a quality improvement activity, not only to improve the process itself, but to better align with the overall agenda of the organization to improve clinical performance and patient safety.

I applied this systems view in my work with Evan Benjamin on the first national study of peer review practices ever conducted in the US.  We found that the key predictors of the likelihood that a program has significant ongoing impact on the quality and safety of care are the same as the key elements of the QI Model.  In other words, the more that peer review looks like QI instead of QA, the more effective it is felt to be.

One important outgrowth of the 2007 survey was a 100-point, evidence-based 13-item Clinical Peer Review Program Self-Assessment Inventory. When applied to the original study population, the mean score was 45, suggesting huge opportunity for improvement across the country. No one scored at A-level, which may explain the dearth of examples of real success stories. These findings were validated and extended by the 2009 ACPE Clinical Peer Review Outcomes Study and the Longitudinal Clinical Peer Review Outcomes Study.

The QI Model continues to evolve. I added the key cultural differentiator of Trust based on the 2009 study results. At that time, I also added case identification via Self-Reporting based a recognition of the incongruity between a physician's Hippocratic obligation to "Do No Harm" and the sad reality that physicians are blocked from learning and sharing experiences of the unintended outcomes of care from fear of recrimination. This was later validated in the 2011-12 study. The current freely available version of the Clinical Peer Review Program Self-Assessment Inventory derives from data collected in the 2015-16 update of the Longitudinal Clinical Peer Review Outcomes Study and reported in the International Journal for Quality in Health Care published by Oxford University Press (publication pending). It contains 20 items.

The Federal protections of the Patient Safety and Quality Improvement Act of 2005 now offer a simple solution to this problem. Provider organizations can report peer review program data as Patient Safety Work Product to a listed PSO and thereby protect the information from discovery and protect the reporter from reprisals.

The QA and QI Models for Peer Review

The following table highlights the important differences between the QA and the QI Models for peer review.

QA & QI Models Compared
Aspect QA Model QI Model
Focus Outliers Shift the curve
Identify Substandard care Learning opportunity:
system & individual
Determine Competence Performance
Inputs Single case Multiple cases
Case Finding Generic screens Self-reporting
Method Judgment Performance measurement
Process Variable Standardized
Reliability Low Good
Leverage Point Expert opinion Aggregate data
Orientation Reactive Proactive
Drivers Fear, punishment Trust, fairness, collegiality
Data Capture "Leveling" against care standards Multiple elements of performance
Relation to Hospital QI Process Isolated Highly interdependent
Accountability for Improvement Low High
Governance Laissez faire Attentive
Ultimate Process Outputs Corrective action System improvement
Recognition of clinical excellence
Performance feedback
ROI Low High
Net Effect Culture of Blame Culture of Safety