Hey, I Just Did a Better Operation! Toward an IDEAL Innovation Model

    loading  Checking for direct PDF access through Ovid


To the Editor:
Sagar and colleagues1 present a simple and accountable approach of institutional governance for surgeons wanting to introduce innovative procedures and/or new medical devices. Although the authors are congratulated for articulating an explicit strategy for the introduction of innovative therapy in their own center, they failed to acknowledge the substantial progress made in the science and thinking of surgical innovation over the last few years.
In 2009, the Balliol Collaboration proposed a model for safe surgical innovation called the IDEAL model: Idea (stage 1), Development (stage 2a), Exploration (stage 2b), Assessment (stage 3), and Long-term monitoring (stage 4).2 An additional stage described, although not part of the original IDEAL acronym, is a preclinical stage or stage 0, which may include simulator and/or animal/cadaveric studies. The IDEAL model provides a structured, rigorous yet non-stifling framework for the evolution and evaluation of surgical innovations (http://www.ideal-collaboration.net/).3 It applies to both “development” of a new procedure (stages 0–2a) as well as “dissemination” of an innovation that has been validated elsewhere (stages 2b and beyond)—the latter scenario being the subject of the commentary by Sagar et al.1
The pathway offered by Sagar and co-workers might make the ethics clearance process less formidable and outcomes reporting more transparent at the local level; however, some potential drawbacks of their model are worth discussing. First, although the administrative processes described by the authors are undeniably non-stifling and innovator-friendly, a consequence of their process might be loss of the unique experience gained in both the success and failure of implementing a new procedure or device from the public knowledge repository. For example, adverse events noted by surgeons (at a given institution) during adoption of a surgical innovation/new device may translate into improved practice patterns at the local level but in absence of peer-reviewed reporting, this knowledge will not help clinicians in other centers seeking to adopt the procedure (as in the framework proposed by the authors—there is no provision for outcomes reporting beyond the institutional level). This is in contrast with the recommendations of the IDEAL collaboration which supports public reporting of outcomes.
Second, although patient safety is central to the governance of surgical innovation in their model, they fail to explicitly mention an objective methodology that monitors it during implementation of a new procedure or device. We have previously shown, in accordance with the IDEAL recommendations, that patient safety during a novel technique of robotic kidney transplantation may be monitored in an objective real-time manner using statistical process control (SPC) methods.4 SPC techniques, which are well established for evaluating the learning curve after surgical training,5 can also be used to track surgeon progress when learning an innovative procedure. Therefore, the authors might consider employing such methods for outcome monitoring in a comparative effectiveness framework—the comparison might be with publicly available outcomes from other institutions/groups for the new procedure implemented, or with outcomes of the approach that is considered the standard-of-care. Incorporating such methods may lead to a more objective oversight of the outcomes.
Lastly, surgeons wanting to adopt a new technique but restricted by financial constraints or the might of the administrative burden may not want to pursue a formal study. In that setting, the pathway offered by the authors is an important step forward. However, to decide when a new technique can be considered non-experimental, and, therefore, can be adopted outside of a research framework, we posit that this critical assumption should only be made when comparative outcomes with intermediate to long-term follow-up are available, showing non-inferiority/superiority of the new operation.
    loading  Loading Related Articles