Introduction: Implantable cardioverter-defibrillator (ICD) and pacemaker (PM) advisories may have a significant impact on patient management. Surveys of clinical practice have shown a great deal of variability in patient management after a device advisory. We compared our management of consecutive patients in a single large university practice with device advisories to the "best" patient management strategy predicted by a decision analysis model.
Methods: We performed a retrospective review of all patients who had implanted devices affected by an advisory at our medical center between March 2005 and May 2006 and compared our actual patient management strategy with that subsequently predicted by a decision analysis model.
Results: Over 14 months, 11 advisories from three different manufacturers affected 436 patients. Twelve patients (2.8%) were deceased and 39 patients (8.9%) were followed at outside facilities. Management of the 385 remaining patients varied based on type of malfunction or potential malfunction, manufacturer recommendations, device dependency, and patient or physician preferences. Management consisted of the following: 57 device replacements (15.2%), 44 devices reprogrammed or magnets issued (11.7%), and 268 patients underwent more frequent follow-up (71.3%). No major complications, related to device malfunction or device replacement, occurred among any patient affected with a device advisory. Concordance between the decision analysis model and our management strategy occurred in 57.1% of cases and 25 devices were replaced when it was not the preferred treatment strategy predicted by the decision model (43.9%, 37.3% when excluding devices replaced based on patient preference). The decision analysis favored replacement for all patients with PM dependency, but only for four patients with ICDs for secondary prevention. No devices were left implanted that the decision analysis model predicted should have been replaced.
Conclusions: We found that despite a fairly conservative device replacement strategy for advisories, we still replaced more devices when it was not the preferred device management strategy predicted by a decision analysis model. This study demonstrates that even when risks and benefits are being considered by experienced clinicians, a formal decision analysis can help to develop a systematic evidence based approach and potentially avoid unnecessary procedures.