This editorial refers to ‘Cost-of-illness study of patients subjected to cardiac rhythm management devices implantation: results from a single tertiary centre’ by J. Fanourgiakis et al., 15: 366–375.
Costs worthy of consideration often do not make it to the picture in cost-of-illness studies. Thus, there is cause for concern when Fanourgiakis et al.'s1 timely analysis of costs associated with implantable cardioverter-defibrillator (ICD) uptake in Greece estimate social cost only as impact on health resources and lost productivity. Cost-of-illness studies have been critiqued for misevaluation of cost from the patients' perspective, yet beyond this, they fail to capture non-economic costs to society, which however may be a by-product of the economic system. In the USA, for example, rapid adoption of a new technology also may have costs for the evidence base, an outcome that results from industry-driven device regulation.
The US regulatory structure struggles to balance innovation with public safety, a practice that, as the history of ICD adoption in the USA shows, has given rise to controversy within and beyond the medical community. The Food and Drug Administration (FDA) have attempted to address this tension via the FDA Modernization Acts I and II, which seek to balance public safety against costs of designing and marketing devices by ‘pursu[ing] the least burdensome means’2 to certify safety and efficacy. Yet, customer protection is actually undermined by this drive to market new devices quickly,3 because pre-market trials tend to be less rigorous, have smaller samples, and contain lower risk patients that are followed up for shorter periods of time.2,4 In the urgency to market new technologies large robust randomized trials generally follow rather than precede FDA approval because study design details affect how quickly a device comes to the market.2 Consequently, market-driven device regulation undermines evidence-based healthcare because rigorous evidence follows—rather than proceeds—federal safety certification. For example, findings from the recently published EXAMINATION5 study showed drug-eluting stents (DESs) to be equally safe compared with bare-metal stents, yet this knowledge is useful mainly for patients with a DES. Following their rapid uptake, clinical use declined over questions about safety, and DESs are now obsolete thanks to newer technologies.
The timing of these findings in relation to clinical use is not accidental, but actually a consequence of regulatory structures. The Center for Devices and Radiological Health (CDRH) regulates the marketing of medical devices. Devices are classified by potential risk to determine the required level of regulatory testing, so that less risky technologies can be marketed more quickly. Devices rated as Class III, i.e. life sustaining or supporting, like the DES, are subject to pre-market analysis (PMA) including a clinical trial, yet for every Class III device that undergoes PMA, ∼60 are marketed under the 510(k) exemption for new devices that are substantially equivalent to a device already on the market.4 Many devices, therefore, bear the FDA label despite limited data on safety and effectiveness.6
Center for Devices and Radiological Health post-marketing surveillance provides a mechanism to detect problems that emerge after approval. Yet, this system relies on clinicians or manufacturers—who may have conflicting obligations to shareholders and patients3—to identify and report problems.3,6,7 The FDA responds to reported adverse events by issuing a safety alert or, more rarely, a market recall.4,7 Yet, poor dissemination means clinicians are often unaware of problems with particular devices.4 Since 1990, there have been 130 device recalls affecting 900 000 major cardiac devices.4 In 10 cases (affecting 25 000 patients), the malfunction was reasonably likely to cause serious health consequences or death.4 Moreover, it is difficult to assign risk values to adverse events because they are underreported and the total number of patients who have a device is unknown.4 Therefore, significant risks often do not emerge until after clinical use—a particularly perilous situation since the US Supreme Court has ruled that (Reigel vs. Medtronic) patients may not litigate manufacturers for FDA-approved devices.
Can the market itself be regulatory? Perhaps, if companies that market unsafe devices lose credibility and customers. However, in fact, recalls have little influence on share prices—although this varies somewhat by company characteristics.8 Ironically, facets of existing regulation are responsible: the FDA label, and protection from personal injury and liability suits, bolster shareholder confidence.8
Pro-industry regulations, rapid technology uptake, and underreported adverse events make risks unclear and emergent. Consequently, ‘evidence’ is one step behind clinical practice. Pro-industry regulatory processes and a disempowered FDA tend to produce evidence of excessively limited generalizability to clinical populations and indirectly exposes patients to unnecessary risk of death. Most clinicians expect evidence to be partial and contextually particular, and rely on some combination of evidence and experience yet regulatory structures which result in stronger and more generalizable evidence are needed.
A.A.C, corresponding author, contributed to initial conceptualization, literature search, and writing (of first and all drafts) of this comment. A.M.C. contributed to refining the conceptualization and writing (second and following drafts) of this comment.
Conflict of interest: none declared.