Regulatory approval of new medical devices, cross-sectional study.

Introduction

The introduction of new medical devices is fundamental to the advancement of healthcare. Historically, such devices have been adopted with little scientific evidence to support their use. Although many devices have greatly improved clinical outcomes, not all are beneficial and some may be harmful. To this end, most jurisdictions have developed regulatory bodies, such as the US Food and Drug Administration, that ensure the safety and effectiveness of new medical devices. These regulatory bodies must also act in an efficient and timely manner such that patients are not deprived of beneficial innovations.

The process by which new high-risk medical devices find their way from bench to bedside is well established: the development of the device resulting in a first-in-human study; the evaluation of the device in clinical trials, culminating in regulatory approval for use; and the adoption of the device. Although high-risk devices warrant considerable scientific evidence for their safety and effectiveness before regulatory approval, the pathway for lower-risk devices is less stringent, allowing for their more rapid approval.

We investigated the use of these distinct regulatory approval pathways for new medical devices.

Methods

We performed a cross-sectional study of new medical devices reported in the literature to determine whether they received regulatory approval and the relative contributions of academia and industry in this process. Before searching for evidence of regulatory approval, we identified clinical studies of devices, allowing us to capture those devices that failed to receive approval.

We defined a medical device according to the FDA definition as an “instrument, apparatus, implement, machine, contrivance, implant, in vitro reagent, or other similar or related article.” If there was no evidence of a previous clinical study of a device in the literature, we considered the device as new.

For each article reporting a clinical study of a new medical device, we defined academia and industry as being involved with the development of the device if a relation was described. If an entry could be found on the FDA medical device databases, we considered a device as having regulatory approval.

Search Strategy

The PubMed database was searched using the Boolean term: (device OR instrument OR apparatus OR implant OR “in vitro reagent” OR system) AND (“first in man” OR “first in human” OR “first experience” OR “first clinical” OR “early clinical” OR “early experience” OR “early human” OR “initial experience” OR “initial clinical” OR “initial human” OR “preliminary clinical” OR “preliminary experience” OR “preliminary human” OR “Phase 1” OR “Phase I”). We selected this search term owing to efficiency and being able to identify the most relevant studies. The search was carried out between 1 January 2000 and 31 December 2004 to allow time for regulatory approval, as previous studies have suggested a long lag between the development of a device and subsequent regulatory approval.

We included articles that reported a clinical study of a new medical device and excluded those that only reported a laboratory study of a device because few such devices ultimately result in a clinical study.9 We also excluded articles if they reported on the novel use of an existing device, as we expected that most such devices would already have received regulatory approval.

Based on a pilot study, we estimated (between 1 January 2000 and 31 July 2000) that this search strategy would select sufficient articles to allow for meaningful analysis.

Two researchers initially screened titles and abstracts to identify relevant articles (HJM and CJP, checked by AHH and APM). We excluded articles if the title or abstract explicitly stated that the article was not original research, related to drug development, related to an existing medical device, or a laboratory study. Full articles were subsequently obtained and further assessed for eligibility. In each instance, we reviewed the reference list and searched the PubMed database using the device name to ensure that we did not miss a related previous clinical study (that would result in their exclusion). Discrepancies were resolved by consensus.

Medical Devices

For each clinical study of a new medical device, we determined the type of device, the target specialty, and the involvement of academia and industry (HJM and CJP, checked by AHH and APM). The types of devices were based on the FDA definition, and the target specialties were drawn from the FDA databases. We considered academia and industry to be involved in the development of a device if a relevant author affiliation, financial support, or provision of technology was described in the author affiliations, main text, or acknowledgments of the article. Discrepancies were resolved by consensus.

Regulatory Approvals

For each new medical device, we searched the FDA databases for relevant regulatory clearance or approval. The FDA recognizes several types of regulatory pathways depending on the nature of the device. Premarket notification (510(k)) is the regulatory pathway if the device is “substantially equivalent” to a predicate device and does not necessarily require clinical data. Premarket approval is the regulatory pathway if the device is “not substantially equivalent,” and requires reasonable evidence of safety and effectiveness. Other regulatory pathways include humanitarian device exemption if the device is for use in patients with rare diseases or conditions. We searched the FDA 510(k), premarket approval, and humanitarian device exemption databases using the device name, applicant name, and relevant keywords (HJM and CJP, checked by AHH and APM). We also searched Google for devices that may have been discontinued, withdrawn, or recalled. Search results were not limited to a date range, allowing for the identification of regulatory clearance or approval before the first published clinical study. All the searches were performed in August 2015, allowing a minimum of 10 years from publication to regulatory clearance or approval. Discrepancies were resolved by consensus.

Statistical Analysis

To compare differences in regulatory clearance or approval between the following groups we used the χ2 test: devices developed by industry alone versus academia alone; devices developed by both industry and academia versus academia alone; and devices developed by both industry and academia versus industry alone. Firstly, we compared the proportion of devices receiving any regulatory clearance or approval (versus no clearance or approval). Secondly, we compared the proportion of devices receiving 510(k) clearance (versus any other approval). We considered differences to be statistically significant if P was less than 0.05. All statistical analyses were performed using SPSS 22.0 (IBM, NY, USA).

Patient Involvement

No patients were involved in setting the research question or the outcome measures, nor were they involved in developing plans for the design or implementation of the study. No patients were asked to advise on interpretation or writing up of results. There are no plans to disseminate the results of the research to study participants or the relevant patient community.

Conclusion

The optimal framework for the regulatory approval of medical innovations remains unclear. This study suggests that many new devices do receive regulatory approval but often lack clinical trial data supporting their safety and effectiveness.

The IDEAL model makes several proposals for the staged introduction of innovations in surgery (and other disciplines that offer complex interventions), including randomized controlled trials to assess safety and effectiveness. At present, few relevant randomized controlled trials are published, and fewer still meet current quality standards for optimal reporting. Changes in the regulatory approval of devices that would require trials for proof of safety and effectiveness might promote adherence to the IDEAL model.