Near misses and research subjects

Short abstract

Near miss reporting used in clinical care should be extended to include clinical research

As in clinical care, the highest priority of clinical research is to protect participants from any undue harm. There are inherent risks involved in any human experiment, but careful analysis of several research related fatalities has clearly shown that human error and system failures contributed to these events.¹,² While clinical medicine has embraced organizational based approaches to patient safety, much of the clinical research safety process remains narrowly focused. This focus has resulted from the traditional “protocol‐centric” approach for managing clinical research risk. Both Institutional Review Boards and Data Safety Monitoring Boards predominately review safety problems on a study by study basis. While these traditional approaches have been remarkably successful, their narrow scope may limit their ability to identify and manage clinical research risks comprehensibly.

Clinical research and clinical care share many similar risks. Both involve analogous procedures, many of which are invasive, and both use pharmacological agents with known and unknown toxicities. Patient safety research has identified system failures as important contributors to adverse drug events, nosocomial infections, and procedural mishaps. These same latent failures may also contribute towards adverse events within clinical research. For instance, environmental factors such as the lack of conveniently placed antiseptic solutions can contribute to poor sterile technique and nosocomial infections.³ Hospital infection surveillance programs seek to identify these environmental factors by reviewing rates of events within the context of the expected rates of infections across an organization. In the “protocol‐centric” model for safety monitoring, the organizational context in which an adverse event occurs becomes lost. Thus, in a single clinical trial an intravenous line infection would be treated as an anticipated adverse event and managed by including this risk within the informed consent document. If an unexpectedly high rate of intravenous line infections were occurring, this pattern would be unlikely to be identified unless all of these events occurred within a single protocol. This amounts to a hospital infection control program following a single employee to determine their rates of nosocomial infections. While this approach may identify problematic individuals, it is unlikely to contribute to overall patient safety.

In order to correct latent failures, organizations must actively seek to identify them. One proposed method to identify these system flaws within clinical medicine is through “near miss” reporting.⁴ A near miss is an error that does not result in any harm. Because these errors often result from the same system failures which can also lead to patient injuries, routing out these latent failures before any harm occurs is an appealing strategy. In addition, because no injury is associated with the error, it is believed that individuals will be more inclined to report near misses because they are less inclined to fear any disciplinary action. Several non‐medical industries have incorporated near miss reporting systems into their safety plans. Within clinical care, transfusion medicine has been a pioneer in this respect.

But near miss reporting should not be limited to clinical care. It may also be able to identify potential latent system failures which may threaten the safety of research volunteers. As such, we have been developing a near miss reporting system for use within a General Clinical Research Center (GCRC).⁵ GCRCs are research centers funded through the National Institute of Health in the USA which conduct clinical research on human participants. The primary targeted users of this near miss reporting system were the GCRC staff nurses. GCRCs employ staff nurses who are typically involved in multiple study protocols involving different human subjects simultaneously. The system allows for both web based and paper based reporting, and reports are anonymous. Some examples of reported research related near misses include medication errors resulting from lack of access to study protocol documents and delays in obtaining appropriate equipment due to a lack of standardization regarding equipment placement. Once identified, as in clinical care, standard operating protocols can be devised and these failures addressed.

Near miss reporting could have an important role in any comprehensive human subject protection program. Although any stakeholder in clinical research would be encouraged to participate, near miss reporting systems would particularly benefit from the input of individuals who work across multiple study protocols. Research nurses and pharmacists would represent an ideal source.

Unfortunately, the success of reporting systems can be greatly hindered by the organizational culture. In clinical medicine several barriers have been described which may hinder reporting, such as fear of reprisals or concerns about going against the “authority gradient”. Whether and how organizational culture might impact safety in clinical research remains poorly evaluated. In a survey we conducted of over 400 GCRC staff nurses we found that 11% reported being made to feel uncomfortable for reporting a protocol violation and 12% suggested they would get into trouble if they refused to carry out a protocol because they thought a subject did not fully understand the study.⁶ This suggests that reporting barriers similar to those reported within clinical medicine may exist within clinical research.

Clinical research shares many of the same risks to patient safety as clinical care. Clinical care has begun to adopt systems based approaches to manage these risks, while clinical research has remained more narrowly focused, managing safety issues on a study by study basis. While this method does successfully address many of the research related risks a study volunteer may experience, we would argue that it is not sufficiently comprehensive to identify and manage all of the potential risks to research subjects. System failures can contribute to research related injuries and must be managed from an organizational perspective. To help facilitate the reporting of events, we propose extending the near miss reporting model to clinical research. Just as in clinical medicine, similar barriers exist which provide a disincentive to report errors and will have to be mitigated. It is time that the same “culture of safety” emphasized in clinical care be brought into clinical research.