Medication Safety Event Reporting: Factors that Contribute to Safety Events During Times of Organizational Stress

Abstract

Background:

Incident reports submitted during times of organizational stress may reveal unique insights.

Purpose:

To understand the insights conveyed in hospital incident reports about how work-system factors affected medication safety during a COVID-19 surge.

Methods:

We randomly selected 100 medication safety incident reports from an academic medical center (December 2020 – January 2021), identified near misses and errors, and classified contributing work-system factors using the Human Factors Analysis and Classification System-Healthcare.

Results:

Among 35 near misses/errors, incident reports described contributing factors (mean 1.3/report) involving skill-based errors (20), communication (8), and tools/technology (4). Reporters linked 7 events to COVID-19.

Conclusions:

Skill-based errors were the most common contributing factors for medication safety events during a COVID-19 surge. Reporters rarely deemed events to be related to COVID-19, despite the tremendous strain of the surge on nurses. Future efforts to improve the utility of incident reports should emphasize the importance of describing work-system factors.

INTRODUCTION

Medication errors cost the health care industry 42 billion dollars annually and are a leading cause of avoidable harm in health systems worldwide. In 2017, in response to this problem, the World Health Organization launched a global initiative to reduce medication-related errors by 50% in 5 years.¹ Despite global improvement efforts, advances in research and technology, and policy changes, medication errors continue to pose a threat to patients in the acute care setting.²

A medication error can be defined as “a failure in the treatment process that leads to, or has the potential to lead to, harm to the patient.”^3(p1013) Nurses play a key role in identifying, reporting, and preventing medication errors because of their role in administering medications and monitoring patients’ responses.⁴ Most hospitals in the United States use incident reporting systems as tools for improving medication safety; however, incident reporting may not always provide actionable insights into conditions that contribute to errors.^5,6

Patient safety experts generally believe that medication errors occur because challenges in the socio-technical work-system (interacting components, humans, organizations, technologies and tools, environments, and tasks) lead to problems with processes that influence outcomes of care.⁷ Examples of work-system challenges include inadequate attention to medication safety procedures, lack of communication and collaboration, lack of training, unclear information, fatigue, miscalculations, environmental challenges, and organizational factors.⁸

Due to disruptions in the socio-technical work-system, medication errors may have become more common during the COVID-19 pandemic.⁹ The pandemic has heightened demands on nurses because of changes in care delivery models (eg, move from primary nursing to team nursing,¹⁰ burdensome patient isolation procedures), increasing work intensity due to heightened disease severity and staffing shortages, and mental and physical fatigue.¹¹ Frontline nurses have also faced unprecedented hazards, such as frequent contact with patients with COVID-19, stigma, and even workplace violence.¹² In 2021, the Institute for Safe Medication Practices revealed that one of the top medication hazards of the year was related to the pandemic: positioning infusion pumps outside of COVID-19 patients’ rooms.¹³ The purpose of this study was to investigate insights conveyed in hospital incident reports about how work-system factors affected medication safety on inpatient nursing units during an unprecedented surge in hospital volume due to COVID-19.

METHODS

Setting and population

The setting was a quaternary referral center in Los Angeles, California, between December 1, 2020, and January 31, 2021, when the hospital experienced the largest surge in COVID-19 patient volume. At the peak, 372/850 (44%) of hospital inpatients were being treated for confirmed or suspected COVID-19. Our analysis included incident reports filed by clinicians on medication errors related to “medication/fluid” and “adverse drug reactions” about any patient admitted to an inpatient nursing unit or seen in the emergency department. The Institutional Human Subjects’ Protection Committee approved this research, and informed consent was not required.

Data sources

The incident reporting system is used for documenting near misses, good catches, and safety events that affect patients, staff, and visitors. Reporters complete an online submission form via the hospital’s electronic health record or intranet (Supplemental Digital Content Appendix). Reporters answer structured questions about each event (eg, medication and route of administration, medication administration stage in which the event originated, event relationship to COVID-19), provide a narrative description of the event, and assess severity according to the National Coordinating Council for Medication Error Reporting and Prevention (NCC MERP).¹⁴ Education on incident reporting at our hospital included a 10 minute online module with information on how to report unsafe conditions, near misses, and adverse events. The module focused on educating staff to include details on timing and location of an event; general information about the event; details about the person affected; a listing of witnesses and others involved; and who was notified. Nurses receive education on incident reporting during orientation both as new graduates and as new nurses to the institution. Once they are assigned units, the use of the reporting system is reinforced as part of the unit orientation checklist. House staff receive education during orientation, and attendings do not receive any formal training; however, they do receive education through various department meetings and in Medical Staff communications. Nurses typically file the majority of incident reports related to medication safety. Our team queried the incident reporting system database for medication safety event reports, and we randomly selected 100 for this analysis.

Classification framework

The research team reviewed each incident report to identify contributing factors, or specific conditions leading to an event. Each contributing factor was then classified using the Human Factors Analysis and Classification System for Healthcare (HFACS-Healthcare).⁶ HFACS-Healthcare is an adaptation of Shappell and Wiegmann’s HFACS, which is based upon James Reason’s Swiss Cheese model of Accident Causation.¹⁵ HFACS-Healthcare consists of 4 tiers of system failures: unsafe acts, preconditions for unsafe acts, unsafe supervision, and organizational influences. Across the 4 tiers, there are 21 contributing factor categories (Supplemental Digital Content Table 1).

Classification process

Two reviewers read and discussed each incident report, including the narrative descriptions of the event as well as structured information submitted by the reporter. One reviewer was a researcher with a PhD in human factors and certified HFACS professional (TC); the other was a physician researcher with expertise in patient safety and training in HFACS (CB). The reviewers categorized each event type (near miss: unsafe patient exposure to a hazardous situation that does not result in harm; error: any act of commission or omission that exposes patient to a potentially hazardous condition, or neither)¹⁶ and identified and classified contributing factors suggested by each narrative (Supplemental Digital Content Table 2). Reviewers further classified contributing factors into sub-categories (Supplemental Digital Content Table 3).¹⁷ Finally, reviewers extracted structured data from event reports, including whether the reporter deemed it related to COVID-19 (yes/no). Reviewers discussed event report content until consensus was achieved. Events that were neither near misses nor errors were not included for analysis.

RESULTS

The incident reporting database contained 211 medication safety events submitted during the 2-month study period. Of the 100 randomly selected reports, 65 were adverse drug events but not near misses or errors (eg, allergic reactions that could not have been anticipated; intravenous contrast extravasation), so contributing factors were not relevant. Among the 35 reports on near misses or errors, we identified 46 contributing factors (mean: 1.3, range 1–4). The most common contributing factors included skill-based errors (ie, ‘doing’ errors that involve highly practiced behavior that occurs with little or no conscious thought; 20/46 = 43.5%), communication challenges (8/46 = 17.4%), poorly designed tools/technology (4/46 = 8.7%), decision errors (ie, ‘thinking’ errors that represent conscious, goal-intended behavior that proceeds as planned but is inadequate for the situation; 4/46 = 8.7%), and problems with coordination (4/46 = 8.7%). (Supplemental Digital Content Table 2).

Skill-based errors primarily consisted of forgotten actions (6/20 = 30%) and administering medication to the wrong patient (3/20 = 15%). Communication challenges consisted primarily of failures to warn/disclose/handoff critical information (5/8 = 63.5%). Tools/technology challenges involved unclear processes in the electronic health record (2/4 = 50%) and electronic health record design features that did not promote safety (eg, alert disappearing) (2/4 = 50%). Decision errors involved not following a process (1/4 = 25%), incorrect selection of medication based on provided information (1/4 = 25%), misinterpretation of information (1/4 = 25%), and overriding/ignoring of warnings (1/4 = 25%). Coordination problems involved poor coordination across different work teams (4/4 = 100%) (Supplemental Digital Content Table 3). Overall, 17.1% (6/35) incident reports related to near misses or errors had zero identifiable contributing factors. Reasons for the inability to discern contributing factors included use of the passive voice (eg, “ordered doses placed does not reconcile number of doses administered”) and a lack of specificity about tasks, tools, or personnel (eg, “device dislodged from patient”). Reporters rated the NCC MERP level of harm as “Temporary harm, Intervention needed” for 8.6% (3/35) of reports; “No Harm – Event Reached Patient, Additional Monitoring Required” for 5.7% (2/35); “No harm – Event Reached Patient, No Monitoring Required” for 60.0% (21/35); and “No harm – event did not reach patient” for 12.8% (8/35) of events.

Reporters deemed 7 of 35 near misses or errors as related to COVID-19. For these events, there were 7 contributing factors: skill-based errors (4/7), resource management (1/7), inadequate supervision (1/7), and coordination (1/7). Supplemental Digital Content Table 4 includes paraphrased report narratives. Reports involving COVID-19 included an average of 1.0 contributing factor (range: 0–2), while non-COVID-19 reports included an average of 1.3 contributing factors per event.

DISCUSSION

The event reports reviewed for this study provided valuable insight into the function of the clinical work-system, during a time of unprecedented stress, workload, and fatigue. The most frequently identified work-system factors involved skill-based errors and communication breakdowns. Given that medication administration includes multiple skill-based tasks (routine and highly practiced activities including reviewing patient orders; obtaining the medication; double-checking the patient’s name, medication name, medication dosage, medication route, and frequency prior to administration; scanning barcodes; administering the medication; and monitoring the patient’s clinical condition),^15,18 it is reasonable to observe skill-based errors (eg, forgetting to administer a dose of a medication, selecting the incorrect medication or dose, failing to check or confirm information). These types of errors are more likely to occur when individuals are experiencing elevated levels of stress, which can impact cognitive processes involving decision making, working memory, and divided attention.^19,20 Interruptions and distractions (eg, patient calls, family questions, staff entering or speaking during administration), which serve to divert much needed attention, are often more common during stressful conditions.²¹ In our study, skill-based errors consisted mainly of omissions. Nurses working in hospitals consistently encounter work-related stressors that may impact their performance, including conflict with other staff, lack of support, high workload, and uncertainty.^22,23 Given that a nurse’s job is challenging on a typical workday, it is likely that stressors occurred more commonly and contributed to more patient safety events during COVID-19 surges.⁹

In our sample, communication breakdowns were also common. Effective communication in health care is challenging, given that it is influenced by technology, people, and processes. Most communication breakdowns resulted from neglecting to warn or disclose critical information to other team members. Information sharing may be impeded by factors like workload, staffing, and trust.²⁴ These challenges are likely to be further exacerbated during the pandemic, since staff absences and inadequately trained temporary workers have been frequently cited as work-related stressors during the pandemic, due to increasing workload and time pressure among team members providing care.²⁵ Moreover, attempts to restructure staffing during surges may lead to increased team unfamiliarity and role ambiguity,²⁵ making information sharing more complex.

How work-system factors commonly contribute to medication errors during a COVID-19 surge “stress test” may help health care organizations identify systemic challenges and intervene before serious adverse events occur. For example, interventions can be put in place to mitigate specific work-system challenges like skill-based errors and communication breakdowns. Skill-based errors can be reduced by identifying and limiting distractions by implementing ‘no interruption zones,’ education on distractions and their impacts, and encouraging patients, relatives and other staff to wait until nurses have completed medication administration before approaching.²¹ Moreover, institutions can also provide emotional skills training to staff involving strategies to mitigate stress.^26,27 Communication breakdowns, especially those involving a lack of information sharing among team members can be improved by implementing verbal communication protocols (eg, Illness, Severity, Patient Summary, Action List, Situation Awareness and Contingency Planning, and Synthesis by Receiver [I-PASS])²⁸ and/or team training programs like the Team Strategies and Tools to Enhance Performance and Patient Safety (TeamSTEPPS)²⁹ program.³⁰

Interestingly, during a time of unprecedented stress on nurses at our institution, medication safety event reporters rarely described working conditions that contributed to events. This pattern persisted across all event types including those related to COVID-19. Given that situational factors likely led nurses to experience extraordinary levels of fatigue and emotional exhaustion while they were also asked to take on clinical activities that may have been unfamiliar, we expected to encounter reports describing work-system circumstances that contributed to near misses or errors. However, our findings did not affirm this expectation. Instead, we found that contextual factors were rarely reported, suggesting there is an opportunity to improve our incident reporting system and the education that is provided to users about describing work-system challenges.

Advances in patient safety may proceed more slowly if work-system factors are not identifiable in incident reports, the primary source of information that hospital and nursing leaders rely on to conceive strategies to improve patient safety. The pandemic introduced unique circumstances that likely influenced reporting practices. It is likely that additional workload caused nurses to have limited time to submit incident reports while in the workplace. Also, nurses may have deprioritized tasks perceived as non-urgent (eg, filing incident reports), and were less adherent to following strict protocols to accommodate changing needs on their units. Moreover, our research team believes that at least 3 challenges associated with incident reporting systems may contribute to underreporting of work-system factors in our study: emotional burdens perceived by event reporters,³¹ lack of education and skill on best practices for incident reporting,³² and fear of adverse consequences due to incident reporting or disclosure.³²

A recent review article called attention to nurse experiences as second victims of medication errors and identified 4 findings: 1) errors may contribute to lasting emotional burdens; 2) the type of support received following an error impact one’s perceptions of the error; 3) a nurse may aim for reconciliation, which is achieved through accepting fallibility and acts of restitution; and 4) a nurse may be confronted with the dilemma of disclosure. Importantly, the decision to disclose and report the error, may be impacted by how one feels about the error, the level of harm the patient experienced, the level of support available, and how errors have been dealt with in the past. A tendency toward non-disclosure means that safety threats may go unreported, making them difficult to investigate and solve.³¹

How nurses are educated to identify, describe, and report incidents also impacts the actionability of the report. We believe there are opportunities to improve education in composing the narrative portion of the incident report. In this study, staff entering incident reports were not educated on the relevance of work-system factors, why they are important, and how they should be described in incident reports.

Our incident reporting system is intended to serve as a window into the underlying work-system; although it does provide insight that would otherwise be undiscoverable, a fundamental tension exists between the purpose of an incident reporting system (to learn from events to create a safer health care system) and the sometimes-punitive perception created. Several high-profile cases have occurred at other institutions where disclosing information after an error has led health care workers to experience additional suffering and criminal charges. Consider, for example, the recent case of a former Vanderbilt University Medical Center nurse who was found guilty of negligent homicide following a medication error that led to the death of a patient.³³ According to the Institute for Safe Medication Practices, “discussions about this case are dominating the health care community—from social media to headline-grabbing news reports.” ^34(p1) Cases like this may lead nurses and other staff to be fearful of reporting safety events, which may ultimately worsen patient safety if health care stakeholders do not respond and address reporters’ fears.

Despite efforts to improve incident reporting practices, challenges persist. Nurses and other health care workers require education and skill to improve understanding of work-system challenges, how interactions within a work-system may contribute to errors, and how to describe work-system factors in reports. Moreover, unless a system is in place to educate, protect, and support nurses and other staff after an error occurs, trainings are likely to fall short. Institutions and their leadership must work to improve patient safety culture, or else frontline workers such as nurses will never provide the kinds of reports needed to protect patients.

To improve reporting practices and protect patients at our own institution, we are launching and studying an intervention, the Safety Action Feedback and Engagement (SAFE) Loop, as a pragmatic randomized controlled trial at the level of the nursing unit. The SAFE Loop intervention has 5 key attributes designed to transform hospital incident reporting systems into more effective tools for improving patient safety: obtaining nurses’ input about which medication safety problems to address; focusing on selected high-priority events; prompting nurses to report high-priority events during a designated period and training them to write more informative reports; integrating information from reports, investigations, and other internal and external sources; and providing feedback to nurses on the problems identified and mitigation plans. If effective, the SAFE Loop will have several benefits: increasing nurses’ engagement with reporting, producing more informative reports, enabling safety leaders to understand problems and design system-based solutions more effectively and efficiently, and lowering rates of medication errors.³⁵

Limitations

Data investigated for this study were obtained from a single site, which may limit generalizability. We did not examine pre-pandemic near misses and errors, which limits our ability to make direct comparisons with our findings during the COVID-19 pandemic. Our conclusion that reported incidents may differ during the pandemic requires further investigation and validation in future studies that involve a pre-pandemic control group. Our ability to understand the role of COVID-19 in safety events is most likely limited for several reasons, including lack of reporter time, insufficient training of reporters, and opportunities to enhance reporting. Finally, incident reports were our only data source, so future work using other data sources, such as interviews with nurses, may give a more complete picture of factors contributing to events.

CONCLUSION

In summary, incident reports revealed that skill-based errors were the most common contributing factors for medication safety events during a COVID-19 surge. Reporters rarely deemed events to be related to COVID-19, despite the tremendous strain that the surge placed on nursing. Future efforts are needed to optimize the utility of incident reports, including improved education of staff and development of patient safety processes that emphasize the importance of reporting and an emphasis on improving the contributing work-system factors.