Effect of improvement measures in reducing interruptions in a Japanese hospital pharmacy using a synthetic approach based on resilience engineering and systems thinking

Work interruptions and distractions are significant contributors to dispensing errors in hospital pharmacies, along with high workloads, low staffing levels, mix-up of look-alike/sound-alike drugs, lack of experience, and urgent deadlines/hurrying of tasks [1]. A high frequency of interruptions has been reported, ranging from every 2 to 6 min [2‐4], making this a high-priority issue in patient safety. Telephone calls are the prominent interruption sources for inpatient pharmacies [2‐4], and nurse inquiries about the prescription status accounting for a substantial part of the interruptions [5]. Human behaviors that might produce interruptions have been rarely studied, partly because conventional patient safety, reacting to incidents, and using a reductionistic approach and a linear causal model, fail to capture the dynamic nature of human behavior in adapting to given conditions to accomplish work [6‐9]. Recognition of healthcare systems as complex adaptive systems has attracted the interest of healthcare providers and researchers toward a synthetic approach to solving complex problems [10, 11]. Both resilience engineering and systems thinking are theories for a synthetic approach to understanding systems’ behavior emerging from human interactions. Resilience engineering, a new paradigm in safety management, provides a perspective on how people work adaptively through performance adjustments in an environment with changes and constraints to achieve their goals [12]. Systems thinking helps develop a causal model of systems’ interactions with feedback loops [13]. However, there is no study about how complex problems such as interruptions in health care systems can be solved using a synthetic approach.

Through daily patient safety activities, we recognized that pharmacists in our university hospital experienced frequent interruptions during their dispensing work for inpatients owing to telephone calls from nurses in inpatient wards and counter services for urgent provisions of dispensed medicines to staff in the inpatient ward. This study aimed to identify a mechanism for these kinds of interruptions in a hospital pharmacy to find improvement measures using a synthetic approach and assess the effect of the implemented measures in reducing interruptions.

This study found that interruptions to pharmacists' work due to telephone inquiries and counter services were a systemic problem that emerged from the adaptive behavior of pharmacists and nurses to the work constraints such as short staffing, limited frequency of medication deliveries, and lack of information about the medication dispensing status. The measures for mitigating cross-system performance adjustments—specifically, the medication dispensing tracking system, additional medication deliveries, and the use of pass boxes—reduced interruptions by 60%. The information flow (of the dispensing status), the rules (for obtaining medicines earlier), and the feedback loops (between the two systems) that we intervened have been effective leverage points to change the system’s behavior [15]. The medication dispensing tracking system reduces interruptions by nurses’ phone calls in a hospital pharmacy [16].

The strengths of this study are as follows. First, to the best of our knowledge, this is the first study to determine how interruptions in a hospital pharmacy were caused and mitigated from a systemic perspective. Second, this study demonstrates the methodological guidance for implementing the theories of a synthetic approach into safety management practices. In practicing Safety-II based on resilience engineering, concrete ways of learning from everyday clinical work, analytical tools for identifying systemic problems, and the scientific evidence for its effectiveness have been challenging [17]. We illustrated how to learn from work-as-done where performance is continually being adjusted, to analyze the non-linear effect of interactions between the different systems with a dynamic thinking tool, and to identify effective measures for reconciling work-as-imagined, defined as work that should be performed according to plans or rules, and work-as-done [12].

The study results illuminate hidden risks of clinicians’ performance adjustments that enable the law of stretched systems and the law of fluency [18]. Among the three adaptive tactics for extending the capacity for maneuver in the IMDU-OT, extending the capacities of individual pharmacists was reported as a first-order problem solving to increasing demand [19] and has a possibility of falling into unacceptable performance [20], such as the criminal case of a medication error [21]. Creating constraints to other systems, which is a typical pattern of adapting to the shortfall in capacity for maneuver [22], may cause additional problems in the interrelated systems through a mechanism of the fallacy of composition or a problematic pattern of systems behavior called “shifting the burden” [23, 24]. The study indicates that solving a systemic problem requires looking into interrelated systems widely, not a single system locally. Reciprocity, which benefits both pharmacists and nurses, is also essential in determining improvement measures to extend the adaptive capacity of interconnected systems [25, 26].

Our study had several limitations. First, the study is conducted in a single Japanese university hospital, and so may limit the generalizability of the study findings. Yet the literature found commonalities in causes or sources of interruptions in other hospital pharmacies with ours [2‐5]. A more important aspect of this study is the process of identifying a systemic problem and effective improvement measures unique to the context using a synthetic approach. Second, the daily prescription shown in Fig. 1 may not represent the regular working situation in the IMDU-OT because it was based on work performance data from one day in August 2017. This could limit the generalizability of the study findings. However, the number of prescriptions handled that day (999) was within the normal range (within mean plus 2 standard deviations, SD), considering the prescription data from all business days in 2017 [daily mean (SD): 859 (94)]. In addition, the staffing support scheme for the IMDU-OT on that day followed a typical patterns based on the time and day of the week. Thus, those would minimize this limitation. Third, the study targeted interruptions regarding oral and topical medicines only, leaving other significant sources of interruptions to pharmacists untouched (e.g., telephone inquiries about injectables for inpatient wards and novice pharmacists’ requests for guidance from skilled pharmacists). Fourth, the study did not assess the long-term effect of the intervention because data collection over time became unperformable due to the deteriorating situation of the COVID-19 pandemic. As of now, at least, the utilization data of the machinery extra delivery system has been at a similar level as the immediate post-intervention, and telephone inquiries about the dispensing status have remained a few, according to pharmacists' recognition.

Our study found that interruptions in the hospital pharmacy due to telephone inquiries by nurses and their use of counter services are a systemic problem that can be reduced by mitigating work-related difficulties that are compensated for by human performance adjustments. It also showed that a synthetic approach to safety management can be useful in finding a systemic problem threatening patient safety and effective interventional points in complex adaptive systems. Our findings have implications for patient safety practices and regulations. Healthcare organizations and patient safety experts need to develop proactive safety management in dynamic work environments, gaining insight into the risks of adversity behind the fluent performance. Healthcare regulators also should promote such practices and research, which have been introduced into health policies in some countries [27, 28]. Further studies are required to validate the effectiveness of a synthetic approach to solving interruptions and other systemic problems in different settings.