For three decades, researchers have documented the adverse effects of hazardous drug exposure.^1-5 Antineoplastic drugs (ADs) are among the most commonly used hazardous drugs identified as carcinogenic by the International Agency for Research on Cancer (IARC).^6-7 Governmental bodies⁸ and professional associations^9-10 have published guidelines on safe handling of ADs, yet guideline adoption is suboptimal.¹¹ These same guidelines have not determined what drug dosages are correlated with health effects. Instead, the guidelines recommend that preventive measures be taken to reduce exposure risk.^9-10 Research efforts have focused on AD exposure assessment and on process developments to reduce exposure, primarily through closed-system transfer devices and personal protective equipment. AD exposures have been monitored through environmental sampling of surface contamination.^12-15 Prospective biological monitoring from humans has been confined to spot urine samples with limited evidence of uptake.¹⁶ Innovative methods have been developed to determine specific urine concentrations of ADs and drug metabolites. These methods have been validated using highly sensitive liquid chromatography electrospray ionization tandem-mass spectrometry to analyze urine samples.¹⁷ These techniques enable prospective studies to examine drug exposure.

Few published studies have examined the relationship between organizational factors and AD exposure. A 2010 study¹⁸ reported 16.9% of surveyed oncology nurses reported dermal or eye exposure to ADs. Self-reported exposures were significantly more likely to occur with poorer nurse staffing levels and decreased performance of two-nurse chemotherapy dosing verification. In addition to adequate staffing, safety behaviors may play an important role in mitigating risk to employees.

AD exposure remains a prevalent problem in oncology settings, limited data are available on acute exposures (i.e., spills of ADs), and prior studies have focused on surveillance and retrospective data collection. The pilot study reported below studies AD exposure prospectively, includes survey reports and biological measures, and explores organizational factors that may influence acute exposures.

Our study used prospective questionnaires linked to urine samples followed by a retrospective survey to address three research questions: (1) what is the context in which AD spills occur in ambulatory oncology settings, (2) do workers who report AD spills have detectable drug levels, and (3) what organizational factors are associated with AD spills? The institutional review boards of the principal investigator's institution and the participating facility approved the research protocol and all participants completed written informed consent.

The prospective study had 40 participants who completed informed consent (see Figure 1 for a study flow diagram). During the 6-month study period, four unique drug spill events were reported, nine participants who were present at the time of the spill completed questionnaires, and nine urine specimens were provided after a reported exposure. The reported drugs spilled were etoposide, cisplatin, pemetrexed, and docetaxel, with a range in volume from 5 to 70 mLs. All spills occurred in the patient care area and all involved the use of a closed system transfer device. Three of the 4 spills occurred in the most congested area of the infusion center. During the spill, all respondents wore one pair of gloves, no respondent wore two pairs of gloves, and 29% wore a single-use disposable gown. Sixty percent reported zero or a low level of personal concern about the spill. The mean (SD) number of personnel who responded to each spill was 4.2(1.3). None of these events was reported using the facility's online risk management software.

Participants could describe the spill in their own words. All reported spills involved loose connections between the chemotherapy bag and the intravenous tubing or between the tubing and the vascular access device. One spill involved a patient with cognitive impairment who may not have understood they were connected to an infusion pump. Exposed workers reported assisting in the cleanup of patients and exposed surfaces.

In this single-site, practice-based pilot study of antineoplastic drug exposure in the ambulatory oncology setting, 4 spills were reported to the study team, all of which involved a hazardous drug.⁷ An important finding of this study showed that a single drug spill exposed multiple workers simultaneously. With the exception of pemetrexed, the drugs reported are used frequently in ambulatory oncology settings. The findings suggest that low, but quantifiable levels of the tested drugs are found in health care workers soon after a spill. Nor is exposure limited exclusively to drug spills. Urine samples from cancer center employees who did not report a drug spill had detectable, but no quantifiable levels of docetaxel. The mostly likely explanation for this finding is contamination of surfaces in the infusion area. The findings from those who did and did not report a drug spill suggests that drug spills pose a greater exposure risk to health care workers than routine environmental exposure. Yet is it important for health care workers to understand that antineoplastic drug exposure occurs in the context of both acute spills and routine drug handling events.

Across a 6-week period with 9,762 drug handling events, Connor and colleagues¹⁶ reported 29 spills or splashes. Over the 6-month study period with 19,284 antineoplastic drugs dispensed, 4 spills were reported by nursing personnel. This suggests a lower spill rate than published previously in a multi-site study where individual personnel were provided with drug handling diaries.¹⁶ There is no standard measurement approach for drug spills. Self-report of drug spills is prone to inherent bias that cannot be overcome without direct, continuous practice observation and/or visual recording. The number of drug spills reported in the study is a low estimate due to underreporting and restriction to consented participants.

Two participants with quantifiable levels of docetaxel reported spills involving drugs other than docetaxel. The facility's latest surface swipe testing report identified docetaxel contamination. During spill cleanup, the participants may have also come in contact with surfaces contaminated with docetaxel. As personal protective equipment use was low by study participants, these workers may have handled docetaxel during their shift without adequate protection. It is unlikely docetaxel was administered concurrently with these agents. The structures and molecular weights of paclitaxel and docetaxel differ enough to eliminate the possibility of errant findings using LC-MS/MS.²² Pre-existing surface contamination with docetaxel likely explains these findings. Surface swipe test monitoring and thorough cleaning of affected surfaces are necessary to reduce environmental exposure.

The participating facility had devoted significant resources for training, supplies, and medical monitoring. Considering the high frequency of unintentional drug exposure reported previously,¹⁸ the current findings are conservative. This only highlights the need for strategies to reduce exposure through environmental modifications and worker adherence to practice guidelines.

We found modest support for a relationship between organizational factors and AD exposure. Nurses who reported drug spills reported significantly less favorable relationships with physicians. Similar trends were observed across other organizational factors, though no other differences were statistically significant. Collegial nurse-physician relationships may serve as a proxy for stronger teamwork, which has documented importance on a variety of safety behaviors.²³ It is possible, though untested, that organizational factors must be favorable to assure adequate adherence to recommended guidelines and heightened situational awareness.

Despite initial receptivity, pharmacy personnel had low participation rates for unclear reasons. The pharmacy physical location and access restrictions limited the ability of study staff to visit to review study procedures and answer questions. The absence of a pharmacist on the study team and low perceived exposure risk are additional explanations. This study was conducted in one facility with a relatively small sample, which limits the generalizability of findings. The absence of statistically-significant differences in the retrospective survey are likely due to this small sample size and warrant confirmation in a larger, multi-site study. It is likely that non-responders to the retrospective survey have differing perceptions than responders. Finally, our LC-MS/MS procedures were performed at two different labs due to pre-existing capabilities. To increase the reliability of biological measures, we recommend all samples be processed for all antineoplastic drugs by one laboratory, followed by independent confirmation by a second laboratory. Finally, the survey findings would be strengthened by baseline collection of survey measures to examine work environments, workloads, and safety behaviors before drug spills are measured. These limitations are presented alongside one of the few published prospective investigations of antineoplastic drug spills and organizational factors in an ambulatory oncology setting.

Our findings have important implications for clinicians, practice leaders, and researchers. Detectable levels of drugs are found in both the presence and absence of acute drug spills. This finding suggests increased vigilance by staff is warranted throughout the drug preparation, administration, and disposal process. Drug spills pose a significant threat to workers, and increased efforts are needed to minimize worker exposure through strict adherence to published guidelines. These efforts should include structural (e.g., closed system transfer devices), individual (e.g., personal protective equipment adoption), and behavioral (e.g., heightened situational awareness) approaches.^8-10 The 2013 ASCO/ONS revisions to chemotherapy safety standards discuss necessary safe handling education and instruction.²⁴ The challenge is how to implement these recommendations and optimize worker protection.

Intravenous tubing connections are a significant contributor to chemotherapy spills. This finding should alert practice managers and clinicians to review products carefully for suitability for hazardous drug preparation and administration. Staff should participate actively in selecting personal protective equipment that balances comfort and safety. Practice leaders must promote a culture of safety and blame-free reporting of spills; incidents should be viewed as learning opportunities. As multiple employees were involved in spills, educational efforts should evolve to practical team-based case study approaches.

Several quality improvement opportunities are available. Practices can design and evaluate initiatives to promote recommended use of personal protective equipment. Audit and feedback programs on drug spills and surface contamination reports can target areas and activities at higher risk. Collaborative efforts can share optimal strategies to protect workers and maintain clinical efficiency.

Finally, prospective research studies that evaluate biological hazardous drug exposure are feasible and yield important insights. Future research efforts should focus on intervention development and evaluation, multi-site studies to compare exposures by organizational factors, and studies that correlate exposure to health outcomes. These approaches will generate the necessary evidence to address a thirty-year old problem.