Since the Institute of Medicine’s (IOM) publication of “To Err is Human: Building a Safer Health System,” there has been greater focus on patient safety in medical care.1 While this document is more widely related to hospitals and standard practice, these same principles are applicable to patients receiving treatment in clinical trials. Compared with standard practice, clinical trials have increased complexity because of additional heightened monitoring and strict control required for trial implementation. These complexities involve regulatory compliance while maintaining operational and clinical equipoise. Detailed interpretation of trial documents is needed to implement a clinical trial successfully and safely.

The International Conference of Harmonization Good Clinical Practice guidelines standardize clinical trial conduct to ensure a trial is executed ethically and with scientific quality while protecting patient rights.2 In addition to trial sponsor oversight, clinical trials are also overseen by institutional review boards and safety monitoring committees.3 These committees face complex regulatory requirements and must meet increasing demand as the volume of available clinical trials increases.4 While Protocol Review and Monitoring System (PRMS) Committees review trials for scientific merit and general feasibility, their scope may not include reviewing for the detailed intricacies of operational feasibility. Operational feasibility encompasses the comprehensive review of clinical trial documents to ensure that trial requirements are reasonable for site-specific implementation. These may include verification and alignment of drug dosing and administration across all documents, ensuring accurate and consistent timepoints for various assessments (e.g., laboratory monitoring, electrocardiograms, biopsies, imaging), and harmonizing procedures across site departments to reduce barriers to protocol compliance. As most site personnel only review portions of documents related to their specific areas of expertise, rather than reviewing all documents in their entirety in granular detail, operational inconsistencies or gaps may occur at the site. Therefore, review of trial documents may still garner PRMS approval despite undiscovered errors, such as drug dosing or administration errors, laboratory monitoring issues, and biopsies or procedures inconsistent with standard of care (SOC) recommendations or protocol, resulting in potential patient harm.

Study documents are created and frequently amended during clinical trial conduct, including the protocol, laboratory manual, and pharmacy/cellular therapy manual. This process introduces the potential for inconsistencies within or across documents, which may lead to implementation error, affect trial integrity, and/or lead to patient harm. If these errors reach the patient, protocol deviations resulting in minor divergences from the research protocol, or protocol violations resulting in errors that affect protocol integrity or patient safety, may occur.5 While making interventions, such as clarifying drug dosing or resolving discrepant assessment timepoints with the sponsor, is standard practice for each institution during clinical trial implementation, to date, little to no information has been published regarding these interventions. Furthermore, to our knowledge, there is no standard for how these interventions are discovered and no defined best practice to confirm detailed operational study feasibility with an emphasis on patient safety.

This study aims to characterize the frequency and type of administrative and potential patient safety interventions (PPSIs) made by centralized clinical content specialists during the review of oncology clinical trial documents for implementation at a single institution.

PATIENTS AND METHODS

The study was conducted at City of Hope National Medical Center, a National Cancer Institute (NCI)–Designated Comprehensive Cancer Center, in Duarte, California, enrolling more than 6000 patients into clinical trials each year. The Protocol Content Administration (PCA) team is composed of 8 pharmacists and 4 nurses. The PCAs are centralized clinical content specialists who develop and maintain study-specific clinical content for oncology interventional clinical trials, which are incorporated for build into the Epic electronic health record (EHR) oncology module. To develop this content, the PCA team reviews study-specific documents, including the protocol, laboratory manual, and pharmacy/cellular therapy manual (henceforth referred to as pharmacy manual). Consent, screening, and follow-up study phases are not included in content development.

From February 1, 2021, through February 28, 2022, interventions were collected for any PCA-reviewed trials and recorded in Research Electronic Data Capture (REDCap) (Vanderbilt University, Nashville, TN).6,7 Interventions were defined as any items that required further review by the PCA team resulting in a clarification from sponsor. Each trial was categorized by phase of study and sponsor type. Trials with more than one phase (e.g., phase 1/2) were grouped based on earliest study phase, and pilot studies were grouped with phase 1 studies. Trial sponsors were classified as industry, investigator-initiated trials, or other (e.g., federally funded, consortium, cooperative group studies).

During review of trial documents, interventions were identified and further assessed by the assigned PCA for potential to cause patient harm. Multiple interventions could be identified per trial. Interventions with potential patient harm were reviewed and confirmed by a panel of 5 PCAs and labeled as PPSIs. Potential patient safety interventions were further subcategorized as medication, laboratory, procedure related, or other (e.g., operational workflow issues). Potential patient safety intervention examples included medication-related interventions, procedures with inconsistent timepoints, and laboratory collection discrepancies potentially resulting in blood draws inconsistent with SOC or protocol. Non–potential patient safety interventions were deemed administrative, such as correction of typos, laboratory processing discrepancies, workflow questions, and any other interventions with no direct patient safety impact.

RESULTS

Over a 1-year period, a total of 585 oncology clinical trials were reviewed, of which 46% required intervention(s) and 18.8% with PPSIs. Of 1001 total interventions, 17.1% were designated PPSIs, with the remaining deemed administrative (Table 1). Potential patient safety interventions were primarily medication related (45.6%), followed by laboratory related (30.4%), procedure related (19.3%), and other (4.7%). Most frequently noted PPSIs included interventions surrounding blood draws inconsistent with SOC or protocol (22.8%), commercial, investigational, and/or supportive care drug dosing discrepancies (14.1%), unclear drug schedule start/stop times (11.1%), discrepant multiassessment procedural safety monitoring involving vitals and/or electrocardiograms (9.4%), laboratory monitoring issues (7.6%), and biopsy timepoints inconsistent with SOC or protocol (6.4%) (Table 2).

TABLE 1 - Distribution of Interventions by Trial Trials* PPSIs Administrative Interventions Total Interventions n % n #† n #‡ n % Phase  Phase 1 320 54.7 113 0.35 630 2 743 74.2  Phase 2 178 30.4 48 0.27 129 0.72 177 17.7  Phase 3 87 14.9 10 0.11 71 0.82 81 8.1 Sponsor type  Industry 407 69.6 113 0.28 654 1.6 767 76.6  Investigator-initiated trials 96 16.4 42 0.44 120 1.25 162 16.2  Other 82 14 16 0.2 56 0.68 72 7.2 Total 585 100 171 0.29 830 1.4 1001 100

*One endocrinology clinical trial was included.

†Number of PPSIs per trial.

‡Number of administrative interventions per trial.

Among medication-related PPSIs, drug-dosing interventions were most frequent (53.8%), with drug administration (43.6%), and drug dispensing (2.6%) (Table 2) also noted. Examples of drug dosing interventions included 2 different investigational drug doses stated in the protocol, a decimal dosing error (i.e., 0.80 mg versus the intended dose of 80 mg), and a carboplatin AUC capped at an incorrect maximum dose. Notable drug administration interventions included erroneous protocol verbiage instructing accelerated drug infusion rate, rather than slowed infusion rate, for patients experiencing an infusion-related reaction, and incorrect dosing directions instructing histamine-2 blockers to be taken within 6 hours of investigational drug, while the intent was to avoid within 6 hours. Drug dispensing errors involved instructions to dispense an insufficient supply of investigational drug bottles needed to fulfill a complete cycle of treatment. Another dispensing error provided directions to dispense excessive investigational drug bottles before a pivotal protocol-required disease assessment necessary for investigational medication continuance.

TABLE 2 - Potential Patient Safety Interventions With Intervention Details PPSIs Intervention Details n % Medication related 78 45.6 Drug schedule start/stop timing discrepancies (n = 19)
Dosing discrepancies (n = 11)
Unclear premedication requirements (n = 9)
Conflicting adverse event management guidance (n = 6)
Age/weight-based dosing ambiguity (n = 5)
Omitted missed/vomited dosing instructions (n = 5)
Infusion duration variances or omissions (n = 4)
Unclear cycle length (n = 3)
Unclear total treatment duration (n = 3)
Conflicting prohibited medications (n = 3)
Flushing discrepancies (n = 2)
Drug sequencing ambiguity (n = 2)
Supportive care administration discrepancies (n = 2)
Incorrect number of bottles to be dispensed (n = 2)
Decimal dosing error (n = 1)
Unclear waiting time between medications (n = 1)  Drug dosing 42 53.8  Drug administration 34 43.6  Drug dispensing 2 2.6 Laboratory related 52 30.4 Blood draw timepoints inconsistent with standard of care or protocol (n = 39)
Laboratory safety monitoring issues (n = 13) Procedure related 33 19.3 Discrepant multiassessment safety monitoring (vital signs and/or electrocardiograms) (n = 16)
Biopsy timepoints inconsistent with standard of care or protocol (bone marrow or solid tumor) (n = 11)
Disease assessment timing issues impacting imaging/efficacy (n = 3)
Lumbar puncture/cerebrospinal fluid volume inconsistent with standard of care or protocol (n = 3) Other 8 4.7 Operational workflow issues (n = 3)
Institutional Biosafety Committee regulatory omissions (n = 2)
EHR-related incorrect medication units (n = 1)
Incorrect radiation schedule (n = 1)
Terminology definition of “biweekly” (n = 1) Total 171 100

Of all 585 trials reviewed, approximately half were phase 1 (54.7%). One phase 4 trial was reviewed and, for calculation purposes, was grouped with phase 3 trials. Industry trials accounted for 69.6% of all studies reviewed. Of 1001 total interventions, the majority were made in phase 1 trials (74.2%). Among sponsor types, industry trials had the most interventions (76.6%) (Table 1).

Phase 1 trials had the highest proportion of PPSIs per trial (0.35:1) and also demonstrated the highest proportion of administrative interventions per trial (2:1). Investigator-initiated trials saw the highest proportion of PPSIs per trial (0.44:1) compared with industry and other sponsor types. Industry trials had the highest administrative interventions per trial (1.6:1) (Table 1).

Among all trial documents, most interventions were found in the protocol (68.4%). Of all PPSIs, 76.6% were found in the protocol. Of note, 1.8% of PPSIs were found in the pharmacy manual (Table 3). Of 52 laboratory-related PPSIs, 36.5% were identified as a discrepancy between the laboratory manual and protocol.

TABLE 3 - Interventions by Clinical Trial Document PPSIs Administrative Interventions Total Interventions n % n % n % Protocol 131 76.6 554 66.7 685 68.4 Laboratory manual 5 2.9 139 16.7 144 14.4 Pharmacy/cellular manual 3 1.8 20 2.4 23 2.3 Discrepancy between documents 28 16.4 99 11.9 127 12.7 Other (not found within documents) 4 2.3 18 2.2 22 2.2 Total 171 100 830 99.9 1001 100
DISCUSSION

The results of this study indicate that phase of study and/or sponsor type may be related to the number of interventions. Phase 1 trials had the highest proportion of PPSIs and administrative interventions per trial. Phase 1 trials are more likely to have complex documents given intricate trial designs including dose escalation structures involving dynamic drug dosing cohorts and detailed pharmacokinetic and biomarker endpoints. These studies also typically incorporate multiple amendments, increasing the inherent risk for both administrative and patient harm errors. Correspondingly, fewer PPSIs were noted in phase 2 and 3 trials as the study phase matured. Phase 3 trials demonstrated more administrative interventions per trial (0.82:1) versus phase 2 (0.72:1); however, results may be affected by the small sample size of phase 2 and 3 trials. Investigator-initiated trials showed the highest proportion of PPSIs of all sponsor types. Investigator-initiated trials typically lack the infrastructure and resources observed in industry-sponsored trials, with challenges including budget constraints, limited monitoring capacity, and lack of dedicated medical writers.8 Industry trials saw the highest frequency of administrative interventions, which is expected, as each trial is often accompanied by several supplemental documents. Differences in authorship of individual trial documents on a given study may contribute to inconsistencies found between these documents. In addition, the higher volume of study documents requires more revision maintenance, increasing the likelihood of error. Furthermore, industry protocol writers may be unfamiliar with how protocols translate into the diverse nature of site-specific clinical practice and settings, potentially contributing to implementation challenges.4

Medication-related PPSIs were most frequently identified among all PPSI categories, consistent with literature confirming that the most common type of preventable patient harm is drug related.9 Approximately half of medication-related PPSIs were linked to drug dosing discrepancies, arguably the most critical aspect of a clinical trial. In addition, the highest number of medication-related interventions were drug schedule start and stop time discrepancies. If drug dosing or drug treatment schedules are not clearly defined, this introduces risk for overdosing or underdosing, potentially resulting in patient harm by way of end-organ toxicity or diminishing treatment effectiveness via lack of therapeutic benefit, respectively. Furthermore, this may confound study objectives for drug efficacy, safety profile, pharmacokinetic data, and other laboratory measurements. Errors in premedication dosing were among the highest medication-related interventions found. Proper premedication dosing is vital to managing common anticipated toxicities of oncology agents including nausea or vomiting or infusion-related reactions; unclear agents, discrepant timing, or incorrect sequencing could alter a drug’s safety profile and impact patient safety.

Other important interventions found were related to laboratory and procedural assessments. Laboratory-related interventions, such as blood draws inconsistent with SOC or protocol and missed laboratory monitoring, were the second highest PPSI category identified. Blood draws inconsistent with SOC or protocol were the most frequent intervention identified of all PPSIs. This demonstrates the lack of scrutiny surrounding research-related blood draws, which should be given careful consideration, because oncology patients have potential for poor vein access and anemia due to an underlying cancer diagnosis, and can add to patient discomfort.10 Laboratory monitoring is critical to ensure a patient meets treatment criteria and could increase the risk for toxicity if missed. While biopsy timepoints or cerebrospinal fluid volume interventions inconsistent with SOC or protocol only accounted for 8% of PPSIs, these are concerning given inherent risks associated with these procedures. Tumor biopsies (solid tumor and bone marrow) and lumbar punctures are painful, invasive, and could lead to increased morbidity and decreased quality of life.

Although 78 medication-related PPSIs were identified, only 3 were found within the pharmacy manual. While pharmacy manual PPSIs were overall low, this could be due to the separate review of pharmacy manuals with potential interventions by investigational drug services, which are not captured by this study. Most medication-related PPSIs were found within the protocol, suggesting careful review of both the protocol and pharmacy manual are required to ensure accurate interpretation. Similarly, of laboratory-related PPSIs, approximately a third were found as a discrepancy between the protocol and laboratory manual, demonstrating that if the same personnel are not reviewing both documents to ensure alignment, these errors may be missed, subjecting patients to unwarranted blood draws. The highest frequency of PPSIs occurred within the protocol and as discrepancies between different documents. In many research institutions, it is likely that different institutional staff review each document as it pertains to their clinical areas of service. However, potentially very few institutions have dedicated personnel reviewing all study documents together in their entirety. This highlights the need for comprehensive cross-referencing between trial documents.

Patient safety literature primarily assesses patient harm severity by the outcome of a safety event after the event has already occurred. Various published tools classifying harm associated with medical/medication errors exist, but there is currently no standardized scale.9,11–13 Harm tools are often used to classify actual harm, while potential harm is often minimally addressed or unclear if the tool truly studied potential harm.12 Ghooi and colleagues14 reviewed protocol deviations against a classification system for impacts on patient safety and data quality, but similar to patient safety literature, deviations reviewed had already reached the patient. The process and interventions made in this study are a preemptive safety measure to ensure any errors or inconsistencies within clinical trial documents are addressed before patient enrollment. The severity of these interventions is not currently evaluable by any existing harm classification tools as these errors never reach the patient. In addition, in most cases, novel agents are being investigated, and the true magnitude of downstream safety ramifications are unknown. However, this study clearly demonstrates a need for severity scaling on captured potential errors or near misses. Literature regarding clinical trial–related near misses and potential errors is scarce because these data may be captured by other mechanisms including internal site reporting or sponsor workflows but is not published. This study supports the value in sharing PPSIs found during clinical trial implementation to highlight potential gaps in care.

Despite thorough review by safety committees, clinical trial documents are likely not reviewed in such specificity to identify clinical errors that can affect patient safety. Using a dedicated centralized team of clinical personnel to comprehensively review study documents is critical for patient safety and operational feasibility. This clinical content specialist can ensure patient safety, align with institutional and clinical standards of practice, and uphold protocol integrity. Embedding such a service within PRMS committees can serve to dually strengthen clinical trial patient safety and operational feasibility.

The process of identifying interventions in this study is based on the PCA review of documents for clinical content development within the EHR. The 2001 Institute of Medicine follow-up “Crossing the Quality Chasm” highlights the benefits of using information technology to improve the quality of health care in both safety and effectiveness.15 Published data also support that using technology has improved the quality and safety of patient care.16 The incorporation of clinical trial content into the EHR can streamline complex protocol information in a safe and accessible manner for study staff and other healthcare providers involved in the care of the research subject. This has served as an additional measure to enhance patient safety by streamlining operational workflows. Developing clinical trial content for EHR operationalization aims to embed requirements of the clinical trial into an institutional framework for modernized prescribing and patient care.17

The PCA team was established at City of Hope in 2017 as a novel role; thus, there was a notable lack of comparator or baseline data. The scope of this study was limited to the PCA role in reviewing oncology interventional trials during the treatment phase. Most oncology trials use active comparator controls, allowing for unique interventional work with SOC references in contrast to placebo-controlled studies. Oncology trials also have larger ramifications for patient safety due to the nature of oncology and the hazardous properties of most oncolytic drugs used in this setting, warranting scrupulous oversight during clinical trial implementation. However, further exploration is warranted to investigate whether similar interventions made in this study may be applicable to other windows within a trial (i.e., screening, follow up, etc) or in complex nononcology trials. Another limitation was the process of internal intervention review and determination of PPSIs. Although this determination was standardized across a panel of dedicated PCAs, an independent, third-party clinician(s) review could provide a more robust evaluation of potential patient harm. In addition, because the PCA team is composed of pharmacists and nurses, there may be unconscious bias toward medication and nursing interventions. As the interventions occurred before reaching a patient, no direct patient harm occurred, and existing harm models are currently insufficient in reporting severity classification of PPSIs or preventable harm. A possible future process could allow the trial’s principal investigator or ethical committee to determine the level of severity if the error had reached the patient.14 Finally, our data showed a high proportion of phase 1 and industry trials, potentially skewing results. Additional studies are needed to identify the true impact of interventions on phase 2 and 3 trials and other sponsor types.

CONCLUSIONS

During oncology clinical trial implementation, errors or inconsistencies in trial documents can potentially harm patients while also reducing the integrity of the trial’s intended objectives, result in regulatory noncompliance, and cause costly financial burden despite rigorous review from institutional regulatory bodies and sponsors. There is a clear gap in patient safety during clinical trial implementation as evidenced by the PPSIs identified in this study. While all interventions were considered near misses because they were identified before reaching a patient, the volume and types of PPSIs identified highlight the benefit of dedicating resources to comprehensively review clinical trial documents before study implementation.

The interventions identified in this study are not unique to the clinical trial realm, as similar types of patient safety interventions are common practices within standard of care, particularly within the oncology pharmacy domain.18,19 While standard of care interventions are well documented, to our knowledge, this is the first study to highlight interventions identified during EHR clinical content development to potentially prevent patient harm. Dedicating a team of centralized clinical specialists to review and translate clinical trial documents for implementation is one method of establishing best practice institutional guidelines for enhancing patient safety.

ACKNOWLEDGMENT

The authors thank the following people for their contributions: Jenny Bardens, Grace Chen, Ashley Johns, Jacklyn Kim, Allison Lam, Joo Hye Lee, Tiffany Luu, Jennifer Neprud, Nancy Nguyen, Monica Peplow, and Kathryn Yee.

REFERENCES 1. Institute of Medicine Committee on Quality of Health Care in America: in, Kohn LT, Corrigan JM, Donaldson MS, eds. In: To Err is Human: Building a Safer Health System. Washington, DC: National Academies Press (US); 2000. 2. ICH Harmonised Guideline: Integrated Addendum to ICH E6(R1): Guideline for Good Clinical Practice E6(R2); 2016. Available at: https://database.ich.org/sites/default/files/E6_R2_Addendum.pdf. Accessed November 30, 2022 3. Godfrey C, Payton M, Tasker S, et al. Ensuring participant safety and trial integrity with clinical trials oversight. J Acquir Immune Defic Syndr. 2014;65(suppl 1):S40–S43. 4. American Society of Clinical Oncology Policy Statement. Oversight of clinical research. 2003;21:2377–2386. 5. Bhatt A. Protocol deviation and violation. Perspect Clin Res. 2012;3:117. 6. Harris PA, Taylor R, Minor BL, et al. The REDCap consortium: building an international community of software platform partners. J Biomed Inform. 2019;95:103208. 7. Harris PA, Taylor R, Thielke R, et al. Research Electronic Data Capture (REDCap)—a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. 2009;42:377–381. 8. Konwar M, Bose D, Gogtay NJ, et al. Investigator-initiated studies: challenges and solutions. Perspect Clin Res. 2018;9:179–183. 9. Panagioti M, Khan K, Keers RN, et al. Prevalence, severity, and nature of preventable patient harm across medical care settings: systematic review and meta-analysis. BMJ. 2019;366:l4185. 10. Eaton KP, Levy K, Soong C, et al. Evidence-based guidelines to eliminate repetitive laboratory testing. JAMA Intern Med. 2017;177:1833–1839. 11. National Coordinating Council for Medication Error and Reporting and Prevention. Available at: https://www.nccmerp.org/sites/default/files/taxonomy2001-07-31.pdf. Accessed December 8, 2022 12. Gates PJ, Baysari MT, Mumford V, et al. Standardising the classification of harm associated with medication errors: The Harm Associated with Medication Error Classification (HAMEC). Drug Saf. 2019;42:931–939. 13. Hoppes M, Mitchell JL, Venditti EG, et al. Serious safety events: getting to Zero™. J Healthc Risk Manag. 2013;32:27–45. 14. Ghooi RB, Bhosale N, Wadhwani R, et al. Assessment and classification of protocol deviations. Perspect Clin Res. 2016;7:132–136. 15. Institute of Medicine Committee on Quality of Health Care in America. Crossing the Quality Chasm: A New Health System for the 21st Century. Washington, DC: National Academies Press (US); 2001. 16. Alotaibi YK, Federico F. The impact of health information technology on patient safety. Saudi Med J. 2017;38:1173–1180. 17. Zon R, Meropol NJ, Catalano RB, et al. American Society of Clinical Oncology statement on minimum standards and exemplary attributes of clinical trial sites. J Clin Oncol. 2008;26:2562–2567. 18. Duarte NC, Barbosa CR, Tavares MG, et al. Clinical oncology pharmacist: effective contribution to patient safety. J Oncol Pharm Pract. 2019;25:1665–1674. 19. Visacri MB, Tavares MG, Barbosa CR, et al. Clinical pharmacy in onco-hematology and bone marrow transplant: a valuable contribution to improving patient safety. J Oncol Pharm Pract. 2021;27:1172–1180.