Jane Ferguson, RN, MN, CCRP
Senior CRA, Westat
Abstract: Targeted monitoring is the on-site component of risk-based monitoring and focuses on the aspects of clinical research that have the most potential to impact participant safety and the credibility of the study’s results. This article provides an overview of targeted monitoring, including the need for targeted monitoring and its role in risk-based monitoring. The basics of conducting a targeted monitoring visit are covered
Most clinical research sites think that targeted monitoring means less oversight and fewer visits. Some sites, however, want more frequent monitoring visits. In the long run, targeted monitoring enables monitors to better monitor protocol compliance, which is the most important part of conducting a study. Targeted monitoring is the component of risk-based monitoring that is conducted on-site and focuses on the aspects of clinical research that have the most potential to impact participant safety and the credibility of the study’s results.
In the past, monitors reviewed 100% of the source data for 100% of the participants. Under risk-based monitoring, this is changing. Instead of, for example, spending a great deal of time reviewing every blood pressure measurement in a participant’s source documents and comparing them to every blood pressure measurement in a case report form (CRF) to ensure that the numbers are exactly the same, the monitor is spending time looking in source documents for any blood pressure measurements that are very high and for high-grade adverse events that should have resulted in stopping dosing or decreasing the dose of the study drug. It is more important to look for errors related to participant safety than transcription errors.
In 1996, the International Conference on Harmonisation Good Clinical Practice (ICH GCP) guidelines defined monitoring responsibilities as overseeing clinical trial progress and ensuring that the study is conducted, recorded, and reported in accordance with the protocol, standard operating procedures, GCP, and applicable regulatory requirements. In 1988, the U.S. Food and Drug Administration (FDA) guidance on monitoring stated that the sponsor needs to ensure adequate monitoring of the clinical trial.
Both ICH GCP and the FDA guidance are vague. Neither ever required 100% review of source data or 100% review of the participants. It is likely that sponsors decided that 100% monitoring surely had to be “adequate.” One sponsor decided to perform 100% monitoring and others followed suit. The FDA did not complain about this.
Thus, traditional monitoring involved 100% source data verification (SDV) on 100% of the participants. Monitors visited each clinical research site every four to eight weeks. At each visit, monitors also reviewed the regulatory records, the pharmacy, and the laboratory specimens. Traditional monitoring separated the monitors, the data management center, and the sponsor.
Over time, the monitoring environment has changed. The author believes that changes in the workforce and time issues are part of the impetus for moving to a risk-based model of monitoring. The shortage of monitors means that many monitors are working too hard and become burned out. Training and education of monitors and clinical research professionals has decreased, and the workforce has less experience and less clinical background. The FDA has issued an increasing number of FDA Form 483s citing inadequate monitoring of the study. With too few monitors, it takes longer to monitor. Time spent waiting for the monitoring increases the amount of time necessary from discovery of the drug to FDA approval of effective drugs.
Other reasons for changes in the monitoring environment are new technology, the futility of SDV, and economics. When clinical research sites used paper CRFs, the monitor only saw them at monitoring visits. Now, most CRFs are electronic, enabling monitors to access data right away, monitor/oversee studies remotely, and work faster.
TransCelerate, a group of 19 pharmaceutical and biotechnology companies coming together to try to improve clinical research, conducted a retrospective analysis of monitoring and discovered that of the queries done at sites during on-site monitoring visits, just 2.4% were related to critical variables. From discovery of the drug to FDA approval, monitoring is the second highest cost in the research and development budget. The cost of trying to achieve perfection is disproportionate to the benefit of doing so.
A 2013 Price Waterhouse study reported that using risk-based monitoring, including targeted monitoring, can reduce monitoring costs by 25%, project management costs by 20%, and data management costs by 20%. The FDA Guidance for Industry Oversight of Clinical Investigations — A Risk-based Approach to Monitoring (August 2013) and the revised ICH GCP E6 (R2) guidelines, published in 2016, provide support for risk-based monitoring. These two documents basically state that monitoring should:
• Focus on the most important aspects of study conduct and reporting
• Use alternative monitoring approaches, and take advantage of technology
• Maintain a flexible approach based on risks and critical findings. Thus, if a problem is seen at a clinical research site or across sites, the monitoring plan must be adjusted.
Shifting to risk-based monitoring is scary because 100% SDV of 100% of participants is what monitors have always done, and change is scary.
Risk-based monitoring is best performed by a cohesive team. The author worked on four Phase 1 studies involving risk-based monitoring where a cohesive team produced the best results. When monitors or data management staff see problems, monitors visit the clinical research sites and train clinical research professionals to fix the problems.
The monitoring plan must be study specific (Table 1). The sponsor and the project manager develop the monitoring plan based on a risk analysis (phase of study, drug profile, experience level of the clinical research sites, etc.). Monitors may review drafts; however, they will not be a major part of monitoring plan development. The monitoring plan is constantly adjusted in response to new risks and critical findings. It has three main components for oversight of the study: centralized, remote, and targeted monitoring.
There is no clear guidance about how to combine centralized, remote, and targeted monitoring. The author tends to think of centralized monitoring as the component of monitoring that is performed alongside data management, including data entry and queries. Centralized monitoring identifies adverse event trends and outliers as well as whether clinical research sites are completing their CFRs and addressing queries in a timely fashion.
Remote monitoring has two parts. The first part is the monitoring that is similar to the work done on-site; however, it is done remotely. The second part is reviewing the overall data for protocol compliance. Monitors can remotely review listings of various types of data. On a leukemia study that the author worked on, she identified through remote monitoring that one clinical research site continually reported that the blast cells in the CBC were not measured. The laboratory always measured the blast cells; however, if there were no blast cells, the laboratory did not report it. Monitors had to train clinical research professionals at this site that the correct answer on the CRF is 0 blast cells, not “not done.” Targeted monitoring is on-site monitoring that focuses on critical processes and critical data and aspects of the study that cannot easily be evaluated remotely.
Risk-based monitoring is a process comprised of Do, Evaluate, Adjust, and Repeat. Monitoring is performed, and then significance of the findings to the validity of the study is evaluated. The monitoring plan is adjusted in response to any problems in study conduct that have been identified. In the four Phase 1 studies mentioned earlier, someone decided that onsite monitoring (with the only review of source documentation) would be done every six months. That was not realistic for Phase 1 studies. The team for these studies was able to change the plan to include more frequent monitoring.
Targeted monitoring generally refers to onsite monitoring that focuses on critical processes and critical data as well as aspects of the study that cannot easily be evaluated remotely. It includes on-site pharmacy review and usually a brief regulatory review. Most of the pharmacy review, such as ensuring that the drug is being maintained correctly and counting the drug, must be done on-site. Many clinical research sites are handling regulatory documents electronically so that most of the regulatory review can now be done remotely.
In performing targeted monitoring, the monitoring plan is the most important piece of “equipment” (Table 2). The author brings a copy of the monitoring plan with her on-site to remind her of what she should and should not be doing. The monitoring plan states the frequency and length of on-site visits. If the monitor finds that more time is necessary, the monitor must discuss this with the project manager. There are many methods for targeted monitoring, such as 100% SDV on 50% of participants, or review of 100% of patients with 100% SDV of Screening and Cycle 1, then review of just dosing, AEs, and outcome for further cycles. On some studies, the author has monitored 100% for the first three participants, and if there were no major deviations, every other participant after that. She has also done 100% of participants, with source data review on only targeted processes and data. The method to be used will be defined in the monitoring plan for the study.
Requirements for regulatory, pharmacy, and/or laboratory review will be outlined in the monitoring plan. The author has monitored some studies where pharmacy was reviewed once a year, and the author has monitored others where if the first pharmacy review was fine, pharmacy was not reviewed again. The monitoring plan states the scope of source data review and the percentage of participants for source data review, the critical variables, and thresholds and triggers for changes to the plan. Monitoring will cover 100% SDV for the critical variables.
Since on-site monitoring is less frequent when using targeted monitoring, the monitor needs to know what is happening between monitoring visits. In one study that the author worked on, on-site monitoring was performed for each participant after Cycle 2, after Cycle 6, at one year, and every year after that. The time that is saved by fewer on-site monitoring visits should be spent performing centralized and remote monitoring.
Source data review is the review of source documentation for protocol compliance, adequacy of site processes, and quality of the records. The author believes that 100% of the source documentation should be reviewed. Monitors should start with the source documentation and read it like a book, one chapter at a time. They should then check the critical variables in the CRFs and then go on to the next chapter. Monitors must also review what happens between study visits, such as participant visits to the emergency room.
Performing source data review requires critical thinking skills and goes beyond looking for transcription errors. Many problems are missed if the monitor only compares the CRFs to the source documentation for transcription errors. That monitor is not checking whether the clinical research site is following the protocol because this is not all captured in the CRFs. For example, one study that the author is working on requires the dose of the drug to be lowered in patients that have a platelet count of under 100,000. One clinical research site that the author monitored did not do this for the two patients she reviewed. Merely checking the CRFs for the dose administered and the platelet counts for transcription errors is much different than looking for protocol compliance.
One hundred percent review of 100% of the data could result in perfect data; however, this does not guarantee that the clinical research site followed the protocol. When the FDA does an audit, it cares most about whether the site followed the protocol and the participants’ response to the study treatment. The CRFs must contain the patients’ response as per the protocol criteria, not the doctor’s definition of response. Thus, monitoring must be done at a higher level. Targeted monitoring provides more time for this.
The FDA defines quality as “the absence of errors that matter.” A temperature of 98.4o F versus a temperature of 98.6 does not matter; however, a temperature of 103 might be important. The FDA wants monitoring to focus on critical variables; those factors that are most relevant to the integrity and reliability of the study results as well as factors relevant to participant safety.
The critical variables can be different for each study, as they depend on the phase of the study, perceived drug/device risks, primary and secondary objectives, etc. Table 3 outlines the critical variables as per the 2013 FDA guidance on risk-based monitoring. These include informed consent, eligibility, and the treatment plan, among others. The treatment plan includes monitoring for the correct drug, dose, and administration; any dose modifications that were necessary; and drug accountability.
The FDA considers unanticipated adverse events a critical variable. From Phase 2 on, monitors will not need to check every detail on every adverse event when using targeted monitoring. Monitors need to review the important adverse events and can ignore the low-grade adverse events, which take a great deal of time to review. Most or all study endpoints are critical variables along with protocol compliance.
Monitors must consider the critical variables and the data points on the CRFs that must be reviewed. In general, all data points on the screening CRF must be reviewed, as they are relevant to eligibility. Laboratory data points can take a great deal of time to review. When using targeted monitoring, the monitoring plan may only require review of certain laboratory data points. The study drug administration CRF must be reviewed, as this contains data on the treatment plan, and adverse events may require dose modification. Monitors must read all of the source documentation; however, for adverse events, they may only review Grade 3 adverse events in the CRFs. The disease response assessment CRF and end-of-treatment CRF must be reviewed.
While it is not possible to be prescriptive about risk-based monitoring, targeted monitoring will always require 100% source review. Targeted monitoring of CRFs focuses on just the critical variables to ensure that the study is being conducted correctly. Table 4 has references related to the article.