Clinical Trial Fraud - Lab Information, General Warning Signs of Misconduct

In my experience, questionable data is readily and easily detected in subject lab information and results. Clinical study staff are trained by education and are expected to know each protocol lab test and understand lab values, IUs and ranges and the proper way to fill out the lab forms. Lab protocols and tests are reviewed extensively during pre-study start Investigator Meetings with staff, monitors and PIs/SCs.

It is my practice as well as my recommendation when reviewing, monitoring and investigating subject lab information and results - do not just log the test...instead, read, read the test and the results, each data entry, read the entire form. A high frequency and incidence of data fraud at the clinical study site will, most definitely, be found in subject lab data.

Where and How?

• Panic values on labs coded "1" and no comment is made by PI
• Labs not evaluated prior to entrance into the clinical study
• Technical units are expressed several different ways by staff and PI/SC.

General Warning Signs of Misconduct - What are these? Easy to identify for a monitor who is clear and present...

• High staff turnover
• Staff are disgruntled
• Staff are fearful
• Staff are anxious
• Staff are depressed
• Staff are defensive
• High pressure work environment
• Obsession with study payments
• Absent PIs
• Absent SCs
• Absent staff
• Lack of GCP training
• Unusually fast recruitment.

Monitoring with Quality Control is a full time job and when done correctly, GCP is clear and misconduct and fraud are easily seen in trends and such. Choose and interview your clinical PIs, SCs, staff, CRO and/or consultant CRAs/CRCs and monitors carefully. Train them on GCP and QC and QA. Monitor their visits, reports, correspondence, communication and performance.

As well, there is clinical study data that falls into the "miscellaneous" subject data category - interestingly, not significant to subject outcomes or endpoints, but will point a finger at data trends leading to the identification of clinical fraud. These categories will be highlighted in my next blog pertaining to clinical fraud at the clinical study site.

Questionable Data in Clinical Fraud - Diary, Source Documents

A Diary is a collection of observations recorded by the subject when home. Source documents record the subject's experience onsite, on-study and are data entered by clinical staff. Interestingly, this is an area where misconduct and fraudulent data is plentiful, the subject most of the time has no vested interest in the fraud and will not therefore record untruthful data. Whereas, data entries into the source documents are at the "mercy" of the clinical staff onsite, conducting the study, who may have a vested interest in the misconduct. Often, when FDA performs an unannounced, onsite audit for cause, Subject Diaries are always asked for and reviewed and compared to the source documents, CRFs, etc.

Diary:

• Diary states "dose not taken"- dosing record CRF page indicates date taken
• Diary card information written over
• Diary and dosing records are out of synch
• Handwriting is the same for several diary cards
• Handwriting of SC/PI is the same as writing on the diary cards
• Date on diary and dosing record is a date that does not exist
• Subject diaries not collected at each visit
• Subject diaries altered by SC/PI to match dosing dates in dosing record.

Source Documents:

• Source document and drug log dates differ
• Data on source documents not present at previous CRA visit appears at future visit
• No SAE reported on source documents but an entry added in different handwriting and different ink indicates "except erythyma, edema"
• Use of pencil in source documents
• Source documents indicates no PI available to do LSR's that day, yet subject had PE performed
• PE information in source documents and CRF differ.

During an onsite for cause audit by FDA, FDA inspectors (DSI, OCI Divisions at FDA) come in and have already determined that fraud exists and will confiscate everything related to data or the collection of data at the clinical study site, including computers, electronic data capture systems, hand and pocket data collections systems are well as all paper trails. In some cases, FDA will order biologic samples taken from the subjects on-study to be carted away to a "forensics" lab for future investigation.

Good Quality Control and GCP Monitoring cannot be stressed at the clinical study site. Each clinical study and the data derived from the subjects at each site serve as a building block to the foundation of what will eventually be an integrated global database, a GIDB, leading to a clinical submission to FDA. Clinical studies identified with fraudulent data often are censored from the mix. This often impacts the ability to effectively evaluate the safety of the clinical drug under study as well as subject exposure measurements, efficacy/therapeutic endpoints, clinical summaries of safety and efficacy, benefit/risk statements, clinical summary overviews and the label in the clinical submission. In my experience, and in many cases, data censored at the last moment for misconduct leading to fraud will delay the clinical submission timelines, delay the filing date of the clinical submission and/or FDA may reject the clinical submission in its entirety, a RTF. Several warning letters will be issued, a clinical hold is ordered for the study to halt immediately, while the investigation continues. After the investigation (and data censor) is completed and if the filing of the clinical submission is accepted by FDA - the approvability of the content of the clinical submission, is still in question and is in serious jeopardy for "data quality, non-reporting and integrity" issues.

Monitor, Monitor and Monitor with GCP and Quality Control.

Questionable Data in Clinical Fraud - Photos, Drug and Supplies

Often the question arises from the sponsor or a concerned clinical site team member - "There is so much data collected per subject per clinical study protocol at the clinical study site, where do we look for questionable data?"

Areas that one may find questionable data are numerous and in my view and years of experience - I can offer hundreds of locations and ideas. As well, I (colleagues) will not be surprised to uncover more areas where questionable data arises in the future.

At this writing, I will share common and not so common areas of questionable data at clinical trial sites suspected of clinical misconduct and fraud, starting with Photos, Drug and Supplies:

• Photos

1. Placards indicated different subject numbers but photos are of the same subject
2. TX area on source documents and CRFs differs from photos - for example cheek vs. back of head
3. Screening photo shows reaction in TX area - subject already dosing?
4. Dates on photos do not agree with visit #
5. Fraudulent photos
6. Missing photos

• Drug Log and Supplies

1. Incorrect subject initials on drug boxes - initials of subject not in our study
2. Date on drug log and drug box different
3. No dates dispensed on drug boxes and/or labels in CRF
4. Drug dispensing log - no entries for several months***
5. Drug dispensing log indicates drug dispensed, however drug boxes are still in inventory
6. Date of drug dispensed pre-dated -----after date of audit
7. Emergency supplies have wrong subject initials on box
8. Drug dispensing forms were pre-dated.

***At the time/visit on dispensing drug to subjects - the drug dispensing log and forms must be filled out at that time. This is GCP regardless of misconduct. This is true for any observation while subject is onsite, on-study. Fill in the appropriate information at the time of that visit - do not wait.

More to come in the next blog - Diary, Source Documents, Lab information.

Strategies for Detecting Clinical Trial Fraud

Monitor the data.
Monitor the monitor.

What do you monitor - what do you ask for when doing an audit? What detection tactics and strategies work?

• Ask for all subject information (data, docs, records) pertinent to the clinical study:

1. CRFs
2. Source worksheets
3. Clinical charts
4. Sign-in sheets
5. Lab requisitions
6. Shipping records.

• Accept no copies -

1. Review originals whenever possible.

• There is no easy way, don't just inventory and log, use your experience and understanding - Read -
  

1. Lab reports
   
2. X-rays.

• Fraud exists, expect fraud on an audit - assume fraud - work backwards.
• Question data, question and follow "open-ended data", "loose-ends" -
  

1. Missing
2. Altered
3. Inconsistent.
   

• Don't be timid - challenge the site to explain suspected fraudulent data.
• "Blame-shifting" during onsite audits "for cause" is pervasive.
  
• Be suspicious and challenge the investigator and remind he or she that clinical study conduct and GCP sits with them.
• Find and engage and cultivate "whistleblowers" - pay attention to -
  

1. Staff complaints
2. Listen to grievances
3. Establish rapport be approachable
4. Follow your leads
5. Don't stray from the data - the data is the data
6. Watch for changes in clinical staff behaviors
7. The list is endless.

GCP is well documented and simple to follow - unless there is serious intent to do otherwise. Detect clinical fraud early in data at a clinical site. The ramifications of doing so saves time and perhaps the integrity of the clinical study and/or clinical submission but most importantly protects subject safety, for those subjects that are real.

Clinical Research Fraud - Investigator Data Identifiers

In my experience, clinical research fraud is pervasive and can be found in many subject data collection systems, source records, subject test results and visit documentation entries at the investigator's clinical trial site. The following are the most common data identifiers:

• Implausible, impossible, in my view, trends and patterns:

1. 100% drug compliance
2. Perfect efficacy responses for all subjects
3. Identical lab results
4. Identical ECG results
5. No SAEs reported
6. Subjects adhering perfectly to visit schedules.

• Clinical site data not consistent with other centers (statistical outliers)
• Perfect diary cards, immaculate subject CRFs
• All source records & CRFs completed with the same pen
• Source records lack an audit trail
  
• No signatures of persons completing documentation
• No dates of persons completing documentation
• Subject handwriting inconsistent across documents (consents, diaries)
• Subject signature inconsistent across documents (consents, diaries)
• Questionable subject visit dates (Sundays, holidays, staff vacations)
• Impossible events (eg, randomization before drug delivery)
• Data contains "digit preference"
• Subject visits cannot be verified in the medical chart, appointment schedule or billing records.

How to avoid clinical trial site, staff and investigator fraud?

How to uncover data identifiers?

• Monitor data.
• Monitor data.
• Then Monitor data again.
• Monitor the Monitors.

FDA - Consequences of Fraud

The consequences of fraud are pervasive, outreaching and forever. If the Sponsor is found guilty of fraud as per FDA, data validity is compromised and the clinical submission is jeopardized. If the Investigator is involved in the fraud, he or she is disqualified, listed, given fines, incarcerated, incur tremendous legal expenses, he or she has just ruined their career. If the Institution, a Hospital, a Phase 1 Facility, a Clinic, a CRO, IRB or otherwise - is found guilty of fraud - lawsuits. What is the consequence to Subjects enrolled in the Clinical Trial - safety and well being at risk, loss of trust in the clinical trial process, lawsuit.

Facts:

• Fraudulent Investigators are often used by multiple Sponsors on Multiple Clinical Trials.
• Since Investigators are often used by multiple Sponsors on Multiple Clinical Trials, these Investigators will have a broad impact on many clinical submissions made by many Sponsors.
• For example, a case in mind - Investigator F, was involved in 91 clinical submissions with 47 different Sponsors during the same calendar year.
• Why? Word gets around that there are Investigators that "want to see their clinical trial site do the best". Which means that prior to selection of that Investigator, he or she understands how to manage the Subject CRF, Subject Informed Consent and Clinical Trial Protocol, for examples, in order to provide the Sponsor with "higher" statistically and clinically significant subject positive outcomes and better endpoints with respect to the efficacy and safety of the drug on clinical trial.

Why does fraud occur and how does it happen? To mention a few of the most obvious reasons:

• Not enough time
• Not enough staff
• Not enough monitoring, frequency or intensity to detail
• Not enough subjects, so subjects are reused and/or subjects are "made-up"
• Lack of GCP
• Lack of the understanding of GCP
• Lack of training GCP
  
• Lack of training FDA requirements and guidance
• Lack of regulatory oversight
• Lack of interest and motivation
• Money
  
• Greed
• Intent
• Pressure to perform
• Pressure to publish
• Pressure from senior management
• Pressure from piers
• Pressure from political arenas within medical societies, institutions, associations
  
• Pressure from CRO
• Pressure from Sponsor.

FDA Regulations (original statements and updates), pertaining to ("early") actions that must be taken by the Sponsor when fraud, misconduct and noncompliance are suspected or proven, can be found:

• 21 CFR 312.56(b)
• ICH E6 - 5.20.1
• ICH E6 - 5.20.2.
  

TRCL: 32KZYNXE6YNW

Clinical Research Fraud - FDA Definition

What does clinical research misconduct, fraud mean? FDA provides a definition that is clear in message and severity. Clinical research misconduct means falsification of data in proposing, processing, designing, performing, recording, supervising, reviewing, analyzing, collecting clinical research or reporting clinical research results, outcomes and endpoints. The manipulation of data and reporting for a self-serving purpose, usually monetary! Simple!

The FDA uses fraud and misconduct interchangeably and includes acts of omission and commission, consciously not revealing all data and consciously altering or fabricating data.
Fraud does not include honest error or honest difference in opinion. Deliberate or repeated non compliance with the protocol and GCP is considered fraud, second to falsification of data which is more severe in penalty and justice and resolve.

Let us review again who commits fraud?

• Investigators
• Study nurses
• Study Coordinators
• Data managers
• CRAs
• Sponsors
• Lab personnel
• IRB staff
• Subjects.

Yes, it is true. Clinical trial subjects have been accused and found guilty of fraud. The reasons are interesting and the content of another blog.

Average breakdown in % of blame:

CRC - 39
Study Nurse - 17
Hospital - 9
Sponsor - 9
Self - 9
Office Staff - 9
PI or Co-PI - 4
CRA - 4.

Monitor and adhere to GCP. Review and monitor repeatedly. Report Fraud.

Clinical Trial Fraud

You are a CRA, CRC, a member of the clinical trial team, CRO, pharma, sponsor, or otherwise - when you become aware of fraudulent data and clinical practices in a clinical trial - what do you do?
Immediately report your suspicion and findings to FDA. FDA will direct the "for cause" call to the appropriate divisions at FDA - Divisions of Scientific and Criminal Investigations. From here, if fraud is proven, any individual or organization participating in fraudulent activities will be held to the full extent of the law governing such in the judicial court system.

Clinical trial fraud is on the rise again.

Major areas riddled with fraudulent clinical trial behavior:

• over enrollment of patients
• faking patient informed consent forms
• faking patient CRFs
• faking visits
• faking lab data
• under reporting of adverse events
• no reporting of "early patient withdrawals"
• no reporting of "lost to follow-up" patient visits.

If fraud is proven, clinical investigators and anyone involved in the fraud will be given heavy fines as well as prison sentences.

Who is at fault?

Principle and co-investigators, clinical team members and organizations.

It is difficult to prevent the intention of fraud - but it is not difficult to identify fraud.

Lack of clinical trial and quality control monitoring allows for the occurrence of fraud.

Monitor. Monitor.

Vaccine Essentials

The development cycle of a vaccine is laborious, time-consuming, costly and long. A vaccine is a biologic and is made from living microorganisms. The development cycle is different that a pharmaceutical product, a drug, but does share a similar Clinical Phase Program, Phase 1,2,3 and 4 (post-approval). The Clinical Trial Protocols for each Phase share safety, efficacy and tolerance in a broad sense, but protocol designs are very different in design, objectives and outcomes.

The essentials:

• Exploratory - in preventing a disease or treating the disease:
  

1. to understand the disease
   
2. to understand the epidemiological data
3. to understand the correct antigens to use.

• Pre-clinical stage:

1. to assess antigen safety
2. to select the best candidate vaccine.

• Clinical Development - all Phases are involved in clinical trials and the 1st lots/batches are produced (clinical batches for trials and industrial batches for compliance):
  

1. Phase 1
2. Phase 2
3. Phase 3

• Regulatory Approval - all data from previous stages are complied, summarized, quality-controlled and assured and are submitted to regulatory health agencies for approval.

• Manufacturing - on an average, it takes just less than 2 years to produce a single batch of vaccine.

• Quality Control - ~70% of the production time is dedicated to quality control of data, process, procedure, R&D, Industrial Operations and Pharmacovigilance.

Requirements for a Vaccine Approval by FDA

Vaccine research, development and approval is a process that requires, time, is costly and is complex.
Approval Requirements:

• Successful review of the BLA for a new indication from FDA
• Successful review of the BLA for a new indication from VRBPAC
• An adequate product label
• Benefit/Risk Statement
• Safe delivery of the vaccine to the public.

After approval of the license, monitoring of the product, production and announced and unannounced manufacturing facility inspections must continue as long as the manufacturer holds the license for the product.
Periodic Inspections (especially for a newly approved BLA license and manufacturer) will include tests for:

• potency
• safety
• purity
• submission of samples from each lot
• continued quality control documentation
  
• continued quality assurance documentation.

After a period of successful test results, FDA CBER, may determine that lot to lot samples are not necessary. FDA CBER may determine that an alternative procedure is acceptable instead of lot to lot sampling. Lot release protocols showing results of applicable tests for example can be submitted instead. The change from one procedure to another must be in writing and documented with the manufacturer, FDA CBER, and the pharma sponsor. The change will only be granted after a long period of successful routine testing of lot to lot samples.

What's next?

Potential adverse events (AEs) cannot be anticipated, therefore there "will be" a committed Phase 4 pharmacovigilance program as well as VAERS, a Vaccine Adverse Event Reporting System to identify problems after marketing begins. Who is responsible for any problems, manufacturing, safety, AE or otherwise ? - the pharma sponsor. Monitor the Phase 4 program, monitor safety, monitor AE Reports, monitor the Phase 4 monitors, monitor manufacturing, monitor quality.

Vaccine Approval Cycle - Hard Facts to Consider

The Center for Biologics Evaluation and Research (CBER) is the division at FDA that receives and reviews vaccine applications in the US. The regulation for vaccine products in the US can be found in Section 351 of the Public Health Service Act and specific sections of the Federal Food, Drug and Cosmetic Act.

• The development of a vaccine follows the same general pathway as for drugs and biologics.
• Pharma sponsor must submit an IND - Investigational New Drug application to FDA.

The IND describes:

• the vaccine
• method of manufacture
• quality control tests (70% of a vaccine's production time is dedicated to quality control)
• safety
  
• immunogenicity of the vaccine in animal testing
  
• proposed clinical protocol(s) for human clinical trials.

Clinical Development Phases: (6 - 8 years in humans)

• Phase 1 Clinical Trials - safety and immunogenicity, 10 - 100 subjects
  
• Phase 2 Clinical Trials - dose-ranging, immune response - 100 - 3000 subjects
  
• Phase 3 Clinical Trials - effectiveness, tolerance and safety - 3000 - 40000 subjects.
  

At the end of a successful Clinical Phase 3 development program and if there are no safety concerns or otherwise by FDA, the pharma sponsor files a Biologics License Application (BLA).

The BLA:

• must provide FDA efficacy and safety information
• risk/benefit statements and assessments
• proposed manufacturing facility "actively" in production of the said vaccine when the inspection takes place.

Vaccine Development Cycle:

• Research & Development - 9 - 14 years

1. Exploratory - 2 - 4 years
   
2. Pre-clinical - 1 - 2 years
   
3. Phase 1
4. Phase 2
5. Phase 3

• Registration - 12 - 18 months
  
• Product Launch
• Industrial Operations
  

1. Bulk Production - 1.5 - 2 years
   

Following FDA's review of a BLA for a new indication, the pharma sponsor and the FDA present their findings to FDA's Vaccines and Related Biological Products Advisory Committee (VRBPAC). This non-FDA committee provides FDA with advice pertaining to the safety and efficacy of the vaccine for the proposed indication.

Monitor the Quality (Data), Monitor the Monitor (CRO)

Pharma Sponsor Data Management as it pertains to Quality Management must regularly assess the quality of the data management process and work at the CRO and request data and metrics to track the quality of the Clinical Study Data. Monitor the Quality and Monitor the Monitors who monitor the quality!

Quality metrics may include:

• frequency of edit checks
  
• number of data changes due to data entry error, obvious error and DCF
• these metrics can be broken down by database and/or clinical study site.
  

Pharma sponsor DM must review these metrics for any possible trends or inconsistencies in the data.

Where can potential issues be found:

• edit check programming
• data entry
• CRF design
• data being provided by the clinical study site
• listings of key clinical study efficacy data
• listings of key clinical study safety data
• CRO data entry rules
• CRO data review rules
• data being provided by the CRO.
  

It is the responsibility of the pharma sponsor DM to review these listings for any possible trends or inconsistencies in the clinical study data. Communication of these findings and any recommendations to the appropriate clinical study team members, for example, CRO, Medical Monitor (Pharma Sponsor), Biostatistician, DM, CRA, CRC, must be immediate and have resolution.

Monitor the quality of your data, ensure use of the DMP and QMP, monitor the monitor, the CRA, the CRO, ensure there is verbal communication as well as written metrics and listings providing data, potential trends and inconsistencies. Once found, the action is immediate - correct and assess the impact on clinical study data quality, data error %, resource and capacity to correct.

Most important, communicate the issue(s) to the team, re-train the team pertaining to appropriate use and adherence to relevant SOP, WI, QMP and DMP and check to see if there is a change in clinical study site personnel.

CRF and DCF Tracking at the Clinical Study Level

The last several blogs pertained to the management of data, discrepancy, quality control, audits, etc., at the Clinical Study Level. This blog pertains to the careful management and tracking of CRFs and DCFs at the Clinical Study Level.

To document data management clinical study progress, while following DM SOPs and WIs, the selected CRO(s) must be required by the sponsor pharma to provide regular metrics reports for CRF and DCF Tracking. The frequency and timing of these reports must be determined on a per study basis.

At a minimum, the following metrics must be tracked for CRFs:

• Complete CRFs received (all expected visits for the subject received)
• Total number visits received
• Total number visits entered
• Total number reviewed

At a minimum, the following metrics must be tracked for DCFs:

• Total number generated/sent
• Total number returned
• Total number resolved
• Number of DCFs outstanding greater that "x" days.

A mistake often made: The tracking tool used should not be a excel spreadsheet, spreadsheets can be changed by error, inadvertently and unintentionally by clinical study personnel. These spreadsheets are convenient to use but are not protected from erroneous revision and do not serve as GCP, Good Clinical Practice, accepted "source" documentation.

On scheduled, announced or unannounced "for cause" clinical audits, the use of excel spreadsheets for CRF and DCF tracking purposes will lead to a "failed" audit with serious regulatory, clinical and investigational repercussions.

Tracking records should be collected and versioned in an agreed upon "safe, lock-out" template, final version saved to an archive at the Clinical Study Site, the CRO, at the sponsor pharma, usually placed in the DM folder on the computer hard drive with paper copies to file in the DM department. All agreed upon metrics and tracking tools must be documented in the Sponsor Pharma DMP - the Data Management Plan.

Managing Data Changes - Clinical Study Level

The sponsor pharma selected a CRO to monitor and process the clinical study data and database. The CRO now must maintain an audit trail to document any data changes that occur after data entry in the clinical study database. What should the data audit trail contain? At the very least the data audit trail should contain:

• sponsor pharma name/protocol number
• investigator number
• patient number
• dataset name
• CRF page
• variable name
• old value
• new value
• date/timestamp of change
• person who made change
• reason for change.

There are generally two types of data changes. Data changes arising from DCFs generated by a CRO monitor, CRA, CRC, DM, QC, QA or quality control process and "obvious errors". Be careful making data changes pertaining to "obvious errors". Be careful of "obvious errors".

Obvious errors are corrections that result in data changes that can be made "without" a confirmatory DCF issued. These corrections are self-evident based on the review of corresponding CRF data. There must be Generic Guidelines as well for Handling Obvious Errors and most importantly, the sponsor pharma is the only authority that can approve Obvious Error corrections, not the clinical personnel at the site, not the investigator. Final approval is only made by the sponsor pharma which is ultimately responsible for the integrity of the final clinical study data and clinical study database.

Monitor the monitor, monitor the data, monitor changes to the data, monitor the audit trail - as a sponsor take an active role in the quality control of the data and database at the clinical study level.

Data Management CRO/Sponsor Pharma Roles

The role of the data management CRO chosen by the sponsor pharma to conduct clinical studies at designated and approved sites, domestic and international, and leading to an integrated clinical submission, CTD, eCTD, Clinical Phases 1,2,3, are as follows:

• Tracks receipt of all data
• Enters data into the database (at the clinical study level)
• Performs data review through execution of programmed edit checks as well as through a manual review of the data
• Maintains an audit trail of all database changes
• Provides adverse event and concomitant medication coding listings for review
• Maintains documentation of the resolution of all discrepancies, including those that did not result in a DCF
• Manages the receipt, loading and processing of any external electronic data received for the clinical study.

1. Interview the CRO.
   
2. Investigate where the CRO conducts studies, North America, South America, Europe, other.
   
3. Investigate the experience of the CRO, what indications, which therapeutic areas, safety, efficacy, BE/BA, Clinical Phases 1,2,3, oncology, neuroscience, respiratory, drug, device, biologic, other.
   
4. Examine the correspondence and communication tactic of the CRO with FDA.
   
5. How well recognized and respected is the CRO in the eyes of FDA?
   
6. Ask for a "track record" of approvals, RTFs, filings, Clinical Submissions, eCTD, CTDs globally.
   
7. Look at the credibility and experience, resumes and CVs of the CRO staff.
   
8. Ask for the resource and capacity available for your clinical studies - some CROs are "over-booked".
   
9. Ask for the client list of the CRO.
   
10. "Interview the CRO" before making a decision.
    
11. Once the CRO is chosen, develop quality management plans to monitor the CRO, contract, conduct, performance, data, etc.
    
12. "Monitor the Monitors".
    
13. The ultimate responsibility of a credible clinical study and a database with data integrity and accuracy is the sponsor pharma.
    
14. Perform Quality Control and Audits periodically during the course of the clinical study.
    
15. Start early, after several patients are eligible and enrolled in the clinical study, quality check the data and documentation and paper trail at randomly selected sites.
    
16. Use Quality Data Management and Quality Control SOPs and WIs for checking selected clinical study sites.
    
17. Monitor the Monitors!