There are certain trends in audits and inspections, which, regardless of the company, phase or type of clinical trial, seem to be ever-present. Some of them are indeed preventable, but require not so much a change in technology or process, but by a change in human behavior.
We have all heard of ALCOA, but what about MALCOA? The bad things that happen during an inspection or audit.
Deadly Sin #1: Monitoring
The (M) here stands for Monitoring – deadly sin number one. With the recent increase in demand for clinical research associates (CRAs), sponsor companies and CROs are experiencing a high turnover rate as competition drives up salaries. As CRAs hop from position to position for sign-on bonuses and pay increases, their loyalty to each employer diminishes (as one might expect). It is not uncommon to see several successive CRAs take over monitoring at a single investigator site.
The result? Loss in continuity and a drop in quality. The site feels less obligated to carry out directives, and sponsor oversight can suffer. Resourcing also suffers as result of re-monitoring, which is a result of a poor handoff from the previous CRA to the incoming one. The previous monitor may be less concerned with leaving a pristine site than starting their new position, and the incoming CRA may not be interested in taking responsibility for what happened before they inherited the site. A well-trained and dedicated monitoring force is imperative to a successful study.
Deadly Sin #2: Attribution
(A) Attribution, it would seem, would not be a problem with the increasing reliance on Electronic Medical Record systems (EMRs), but there is still an area of concern here. It is a common practice for a principal investigator (PI) or sub-investigator (Sub-I) to provide a verbal communication about a lab result, assessment of adverse event, or perhaps a dosage change which is then entered into the EMR by the study coordinator. Since it’s attributable to the person logged into the EMR, it is linked to the study coordinator logged into the EMR who is usually not delegated to make these medical decisions.
If there is no other source (often not) then it has to be explained to an auditor or inspector that this is the practice. To address this gap, some sponsors may require a source worksheet or form be filled in for documentation of the assessment. In that scenario, the sponsor is either asking the (1) site to change its normal practice by completing the form first, or (2) asking them to fill it in after the fact because the EMR is the true source. Either of these requests is “unnatural” to the study site and results in incomplete entries or transcription errors.
Deadly Sin #3: Legal
(L) As we consider the use of EMR – legibility does indeed improve. Gone are the dangerous abbreviations and indecipherable scribbles. Instead, the “L” sin in this article will represent Legal issues. As institutions (rightly) place restrictions on access to medical records by CRAs, there arise delays in monitoring while institutional forms are reviewed and signed, and information technology departments grant access (sometimes after requiring proof of training).
Also, monitors may get only partial access to data because they (or even the site staff) don’t know exactly what roles and databases to request access for, and perhaps the data are partitioned (due to institutional departments, affiliated providers, or other reasons). As a result, there may be medical history, eligibility criteria, or medications that do not get into the clinical trial data.
ICH GCP 4.1.4. requires direct access. However, the level/scope of access may differ depending on who is requesting access, and what a PI may feel is “relevant” to the clinical trial. A radiology department may be considered a different “legal entity,” and the institution may not allow direct access, again causing delays in thorough monitoring, and an incomplete picture of the subject.
Deadly Sin #4: Contemporaneous
(C) A lack of Contemporaneous documentation and sign-off is always a sin. The old adage in research that “if you didn’t document it, it didn’t happen” should probably have a corollary that “if you didn’t document it WHEN it happened, it didn’t happen.” Of course backdating is a cardinal sin, but almost as bad is a late entry (i.e. days or weeks later). Even though it is done with the best of intentions (to ensure the record is complete) – it still brings about suspicion and additional explanation. How much longer would it have taken to sign/date when the original entry was made? Rarely in the clinical research setting would taking a split second longer to initial/date an entry result in any additional risk to patient safety, and it certainly takes five times longer to make a late entry and explain why it was late. Additionally, medical staff are trained throughout their schooling and career to document this way – so why is it so difficult within a clinical trial?
The problem here is there’s often a disconnect between normal practice versus what a sponsor requires in a clinical trial. Additionally, EMR systems are usually not well-designed for completion during the encounter with the patient. Be with the patient while the patient is there. Patients have questions, patients are not always compliant, and patients have adverse events. Site staff can certainly be forgiven for taking care of the patient now, and doing the documenting in the EMR later. What patient wants to be ignored while the endless EMR screens are being completed? The PI/Sub-I may review lab data in almost “real-time” within the EMR, but then the sponsor then asks for hard copy sign-off later. The printed reports can sit in an inbox on someone’s desk waiting for review and sign-off, making the contemporaneousness unverifiable. How does a monitor, auditor, or inspector determine if that lab report was reviewed during the patient visit or not?
Deadly Sin #5: Original
(O) Original – no, not original sin, but the sin of originality. Less and less common is anything original. Forty to fifty years ago, if something was typed, it was special, original, important. Now the opposite is true – something handwritten draws the eye for closer inspection (pun intended).
Photocopy machines, printers, faxes, and scanners have made the replication and duplication of paper so fast and easy (and of great quality) that it is now almost impossible to distinguish between them. Is this the copy of a fax, or the fax itself? When was it printed from the EMR, and were changes made and it re-printed and filed somewhere else (such as in the dreaded “shadow chart”)? Is the site staff really willing to spend the time to allow an over-the-shoulder spot check to ensure that what is being printed truly is a certified copy? We are no longer able to determine what is the original (Think of Chicago’s song “Does anyone really know what time it is?”) and what is a copy, a fax, or a re-print. When sites print out of EMR for an upcoming monitoring visit, are they checking for original hand notations or changes from the last visit?
Are they including an audit trail, or just assuming that “it is what it is?” Does the monitor check, or simply accept whatever the site gives them to use for monitoring?
Deadly Sin #6: Accuracy
(A) Accuracy of the data in clinical trials is paramount, but here again we have a problem. The EMR has lots of pre-programmed word lists – but is the study coordinator in a hurry? Was it metoprolol tartrate or metoprolol succinate; and what was the dosage? The patient was saying they take a “heart pill” that is round and is either white or tan in color.
Is a nurse assistant from a different clinic within the same EMR system entering the information? Will the study coordinator bother to re-question the patient? Would the patient know the difference? Maybe the salt version of metoprolol is not critical, but what about other meds? Did the patient say Tramadol or Trazadone?
Deadly Sin #7: Checklists
(C) Checklists – Yes, the last C usually stands for complete – but verbal feedback paraphrased from a recent FDA inspection reveals current concerns with the use of checklists, especially in the area of informed consent. The inspector made the following suggestions:
- All re-consents need the same documentation describing what happened
- The expectation is that specific documentation of the Informed Consent Process is in the physician Progress Note
- The use of templates that are not edited and specific is insufficient
- It should always be noted that a signed and dated copy of the informed consent was given to the patient
- It is obvious when a template or smart phrase has been utilized because then every patient’s note looks the same
So in this example the use of a checklist (intended to ensure that all the steps in the informed consent process took place) is actually viewed in a way as incomplete, unoriginal, and not attributable when technically it is. The inspector wants more of a picture of the story of the process itself. Understandable, perhaps, but it goes beyond ALCOA. One could have an original checklist, completed contemporaneously, attributed to the person who conducted the discussion, legible and accurate, but still considered lacking.
Is this all leading to a still longer acronym? Should D (descriptive) and R (realistic) be added to the end (or perhaps to the beginning) for ease of recall; DRALCOAC? Is more always better?
Avoiding the 7 Deadly Sins
Here are some overall concepts to consider in avoiding the seven deadly sins discussed above:
- Strive for simplicity. Remember that the more elaborate the process (steps, people, systems), the more opportunities for error. Consider how less complexity can still meet the regulatory requirements.
- Go with the flow. Each clinic has a “normal” way of doing things. Develop the requirements around the normal flow, not vice versa. The more habits are forcibly changed only for the clinical trial (and not how the regular clinic patients are treated), the more unnatural and the less the compliance. That said, however…
- Develop good habits. Employees must wash hands, right? It’s difficult to wash them when you’ve left the bathroom. Sign and date entries before you hand off that paper to the next person; but also look for the signature and date before you accept it (built-in QC) from someone else. It’s easier to catch it now than to chase down someone later for a signature.
- Don’t be too minimalist with your checklists. Sure, they are useful in ensuring everything was done and in the right sequence, but not to the exclusion of full documentation. Seldom is the eligibility criteria page of the protocol the actual source. Where is that lab report, where is the medical history coming from? Rather than robotically stating “the consent was obtained prior to any study procedures being conducted” (of course, what else would be written there?) a consent checklist could ask what time was the consent given, and what (and when) was the first study procedure conducted, and by whom. Was the spouse, child, or other family member there during the consent process? What did they ask during the discussion?
- Provide adequate training, support, and oversight of monitoring. Monitoring is not checking paperwork, but ensuring that the study is conducted properly. It’s not just what’s written down, but also what isn’t.
- Be able to explain (and defend) what you do and why you do it. This goes a long way in giving the auditor or inspector confidence in your work. If you have discovered a deficiency or error, put a CAPA in place and assess how it worked. Be concise and detailed. It’s usually less important that you made a mistake, than what you did to correct and prevent
- Finally, as Alexander Pope was not quoted as saying. “To err is human; to document is divine.”