Conduct, ethics and safety provisions in the plan
A clinical investigation plan for a medical device sets out how the study is governed once it runs: how its data are handled and verified, how deviations are dealt with, the statements of compliance and ethics, how consent is obtained, the definitions and reporting of safety events, the arrangements for stopping the study and following up its subjects, and the policy for publishing its results.
Data handling, monitoring and source verification
The plan describes the procedures for data management: data review, database cleaning, and issuing and resolving data queries; the verification, validation and security of electronic clinical data systems where applicable; data retention, with the retention period specified; and any other aspects of clinical quality assurance as appropriate. Where data entered in the case report form are to count as source data, the plan says so, as is sometimes done for height, weight and questionnaires.
The plan also specifies how much source document verification is to be done. Source document verification compares the entries in the case report form with the patient files, to make sure that what is in the case report form is reflected in the source document. Sites do not simply do their job without mistakes: they make a never-ending variety of them, from height entered in the weight field and the reverse, to miscalculated units, to overlooked medical conditions. It may be worth spending some time working out how much verification is needed. Every case report form likely holds questions that are not needed and are kept just in case. The plan can, for example, state 100% verification for the medical history, the primary and secondary endpoints and all adverse events, 10% verification of the remaining questions, and only a consistency check for the rest.
The monitoring plan should be kept separate from the clinical investigation plan and referenced from it, because it usually changes frequently. The clinical investigation plan gives only a general outline of the monitoring to be followed, including access to source data and the extent of source document verification planned.
Deviations from the plan
The plan describes the procedures for amending it and for handling deviations from it: a statement that the investigator may not deviate from the plan except in an emergency, the procedures for recording, reporting and analysing deviations, the notification requirements and their timeframes, corrective and preventive action, and the criteria for disqualifying a principal investigator.
There will always be non-compliant sites, and they may be very important key opinion leaders, so a cleverly worded deviation section gives the clinical research team enough power to put pressure on them. One wording for it cites the standard: according to ISO 14155, the sponsor shall consider terminating or suspending the participation of a particular investigation site or investigator if monitoring or auditing identifies serious or repeated deviations on the part of an investigator.
Compliance and ethics statements
The plan states that the investigation will be conducted in accordance with the Declaration of Helsinki, ISO 14155 and any regional or national regulations as appropriate, and that it will not begin until the required approval or favourable opinion from the ethics committee or regulatory authority has been obtained, where appropriate. It also states that any additional requirements imposed by the ethics committee or regulatory authority will be followed as appropriate, and what type of insurance will be provided to the subject, where appropriate, which is usually patient insurance.
The Declaration of Helsinki requires that research conform to generally accepted scientific principles and rest on thorough knowledge of the scientific literature. A plan usually shows this in its introduction, which gives an overview of the current literature and the state of the art. The Declaration also requires that the study design and performance be clearly described in a research protocol, which should contain a statement of the ethical considerations involved, information on funding, institutional affiliations, other potential conflicts of interest, and incentives for subjects. On incentives, ISO 14155 allows compensation for subjects only within a range that does not unduly influence a person, so, for example, travel expenses to the follow-up visit can be compensated, but nothing more. The Declaration's provisions for treating and compensating subjects harmed as a consequence of their participation basically mean a paragraph on patient insurance.
Under the Declaration, the protocol should also describe arrangements for post-study access by study subjects to an intervention identified as beneficial in the study, or access to other appropriate care or benefits. For example, if a new insulin drug is tested and the patient is doing well with it, going back to the old treatment could be harmful, so if the new drug is not commercially available at the end of the study, the plan describes what the patient should do in between. Companies usually provide those patients with the medication to cover the gap between the end of the study and the commercial availability of the product. The plan shall also specifically address what medical care, if any, will be provided to subjects after the investigation has been completed. That is basically what the Declaration of Helsinki asks: continued access to a treatment until it is commercially available.
Informed consent and vulnerable populations
The plan describes the general process for obtaining informed consent, the process for providing subjects with new information as needed, and the consent process where a subject is unable to give consent.
Vulnerable populations are people who cannot provide consent or who are in a dependent relationship, and they can be included in a study only if the disease cannot be assessed in a non-vulnerable group. So a treatment against dementia has to be tested in mentally impaired people, a treatment for a very specific disease of newborns in newborns, and a treatment for unconscious shock patients in unconscious shock patients. A heart valve study, by contrast, can be done in a non-vulnerable population: it does not need patients with severe dementia, who can receive a commercially available heart valve instead.
Safety definitions
The plan defines adverse event, adverse device effect, serious adverse event, serious adverse device effect and unanticipated serious adverse device effect, and also device deficiency, a term often misunderstood or not known at all. The adverse event definitions themselves are basically the same as for drug studies, but in device research the term device deficiency is also relevant.
A device deficiency is an inadequacy of a medical device with respect to its identity, quality, durability, reliability, safety or performance, including malfunctions, use errors and inadequate labelling, such as a balloon rupture, unsterile packaging or kinking of the device. It is a device deficiency whether or not it harmed the patient: the term covers product issues that did not necessarily lead to an adverse event. A device deficiency that caused no harm but could have done had circumstances been less fortunate receives the same safety reporting as a serious adverse device effect.
For example, the packaging of a heart valve may not be good enough and the valve becomes unsterile. An investigator who sees a bacterial smear on the valve will not implant it, and the patient comes to no harm. Had the investigator not detected it, for instance because there was no smear or because he was not attentive, it could have caused serious harm, because the patient could have developed sepsis or endocarditis. Likewise, a balloon rupture usually harms nobody, but in theory part of the balloon could embolise to the brain and cause a stroke.
Classifying safety events is easy in an established area, but in a completely novel area, which is rarely the case, it can get more sophisticated. When the thirtieth patient worldwide had received a transcatheter heart valve, paravalvular leaks were being seen. In conventional heart valve replacement, any paravalvular leak is undisputedly unwanted and counts as an adverse event, even a small leak with no medical consequence. So whether to report a grade 1+ leak in the new therapy as an adverse event, when it did not really harm the patient, was unclear. It would later have been classified as a device deficiency, but that term did not exist at the time.
Safety reporting and oversight committees
The plan sets the reporting timeframe, such as 24 hours for serious adverse events. It describes the reporting process for adverse events, device effects and the other categories, lists the foreseeable events and anticipated adverse device effects with their likely incidence, mitigation and treatment, gives emergency contact details for reporting serious adverse events, and gives information on the data monitoring committee where there is one.
A data monitoring committee in this sense is the same as a data safety monitoring board: usually three to five independent physicians who supervise the adverse event numbers and can recommend stopping the study. Not every study has one, but early studies of novel or invasive devices often do.
A clinical events committee, which falls under the data monitoring committee, adjudicates adverse events, and the plan specifies to what extent it does so. Sometimes it adjudicates all adverse events, but mostly it adjudicates the events of special interest to the study, such as the endpoint-related events or, for transcatheter heart valves, the endpoints or events related to the Valve Academic Research Consortium.
Stopping the investigation and following up subjects
The plan gives the criteria and arrangements for suspending or prematurely terminating the investigation or a single centre, for example on the recommendation of a data safety monitoring board or for site non-compliance. Where the investigation involves blinding or masking, the plan gives the criteria for access to and breaking of the code in that event. It also gives the requirements for following up subjects: enrolment may stop during the study because the device does not meet its promises, and the patients who received an implant should still be followed up.
In a drug study, follow-up continues for about 30 days after the last follow-up visit, to see whether the medication may still be in the body after the therapy stops. A device study often has a treatment or an implant followed by the follow-up assessment, so it can make a hard cut after the last follow-up assessment: a valve with a five-year follow-up assessment does not need five years plus 30 days. A device that still has effects after the last follow-up assessment is very rare. An implant remains in place, but after five years the patient is followed under routine standard of care.
Publication policy
The publication policy states whether the results will be submitted for publication, and under what conditions they will be offered for publication, for example that the first publication will be a multicentre publication once the primary endpoint has been reached. Publishing study results is generally mandatory, and in theory posting them only on a specific website such as ClinicalTrials.gov would meet that. In practice, though, publication of the research is always wanted.