Explanation

Evaluation of harms plays a key role in monitoring the well-being of participants during a trial and in enabling appropriate management of adverse events. Documentation of trial-related adverse events also informs clinical practice and the conduct of ongoing and future studies. Use of the term “harms” is preferred over “safety” to better reflect the negative effects of interventions,(274) although we recognise that regional and organisational practices vary in the terminology used.

An adverse event refers to an unwanted occurrence during the trial, which may or may not be causally related to the intervention or other aspects of trial participation.(274, 275) This definition includes unfavourable changes in symptoms, signs, laboratory values, or health conditions. An adverse effect is a type of adverse event that is attributed to being caused by the intervention. In the context of clinical trials, it can be difficult to accurately attribute causation for an adverse event experienced by an individual participant.(276)

How harms outcomes are assessed will impact the results obtained.(277-280) Harms outcomes can be either systematically assessed (i.e., solicited by active or targeted surveillance) or non-systematically assessed (i.e., unsolicited or spontaneous declaration; passive surveillance). To increase statistical power, harms outcomes are sometimes grouped in a composite outcome (e.g., cardiovascular events).

Specifying the methods of assessing, grouping, and analysing harms in the protocol helps to identify and deter the selective reporting of results in publications and registries.(281, 282) Substantial discrepancies have been observed between protocol-specified plans for harms data collection and reporting, and what is described in publications of completed trials.(283)

The protocol should list the systematically and non-systematically assessed harms outcomes that will be collected, as well as their time points of assessment and overall surveillance time frame. It is important to state who will perform the assessment, coding, severity grading, and grouping, and whether those individuals will be blinded to the allocated trial group.

For each systematically assessed harms outcome, the information outlined in item 16 should be reported, including the outcome definition, instrument used (e.g., name of a validated questionnaire), analysis metric (e.g., final value), method of aggregation (e.g., proportion), and the time point of interest for analysis. The data collection methods should also be described as outlined in item 25a.

For non-systematically assessed harms outcomes, protocol authors should report how data will be collected (item 25a), including any standardised questions used to solicit information. It is important to describe the process for coding adverse events (e.g., using standardised terminology) and grading their severity, including the coding system (e.g., Medical Dictionary for Regulatory Activities) and severity grading system (e.g., Common Terminology Criteria for Adverse Events).

Harms outcomes are often grouped by seriousness, severity, body system, discontinuation of interventions due to harms, and causality attribution.(277) If aggregation of adverse events is planned, then protocol authors should define the grouping categories. If harms will be categorised as being causally related to the intervention or not, authors should describe the attribution methods, including who will make the causal attribution and whether they will be blinded to the assigned trial group.(284)

If applicable, the protocol should address how important adverse events will be reported to relevant groups (e.g., sponsor, research ethics committee/institutional review board, data monitoring committee, regulatory agency), a process that is defined by local regulation.(285) Key considerations determining the need to report an adverse event to external groups include its severity, potential causality, and whether it represents an unexpected event.

Summary of key elements to address

For each systematically assessed harm (active/targeted surveillance):

· Definition and measurement (e.g., name of validated questionnaire)

· Where appropriate, the metrics, method of aggregation, and time point of interest for analysis

· Procedures for harms assessment , including:

o Who will do the assessment, and whether they will be blinded to the allocated trial group

• Assessment time points and overall time period for recording harms

For each non-systematically assessed harm (passive surveillance):

· How data will be collected

· Assessment time points and overall time period for recording harms

· Process for coding each adverse event and grading its severity, including:

• Who will do the coding and severity grading, and whether they will be blinded to the allocated trial group

• Which coding and severity grading systems will be used, if any

For grouping of harms by seriousness, severity, body system, discontinuation of intervention (due to harms), and causality:

· Definitions of grouping categories

· Who will do the grouping, and whether they will be blinded to the allocated trial group

If relevant:

· Process of reporting important adverse events to applicable groups (e.g, sponsor, regulator, data monitoring committee)