Table 2.
Definitions of reporting items
| Item and number | Definition of “yes” |
|---|---|
| Title | |
| 1a) Specifically mentions “harms” or other related term | Should contain a word or phrase related to harm, such as “adverse effect,” “adverse event,” “complications,” and “risk” |
| 1b) Clarifies whether both benefits and harms are examined or only harms | Mentions whether benefits are reviewed or not |
| 1c) Mentions the specific intervention being reviewed | Mentions the intervention being reviewed |
| 1d) Refers to specific patient group or conditions (or both) in which harms have been assessed | Clearly states the specific group of patients or conditions being reviewed |
| Abstract | |
| 2a) Specifically mentions “harms” or other related terms | Should contain a word or phrase related to harm, such as “adverse effect,” “adverse event,” “complications,” and “risk” |
| 2b) Clarifies whether both benefits and harms are examined or only harms | Mentions whether benefits are reviewed or not |
| 2c) Refers to a specific harm assessed | Clearly states the adverse event being reviewed. General descriptions are accepted, for example, “cardiovascular events” or “maternal complications.” Also acceptable if the report is searching for any adverse event and it does not refer to a specific harm |
| 2d) Specifies what type of data was sought | Clearly names the kind of studies searched (for example, randomised controlled trials, cohort studies, case reports only, all types of data) |
| 2e) Specifies what type of data was included | Clearly names the kind of studies included in the analysis or results (that is, randomised controlled trials, cohort studies, case reports, all types of data) |
| 2f) Specifies how each type of data has been appraised | Clearly states the method of quality appraisal for included studies (that is, Jadad score, Cochrane risk of bias tool, Newcastle-Ottawa scale) |
| Introduction | |
| 3a) Explains rationale for addressing specific harm(s), condition(s) and patient group(s) | Reports reasons for proceeding with the systematic review |
| 3b) Clearly defines what events or effects are considered harms in the context of the intervention(s) examined | Clearly states the adverse event being reviewed. General definitions are acceptable (for example, “cardiovascular events” or “maternal complications”). Also acceptable if the report is searching for any adverse events, as in a scoping review, and does not refer to a specific harm |
| 3c) Describes the rationale for type of harms systematic review done: hypothesis generating versus hypothesis testing | Clearly states whether a specific adverse event is being reviewed (hypothesis testing review; for example, cardiovascular deaths) or the authors are searching for any adverse events related to the intervention (hypothesis generating) |
| 3d) Explains rationale for selection of study types or data sources and relevance to focus of the review | Clearly states which study designs are included (for example, randomised controlled trials, cohorts, case reports). Rationale is not necessary for “yes” |
| Objectives | |
| 4a) Provides an explicit statement of questions being asked with reference to harms | Clearly defines, preferably at the end of introduction section, the intervention and the adverse events being reviewed. For this review, it was considered acceptable if intervention and outcome were clearly stated |
| Methods | |
| Protocol and registration | |
| 5a) Describes whether protocol was developed in collaboration with someone with clinical expertise for field or intervention under study | Mentions whether a protocol was developed previously. For this review, it was not necessary to state the clinical expertise of who developed it |
| Eligibility criteria | |
| 6a) Clearly defines what events or effects are considered harms in the context of the intervention(s) examined | Provides a clear definition of the adverse event being reviewed. If a general description was provided previously (that is, “cardiovascular events”), now the specific events need to be defined |
| 6b) Specifies type of studies on harms to be included | Defines which kind of study designs will be included (for example, randomised controlled trials, cohort studies, case-control studies) |
| Information sources | |
| 7a) States whether additional sources for adverse events were searched (for example, regulatory bodies, industry); if so, describes the source, terms used, and dates searched | Reports whether any other sources of adverse events (other than regular peer reviewed journals) were searched. For this review, it was not necessary to have the terms used and dates searched |
| Search | |
| 8a) Presents the full search strategy if additional searches were used to identify adverse events | Reports search strategy used. Adverse events terms and databases searched |
| Study selection | |
| 9a) States what study designs were eligible and provide rationale for their selection | Reports the study designs included in the review (for example, randomised controlled trials, cohort studies, case-control studies). No rationale is required |
| 9b) Defines if studies were screened on the basis of the presence or absence of harms related terms in title or abstract | Reports whether the study screening is based on the report of adverse events or not (for example, review of mortality associated with antiglycaemic drugs; only studies reporting on mortality were included) |
| Data collection process | |
| 10a) Describes method of data extraction for each type of study or report | States whether a data extraction form is used, and how it is done (in duplicate, checked by a second author) |
| Data items | |
| 11a) Lists and defines variables for which data were sought for individual therapies | Reports variable(s) sought for the intervention reviewed (for example, whether the outcome is kidney failure, what variable was used to define kidney failure—creatinine level, creatinine clearance, need of dialysis) |
| 11b) Lists and defines variables for which data were sought for patient underlying risk factors | States whether any potential patient risk factors or confounders are sought (for example, age, sex, comorbidities, previous events) |
| 11c) Lists and defines variables for which data were sought for practitioner training or qualifications | States any variable(s) sought for healthcare personnel (practitioner) (for example, relevant degree, years of experience, other qualifications) |
| 11d) Lists and defines harms for individual therapies | Provides the definition for the harm(s) sought (for example, side effects of propranolol and atenolol defined as severe bradycardia (heart rate ≤45 beats per min) and severe hypotension (systolic blood pressure <80 mm Hg) |
| Risk of bias in individual studies | |
| 12a) For uncontrolled studies, describes whether causality between intervention and adverse event was adjudicated and if so, how | Specifies whether review authors adjudicate if the intervention can cause the harm (for example, use of Bradford Hill criteria for causality32 |
| 12b) Describes risk of bias in studies with incomplete or selective report of adverse events | Describes how studies not reporting adverse events are handled regarding risk of bias. For example, adverse events reported in a randomised controlled trial investigating glucose control in type 2 diabetes included hypoglycaemia but not mortality. Did the authors assess whether the event did not occur (zero deaths), was not measured (and may have occurred), or was measured but not reported |
| Summary measures | |
| 13a) If rare outcomes are being investigated, specifies which summary measures will be used (for example, event rate, events or person time) | Defines the summary measures used for rare events |
| Synthesis of results | |
| 14a) Describes statistical methods of handling with the zero events in included studies | Clearly defines how studies with no adverse events reported are reported and analysed (for example, when zero was an outcome for a 2×2 table, reviewers added a 0.5 value to it) |
| Additional analysis | |
| 16a) Describes additional analysis with studies with high risk of bias | Describes whether studies with high risk of bias are analysed separately (for example, subgroup analysis with high and low risk of bias studies analysed separately) |
| Results | |
| Study selection | |
| 17a) Provides process, table, or flow for each type of study design | Provides a table (or text) containing included studies and a clear reason for exclusion of studies |
| Study characteristics | |
| 18a) Reports study characteristics, such as patient demographics or length of follow-up that may have influenced the risk estimates for the adverse outcome of interest | Reports any potential confounder or patient risk factors that can affect the outcome (adverse event) |
| 18b) Describes methods of collecting adverse events in included studies (for example, patient report, active search) | Reports how adverse events are investigated in included studies (for example, voluntary report, active search) |
| 18c) For each primary study, lists and defines each adverse event reported and how it is identified | Clear definition of adverse event under investigation and how it is identified (method of measurement, assessment or identification) |
| Conclusion | |
| 26a) Provides balanced discussion of benefits and harms with emphasis on study limitations, generalisability, and other sources of information on harms | Clearly discusses the benefits of the interventions and the harms identified in the review |