Table 2. Quality of included means restriction systematic reviews (AMSTAR-2).
| Intervention type studied | Author | Summary of quality assessment* | Domains of critical weakness and areas of strength |
|---|---|---|---|
| Roads | Okolie et al36 | High | Areas of critical weakness: none The assessment could only be partially completed because the review did not identify any studies meeting the inclusion criteria. Therefore, some parts of the assessment were conducted based on what the study protocol had planned. |
| Jumping | Okolie et al16 | High | Areas of critical weakness: none. |
| Jumping | Pirkis et al2 | Critically low † (two areas of critical weakness) | Areas of critical weakness: no preregistered protocol; no reasons given for excluding each potentially relevant study (a summary of reasons for exclusion was provided). Areas of non-critical weakness: no duplication of data extraction and no formal assessment of study quality/risks of bias. Notes: heterogeneity and bias due to study designs and limitations of observational nature of studies were discussed. |
| Railway | Barker et al21 | Critically low (two areas of critical weakness) | Areas of critical weaknesses: no preregistered protocol, no reasons given for excluding potentially relevant studies. No systematic quality/risk of bias assessments and no duplication of searches or data extraction. |
| Poisoning | Lim et al27 | Moderate–high | No areas of critical weakness. Areas of non-critical weakness: limited detail on study interventions and settings, limited discussion of publication bias. |
| Paracetamol poisoning | Morgan et al28 | Critically low † (two areas of critical weakness) | Areas of critical weakness: no preregistered protocol, no reasons given for excluding each potentially relevant study. The study includes discussion of sources of bias arising from study design and heterogeneity. Non-critical weaknesses: no duplication of data extraction and searching and no formal assessment of study quality/risks of bias. |
| Pesticide poisoning | Gunnell et al4 | Moderate (one area of critical weakness) | Areas of critical weakness: while a summary of reasons for excluding potentially relevant studies was provided, reasons given for excluding each study were not available. Notes: there were several areas indicating a high level of confidence in quality, including a preregistered protocol, risk of bias assessed using Cochrane Effective Practice and Organisation of Care criteria and duplication of data extraction and reviewing. |
| Pesticide poisoning | Reifels et al30 | Critically low (two areas of critical weakness) | Areas of critical weakness: no preregistered protocol and no clear criteria for/record of excluding potentially relevant studies. Notes: risks of bias were discussed but no systematic quality/risk of bias assessment was conducted. |
| Jumping | Pirkis et al17 | Critically low (two areas of critical weakness) | Areas of critical weakness: no preregistered protocol and no reasons given for excluding each potentially relevant study (however, a summary of reasons for exclusion was provided). Non-critical weakness: no duplication of data extraction and searching. |
| Jumping | Cox et al22 | Critically low (two areas of critical weakness) | Areas of critical weakness: no preregistered protocol and no reasons given for excluding potentially relevant studies. Non-critical weaknesses: risks of bias were discussed, though no systematic quality/risk of bias assessment was conducted. No duplication of searches or data extraction. |
| Multiple: Jumping Domestic gas, motor vehicle gas and charcoal poisoning |
Linskens et al19 | Moderate (one area of critical weakness) | Areas of critical weakness: no record of reasons for excluding potentially relevant individual studies. Notes: the study included a preregistered protocol, comprehensive quality assessment and duplication of searches and data extraction. No justification for focussing on a specific time period. |
| Multiple: Jumping, pesticides, domestic gas, motor vehicle gas and charcoal poisoning |
Ishimo et al20 | Moderate (one area of critical weakness) | Areas of critical weakness: no reasons given for excluding individual potentially relevant studies. Notes: study quality was assessed using the Effective Public Health Practice Project Quality Assessment Tool. Searches and data extraction were duplicated and interventions were described in detail. |
| Multiple: Jumping, pesticides, domestic gas, motor vehicle gas and charcoal poisoning, firearms, medication poisoning |
Zalsman et al23 | Critically low (two areas of critical weakness) | Areas of critical weakness: no preregistered protocol, excluded potentially relevant studies, including those rated as ‘very low evidence’ were not listed. Notes: a group of 18 suicide prevention experts worked to reach consensus on the evidence ratings, which studies to include and the study conclusions. The Oxford Centre for Evidence-Based Medicine criteria were used to rate the level of evidence, though this was limited to identifying the study design rather than the quality of the study. |
| Multiple: Jumping, pesticides, domestic gas, motor vehicle gas and charcoal poisoning, firearms, medication poisoning |
Mann et al24 | Critically low † (two areas of critical weakness) | Areas of critical weakness: no preregistered protocol, though the study parameters were designed with a group of 15 suicide experts. Absence of reasons for excluding potentially relevant studies. Notes: The Oxford Centre for Evidence-Based Medicine criteria were used to rate the level of evidence, though this was limited to identifying the study design rather than the quality of the study. The methodological quality and risks of bias of the studies were discussed. |
| Multiple: pesticides, domestic gas, motor vehicle gas and charcoal poisoning, firearms |
Mann et al31 | Critically low (two areas of critical weakness) | Areas of critical weaknesses: no preregistered protocol, though the study parameters were designed with a group of 15 suicide experts, and no list of excluded potentially relevant studies. Notes: the authors focused on RCTs and epidemiological time-series study designs. The methodological quality and risks of bias of the studies were discussed, but no formal assessment was conducted. The list of included means restrictions studies was not provided. |
| Multiple: Pesticides, firearms, medication poisoning |
Robinson et al29 | Critically low (two areas of critical weakness) | Areas of critical weakness: no preregistered protocol and no list of individual potentially relevant studies that were excluded, though summary reasons for exclusion were provided. Notes: duplication of searches and data extraction was conducted. |
| Pesticide poisoning | Bailey et al32 | Critically low (two areas of critical weakness) | Areas of critical weakness: no preregistration or protocol (though this was an update to Robinson et al and methods were based on the original systematic review) and no list of individual potentially relevant studies that were excluded. Notes: risks of bias were assessed by the Cochrane risk of bias tool. |
| Firearms | Hahn et al34 | Critically low † (two areas of critical weakness) | Areas of critical weakness: no preregistration or protocol, no list of individual potentially relevant studies that were excluded. Notes: a thorough assessment of bias and study quality, and duplication of data extraction and quality assessment was conducted. |
| Multiple: Jumping, firearms |
NICE18 | Moderate (one area of critical weakness) | Areas of critical weakness: no statistical tests for publication bias in the meta-analysis. Notes: searches were comprehensive, with each excluded study recorded and quality assessments of included studies conducted. |
| Pesticide poisoning | Rubbo et al33 | High | Areas of critical weakness: none. Notes: The review was preregistered; screening and study assessment were conducted in duplicate; and excluded studies were described. The risks of bias in primary studies were assessed using a Cochrane-recommended modified version of the risk of bias criteria for interrupted time series studies. |
*
According to the AMSTAR-2 guidance, studies were downgraded to ‘moderate’ if 1 domain of critical weakness was identified, and to ‘critically low’ if 2 or more domains of critical weakness were identified.
†
The study was published prior to, or within 2 years of, the launch of PROSPERO and therefore would not be expected to have preregistered a protocol.
AMSTAR-2, A MeaSurement Tool to Assess systematic Reviews-2.