Hoel 2006.
| Methods | Five‐arm cluster randomised trial | |
| Participants | 272 participants engaged in focus groups pre‐survey; 2505 questionnaires distributed to workers from 5 public sector organisations at pre‐intervention stage; return rate of 41.5% (N = 1041 questionnaires) Gender: 36.2% male; 63.8% female. 150 workers (in total) allocated to one of five intervention groups in each organisation (including one group that acted as a control and did not have an intervention) Post‐intervention 2499 questionnaires distributed, with a return rate of 35.4% (N = 884 questionnaires) Gender: 36.4% male, 63.6% female Age: mean age of participants at both time points was 43 years Eight focus groups six months post‐intervention; number of participants not stated Geographical Setting: London & North & South of England |
|
| Interventions | Programme of interventions: 1. One policy communication session of 30 minutes duration (we judged this at organisation/employer level) 2. One policy communication session of 30 minutes and one stress management training session of three hours duration (at organisation/employer and job task levels) 3. One policy communication session of 30 minutes and one negative behaviour awareness training session of three 3 hours duration (at organisation/employer and individual/job interface levels) 4. One day‐long event comprising of a policy communication session, stress management and negative behaviour awareness training (at organisation/employer, job task and individual/job interface levels) |
|
| Outcomes | Self‐report of bullying using Bullying Risk Assessment Tool (BRAT); witnessing of bullying, sickness absence, measured approximately six months post‐intervention | |
| Notes | Broad theoretical underpinning: intervention designed using literature review and knowledge of local context Funding source: British Occupational Health Research Foundation (BOHRF) Declarations of interest: none stated We requested raw data from the authors to conduct proper analysis on it but they did not respond. |
|
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Blinding Subjects | High risk | no blinding |
| Blinding Outcome Assessors | High risk | no blinding |
| Retrospective unplanned subgroup analyses | Low risk | no evidence of data dredging |
| Follow‐up | Unclear risk | approximately six months; based on unmatched self‐report of behaviour |
| Statistical tests | Unclear risk | appropriate but mainly descriptive |
| Compliance | Unclear risk | problems with compliance reported; "unwillingness/resistance on behalf of participants to engage" |
| Outcome measures | High risk | self‐reported outcome measures susceptible to social desirability; descriptive & qualitative data reported; " showing increases in scores as +; decreases as ‐; and no changes as 0". |
| Selection bias (population) | Unclear risk | employees from different types of public sector organisations |
| Selection bias (time) | Low risk | all participants recruited within the same timeframe |
| Randomisation | Low risk | cluster randomisation |
| Allocation concealment | Unclear risk | Unable to determine (UTD), assignment not reported |
| Adjustment for confounding | High risk | Influencing factors have been described but not taken into account |
| Incomplete outcome data | Unclear risk | loss indicated but not possible to determine if taken into account |