Author, date and country | Patient group | Study type (level of evidence) | Outcomes | Key results | Study Weaknesses |
---|---|---|---|---|---|
Bullard MJ 1993, Taiwan | 96 adult patients presenting to an ED in Taiwan following known or perceived self-harm. | Prospective cohort study using a "suicide risk evaluation form" including a subjective assessment and formal SADPERSONS scale. | Completed suicide and disability following further self harm.Compliance of healthcare staff with recommended management plan according to measured risk | The SADPERSONS scale underestimated suicide risk and placed patients in a lower risk group than subjective assessment.Healthcare staff rarely complied with the recommended management plan. | Pattern of self-harm may be different to other countries (e.g. main cause of completed suicide was agricultural poisoning). Most patients were sent home, even those considered intermediate and high risk. Follow-up was incomplete. |
Hockberger RS, Rothstein RJ 1998 USA | 100 consecutive adult patients attending Los Angeles County Harbor-UCLA Medical Center Emergency Department | Prospective cohort study. The assessment of ED medical staff, using a modified SADPERSONS scale (MSPS), was compared to an immediate formal psychiatric evaluation, from which a second MSPS score and disposition decision (admit or discharge) were determined. | Comparison between the ED assessment and formal psychiatric evaluation, including the MSPS score and disposition decision. Patients were also followed up to determine compliance with future appointments, further hospitalization and completed suicide. | There was reasonable agreement between the ED and psychiatry MSPS scores, but neither accurately predicted patient disposition. A weighted MSPS was developed and prospectively validated in a further 82 patients, with a sensitivity of 31% and specificity of 94% for hospitalization. No patient died during 6 month follow-up. | Small sample sizes, particularly in the second validation cohort. The outcome of most clinical interest (completed suicide) did not occur during the study. |
Cochrane-Brink KA, et al, 2000 Canada | 55 non-consecutive adults requiring psychiatric evaluation in a Canadian ED setting. | Semi-structured interviews using the modified SADPERSONS scale (MSPS) and five other scales designed to assess the risk of further self-harm | Admission to hospital as a result of assessed suicide risk was used as a proxy measure for further self harm or completed suicide. | All clinical rating scales had high sensitivity (92 -100%) and negative predictive value (96% - 100%), but lower specificity (38% to 90%). The MSPS predicted hospital admission (p<0.5), but other scales performed more effectively. | Small and non-consecutive sample excluding intoxicated patients. No follow-up to determine actual risk of further self-harm. |
Herman SM 2006 USA | 271 veteran patients at a large Midwest Veterans Affairs Medical Center (VAMC) | Semi structured interviews using the SADPERSONS scale. | Results of SADPERSONS scale were compared to other assessments of suicide risk | A cutoff score of five (recommended by the scale's developers) resulted in failure to identify 14% of individuals considered to be actively suicidal, with a false positive rate of 87%. | Veterans population only, so of limited generalisability. Single evaluation relying on self-reported data. No follow-up to determine actual risk of further self-harm |