![[ Visit Client Website ]](images/banner.gif)
Rural communities in the U.S. have disproportionately high suicide rates, but limited data are available on best methods for risk screening in rural primary care. This study of rural adults attending primary care appointments evaluated the sensitivity and specificity of the Ask Suicide-Screening Questions (ASQ) tool, compared it to the ninth item on the Patient Health Questionnaire (PHQ-9), explored whether screening for opioid misuse and pain helped identify individuals at risk for suicide, and gathered follow-up data on suicidal symptoms.
Methods
Adults attending appointments at a federally qualified health center (FQHC) in West Virginia were asked to complete an electronic survey followed by a telephone survey one month later. Screening measures included the ASQ Tool, the PHQ-9, the Adult Suicidal Ideation Questionnaire (ASIQ), the PEG Pain Screen, and a modified version of the Alcohol, Smoking, and Substance Involvement Screening Test (ASSIST). The sensitivity, specificity, positive predictive value, and negative predictive value of the ASQ and PHQ-9 item were calculated using the ASIQ as the gold standard. To compare the sensitivity and specificity of the ASQ and the PHQ item, McNemar’s test of proportions was used. Bivariate statistics and linear regressions were used to evaluate whether PEG and ASSIST scores were associated with ASIQ results.
Results
The initial survey sample included N = 214 participants. Of these, 3.7% (N = 8) screened positive for suicide risk on the ASIQ. The ASQ demonstrated moderate sensitivity (62.5%, 95% CI: 30.6% – 86.3%), and strong specificity (91.3%, 95% CI: 86.6%-94.4%) and negative predictive value (98.4%, 95% CI: 95.1% – 99.2%) The PHQ-9 item 9 demonstrated poor sensitivity (50.0%, 95% CI 21.5%-78.5%), and strong specificity (98.1%), 95% CI: 95.1% – 99.2%) and negative predictive value (98.1%, 95% CI: 95.1% – 99.2%). The PHQ-9 item 9 had significantly higher specificity than the ASQ (McNemar’s chi-square = 9.389, p < 0.001). Sensitivity was not significantly different between the two measures (McNemar’s chi-square = 0, p = 1.000), likely due to insufficient power. A relatively small (r = .17), but significant positive correlation was detected between PEG and ASIQ scores (p < .05), but PEG scores did not predict suicidality beyond the ASQ. The low rate of positive opioid screens (N = 3) prevented evaluation of the association between opioid misuse and suicidal ideation. About half (45.7%) of participants completed the follow-up survey. Of these, 6.7% reported suicidal ideation and none reported suicide attempts. The ASQ and ASIQ results were significantly associated with suicidal ideation at follow-up (p’s < .001), but the PHQ-9 item was not. However, the low rate of suicidal ideation at follow-up limited statistical power.
Conclusions and Implications
The ASQ appears to be a valid screening tool for suicide risk in the rural, adult primary care setting. Results from this small study indicate differences in specificity and sensitivity between the ASQ and the PHQ-9 item. Future research with larger sample sizes is needed to evaluate the predictive power of the ASQ tool as a suicide risk screen in rural primary care.