Responses to item 9 of the PHQ-9 accurately identify those patients who  are at increased risk of suicidal behaviour, an American study has found.¹

Data were obtained from the electronic records of patients registered with a large integrated healthcare system. Both primary care and specialist mental health providers had been asked to complete a PHQ-9 whenever patients attended for depression treatment. During the five-year study period, from 2007 to 2011, 84,418 outpatients aged ≥ 13 completed 207,265 questionnaires. Suicide attempts were identified from the electronic records and insurance claims. Deaths by suicide were identified from state death certificates.

Item 9 asks: ‘Over the last two weeks, how often have you been bothered by thoughts that you would be better off dead, or of hurting yourself in some way?’ Those who responded ‘not at all’ had a 0.4% (1 in 250) cumulative risk of attempted suicide over the following 12 months. Those who responded ‘nearly every day’ were ten times more likely to attempt suicide, a 4% (1 in 25) cumulative risk.

Similarly, those who responded ‘not at all’ had a 0.03% (1 in 3,000) cumulative risk of dying from suicide over the following 12 months, whereas those who responded ‘nearly every day’ had a 0.3% cumulative risk (1 in 300).

There were 709 suicide attempts and 46 suicides during the study period. The risk of death by suicide was significantly higher in men. After adjustment for age, gender, previous history of specialist mental health treatment or hospital admission, and severity (measured by the PHQ-9), the response to item 9 remained a strong predictor of suicide attempts and a moderate predictor of death from suicide.

Misclassification of suicide attempts and deaths may have reduced the association between patient responses and suicidal behaviour. Even so, the low absolute risk of suicide (there were only 46 suicide deaths in this large sample) means that by itself, item 9 has only limited predictive value.

I have previously suggested that a PHQ-9 score ≤ 10 can be used to rule out moderate depression. In the same way, a ‘not at all’ response to item 9 could be used to rule out suicide risk. A more detailed risk assessment would then be undertaken only for those who reported thoughts of death or self-harm. However, some patients may misunderstand the question or conceal their suicidal thoughts, suicide is often impulsive and unpredictable, and there is some evidence that suicidal ideation is only an important risk factor in female patients.²

I think it is better to consider suicidal ideation as just one of a number of risk factors. A suicide risk assessment should be undertaken for all patients and include:

• Family history of mental illness and suicide

• Personality traits e.g. impulsivity

• Previous suicide attempts

• Quality of family relationships and social support

• Employment status

• Smoking status

• Alcohol or substance misuse

• Protective factors e.g. responsibility for children

• Precipitating factors e.g. recent bereavement or separation^2,3

A bio-psychosocial assessment4 captures most of the relevant information but this is due to be retired from next year’s QOF indicators.

In the present study, the increased risk developed over a number of months. Risk assessment should not be a one-off event: it requires active follow-up and continuous monitoring (case management).

Item 9 of the PHQ-9 can be used to identify patients with suicidal thoughts who are at increased risk of suicide. However, this is just one component of the comprehensive and ongoing risk assessment that should be made for all patients with major depression.