Clinical question — Is there an accurate question to use for depression screening in primary care?
The evidence. Screening for depression in primary care with two verbally asked questions: cross-sectional study. BMJ 2003; 327: 1144-1146.1
Background. Undiagnosed depression can lead to long-term disability and yet it is often under-recognised in general practice. It has been estimated that up to half of patients with treatable depression in primary care remain undetected.2
To date, two approaches have been advocated to seek to improve the recognition of depression in primary care: the use of depression screening questionnaires or the detailed evaluation of patients at increased risk of depression (for example, those with chronic medical illnesses, chronic pain syndromes, recent life changes/stresses, fair/poor self-rated health or unexplained physical symptoms).3
Unfortunately, the various screening questionnaires available are often not easily accommodated during the course of a busy GP surgery. There is a need for a simple, but accurate, question to be identified that could be applied quickly and easily in the context of a routine consultation.
Study design. A consecutive series of 670 patients not already taking psychotropic drugs were invited to become involved in a cross-sectional validation study. These patients were spread among 15 general practices in New Zealand.1 The 476 patients who agreed to participate were asked two questions by their GP:
During the past month have you often been bothered by feeling down, depressed or hopeless?
During the past month have you often been bothered by little interest or pleasure in doing things?
Subsequently, the patient completed the mood module of the computer assisted composite international diagnostic interview. Complete information was available for 421 patients out of the original 670.
Outcome measure. By comparing the two question responses (individually and combined) against the computer assisted interview — the ‘gold standard’ — sensitivities, specificities and positive predictive values were calculated.
Results. See Table 1.
Commentary. In assessing the accuracy of any screening question for depression it is important to appreciate that unlike, for example, cancer detection there is no ideal gold standard. It is therefore reassuring that the authors were able to cite other evidence to support their contention that the computer assisted interview was a reasonable approximation to the ‘truth’ (in terms of validity and reliability).
In terms of the population studied it is likely that there was some selection operating between those that participated and those that did not. The patients were stated to have been consecutively recruited from 15 practices. However, complete information was only available on 63% of the 670 invited and this amounts to an average of 28 patients per practice. It would be interesting to know whether the patients evaluated in this study were skewed more towards ‘high-risk patients’ as happens in many cross-sectional diagnostic studies of this type.
A further development of this study would be to assess the reliability (reproducibility) of the two questions for particular patient groupings. As GPs we are well aware that we adjust our psychological questioning according to, for example, age, sex, social class and ethnic origin in order to enhance response reliability. The median age in the current study was 46 years and one-third were men.
The bottom line: Two brief verbal questions for screening for depression had reasonable sensitivity and specificity in a younger primary care population.
Vignette 1
I was coming to the end of surgery one evening when our receptionist buzzed through. An elderly lady had phoned to say her husband had fallen and was stuck on the bathroom floor. I was near to finishing and I asked my receptionist to find out if it was urgent, or if it could wait. My receptionist got back to say it could wait.
It was probably about 20 minutes or half an hour before I left. I knew the couple slightly as they were my father's friends. The man lay wedged across the doorway between the bathroom and the corridor. He was clearly very ill. Examining him, I found he was deeply unconscious, sweating, and his blood pressure was low. He had no femoral pulsations, and I realised this was a ruptured aortic aneurysm. I called the ambulance and fretted until it arrived.
He was operated on at the district hospital the same night, but he did not recover consciousness. I visited him in intensive care. He was covered with monitors, and I felt gloomy about whether he would survive. He died, and my parents went to his funeral.
I regretted not leaving the surgery earlier. Thirty minutes less of shock and cerebral anoxia might have helped.
How was I to choose between leaving the worried well, who had booked an appointment, and going out to an emergency? I have visited many ‘emergencies’, which were not, but how do you tell? What makes people sitting in the waiting room with an appointment so important? These things seemed easier to decide on in a hospital, than as the sole doctor working in a practice surgery one night.
Leone Ridsdale
- © British Journal of General Practice, 2004.