Errors in diagnosis of depression lead to over and under diagnosis in pr...( A meta-analysis of more than 50000 p...),Health

A meta-analysis of more than 50,000 patients has shown that general practitioners (GPs) have great difficulty separating those with and without depression, with substantial numbers of missed and misidentified.

GPs looking for depression make more misidentifications (false positives of depression) than the number of depressions they correctly spot following an initial consultation but accuracy could improved by re-assessment of people suspected of having depression.

These are the conclusions of an article published Online First and in an upcoming edition of The Lancet, written by Dr Alex Mitchell of University of Leicester together with Dr Amol Vaze, and Dr Sanajay Rao of Leicester Partnership Trust. The study pooled 41 trials from nine countries that used robust outcome standard of a semi-structured interview to assess depression. The researchers found that GPs were able to recognize about half of people who had clinical depression and correctly reassured 80% of healthy people.

Dr Alex Mitchell said "Imagine a typical GP who is trying to spot depression in a rural practice. He or she might see 100 people over five days. If all the people with depression came to see the GP at once, they would fill the surgery for at least half a day. However the hard pressed GP would actually only spot half of these cases and half would be missed. On four days the GP would see people with other complaints but he or she would mistakenly diagnose up to one in five as depressed, equivalent to almost one full day of contacts. In the worst case scenario false diagnoses could outnumber true diagnoses three to one."

Writing in the journal, the researchers said: "Our results should not be interpreted as a criticism of GPs for failing to diagnose depression but rather a call for better understanding of the problems that non-specialists face."

Dr Mitchell commented further that "research also suggests equivalent errors in the diagno
'/>"/>