The Patient Paradox part 2: Depression screening by numbers
In this week's installment of The Patient Paradox, Dr Margaret McCartney asks: does depression by numbers do any good?
Depression is, truly, a miserable affliction: I have encountered many people who have had both depression and unpleasant treatment for medical conditions and who will say the depression was easily worse. What follows does not denigrate this illness.
The problem for screeners is that not all low mood is depression. Questionnaires, when done as a reflex adherence to protocol, do not account for the circumstances people find themselves in. Life is not a straight emotional arrow. Our mood can sink when we are faced with bad news, an undesirable change in our circumstances, even just a stretch of jet lag-broken sleep. Normal people have moods that change according to what is happening to them. An enormous part of literature, art and music over the past few thousand years has been an attempt to understand and to share what life means through shared emotions.
The advent of the protocol-based questionnaire removes all context from assessing patients' mental states. From the beginning, there is no option for patients to say that they are distressed because their dog has died or they are feeling awful because they have flu. Instead, the questionnaire is administered, high levels of distress recorded and then the doctor or nurse deals with the result. Normal discourse between doctor and patient is relegated to second place behind the paperwork. The questionnaire-based screening for depression is capable of removing human understanding from the encounter between doctor or nurse and patient.
It's known that in some surveys, up to 95% of people, if asked, will admit to feeling low on a regular basis.6 And it is also clear that most people are not incapacitated by ‘depression' as diagnosed by such questionnaires. Indeed, true depression is relatively rare. When the very first antidepressant was developed in 1959, the manufacturers were disinclined to market it to doctors since the pharmaceutical company thought depression was an uncommon disorder and they were not likely to recoup the costs.7
All change. More than 40 million prescriptions for antidepressants – that's forty million – were written in the UK in 2010.8 The question now becomes: how likely is the routine use of depression screening questionnaires to help patients? And how much harm does their use cause?
Indeed, before the instigation of the QOF, it was well known that screening for depression generally resulted in picking up low mood because of life events, and wasn't terribly helpful in finding new depression cases. In one study, researchers found that patients scoring high on questionnaires turned out not to be depressed when they interviewed them.9 The ongoing problem has been this misunderstood differentiation. Studies that look at levels of distress tend to find lots of unhappiness, and conclude that depression is therefore underdiagnosed. Search for ‘depression' and ‘underdiagnosed' on any medical search engine, and there are hundreds of papers supporting that view.
There is a problem with this outlook. Surveys collect a snapshot, data taken at a single point in time. This is contrary to the usual way patients and doctors interact. So, patients who have a diagnosis of heart disease are the kind of patients who are regular attenders at the surgery, returning to have blood pressure checked or blood tests done, and who may well have other conditions too. Real life medicine is not a ‘point in time', paper-based exercise. It is a relationship flowing over months and years.
And indeed, other studies have shown that most true cases of depression found at these ‘point in time' studies have a habit of finding their way in the future to appropriate diagnosis and treatment anyway.10
So how accurate are the questionnaires? The ideal questionnaire with no false positive or negative results does not exist. One commonly used questionnaire, the PHQ (Patient Health Questionnaire), has been noted to be truly correct for depression only around half the time. The bottom line is this: ‘good sensitivity but poor specificity'. Only between 30% and 60% of the time does a positive questionnaire screening score mean that the person really is depressed.11,12
Yet one of these studies concludes that the survey can be a ‘reliable and valid measure of depression severity.' This is quite a jump, considering that the authors have just told us that their tool will get it wrong the majority of the time.
More, there is evidence that the different questionnaires that are still clinically validated and used by GPs for the purposes of cash by QOF are not themselves comparable – meaning that depression of different severity could be diagnosed in the same person with the same set of symptoms, just by the use of different questions.13
And here is the crunch point. Despite all the bits of paper flying around and patients being asked to tick boxes and practice staff being asked to type them into computers, this may all be a wasteful distraction. Doctors don't find them useful. Instead, they listen to their patients, ask them how life is, and try to put everything back in context. If the questionnaires were all that mattered, and the protocols for treating depression from NICE were followed to the letter, then three-quarters of people scoring as depressed would be started on antidepressants. But in practice, far fewer actually do end up on medication.14
So does depression by numbers do any good? Or does our eagerness to hone general practice down to an ‘evidence based' set of protocols and ticksheets create fundamental departures from what the patient might actually want to talk about?
To buy The Patient Paradox for £7.49 and free UK delivery (RRP £9.99) visit http://www.pinterandmartin.com and enter Pulse at checkout or call 020 77376868
6. Parker G. Is depression overdiagnosed? Yes. BMJ, 2007;335:328 August 2007. www.bmj.com/content/335/7615/328.full
7. Healy D. The anti-depressant era. Cambridge, MA: Harvard University Press,1997.
8. The NHS Information Centre Prescribing Support Unit. Prescriptions dispensed in the community, statistics for 1999 to 2009: England. The Health and Social Care Information Centre, 2010. www.ic.nhs.uk/
9. Coyne JC, Schwenk TL. The relationship of distress to mood disturbance in primary care and psychiatric populations. J Consult Clin Psychol1997; 65(1):161-168 psycnet.apa.org/journals/ccp/65/1/161/
10. Kessler D, Bennewith O. Detection of depression and anxiety in primary care: follow up study. BMJ2002; 325(7371):1016-1017 www.bmj.com/content/325/7371/1016.1.full.pdf
11.Phelan E, Williams B. A study of the diagnostic accuracy of the PHQ-9 in primary care elderly. BMC Fam Pract2010; 11:63 www.biomedcentral.com/1471-2296/11/63/abstract
12.Kroenke K, Spitzer RL. The PHQ-9 Validity of a brief depression severity measure. J Gen Intern Med. 2001; 16(9):606-613 www.ncbi.nlm.nih.gov/pmc/articles/PMC1495268/?tool=pubmed
13.Cameron IM, Crawford JR. Psychometric comparison of PHQ-9 and HADS for measuring depression severity in primary care. Br Gen Pract. 2008; 58(546):32-36. www.ncbi.nlm.nih.gov/pmc/articles/PMC2148236/?tool=pubmed
14.Cameron IM, Lawton K. Appropriateness of antidepressant prescribing: an observational study in a Scottish primary-care setting. Br Gen Pract2009; 59(566):644-649 www.ncbi.nlm.nih.gov/pmc/articles/PMC2734353/?tool=pubmed