“One in four” with a mental health problem: the anatomy of a statisticBMJ 2012; 344 doi: https://doi.org/10.1136/bmj.e1302 (Published 22 February 2012) Cite this as: BMJ 2012;344:e1302
All rapid responses
Smith  in an interesting comment writes that the Scottish “Well, what do you think?” survey  is evidence that the prevalence figure for mental ill health in the UK of “1 in 4” has “reasonable empirical support”.
We were not aware of this survey when we wrote our paper, but we do not think that this survey’s results undermine our conclusion that “1 in 4” is poorly supported. In our analysis we looked at studies that aimed to estimate the total burden of mental ill health - both diagnosed and undiagnosed - in a population.
The “Well, what do you think” survey does not actually do this. The survey found that 27% of participants responded they had been diagnosed with a mental health problem by a doctor or health professional. This is therefore an estimate based on diagnosed cases. As many cases of mental ill health go undiagnosed, it is likely to be an underestimate of the number of people who meet criteria for mental disorder.
Further, as we point out in our paper  retrospective estimates of lifetime prevalence ("have you ever...") are likely to underestimate true lifetime prevalence. Past disorder is forgotten, while disorders that people will develop in the future are not recorded. It is also possible that if even more disorders included in the Scottish survey question, then an even higher estimated prevalence would be reported. A recent Department of Health survey  found only 4% of people reported they had a mental health problem. This difference may be accounted for by considering that the Scottish survey used more detailed questions. It prompted for several disorders by name, while the English survey simply asked about "mental health problems".
Therefore, if 27% of Scottish people receive a diagnosis of mental health problem, the true prevalence of mental ill health in Scotland will be far in excess of “1 in 4”.
1. Smith M. Re: “One in four” with a mental health problem: survey data overlooked
2. Glendinning R., Buchanan T., Rose N., Well? What do you think? A national Scottish Survey of public attitudes to mental health, well being and mental health problems
3. Horder J., Ginn S., “One in four” with a mental health problem: the anatomy of a statistic BMJ 2012
4. Moffitt, T., Caspi, A., Taylor, A., Kokaua, J., Milne, B., Polanczyk, G., & Poulton, R. (2009). How common are common mental disorders? Evidence that lifetime prevalence rates are doubled by prospective versus retrospective ascertainment Psychological Medicine
5. Attitudes to Mental Illness survey research report. Department of Health 2010. http://www.dh.gov.uk/en/Publicationsandstatistics/Publications/Publicati...
Competing interests: No competing interests
The authors state “we are unaware of any evidence that straightforwardly supports a UK lifetime prevalence of 25%”, but seem to be unaware of data from Scotland that is consistent with the use of this statistic.
When we launched Scotland’s campaign against the stigma of mental illness in October 2002, we promoted the “one in four” idea (http://www.seemescotland.org/images/pdfs/1.6.seemesofar.pdf). This was based on data collected for a national Scottish survey of public attitudes to mental health, wellbeing and mental health, published by the then Scottish Executive as “Well? What do you think?” earlier that year (http://www.scotland.gov.uk/Publications/2002/12/15967/15270).
The survey was based on face-to-face interviews with a representative sample of 1,381 adults living in Scotland. Participants were asked “Have you ever been told, by a doctor or other health professional, that you personally have had one or other of these kinds of specific mental health problems?” and asked to choose from one or more of eight specified mental health problems (eg depression, anxiety, schizophrenia), plus “severe stress” and “any other mental health problem” (http://www.scotland.gov.uk/Publications/2002/12/15967/15290). 27% of participants responded in the positive.
This statistic has been replicated in subsequent surveys in 2004, 2006 and 2008, and the 2008 report discusses these findings in relation to equivalent health surveys conducted in other parts of the UK (http://www.scotland.gov.uk/Publications/2009/09/15120147/9).
The question was chosen to be a meaningful and pragmatic measure of the population experience of mental health problems. From a Scottish perspective, “one in four” isn’t a marketing-based guesstimate, and in fact has reasonable empirical support.
Competing interests: I am on the Management Board for "see me", Scotland's anti-stigma campaign.
Ginn & Horder (BMJ 2012;344:e1302) are correct in pointing out that there are numerous problems in deciding what is to be counted as a mental disorder, but most epidemiologists have contented themselves with using an official classification such as the International Classification of Disease.
We are responsible for providing evidence that the one year period prevalence of mental disorders in community samples was approximately 250 per thousand at risk (Goldberg & Huxley 1980 p10) , a figure obtained by combining figures for cross sectional prevalence with admittedly speculative estimates of annual inceptions, so that a cross sectional rate of 180/1000 was inflated by assuming that about one third of that figure would develop a new episode during the following year. Even at that time, we had excellent evidence that the vast majority of such episodes are of short duration (less than 3 months). Since that time surveys have become available asking people to remember their health over the previous year, so that by 2002 it was possible to show that these figures were slight under-estimates, and the figure for the UK was revised upwards to 270 (Goldberg & Goodyer 2005, p18), and were
combined with a consideration of rates reported by the Office of National Statistics.
These rates do not include either severe mental disorders such as schizophrenia bipolar disorder or dementia, nor did they include alcohol and drug dependence. It should be noted that these are annual rates, not life-time rates. The concept of lifetime prevalence is necessary for studies of the genetics of mental disorders, but is a highly questionable concept where common mental disorders are concerned, since it assumes that people not only can, but will reveal information about minor disorders that occurred many years ago, but that they have either forgotten or suppressed. For this reason, we have never quoted figures for life-time rates.
However, for those that like to think in these terms, we would suggest that the figure of “at least 25%” is almost certainly a conservative figure.
Goldberg DP, Huxley PJ (1980) Mental Illness in the Community. The Pathway to
Psychiatric Care. London, Tavistock
Goldberg DP, Goodyer I. (2005) The Origins and Course of Common Mental
Disorders. London, Routledge
Competing interests: No competing interests
One in Four: a useful shorthand and a meme that went viral
Ginn, Robinson and Horder (1) are having their epidemiological cake and eating it. I have yet to read weekly and yearly prevalence rates for physical illness in its entirety (definitions would be difficult, as would rates in young and older people) and the complexity with mental disorders increases with its spectrum from the “1% disorder” of schizophrenia to the ubiquitous but self-limiting “diagnoses” of caffeine addiction, some phobias, acute stress reaction, and adjustment disorder including bereavement (2). It is in no one’s interests to bring these latter conditions into the totality of severe psychological distress – neither the health services nor an informed public would believe us.
Beginning 1998, I was one of many health professionals, users and carers involved in the Royal College of Psychiatrists’ 5 year anti-stigma campaign, Changing Minds. Its chair, the late Professor Arthur Crisp, was keen to tag the campaign as “every family in the land”. The counterargument was that many people with severe mental illness had become separated from their families, or in some cases had been alone since childhood. As the campaign media lead, my brief was to highlight many common disorders (depression, anxiety, dementia and alcohol or drug misuse) as well as the rare but more stigmatised disorders of schizophrenia and anorexia nervosa. In populations that live long enough, dementia rates approach 20%, as do combinations of alcohol misuse and anxiety-depression rates in people living in impoverished conditions in the developed world. The most cited study then excluded cognitive decline to calculate a one month adults’ prevalence of 15.4% (3). Back in 2000, our calculation of “1 in 4” led to the short film by that name that deliberately depicted all 6 disorders in its less than 2 minutes’ running time. It played as a cinema trailer for 4 months and is still available to view: http://www.rcpsych.ac.uk/files/stigma_video.mpg
The challenge for Psychiatry is to communicate prevalence in a clear way. A generalisation of “between 23 and 27%” would fail to impress – their quoted statistics for yearly EU rates and past week UK prevalence respectively (1). Without knowing what we were doing (though I must credit colleagues from the advertising industry and the film’s director John Selby), “one in four” became a meme (http://en.wikipedia.org/wiki/Internet_meme ) and went viral. And rightly so. Just this month, public health colleagues have called for a common language among global mental health advocates to make the arguments that treatments work and yield measureable economic benefits (4). I agree with the authors’ summation (1) that the phrase is “not too big, not too small”, and we should avoid using one in four as either point or short term prevalence rate.
1. S Ginn, R Robinson and J Horder “One in four” with a mental health problem: the anatomy of a statistic. BMJ 2012; 344.
2. A Kleinman Culture, bereavement and psychiatry. The Lancet, 379, 9816: 608 – 609
3. DA Regier, JH Boyd, JD Burke et al. One-Month Prevalence of Mental Disorders in the United States – based on five epidemiologic catchment area sites. Arch Gen Psychiatry. 1988; 45:977-986.
4. M Tomlinson and C Lund Why Does Mental Health Not Get the Attention It Deserves? An application of the Shiffman and Smith framework. PLoS Med 9(2): e1001178.
Competing interests: I am Director of Public Education for the Royal College of Psychiatrists