Medicalization explained: the DSM stopped growing. Diagnosis rates didn’t

Medicalization is the process by which ordinary parts of life, such as grief, restlessness, shyness or ageing, come to be defined and treated as medical problems. The word sounds like an accusation, and for half a century it has mostly been used as one. Its history is usually told through psychiatry’s diagnostic manual, the DSM, which grew from a slim booklet into a volume the size of a phone directory.

That story is true, but it stops too early. The manual’s list of disorders stopped growing a long time ago. What kept expanding were the thresholds, the definitions and the public’s readiness to reach for a label. The fight over what counts as normal has moved from the table of contents to the cut-off points, and that is where the real decisions about who is ill now get made.

A reader who has never opened a diagnostic manual still has a stake in this. A diagnosis is more than a description. It decides who gets treatment, who gets support at school or cover from an insurer, and how people explain their own lives to themselves. Getting the line wrong costs something in both directions: set it too tight and suffering goes untreated, set it too loose and ordinary distress becomes a condition to be managed for life.

Where the word medicalization came from

The term belonged to sociology before medicine adopted it. In 1972 the medical sociologist Irving Zola described medicalization as “making medicine and the labels ‘healthy’ and ‘ill’ relevant to an ever increasing part of human existence”, and he treated medicine as an institution of social control, a body that decides what counts as deviant. Peter Conrad, who became the concept’s leading exponent, supplied the working definition still used today: “a process by which non-medical problems become defined and treated as medical problems, usually in terms of illness and disorders.” Conrad’s early case study, published in 1973, was hyperkinesis in children, the condition later renamed ADHD.

The polemical peak came with Ivan Illich’s Medical Nemesis, published in 1975, which opens with the line “The medical establishment has become a major threat to health.” Illich called the damage iatrogenesis and argued that it was cultural as well as clinical. A society that hands pain, ageing and death over to professionals, he wrote, loses its own ways of bearing them.

How the DSM grew, and when it stopped

The numbers behind the familiar story come from the manual itself. According to a history of psychiatric classification by Alina Surís, Ryan Holliday and Carol North in the journal Behavioral Sciences, the first DSM in 1952 listed 106 diagnoses in 130 pages. The third edition in 1980, which rebuilt psychiatry around checklists of symptoms, jumped to 265. After that the curve flattened: 292 diagnoses in the revised third edition of 1987, 297 in DSM-IV in 1994, and 298 in DSM-5 in 2013, a book that nonetheless ran to 992 pages.

Growth was not a one-way ratchet either. The clearest case runs in the opposite direction. Homosexuality was dropped from printings of the second edition from 1973 onward, replaced first by “sexual orientation disturbance” and then removed entirely by 1987. That is demedicalization, and it belongs to the same history. The manual has always recorded social judgement as much as biology.

So the popular picture, a manual that swallows a new slice of normal life with every edition, is about three decades out of date. Since the late 1980s the count has barely moved. Diagnosis rates kept rising anyway. The expansion found other routes.

Three routes from normal to diagnosis

The first route is the threshold. A condition stays on the list, but the line that defines it moves. Cardiology offers the cleanest example, because blood pressure is a number. When the American College of Cardiology and the American Heart Association lowered the threshold for hypertension to 130/80 in 2017, Paul Muntner and colleagues calculated in the Journal of the American College of Cardiology that the share of US adults with the condition would rise to about 46%, from about 32% under the old definition. Nobody invented a new disease. Millions of people crossed a line that had moved towards them.

The second route is the widening definition, and grief is psychiatry’s most contested example. DSM-5 removed the so-called bereavement exclusion, which had kept recently bereaved people from being diagnosed with major depression. Allen Frances, who chaired the task force behind DSM-IV, counted it among the worst changes in the new edition and made grief a centrepiece of Saving Normal, an insider’s attack on DSM-5, the drug industry and what his subtitle calls “the medicalization of ordinary life”. In 2022 the text revision, DSM-5-TR, added prolonged grief disorder, with a safeguard written in response to public objections: for adults, at least twelve months must pass after the death before the diagnosis can be made.

The third route is awareness. In a paper in New Ideas in Psychology in 2023, the psychologists Lucy Foulkes and Jack Andrews set out what they call the prevalence inflation hypothesis. Awareness campaigns, they argue, can raise reported rates in two ways at once: by helping people recognise real problems that had gone unnoticed, and by encouraging some to read ordinary distress as illness, after which the label can shape the experience it names. Their paper is a call to test that idea, not a proof of it, and the authors present it that way.

Underneath all three runs a commercial current. In Selling Sickness, published in 2005, the journalist Ray Moynihan and the researcher Alan Cassels examined ten common conditions whose definitions had been broadened in ways that widened the market for treatment, and they popularised the term “disease mongering” for the practice.

The case for the diagnosis boom

The strongest version of the opposing argument deserves its own space, because serious clinicians hold it, not only the pharmaceutical industry. It runs like this. Rising diagnosis rates mostly reflect conditions that were always there and went unnamed. Autism in girls, ADHD in adults and depression in men were missed for decades because the criteria were built around narrow stereotypes. A label is often the only door to support at school, at work or in the health system, and calling that medicalization can sound like telling people their suffering is not real.

Some psychiatrists go further and argue the manual is too cautious, not too generous. Writing in Psychiatric Times in March 2026, the psychiatrist Awais Aftab of Case Western Reserve University described DSM criteria as “rather conservative” and argued that “large swaths of psychopathology simply are not named.” Critics of the overdiagnosis thesis also go after its weakest examples. In BJGP Life, the general practitioner Elke Hausmann faulted the neurologist Suzanne O’Sullivan’s book The Age of Diagnosis for treating long Covid as largely psychosomatic and, in Hausmann’s reading, for selecting the studies that fit. Reviewing the same book for Women in Neuroscience UK, Ciara Lee wrote that “O’Sullivan presents the overdiagnosis of autism as fact, despite this being a contested and widely debated topic.”

Why both camps are partly right

MCM’s reading is that both camps describe the same machine from different ends. Underdiagnosis and overdiagnosis can rise at the same time, because they happen to different people. The girl whose autism was missed and the widower handed an antidepressant a few weeks after a funeral are not competing anecdotes. They are what a diagnostic system produces when its thresholds are set without a stated purpose.

That is the useful part of Aftab’s criticism, and it cuts both ways. “DSM thresholds do not seem to clearly optimize anything,” he wrote. “Some thresholds seem semi-arbitrary by design.” A cut-off that was never chosen to achieve anything cannot be defended against either charge: that it leaves people out, or that it pulls too many in.

O’Sullivan, a British neurologist, sharpens the other half of the problem. Her book separates over-detection, finding disease earlier than any treatment can help, from expanded disease definitions, which she blames for much of the recent rise in ADHD and autism diagnoses. Reviewing the book for Undark, the journalist Lola Butcher summarised the core charge as patients receiving diagnoses that are technically correct but do not necessarily benefit them. The figures cited are large: ADHD diagnoses among US children rising from under 6% to nearly 10% over two decades, and US autism estimates moving from one in 150 children in 2000 to one in 31 in 2022. Neither number, on its own, settles whether the new cases are recognition or inflation. The system as it stands was never designed to answer that question.

The next DSM wants to blur the line itself

The debate now has a live anchor. In January 2026 the American Psychiatric Association published a roadmap for the future of the manual, drafted by a strategic committee chaired by Maria Oquendo, chair of psychiatry at the University of Pennsylvania. It proposes renaming the book the Diagnostic and Scientific Manual and organising assessment in four parts: contextual factors such as social and economic conditions, biological markers from genetics to wearable devices, diagnoses of varying specificity and severity, and transdiagnostic features that cut across conditions. The stated aim is to “advance scientific rigor, cultural inclusivity, and adaptability while ensuring that the DSM remains useful to clinicians.”

Measuring distress along dimensions changes the terms of the medicalization debate. If symptoms sit on a continuum rather than in present-or-absent boxes, there is no single border between normal and disordered left to argue about, only a scale and a decision about where treatment begins. That could end the category wars. It could also make the cut-off points even less visible than they are today, and the cut-off points are where most of the expansion has happened since the list stopped growing.

What the record shows, and what is still argued

The documented part is solid. The number of DSM diagnoses grew two and a half times between the first and third editions and has barely changed since the late 1980s. Homosexuality was removed. The bereavement exclusion was dropped in DSM-5, and prolonged grief disorder arrived in DSM-5-TR. The 2017 US blood pressure guideline moved roughly one in seven American adults into the hypertension category on paper. The APA is now designing a manual built partly on dimensions.

The argued part is just as clear. Nobody has settled whether rising rates of autism, ADHD, anxiety and depression mainly reflect recognition or inflation. Whether awareness campaigns help or feed the problem remains, in Foulkes and Andrews’ own framing, a hypothesis to test. And whether a dimensional manual would make thresholds more honest or harder to see depends on choices the APA has not yet made.

MCM’s verdict: the useful question is no longer whether medicine has colonised normal life, a charge that fit a manual doubling in size and fits less well one that has stood still. The question is who sets each threshold, what that threshold is meant to achieve, and whether anyone checks that the people it captures end up better off. Until the next manual answers that in writing, normal will keep being defined in the fine print.

Tags: Folarin Balogun, Lili Reinhart, DSM, Medicalization, mental health diagnoses, Overdiagnosis