In the 70s and early 80s, right before a little known manual called the DSM (then on its 2nd edition), came out with its 3rd edition that created mental illnesses on the basis of statistical classification and clustering of symptoms, psychiatrists were considered a joke by other doctors. Too many were still doing therapy, relied on Freudian-based psychodynamic theories, didn't use any of their medical training, etc., they came under attack from three sides.
First, there were other doctors and potential doctors. The medical sciences had long since moved to a biomedical diagnostic model. Psychiatrists had couches and knew less about the physiological make-up and dynamics of the brain than various non-medical doctorates. Pre-med students weren't going into psychiatry and other doctors didn't consider psychiatrists any different than various other mental health workers.
That's where the 2nd attack came from. Despite the fact that Freud saw no need for psychoanalysts to have the medical background he did, nonetheless psychiatrists had for decades been medical doctors who didn't practice medicine. As a result, social workers and clinical psychologists argued that any pride-of-place granted psychiatrists by their medical background was irrelevant. Psychiatric research itself and psychiatric practice demonstrated how little what psychiatrists did was "medicine". So other mental health workers became a serious economic (as well as academic/research) threat to psychiatry.
Finally, there were the insurance companies. Before the DSM-III, mainstream psychiatric theories frequently held that mental health existed along a continuum and wasn't a matter of discrete diagnoses or disorders. This meant that insurance companies had no recourse when they were called to pay for therapy sessions, as they couldn't depend upon professionals to tell them what conditions were or weren't medical. After all, even schizophrenia (a disorder one would think clearly distinguishable from mood disorders) was so confused with bipolar disorders that US and UK psychiatrists tended to diagnose both oppositely.
So, an elect few among psychiatrists created the biomedical model. The idea was mainly to regain respect from other doctors, regain supremacy, and save their profession. This wasn't, however, a selfish, economic move (at least not mostly). Feighner, Spitzer, etc., really believed in the biomedical model. However, those who drafted and made the diagnostic "bible" of mental disorders didn't have any biological basis for their classification system. So they developed strict criteria based upon symptoms that ensured their internal reliability by e.g., including as a diagnostic criterion for X disorder that it couldn't be Y disorder: thus two disorders which often presented nearly identically could not, according to psychiatry, be conflated because they were defined as distinct, thus any failures in inter-rater reliability were the fault of the practitioner. The diagnostic model ensured that while there could be (and was and is) unbelievably prevalent co-morbidity, every diagnosis had a discrete, specific underlying pathology.
The hope was that we would find the evidence for underlying biological mechanisms by which psychiatrists could use the kinds of tests all other doctors do for differential diagnoses. Instead, we found that the brain was far more plastic than thought, that neurotransmitters first identified and first said to cause disorders like schizophrenia or major depression were implicated in virtually the entire range of diagnoses and in symptoms independent of particular diagnoses (such as anxiety). The Human Genome Project proved underwhelming and epigenetics has replaced it, and once again psychiatry is far behind the other medical sciences, neurosciences, etc.
