The North Wales asylum made its way into my life by accident. The history department at Bangor University secured a grant to look at the social impact of the asylum. Looking at the records they collected, it was striking how people declared their madness a century ago – they tore off their clothes and escaped through windows, which they never do now.
But when we set about entering North West Wales records from the asylum in 1896 to compare with admissions in 1996, something more startling became apparent – we admit 15 times more people with serious mental illness now, and compulsorily detain 3 times more people (Healy et al 2001). This prompted a quixotic adventure – why not enter all records from North West Wales between 1875 and 1924 and build a modern database covering admissions from 1994 to 2010?
This was quixotic because no-one wanted to fund us. Grant-giving bodies in history were not able to see the value in quantifying records or in having modern data. Modern epidemiologists could not see the value in historical records. So over 15 years and without support we put together easily the largest body of historical epidemiology in existence – we had no competition, no-one else thought this made sense.
North West Wales is uniquely suited to what we did. It’s surrounded by the sea on 3 sides and the mountains on the fourth. The population remained almost exactly the same between 1896 and 1996. It was then and is still now almost exclusively Welsh. It was rural and never urbanized. The people are poor and so there is no private care. There was then and is now only one place to go for a hospital admission. In the case of the few people who left the area, through the National Health Service we have been able to track everyone down so not a single person is lost to follow-up.
The response from some experts to our findings has not been unlike the dismissive attitude of Orcs or Black Riders to Hobbits. They sneeringly fall back on the fact that we have made diagnoses based on century old records. One hostile reviewer argued that when black men in America in the 1960s could be detained compulsorily and diagnosed as having schizophrenia, what chance is there our records have it right. Others are sure women with unwanted pregnancies were incarcerated in mental asylums and diagnosed as schizophrenic.
The ultimate proof that out methods are right lie in a set of bricks and mortar developments, as will become clear.
Postpartum (or puerperal) psychosis was one of the severest mental illnesses ever. It accounted for the admissions of one in ten of the women of child-bearing years admitted to the old asylums between 1875 and 1924. It came in two forms. The most dramatic form happened in women who had never had a hint of mental illness before, who were stably married and well-placed who within a week or two of giving birth to their first or fourth or other child went floridly mad. These women with de novo onset postpartum psychoses accounted for 4 out of every 5 cases. Their psychosis looked completely different to either schizophrenia or manic-depressive illness.
The second group were women who had had a previous mental illness who after giving birth might have a further episode of essentially the same illness they had before. Their illness typically looked like a further episode of manic-depressive illness.
When the asylum closed it was replaced by 3 district general hospital units, across North Wales, the Hergest Unit in the West, Ablett and Llyn-y-Groes in the middle and East. All 3 had mother and baby units. These units were opened up because of the rates of admission for both kinds of postpartum psychosis in our asylum records – these map directly onto what was in 1993 the accepted rate for the occurrence of post-partum psychoses.
Hergest was the first of these general hospital units to open but within 3 years of opening, the mother-and-baby unit closed. There were no cases. The Ablett mother and baby unit had just opened up and this is where Cora went (see Cora’s Story)– and going there may have partly caused her loss of life. All 3 mother-and-baby units have now closed.
Our historical and contemporary databases bring this out perfectly. De novo onset postpartum psychoses have vanished – the manic-depressive type remains (see Tschinkel et al 2007). These are not disorders you can treat in the community. They are the most high risk in all psychiatry.
But when we came to report the findings we were in for a surprise. No-one it seems wanted to hear about a disorder vanishing – not the postpartum experts whose careers depend on it and are still busy portraying it as commoner than ever. Not the health service managers whom one might have thought would have an interest because of the budgetary implications. Nor the researchers you might have thought would be interested in the implications for theories of mental illness. Not the historians specializing in postpartum psychoses or women’s mental health. It was difficult to get published.
What our critics fail to realize is that we are not going on historical or contemporary records. Doctors in the nineteenth century made diagnoses based on how the patient looked when they walked in the door but we have had the entire lifetime medical record of these patients available to us and could base our diagnoses on these.
Doctors today don’t make diagnoses on the patient as they walk in the door. When the patient is discharged, the ward clerk or someone with no medical training typically enters the diagnostic code into an administrative database. This can work very well in medicine, but doesn’t make sense in psychiatry. A great deal of modern psychiatric epidemiology is based on diagnoses like these – close to useless. Other studies are based on diagnostic interviews instead, which are also close to useless. Almost 50% of cases of schizophrenia, as clear cut a condition as psychiatry has, get a different diagnosis during their first year or two of admission. It takes time for the nature of the illness to become clear. And this is where our approach scores for both historical and modern cases – we make our assessments based on the full clinical record with input from everyone who knows the patient.
There has been extensive work by others on historical records at this point. These have revealed several famous examples of individuals who may not have been mad and who have protested against their incarcerations. But these are exactly the cases that we do not diagnose as having had a psychosis. If a Black or Irish man had been inappropriately incarcerated in North West Wales in 1896 we would have spotted it. In fact the asylum records show clinicians sensitive to issues of whether this patient was “a knave or a fool”. They were very reluctant to admit and quick to discharge knaves. There was not a single admission for unwanted pregnancy or ethnic factors, and in cases where there was domestic or other abuse, these issues were clearly flagged up, making it possible for us to take this into account retrospectively.
The abuse of the asylum by doctors was a mid-twentieth century phenomenon – a telling example of how medicine can turn to the dark side rather than sustain progress. But just because the Nazis exterminated mental patients and psychiatry acted like an agent of the State in both the Soviet Union and the West in mid-century doesn’t mean it was always this way. Life in the Shire, if not perfect, was once good enough. The challenge has always been to keep it that way – although a case can be made that things are getting even worse now.
So what happened postpartum psychoses? We have no idea. This is in complete contrast to catatonia which has also vanished and we are certain we can explain this. It’s also in contrast to schizophrenia, where we have some good ideas.
What could have happened? The postpartum psychoses look more like steroid psychoses than schizophrenia. Perhaps their disappearance stems from the active management of labor now which means few women get as fatigued as once happened, or the use of pain relief like Heroin (see Tschinkel et al 2007). If post-partum psychoses are related to catatonia, the availability of benzodiazepines might be expected to make a big difference. Whatever the reason, our best guess is this condition began to disappear somewhere between the late 1970s and the 1980s. We want any suggestions anyone can offer.