The Confidential Enquiry into Maternal and Child Health (CEMACH) 2003-2005 ^{Reference Lewis1} identified psychiatric illness as the third leading cause of maternal deaths, most commonly death by suicide. The well-known and tragic case of Dr Daksha Emson, a consultant psychiatrist with bipolar disorder who killed her 3-month-old daughter and died by suicide while suffering from puerperal psychosis, ² highlighted another important finding. The suicide profile of childbearing women in the Confidential Enquiries ^{Reference Lewis1,3} is different in many respects than that of other women and men. The majority of the women came from comfortable backgrounds, were in stable relationships and had partners who were in employment. Many had completed higher education and several had higher professional qualifications. Typically, their suicides were by violent means. Of the women who died by suicide, 64% had a previous history of psychiatric illness, but less than half of them had their past history correctly identified during pregnancy and only 20% had an adequate management plan put in place.

It is therefore useful to study the profile of women referred to a specialist perinatal mental health service in order to identify potential risk factors and indicate preventative interventions.

Method

The Glasgow Perinatal Mental Health Service was established in 2004 to provide in-patient mother and baby care to west Scotland (25 000 births per year) and community/maternity liaison services to the Greater Glasgow area (10 000 births per year). The service has approximately 50 admissions and several hundred community and maternity liaison referrals each year. The Greater Glasgow area includes all sociodemographic groups, but there is a high level of deprivation, unemployment and disability. The community service is provided by a small multidisciplinary team comprising psychiatric nurses, psychiatrists, a social worker and a nursery nurse.

A standardised new patient information form was devised to identify demographic information, referral sources, risk factors for postnatal mental illness, and prescribed and non-prescribed drug use in pregnancy and breastfeeding (Box 1; for full patient information form, see the online supplement).

A time period of 1 year (1 April 2005–31 March 2006) was chosen to give a representative sample. A list of all new patients seen by the service during this period was obtained from the patient information management system database, a central computerised record of all patients, and the information on the patient information form was cross-checked against this database.

Of 277 new patients, 219 (79%) had forms with some sections incomplete. The data were audited and the information was subsequently analysed to provide a profile of patients seen by the service.

Results

ICD–10 diagnosis

Perhaps unsurprisingly, 51% of women had an ICD–10 F3 diagnosis at initial assessment. ⁴ The majority had depressive disorder, but 8% had bipolar affective disorder. This was followed by F4 (25%), predominantly anxiety and adjustment disorders, and F2 (7%). Emotionally unstable personality disorder made up 5% of the total. Approximately one in seven women referred had no current mental illness (a full list of diagnoses in our sample group is given in Table 1).

Table 1.

ICD–10 diagnosis at initial assessment (n=183)

Diagnosis	Service users n (%)
Mood [affective] disorders
F32 Depressive episode	41 (22)
F33 Recurrent depressive disorder	33 (18)
F31 Bipolar affective disorder	14 (8)
F34 Persistent mood [affective] disorders	5 (3)
Neurotic, stress-related and somatoform disorders
F41 Other anxiety disorders	18 (10)
F43 Reaction to severe stress, and adjustment disorders	15 (8)
F42 Obsessive—compulsive disorder	8 (4)
F40 Phobic anxiety disorders	3 (2)
F45 Somatoform disorders	1 (<1)
No diagnosis of mental illness	25 (14)
Schizophrenia, schizotypal and delusional disorders
F20 Schizophrenia	4 (2)
F23 Acute and transient psychotic disorders	4 (2)
F25 Schizoaffective disorders	3 (2)
F21 Schizotypal disorder	1 (<1)
F22 Persistent delusional disorders	1 (<1)
Disorders of adult personality and behaviour
F60.3 Emotionally unstable personality disorder	9 (5)
F60.6 Anxious [avoidant] personality disorder	1 (<1)
F1 Mental and behavioural disorders due to psychoactive substance use	7 (4)
F0 Organic, including symptomatic, mental disorders	1 (<1)
F5 Behavioural syndromes associated with physiological disturbances and physical factors	1 (<1)
F7 Mental retardation	1 (<1)

Discussion

This study highlights many of the challenges in dealing with patients with mental illness related to childbirth. In particular, a significant proportion of women have a history of psychiatric illness, among whom are a group at particularly high risk of relapse, including 29% in our sample who have a recurrent mood disorder which places them at 1:2 to 1:3 risk of relapse. ^{Reference Cohen, Altshuler, Harlow, Nonacs, Newport and Viguera5,Reference Robertson, Jones, Haque, Holder and Craddock6} Focusing on past psychiatric history is important in ensuring prompt and effective follow-up, prophylaxis and early intervention for women known to be at increased risk. ^{7-Reference Jones and Craddock9} In our study, most patients were seen during pregnancy (81%), allowing a real opportunity to identify risk factors and formulate postnatal care plans to reduce risk of relapse. This is especially pertinent as CEMACH (2003–2005) ^{Reference Lewis1} found that 81% of deaths from psychiatric causes had a prior psychiatric history which, in a number of cases, could have been used as an indicator of high risk for postpartum relapse. Robertson et al ^{Reference Robertson, Jones, Haque, Holder and Craddock6} showed that in a cohort of women with one episode of bipolar affective puerperal psychosis, 57% had a subsequent puerperal psychosis and 62% experienced at least one non-puerperal affective episode.

We also found that a surprisingly high proportion of women referred to the service were taking prescribed psychotropic medication at conception. There may be potential teratogenic risks with some SSRIs, ^{Reference Bar-Oz, Einarson, Einarson, Boskovic, O'Brien and Malm10,Reference O'Keane and Marsh11} the most commonly prescribed drugs in our study. A proportion of women discontinued medication abruptly on discovering the pregnancy, placing them at risk of relapse. Both these findings highlight the importance of counselling women of childbearing age who are taking psychotropic medication about the need for adequate pregnancy planning.

Substance misuse remains a risk factor for women and requires specialist integrated care. ^{Reference Lewis1} We found that nearly a third of women in our study continued to smoke during pregnancy, 7% used illicit drugs of which over a third used multiple drugs, and one in seven continued to consume alcohol despite published guidelines ¹² stating the safest approach is abstinence or no more than two units of alcohol twice a week. Identification and appropriate management of women at high risk should be a key priority as they are typically difficult to engage, have co-existing mental health and medical problems, are prone to domestic violence and child protection procedures, and are often poor attenders for antenatal care.

Conclusions

In summary, our key findings suggest that perinatal mental health services are in a prime position to identify and intervene with women at high risk of postnatal mental illness. However, women may be missing out on adequate pre-pregnancy advice in relation to prescribed psychotropic medication.

Declaration of interest

None.