Technical Report: Feasibility of evaluating perinatal mental health services using linked national maternity and mental health data sets, based on births between 1 April 2014 and 31 March 2017 in Scotland

Introduction In this short report, we describe the feasibility of using linked national data sets to evaluate perinatal mental health services. Perinatal mental health conditions are common. About 10% of pregnant women and 13% of women who have just given birth experience a mental health problem. Some perinatal mental health problems can, if not adequately treated, have significant and long-lasting effects on a woman and her baby. For this report, we only used Scottish data sources. The data sets include episodes of admission to secondary care, including hospital admission for perinatal mental health conditions. The results based on Scottish data are expected to inform future analyses of similar data from England and Wales. Specific data sets on mental health services in Wales were not yet available at the time of this study. The report consists of three parts. First, we describe the data sets that were used and how they were linked. Second, we present a grouping of mental health diagnoses that are similar with respect to their prognosis and treatment (to maximise the clinical relevance) while limiting the number of diagnosis groups (to maximise statistical power). Third, we use the results of this preparatory work to demonstrate the clinical relevance of the linked data sets by describing a number of clinical outcomes according to the timing of the perinatal mental health admissions. Methods We used linked national maternity and mental health data for Scotland on all births that took place between 1 April 2014 and 31 March 2017, and inpatient admissions for mental health conditions between 1 April 2000 and 31 March 2018. Births records were identified in the National Records of Scotland (NRS). These records were used as a ‘spine’ against which records from all other Scottish Morbidity Record (SMR) data sets were linked: general/acute inpatient records (SMR-01), maternity inpatient records (SMR-02), mental health inpatient records (SMR-04) and the Scottish Birth Record (SBR). Women who had a mental health admission were identified in SMR-04 data as well as in SMR-01 data if their admission record contained a diagnosis code from Chapter V (‘Mental and behaviour disorders’) of the International Classification of Disease, 10th Revision (ICD-10). Findings Both mental health inpatient data (SMR-04) and general/acute inpatient data (SMR-01) need to be used to identify women who had a hospital admission for mental health indications. We identified 3457 births in women who had a mental health admission. About two-thirds of the mental health admissions were identified in SMR-04 and about one-third in SMR-01. 163109 births were identified. 3043 (2.1%) of these births were in women with a prepregnancy history of a mental health admission. 176 (5.8%) of the women with prepregnancy mental health admission were also admitted during the perinatal period (during pregnancy or in the first year after giving birth). In comparison, only 414 (0.3%) of the 160066 births of women without a prepregnancy mental health admission had a perinatal mental health admission. Therefore, in the majority of cases Evaluating perinatal mental health services using linked national maternity and mental health data sets – 414 of the 590 perinatal mental health admissions (70.2%) – the perinatal mental health admission was a women’s first mental health admission. Diagnostic codes were grouped into eight diagnosis groups aiming to maximise the clinical relevance and statistical power. Based on this grouping, we found that major depressive disorders were the most frequently observed diagnoses (22.9%) among the 590 women with a perinatal mental health admission, followed by admissions for anxiety and post-traumatic stress disorders (19.3%). However, if we only considered the 176 women who had a perinatal mental health admission after a prepregnancy mental health admission, the most frequently observed diagnoses were related to psychoactive substance use (25.0%). Following this preparatory work, we demonstrated the clinical relevance of these data. Babies born to women with a prepregnancy history of perinatal health admission were found to be more likely to be preterm (12.0% born before 37 weeks), to have low birthweight (4.3% with birthweight below 2500 g in term babies) or to need some medical help (2.6% with an Apgar score less than 7 at 5 minutes after birth) than babies born to women without such a history (7.1%, 2.0%, and 1.7%, respectively). Outcomes in babies of women who had a perinatal mental health admission (590) were similar to those of women with a prepregnancy history of mental health admission (3043). Admission to an inpatient psychiatric mother-and-baby unit (MBU) was most frequent in women who had a mental health admission in the first 12 weeks after giving birth (79.5%) and considerably lower in women who had a mental health admission during pregnancy (23.7%) or between 13 and 52 weeks after giving birth (38.1%). Conclusions This study demonstrates the feasibility as well as the clinical relevance of using linked national maternity and mental health data sets from Scotland to assess the care that women with perinatal mental health problems receive. Despite only identifying women with severe perinatal mental health conditions, linkage of data sets of secondary care admission will offer an important opportunity to monitor the impact of national initiatives to improve perinatal mental health services in all four nations of the UK