The natural history and NHS service use of women with depression and impact on children in a multi-ethnic inner-city population

Between 5 to 13 percent of women have symptoms of depression during or after a pregnancy. Some of these women go to their GP with symptoms, are correctly identified as depressed and undergo treatment. However, for various reasons some women who feel depressed do not go to their GP. Some will go to their GP but will fail to be recognised as depressed and so will not be offered treatment. Older research indicates that about half of all women at risk of depression remain undiagnosed (either do not go to the GP or fail to be diagnosed by their GP). This estimate may be out of date. Untreated, in addition to the poor quality of life experienced by the woman, it is well documented that maternal depression can affect some children and families; leading to relationship breakdown, poor child behaviour and emotional well-being and reduced cognitive development. For some children, this translates in later life to difficulties in school, problems with relationships, increased levels of aggression and poor mental health themselves. These potentially preventable disabilities come at great personal cost, as well as high financial costs to the NHS and social care agencies, the criminal justice system, and wider society.

Our research involves examining electronic GP records for a large number of mothers (~13,000) who have already elected to participate in a research study. The women live in an ethnically diverse economically deprived city in the UK; factors that make them more at risk for depression than the general UK population. They have consented for us to look at their NHS records. Using record linkage in this way allows us to look at routine NHS data in addition to extensive social, economic data and measures of mental health that have been collected through the research study. This is called an enhanced dataset and will allow us to explore a number of issues that are important to the health of mothers and their families.

Using this enhanced dataset we can calculate the number of women who have been identified by the research study as being depressed during their pregnancy but have no GP record of depressive symptoms. We can then provide an updated estimate of the percentage of women whom we already know to be depressed, but who are unidentified to the NHS as depressed. We will then look at the economic and social profiles of three groups of women; those who are unidentified as depressed, those identified as depressed and those who are not depressed. This will enable us to understand in what circumstances unidentified women live and may provide clues about how the NHS can provide services to reach them.

One in ten of these women (~1,300) have been followed up with questionnaires about their lifestyle, circumstances and mental health every six months or so since the birth of their child. These children are now around 3-4 years old and their mothers have recently assessed their child's behaviour using a standardised questionnaire. We can combine this assessment of child behaviour with our classification of the mother's mental health (not depressed, treated for depression, not identified as depressed) to estimate any impact on children. This information might help future researchers more accurately predict how much untreated depression costs the NHS. We can also use the enhanced dataset to understand whether the circumstances of some of the women, for example their ethnicity, their marital or economic situation, and their parenting skills protect their children from having behavioural problems. This will help us understand mental health inequality and identify groups of children most at risk of developing behavioural problems so services such as parenting programmes can be targeted.

We aim to describe the natural history of depression and its consequences for women pre- and post-pregnancy in a multi-ethnic and economically deprived community in England.

We plan to build an enhanced dataset composed of a suite of baseline measures taken at ~28 weeks gestation from the 13,000 women enrolled in the Born in Bradford Cohort; uniquely linked to the local GP database, SystmOne. To extract the NHS data we will devise and test an algorithm of previously validated Read codes for depression diagnoses, screening, symptoms, and treatment referrals along with anti-depressant prescribing and counts of consultations for any reason. NHS data will be drawn 15 months prior to the index birth and for 3 years post-natally. Further measures of mental health (GHQ-28 and Kessler Scale for Psychological Distress) and parenting practices are available for 5 time points up to 3 years post-partum for a subsample of ~1300 women along with child behavioural outcomes (SDQ) at 3 years.

After data cleaning and validation, we will use rates and ratios to describe the data and population in terms of demographic characteristics, mental health and presentation to primary care. Receiver operator characteristic (ROC) curves will be used to calibrate the relative accuracy of cohort mental health measures against each other and NHS diagnoses. We will use structural equation modelling to explore the following; latent class analysis for characteristics related to presentation/non-presentation to primary care and latent transition for fluctuations over time, path analysis to explore moderators and mediators of child outcomes and confirmatory factor analysis to examine fit of latent measures.

The results will describe social inequality in service use and quantify impacts of health seeking behaviour on children. Findings will be disseminated through academic routes, exploitation of local public involvement and engagement with NHS and other agencies via dissemination specialists.

We have identified four main groups of potential beneficiaries of the proposed research:

1. The maternal population and their families
2. The community of Bradford
3. Health and social care professionals (visitors, midwives, GP's, commissioners and providers of mental health services, social care agencies, local authorities, acute trusts, third sector agencies, the National Screening Committee
4. Academics in fields of mental health, inequality, child development, social care and ethnicity.

Our strategies for impact with all four groups have been developed in collaboration with experts in knowledge exchange and dissemination and are detailed in our Pathways to Impact statement.

We anticipate our findings will benefit the general population, and the maternal and child population, indirectly by quantifying any consequences of not seeking treatment for depression, calculating the effect of seeking treatment, and understanding risk and resilience in relation to child behaviour outcomes. This should enable improvements in targeted provision of preventive, primary and secondary care, as well as cross-sector response to maternal depression. Research conducted in BiB offers potential benefits to the population of Bradford and helps to enhance local research capacity. NHS health agencies such as primary, specialist and maternal care will benefit from our longitudinal, long term results as the proposed research is designed to provide direct, robust evidence of any under-reporting of depression and so help them identify and target groups of vulnerable women. Other non-health government agencies such as local authorities, and public health agencies can use results to further refine targeted health and social inequality plans and this is likely to be of interest for third sector agencies, who we also anticipate will use our results to promote mental wellbeing.