Introduction

The Office of Health Improvement and Disparities (OHID), part of the Department of Health and Social Care (DHSC), has developed local modelled prevalence estimates for perinatal mental health (PMH) conditions in England. This is the first time that local level estimates have been produced for PMH conditions.

These prevalence estimates have been derived using a cohort of females who had a birth event in 2016, identified from the Clinical Practice Research Datalink (CPRD), a primary care sample database. The results from the CPRD cohort were produced for 2016 and then applied to 2016 and 2019 maternities data to produce prevalence estimates. The estimates for 2019 are presented in this report. Estimates were not produced beyond 2019 because the data to measure the impact of the COVID-19 pandemic is not yet available. See the ‘Methodology’ section for further detail.

See further local area analysis and details in the perinatal mental health profile on Fingertips.

The purpose of this report is to:

• present national PMH estimates separated into:
  • common mental disorders (CMDs)
  • a combined group of severe mental illness (SMI), eating disorders and personality disorders
• present analysis of modelled estimates by different administrative geographies in England
• describe how the new PMH condition prevalence estimates were derived, detailing what factors (age and deprivation) have been included to model local area estimates
• provide guidance to local areas on how they can use these estimates to understand need in their population
• review the likely impact of ethnicity on the available data and variation of PMH conditions

This report is for:

• national organisations that lead and develop strategy, policy and guidance for the mental health of females with PMH conditions
• local organisations that plan and deliver services for females with PMH conditions, including NHS and non-NHS providers of care

Why perinatal mental health is important

PMH conditions are defined as mental illnesses that affect females during the perinatal period. For the purpose of this work, the perinatal period is the 36-month period that spans:

• the 3 months prior to conception
• the duration of the pregnancy
• the 2 years post birth event

This definition of the perinatal period has been chosen to align with NHS England’s NHS Long Term Plan. This timeframe of 36 months is the period in which the prevalence of PMH conditions is estimated, using a period prevalence methodology. Throughout this report, we refer to the period prevalence as ‘prevalence’.

The perinatal period brings significant change and can be a vulnerable time. There is increased risk of poorer PMH if parents experienced difficulties in pregnancy, trauma in childbirth or an unexpected event in their baby’s development. It’s also a time when health inequalities can be exacerbated for those who experience discrimination, adversity and poor access to services.

Perinatal mental illnesses can have significant and long-lasting effects on the mother, child and the wider family if left untreated, resulting in wider costs to society, according to the Centre for Mental Health report Costs of perinatal mental health problems.

Evidence from the House of Commons Library debate pack briefing on Perinatal mental illness shows that:

• maternal depression and anxiety disorders have an adverse impact on the infant’s health and development
• psychotic illness in pregnancy is also known to be associated with an increased risk of pre-term delivery, stillbirth, infant mortality and neurodevelopmental disorder in the child
• perinatal psychiatric disorders are associated with an increased risk of mortality and morbidity in both mother and child

Evidence from MBRRACE-UK’s Perinatal confidential enquiry reports show that:

• over the past 2 decades, psychiatric disorder contributed to 15% of all maternal deaths in pregnancy and 6 months after birth
• suicide remains one of the leading causes of maternal mortality in the UK

Evidence from the Royal College of General Practitioners’ position statement on perinatal mental health shows that parental depression has a negative impact on how parents interact with children.  

Evidence shows that the risk of self-harm increases after childbirth, particularly for 15 to 24 year olds (see reference 1).

There is also evidence to suggest that the number of females in the UK with perinatal mental illnesses is increasing over time (see reference 2).

Summary of findings

The main findings show for females in England that had a birth event in 2019:

• the prevalence of PMH conditions is 25.8%
• the prevalence of CMDs is 25.3%
• the combined prevalence of SMI, personality disorders and eating disorders is 1.2%
• regionally there is higher prevalence of PMH conditions in the north (27.5% in the North East, 27.1% in Yorkshire and the Humber) than in the south (24.7% in London and 24.6% in the South East)

In addition, the review identified that:

• PMH need may be underestimated in areas with a higher proportion of females from ethnic minorities, who have had a birth event
• PMH need is higher in areas with younger populations and higher levels of deprivation

Note: PMH need is the proportion of females with a PMH condition.

Findings and interpretation

It’s important to explain the approach taken to develop the prevalence estimates. The prevalence estimates are derived from an analysis of primary care records from the CPRD database. A cohort of 128,070 females aged 15 to 55 years with a birth outcome in 2016 in England were extracted from CPRD.

Cases of PMH conditions were identified based on any of the following:

• a record of diagnosis of a mental health condition during the perinatal period
• a record of mental health condition symptoms or medication usage during the perinatal period
• combinations of a record of a historical diagnosis, with a record of mental health condition symptoms or medication use during the perinatal period
• a historical diagnosis of SMI at any point in time prior to the perinatal period.

PMH condition prevalence estimates were modelled by using the prevalence derived from the CPRD cohort disaggregated into 30 strata, by area deprivation quintiles and 6 age groups. These were applied to comparable strata in the England, regional, integrated care board (ICB) and upper tier local authority (UTLA) local populations and aggregated to model estimates for each geographical area.

Although evidence suggests that females from some ethnic minorities may experience a higher level of PMH conditions, ethnicity was not included in the model, as primary care data may not provide a complete picture of PMH need consistently among all ethnic minority groups (see the ‘Variation in PMH conditions by ethnicity’ section).

Selecting the 2016 calendar year allowed for the 2-year follow-up after birth for each female in the study. As some females had more than one pregnancy outcome or event recorded during 2016, for consistency, a female’s earliest record of a pregnancy birth event was used.

Modelled prevalence for PMH conditions in England

PMH condition prevalence in England is estimated to be 25.8% (95% credible interval (CrI) 25.5% to 26.1%) in the 2019 cohort (the maternities in 2019), see table 1.

Table 1: estimates of PMH condition prevalence in England by condition type, 2019 cohort

PMH conditions group	Estimated prevalence % England	95% CrIs	Estimated number of females affected	Range of estimated number of females affected
All PMH conditions	25.8	25.5 to 26.1	155,660	153,911 to 157,450
Common mental disorders (CMDs)	25.3	25.0 to 25.5	152,358	150,681 to     154,103
Other conditions, including SMI, eating disorders and personality disorder	1.2	1.1 to 1.2	6,960	6,600 to 7,455

Data source: OHID, based on CPRD and Office for National Statistics (ONS) births extract.

Table 1 provides the prevalence of the PMH condition subgroups. SMI, personality disorders and eating disorders have been grouped together as the case numbers were too small for these conditions to be reported individually. It should be noted that some females may have both CMD and a condition within the SMI, eating disorders and personality disorder subgroup.

Modelled prevalence for PMH conditions by geography

Prevalence of PMH conditions was modelled at 4 different sub-national geographies: UTLA, ICB, statistical region and NHS region (see table 2).

In this section of the report data for NHS region has not been presented because it is similar to statistical region figures. However, prevalence by NHS region can be found on Fingertips.

Table 2: estimates of PMH condition prevalence in England by geography, 2019 cohort

Geographical level	PMH condition prevalence, England​ females 15 to 55 years
Range at UTLA​ level	20.8% to 29.4%
Range at ICB level	22.6% to 28.2%
Range at statistical region	24.6% to 27.5%
Range at NHS region	24.6% to 27.2%

Data source: OHID, based on CPRD and ONS births extract.

Figure 1 below shows higher prevalence of PMH conditions in the north of England than in the south, and high prevalence in urban populations (though London is a notable exception to this).

Figure 1: map of variation in PMH condition prevalence in England by UTLA, 2019

This data can be found in table 1 of the accompanying data tables on the Perinatal mental health condition prevalence page.

Data source: OHID, based on CPRD and ONS births extract.

Tables 3 and 4 detail the UTLAs in the 2019 cohort where PMH condition prevalence was estimated to be the highest and lowest.

Table 3: highest prevalence of PMH conditions by UTLA in England, 2019 cohort

UTLA	Number of females with birth event	Estimated prevalence %	95% CrIs	Estimated number of females affected	Range of estimated number of females affected
Blackpool	1,548	29.4	29.0 to 29.9	456	448 to 463
Kingston upon Hull	3,151	29.2	28.8 to 29.7	921	907 to 936
Middlesbrough	1,744	28.9	28.4 to 29.3	504	496 to 512
North East Lincolnshire	1,646	28.8	28.4 to 29.2	474	467 to 480
Stoke-on-Trent	3,228	28.7	28.3 to 29.2	927	913 to 941
Walsall	3,672	28.6	28.2 to 29.1	1,051	1,035 to 1,068
Hartlepool	937	28.6	28.1 to 29.0	268	264 to 272
Oldham	3,112	28.5	28.0 to 28.9	886	872 to 900
Doncaster	3,423	28.4	28.1 to 28.8	973	960 to 987
Sandwell	4,333	28.4	28.0 to 28.9	1,231	1,212 to 1,251

Data source: OHID, based on CPRD and ONS births extract.

The estimated numbers of females affected for each UTLA in table 3 demonstrate that while areas may have the same prevalence rate, the numbers of females affected may vary greatly, due to differences in the population structure. For example, 28.6% prevalence in Walsall affects 1,051 females, whereas 28.6% prevalence in Hartlepool affects 268 females.

Table 4: lowest prevalence of PMH conditions by UTLA in England, 2019 cohort

UTLA	Number of females with birth event	Estimated prevalence %	95% CrIs	Estimated number of females affected	Range of estimated number of females affected
Richmond upon Thames	2,147	20.8	20.4 to 21.2	446	437 to 455
Wokingham	1,683	22.1	21.6 to 22.7	372	364 to 381
Kingston upon Thames	1,917	22.3	22.0 to 22.7	428	421 to 435
Wandsworth	4,407	22.3	22.0 to 22.7	984	970 to 999
Windsor and Maidenhead	1,497	22.4	22.0 to 22.8	335	330 to 341
Surrey	12,023	22.6	22.2 to 22.9	2,713	2,672 to 2,755
Bracknell Forest	1,290	23.0	22.6 to 23.3	296	291 to 301
Bromley	3,802	23.0	22.7 to 23.3	874	863 to 886
Merton	2,881	23.0	22.7 to 23.3	663	655 to 673
Buckinghamshire	5,568	23.1	22.8 to 23.5	1,287	1,268 to 1,306

 Data source: OHID, based on CPRD and ONS births extract.

Table 4 shows that PMH condition prevalence is lowest in South West London, Wokingham, and Windsor and Maidenhead. 

Figure 2, which presents variation in PMH condition prevalence by ICB, shows higher prevalence of PMH conditions in the north of England than in the south.

Figure 2: map of variation in PMH condition prevalence in England by ICB, 2019 cohort

This data can be found in table 2 of the accompanying data tables on the Perinatal mental health condition prevalence page.

Data source: OHID, based on CPRD and ONS births extract.

Tables 5 and 6 detail the ICBs in 2019 where PMH condition prevalence was estimated to be the highest and lowest. The patterns identified in the analysis of UTLAs is also reflected in ICBs.

Table 5: highest prevalence of PMH conditions by ICB in England, 2019 cohort

ICB	Number of females with birth event	Estimated prevalence %	95% CrIs	Estimated number of females affected	Range of estimated number of females affected
NHS Black Country	14,637	28.2	27.8 to 28.6	4,129	4,069 to 4,190
NHS South Yorkshire	14,712	27.6	27.3 to 28.0	4,065	4,013 to 4,119
NHS Birmingham and Solihull	17,515	27.6	27.1 to 28.0	4,826	4,754 to 4,900
NHS North East and North Cumbria	28,194	27.5	27.1 to 27.8	7,741	7,645 to 7,839
NHS West Yorkshire	27,664	27.1	26.8 to 27.5	7,500	7,404 to 7,598
NHS Lancashire and South Cumbria	17,218	27.1	26.8 to 27.4	4,665	4,608 to 4,723
NHS Greater Manchester	34,061	27.1	26.7 to 27.4	9,221	9,099 to 9,346
NHS Lincolnshire	66,96	26.9	26.5 to 27.2	1,798	1,777 to 1,820
NHS Staffordshire and Stoke-on-Trent	11,427	26.8	26.5 to 27.1	3,062	3,027 to 3,097
NHS Nottingham and Nottinghamshire	11,547	26.8	26.4 to 27.1	3,090	3,054 to 3,127

Data source: OHID, based on CPRD and ONS births extract.

Table 6: lowest prevalence of PMH conditions by ICB in England, 2019 cohort

ICB	Number of females with birth event	Estimated prevalence %	95% CrIs	Estimated number of females affected	Range of estimated number of females affected
NHS Surrey Heartlands	10,729	22.6	22.2 to 22.9	2,420	2,384 to 2,457
NHS South West London	19,126	23.3	23.0 to 23.5	4,446	4,391 to 4,504
NHS Buckinghamshire, Oxfordshire and Berkshire West	17,910	23.5	23.1 to 23.8	4,205	4,146 to 4,266
NHS Hertfordshire and West Essex	16,749	23.5	23.2 to 23.9	3,942	3,890 to 3,996
NHS Frimley	8,295	23.7	23.4 to 24	1,965	1,939 to 1,991
NHS South East London	22,867	24.5	24.1 to 24.8	5,590	5,519 to 5,664
NHS North West London	27,011	24.6	24.3 to 24.9	6,649	6,566 to 6,738
NHS Gloucestershire	6,031	24.7	24.4 to 25	1,491	14,72 to 15,10
NHS North Central London	18,009	24.8	24.5 to 25.1	4,458	44,04 to 45,14
NHS Bristol, North Somerset and South Gloucestershire	10,512	24.8	24.5 to 25.1	2,606	2,576 to 26,37

Data source: OHID, based on CPRD and ONS births extract.

Figure 3 shows the estimated PMH condition prevalence for the 2019 cohort by statistical region. It shows that prevalence of PMH conditions across the regions in England varies between 24.6% and 27.5%.

Figure 3: estimates of PMH condition prevalence (%) in England by statistical region, 2019 cohort

Data source: OHID, based on CPRD and ONS births extract.

Regionally there is higher estimated prevalence of PMH conditions in the north (27.5% in the North East, 27.1% in Yorkshire and the Humber), and lower prevalence in London (24.7%) and the South East (24.6%).  

PMH condition prevalence by age and deprivation

The data from CPRD was analysed to examine the association of age and deprivation with PMH condition prevalence within the study sample. Note the prevalence proportions detailed in this section are for the CPRD 2016 birth cohort only, they are not modelled prevalence estimates that extrapolate beyond the observed population. 

Age

Figure 4 shows that PMH condition prevalence varied by age group. There is a ‘U-shaped’ relationship between age and PMH condition prevalence in this cohort, with the highest proportions in those aged 15 to 19 years (39.6%) and lowest in ages 35 to 39 years (20.3%), and prevalence increases again for the 40 and over age group (22.3%).

Figure 4: PMH condition prevalence (%) in CPRD by age 15 to 55 inclusive, 2016

Age group	Prevalence of PMH (%)
15 to 19	39.58%
20 to 24	36.79%
25 to 29	28.22%
30 to 34	21.27%
35 to 39	20.30%
40 and over	22.32%

Data source: OHID, based on CPRD.

Area deprivation

There is a relationship between PMH condition prevalence and area deprivation; with higher levels of need (32%, 95% CrIs 31.1% to 32.9%) in the most deprived areas and lowest prevalence (20.2%, 95% CrIs 19.3% to 21.0%) in the least deprived areas (see figure 5).  However, the difference between deciles is not always the same, for example, the absolute difference between the most and second most deprived decile is higher than it is for the other deciles.

Figure 5: PMH condition prevalence (%) in CPRD by deprivation decile, 2016

IMD (2015) decile	Prevalence of PMH (%)
1 - most deprived	31.98%
2	27.31%
3	25.70%
4	26.15%
5	24.48%
6	25.08%
7	23.48%
8	23.21%
9	21.49%
10 - least deprived	20.15%

Data source: OHID, based on CPRD.

The relationship between area deprivation and PMH condition prevalence varies by age group. For the youngest age group with teenage pregnancies, which have the overall highest prevalence of PMH conditions, there is evidence of greater need in those living in the least deprived areas. This contrasts with age groups from age 25 onwards, where there is higher prevalence in the more deprived areas (see figure 6).

Therefore, targeting support for the most vulnerable groups must take into account both age, and social and economic circumstances.

Figure 6: PMH condition prevalence (%) in CPRD by age group and Index of Multiple Deprivation (IMD) quintile, 2016

This data can be found in table 3 of the accompanying data tables on the Perinatal mental health condition prevalence page.

Data source: OHID, based on CPRD.

Variation in PMH conditions by ethnicity

Initially the model for producing local level estimates of PMH conditions included ethnicity in addition to deprivation and age. However, after reviewing related evidence, the decision was made to exclude ethnicity from the model on the basis that there was a risk that PMH need in minority ethnicity females may be under-represented in primary care data.

This section summarises the evidence on variation in PMH conditions by ethnicity in primary care.  It is drawn from a rapid review of research exploring inequalities in mental health for females from ethnic minority backgrounds. Evidence shows inequalities in the prevalence of PMH conditions (see reference 3) and in the recording of PMH conditions (see references 4 and 5) for females from ethnic minority backgrounds. The following factors contribute to this: 

• females from ethnic minority backgrounds are less likely to register with a GP practice (see reference 6)
• females from ethnic minority backgrounds are less likely to seek support from their GP and less likely to disclose their condition or concerns (see references 3 and 6)
• females from ethnic minority backgrounds are less likely to access and utilise mental health services in the perinatal period (see reference 4), which may lead to a lower level of recording in primary care notes based on communication between services
• females from ethnic minority backgrounds have lower satisfaction with primary healthcare services which may correlate with accessibility challenges (see reference 7)
• information on females from ethnic minority background and their mental health during the perinatal period are less likely to be recorded (see reference 8), with some evidence suggesting they are less likely to be asked about their PMH (see reference 9)
• females from ethnic minority backgrounds are less likely to have CMDs diagnosed in primary care during the perinatal period (see reference 10)

However, evidence relating to the density of ethnic populations is worth considering.  Some studies suggest that ethnic minorities living in communities with a higher proportion of co-ethnic populations report better mental health. However, the impact is not the same for all communities and residential segregation may play an important role alongside density of ethnic populations (see references 11 and 12).

Evidence also suggests that when services are made accessible, females from ethnic minority backgrounds show higher levels of utilisation, which may reflect a greater need for support provided by these services (see reference 4).

Figure 7 is included to show the variation in diagnosed and self-reported PMH condition prevalence by ethnicity at 9 months and 3 years after birth. The results show that White females have higher diagnosis rates, whereas females from ethnic minority backgrounds have higher levels of self-reporting of mental health symptoms.

Figure 7: diagnosed and self-reported PMH condition prevalence (%) by ethnicity at 9 months and 3 years after birth, 2000 to 2004

This data can be found in table 4 of the accompanying data tables on the Perinatal mental health condition prevalence page.

Data source: Millenium Cohort Study (MCS) - wave 1 (2000) and wave 2 (2004).

It is important to note that the figure presents results from females 3 years after birth. This is beyond the end of the recognised perinatal period but is indicative of on-going mental health need.

Guidance on how to use these findings

Considering the estimates in relation to your local area

Local area estimates for PMH conditions are provided at a range of geographies for service planners to use directly from the perinatal mental health profile on Fingertips.  It should be noted that these are modelled prevalence estimates, based on a CPRD sample and adjusts for variation in age and area deprivation of the local primary care population.

The report also provides prevalence proportions for PMH conditions by age and deprivation taken directly from the CPRD cohort. They are not modelled England level estimates and therefore do not give a direct estimate for PMH need by age and deprivation for England. However, they do provide an indication of what variation in PMH conditions by age and deprivation may look like in local areas.

When considering local PMH need it is important to recognise that need is dependent on the make-up of the local population that have had a birth event and not the population as a whole. For example, an area may have a large elderly population and a low birth rate, but amongst those who have given birth, need may be high.

The report also provides prevalence estimates according to the types of PMH condition, one being CMD and the second being a grouping of SMI, personality disorders and eating disorders. These estimates will provide an indication of the different types and level of services required to address the different conditions.

These are the best available estimates of PMH condition prevalence across the span of the perinatal period for females with a birth event in a specific year. The 2019 PMH prevalence figure (25.8%) estimates need in females who had a birth event in England in 2019, although that need may occur at any point in the perinatal period. However, the 25.8% figure is a reasonable indication of total need in 2019. This is because in 2019 PMH services are available to females who are in the perinatal period in 2019, which amounts to a 4-year period. This is females who had a birth event between 1 January 2017 and 31 December 2019, and those females in the 1-year pre-birth period who are likely to have a birth event in 2020.

Considering ethnicity

Ethnicity has not been included in the model due to data limitations. However, although not all females from ethnic minority backgrounds have the same experience, there is evidence that many experience inequalities in access to services (primary care and specialist mental health) during the perinatal period.

It is important for service planners and providers to take into consideration inequalities. They should consider that the PMH need in local populations with higher proportions of females from ethnic minority groups who have a birth event may be higher than those who access services and also potentially higher than these local estimates suggest. It is important for local services to work with their communities to co-design culturally sensitive services that can be accessed by different ethnic minority groups.

Methodology

This section describes the methodology followed to produce the prevalence estimates. The development of the estimates was advised by an external expert reference group (ERG) made up of academics, clinicians, and mental health policy leads with expertise in data, service and clinical areas relating to perinatal mental health.

Data source

CPRD was chosen as the most appropriate available data set to estimate prevalence of PMH conditions across England. CPRD is a primary care sample data set of anonymised patient records including medical observations (diagnosis, symptoms and treatments), medication prescriptions, and practice details. The data set covers around 20% of the total registered practice population in England. CPRD was selected for this study because:

• of the large sample size
• it is broadly representative of the population in England
• its good data quality in terms of recording and high levels of completeness (see reference 12)

Terminology and definitions

Perinatal period

The perinatal period is defined as starting from 1 year before birth (3 months pre-conception and up to 9 months of pregnancy) and ending 2 years after the birth event.

Prevalence

Prevalence in this work was defined as a recorded mental health condition that was present during the perinatal period (see the ‘Case identification’ section for further detail).

Period prevalence

This is prevalence recorded over a period of time. The period of time here was the perinatal period, which consists of the 12 months prior to birth, and 24 months after birth. The study population was based on the perinatal period of females who had a birth event in 2016. The perinatal period for this cohort starts from 1 January 2015 (12 months before the first birth on 1 January 2016) and ends on 31 December 2018 (24 months after the last birth on 31 December 2016).

Birth event

This is defined in CPRD as either a direct or indirect birth outcome. Direct birth outcomes are medical observations that record an actual birth event or delivery. Still and premature births at 24 weeks gestation or later are included. Indirect birth outcomes are medical observations that record a procedure or process indicating a birth event or delivery, for example, breast-feeding initiation or a post-natal GP check.

Maternities

Using ONS birth extracts, a maternity is a pregnancy resulting in the birth of one or more children, including live and still births.   

Perinatal mental health condition

For the purposes of this work, a ‘prevalent case’ for a perinatal mental health condition is defined by National Institute for Health and Care Excellence (NICE) clinical guideline Antenatal and postnatal mental health: clinical management and service guidance (CG192) as a female who has at least one of the mental health conditions outlined below:

• CMDs including:
  • depression
  • anxiety (generalised anxiety disorder, panic disorder)
  • obsessive compulsive disorder (OCD)
  • post-traumatic stress disorder (PTSD)
  • antenatal and postnatal depression
  • social anxiety disorder (also known as social phobia)
  • seasonal affective disorder
• personality disorders
• eating disorders including:
  • anorexia nervosa
  • binge eating disorder
  • bulimia nervosa
• severe mental illness (SMI) including:
  • bipolar disorder
  • psychosis
  • schizophrenia
  • perinatal mental health psychosis

Case identification

There is anticipated variation in how GP practices manage the recording of patient mental health conditions. Therefore, the project had to account for this in its approach to identifying cases of perinatal mental health conditions from CPRD. For some females with PMH conditions, GPs will record symptoms and medication rather than actual diagnosis.  Therefore, case identification for this study was based on a mixture of recorded diagnoses, symptoms and medication - outlined below.

The first stage of case identification of females with a PMH condition for this study was to select the birth cohort from CPRD in 2016. This comprised of females, aged 15 to 55 in England who had a birth event between 1 January 2016 and 31 December 2016, for whom a full medical record was available for the whole perinatal period. From this cohort (128,070), cases were identified in the following order:

• group 1: females (23,205) with a diagnosis code for a mental health condition in the perinatal period
• group 2: if not captured in group 1, females (4,119) with a mental health condition symptom code and relevant medication code in the perinatal period
• group 3: if not captured in group 1 and 2, females (4,497) with a historical mental health condition diagnosis code (before perinatal period) and a relevant medication code in the perinatal period or with a historical mental health condition diagnosis code (before perinatal period) and relevant symptom code in the perinatal period
• group 4: if not captured in group 1,2, and 3, females (45) with a historical SMI diagnosis code (before perinatal period)

The prevalent cohort of females with a PMH condition is all of these 4 different groups added together, which totalled 31,866.

Breakdown of the CPRD cohort

Table 7 provides a breakdown of females with a birth event from CPRD in 2016, and compares it to ONS births in 2016.  It shows how the cohort was split across the 9 regions of England.

Table 7: CPRD 2016 birth cohort breakdown and ONS 2016 births by statistical region

Statistical region	Number of females with a birth outcome count in CPRD	Percentage of females with a birth outcome in CPRD	Number of maternities in total population	Percentage of maternities
East Midlands	3,042	2.4	52,699	8.0
East of England	6,229	4.9	71,337	10.9
London	27,105	21.2	127,217	19.4
North East	4,132	3.2	28,206	4.3
North West	20,545	16.0	85,190	13.0
South East	27,174	21.2	100,728	15.4
South West	15,250	11.9	56,606	8.6
West Midlands	20,209	15.8	70,268	10.7
Yorkshire and the Humber	4,384	3.4	63,201	9.6
Total	128,070	100%	655,452	Not applicable

Data source: OHID, based on CPRD and ONS births extract.

Table 8 provides a breakdown of the different ages of the females with a birth event from CPRD in 2016, and compares it to ONS births in 2016. The smallest counts were in age groups 15 to 19 years (2,467 females) and 45 years and over (904 females). The highest count was in age group 30 to 34 years (42,554 females).

Table 8: CPRD 2016 birth cohort breakdown and ONS 2016 births by age

Age group	Number of females with a birth outcome count in CPRD	Percentage of females with a birth outcome in CPRD	Number of maternities in total population	Percentage of maternities
15 to 19	2,467	1.9	20,815	3.2
20 to 24	15,023	11.7	96,053	14.7
25 to 29	32,117	25.1	184,258	28.1
30 to 34	42,554	33.2	208,078	31.8
35 to 39	28,066	21.9	118,320	18.1
40 to 44	6,939	5.4	25,899	4.0
45 and over	904	0.7	2,029	0.3

Data source: OHID, based on CPRD and ONS births extract.

Table 9 provides a breakdown by IMD 2015 decile, of the females with a birth event from CPRD in 2016 and compares it to ONS births in 2016.

Table 9: CPRD 2016 birth cohort breakdown and ONS 2016 births by IMD decile

IMD decile	Number of females with a birth outcome count in CPRD	Percentage of females with a birth outcome in CPRD	Number of maternities in total population	Percentage of maternities
1 - most deprived	14,392	11.2	92,192	14.1
2	15,082	11.8	84,587	12.9
3	14,459	11.3	77,964	11.9
4	13,538	10.6	70,742	10.8
5	11,795	9.2	63,599	9.7
6	11,445	8.9	60,506	9.2
7	11,463	9.0	55,749	8.5
8	11,757	9.2	54,079	8.3
9	11,447	8.9	50,792	7.8
10 - least deprived	12,692	9.9	45,242	6.9

Data source: OHID, based on CPRD and ONS births extract.

Weighting CPRD to make the cohort representative of England

Weighting was used to make the cohort, and therefore the prevalence estimates, representative of the England population. CPRD is a large, non-random sample of GP practices and not all regions are equally represented. Table 7 shows the comparison between ONS maternities data and CPRD data by statistical region. The table shows that not all areas are equally represented within this cohort; and that some are over-represented (for example, the West Midlands, London and the South West) and some under-represented (for example, Yorkshire and the Humber, and the East Midlands). Therefore, a process of weighting was undertaken to ensure the results from CPRD were more representative of England. Using ONS maternities data as the reference, weighting of the birth cohort was carried out to account for regional differences in CPRD and ONS maternities.

These weighted prevalence proportions were used to model prevalence for England, government office regions, NHS regions, UTLAs and ICBs.

Although CPRD is not regionally representative, comparison to ONS births data shows that it aligns well with females giving birth in England by age groups and deprivation deciles, hence no weighting was needed to be carried out for age or deprivation.

Modelled prevalence estimates

The model used to calculate PMH condition prevalence in England, and for each UTLA, ICB, NHS and government office region was based on the following numerator and denominator:

• numerator - this is the estimate of the total number of females (aged 15 to 55 years) expected to have a PMH condition, based on the number of females by age and area deprivation of giving birth (ONS maternities) for that area and weighted prevalence proportions from CPRD
• denominator - this is the number of maternities in females (aged 15 to 55 years) derived from ONS maternities data for England and each UTLA, ICB, NHS and government office region

The final estimates are presented as a proportion of females with a birth event in an area with a PMH condition.

The prevalence of PMH conditions at a local level is calculated by applying age- and deprivation-specific prevalence proportions derived from CPRD to the local age- and deprivation-specific population of females who had a birth event. Using simulation, the local estimated prevalence proportions have credible ranges produced which take into account the uncertainty around the CPRD derived prevalence. These credible ranges are presented as confidence intervals around the estimate.

Local level estimates are modelled by UTLA, ICB, NHS region and statistical region.

Refreshing the model for 2019 local ONS population data

The 2016 model for calculating perinatal mental health period prevalence was applied to the 2019 local ONS population to provide as up to date figures as possible. Therefore, the estimates for 2019 reflect changes in the age and deprivation structure of females aged 15 to 55 who are giving birth rather than changes in the level of perinatal mental health needs.

Development of a post-COVID-19 pandemic prevalence estimate

A post COVID-19 pandemic update to the prevalence estimates was explored. The COVID-19 pandemic had an impact upon the numbers of females accessing GP services, particularly during the 2 lockdowns, with many females not accessing GP services for problems they would have otherwise done before the pandemic.  As a result, primary care data in 2020 and the start of 2021, will not reflect the true number of females who had a mental illness and so there is a likelihood that data taken during this period would underestimate prevalence of PMH.

The earliest CPRD cohort that can be used post COVID-19 pandemic would be the 2023 birth cohort. Since the perinatal period for the 2023 birth cohort includes the 2 years post birth, data for all of 2025 would be needed. Once full year CPRD records for 2025 have been obtained the feasibility of a post-pandemic update will be pursued.

Limitations of the indicator

Please note the following limitations regarding the data:

• source data is from the 2015 to 2018 perinatal period and 8 or more years old, though the prevalence estimates will be updated once full year CPRD records for 2025 have been obtained
• source data is from before the COVID-19 pandemic (and lockdowns), and does not account for the impact this will have had on the prevalence of PMH conditions, this will be addressed once full year CPRD records for 2025 have been obtained
• only females registered with GPs are considered in this analysis, nevertheless it has to be acknowledged that only a very small proportion of the population is not registered during pregnancy
• this study analysis relies on the medical record history covering the full perinatal period and for the female to remain with the same GP practice until the end of the period.

References

1. Hope H, Pierce M, Cemre SO, Morgan C, John A and Abel KM. Self-harm risk in pregnancy: recurrent-event survival analysis using UK primary care data. The British Journal of Psychiatry, volume 221, issue 4 (2022): pages 621 to 627.

2. Abel KM, Hope H, Swift E, Parisi R, Ashcroft DM, Kosidou K and others. Prevalence of maternal mental illness among children and adolescents in the UK between 2005 and 2017: a national retrospective cohort analysis. The Lancet, Public Health, volume 4, issue 6 (2019).

3. Lee SI, Azcoaga-Lorenzo A, Agrawal U and others. Epidemiology of pre-existing multimorbidity in pregnant women in the UK in 2018: a population-based cross-sectional study. BMC Pregnancy Childbirth 22, 120 (2022).

4. Jankovic J, Parsons J, Jovanović N and others. Differences in access and utilisation of mental health services in the perinatal period for women from ethnic minorities - a population-based study. BMC Medical 18, 245 (2020).

5. Catalao R, Dorrington S, Pritchard M and others. Ethnic inequalities in mental and physical multimorbidity in women of reproductive age: a data linkage cohort study. BMJ Open, volume 12, issue 7 (2022).

6. Bansal N, Karlsen S, Sashidharan SP, Cohen R, Chew-Graham CA and Malpass A. Understanding ethnic inequalities in mental healthcare in the UK: A meta-ethnography. PLoS Med. 13 December 2022.

7. Magadi JP and Magadi MA. Ethnic inequalities in patient satisfaction with primary health care in England: Evidence from recent General Practitioner Patient Surveys (GPPS). PLoS One. 21 December 2022.

8. Prady SL, Pickett KE, Petherick ES and others. Evaluation of ethnic disparities in detection of depression and anxiety in primary care during the maternal period: Combined analysis of routine and cohort data. British Journal of Psychiatry, volume 208, issue 5 (2016): pages 453 to 461.

9. Harrison S, Pilkington V, Li Y and others. Disparities in who is asked about their perinatal mental health: an analysis of cross-sectional data from consecutive national maternity surveys. BMC Pregnancy and Childbirth 23, article number 263 (2023).

10. Prady SL, Pickett KE, Gilbody S and others. Variation and ethnic inequalities in treatment of common mental disorders before, during and after pregnancy: combined analysis of routine and research data in the Born in Bradford cohort. BMC Psychiatry 16, article number 99 (2016).

11. Shankley W and Laurence J. Community ethnic density, ethnic segregation, and ethnic minorities’ common mental disorders in the UK. Health & Place Volume 73, 102723 (2022).

12. Becares L, Dewey ME and Das-Munchi J. Ethnic density effects for adult mental health: systematic review and meta-analysis of international studies. Psychological. Medicine 48 (12) (2017).

13. CPRD Useful publications on the quality of CPRD data for research.