Parts of this chapter are reproduced with permission from Gurol-Urganci et al.93 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) licence, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/. The text includes minor additions and formatting changes to the original text.
Background
In 2014, a guideline for the NHS in England recommended that women with pre-existing complex and severe mental illness should be referred to specialist mental health services during pregnancy and postnatal period, preferably set up as a CPMHT for timely assessment and treatment treatment.16,19
The NHS in England announced in the 2016 5 Year Forward Plan an investment of £365M to be spent over a 5-year period in order to improve overall access to PMH maternity services equitably across England, and to improve the treatment of women with new mental health problems and to reduce the risk of relapse of women with pre-existing mental health problems, particularly women at risk of severe acute relapses needing hospitalisation or the input of a specialist mental health CRT which provides intensive support in the community input. Before 2016, there had only been a few CPMHTs in England, but there has been a gradual increase in their coverage following this investment.
Aim
The primary aim of this study was to:
- test the hypothesis that the gradual implementation of CPMHTs was associated with an increase in the overall use of ‘specialist mental healthcare services’, defined as any contact with CPMHTs and other type of specialist (secondary) mental healthcare services (including day care, outpatient visits, and community-based care) in the perinatal period (i.e. pregnancy and the year after birth)
- and a reduction in the risk of ‘acute relapse’ in the postnatal period (i.e. the year after birth), defined as an admission to a psychiatric ward or psychiatric MBU or the involvement of a CRT.
In this chapter, we report that the risk of adverse obstetric and neonatal outcomes was increased in women with at least one pre-pregnancy contact with specialist mental health care. We showed that the risk of adverse outcomes was higher in women with mental healthcare contacts related to more severe mental illness and in those with more recent mental healthcare contacts. Therefore, we include as a secondary aim to evaluate the association of the implementation of CPMHTs with the risk of adverse obstetric and neonatal outcomes.
Subgroup analyses
In the first subgroup analysis, we only included women who had received pre-pregnancy mental healthcare contacts that reflect a more severe pre-existing mental disorder. In the second subgroup analysis, we only included women who had received more recent pre-pregnancy mental healthcare contacts. These subgroups were chosen as they define women who are thought to be at an increased risk of a relapse in the postnatal period.
Methods
Data sources and linkage
The data sources described for this study are already described earlier in this chapter.
Cohort selection and comparison groups
Maternity episodes for women, aged 18 and above, with a recorded gestation of at least 24 completed weeks, with an onset of pregnancy after 1 April 2016 and a singleton birth before 31 March 2018 in the English NHS were identified from HES records and PDS birth notifications (see Report Supplementary Material 1) because for these women the mental health outcomes were captured in the most recent version of the specialist mental healthcare data set.
Women were considered to have a pre-existing mental disorder if they had a record of a contact with any form of specialist (secondary) mental health care within 10 years before the onset pregnancy. Please note that in this chapter, we could use a ‘look-back period’ to detect a pre-existing mental disorder of only 7 years. (See Chapter 5 for a description of how we managed multiple maternity episodes, onset of pregnancy and identification of pre-existing mental disorder, and distinguished levels of specialist mental health care.)
Chapter 2 explained how we collected information on the regional presence of CPMHTs at the level of the CCGs. In this study, we define the regions according to the 207 CCGs present in July 2017. Typically, these regions cover areas with a population of 250,000 people. Women were considered to have had access to a CPMHT if the date of CCG implementation in their region was before the onset of pregnancy.
Study outcomes
The primary mental health outcome was ‘acute relapse’ (‘psychiatric hospital admission’ or a ‘CRT contact’) in the postnatal period (i.e. in the first year after birth). The secondary mental health outcome was ‘any specialist care contact’ defined as at least one specialist mental healthcare contact during the perinatal period (i.e. during the pregnancy and the first year after birth), also including a specialist community mental healthcare contact.
The obstetric and neonatal outcomes are described in this chapter in the Study outcomes section.
Statistical analysis
The analysis approach follows the approach described in this chapter.
Results
The flow of data and description of participants is demonstrated in Figure 10. Figure 11 demonstrates that we identified 780,026 eligible women with a singleton birth and that 70,323 (9.0%) women with a pre-existing mental disorder were included.
FIGURE 10 Flow diagram indicating number of women with and without pre-pregnancy mental healthcare contact, by type of contact. (Figure not reproduced here; see the original article.)
FIGURE 11 Flow diagram (CPMHT). (Figure not reproduced here; see the original article.)
Figure 12 shows that in April 2016 a CPMHT was available in 81 of the 207 CCG regions (39.1%). The number of regions with a CPMHT increased to 130 (62.8%) in June 2017, just before the onset of pregnancy of women who gave birth in March 2018, the most recent month of our cohort (more information can be found in Chapter 2).
FIGURE 12 Availability of CPMHTs as present in July 2017 in 207 regions defined according to areas covered by NHS CCGs (orange: CPMHT available; blue CPMHT not available). (Figure not reproduced here; see the original article.)
Of the 70,323 included women, 31,276 (44.5%) gave birth when a CPMHT was available in their region at the onset of pregnancy and 39,047 (55.5%) when a CPMHT was not available. The characteristics of women with and without a CPMHT available were broadly similar (Table 16), with a lower proportion of white women in regions with a CPMHT than in regions without.
| All women | No CPMHT at onset of pregnancy | CPMHT at onset of pregnancy | |
|---|---|---|---|
| 70,323 | 39,047 (55.5) | 31,276 (44.5) | |
| Maternal age missing n = 6; 0.01% | |||
| 18–24 | 17,463 (24.8) | 9971 (25.5) | 7492 (24.0) |
| 25–34 | 40,031 (56.9) | 22,219 (56.9) | 17,812 (57) |
| 35–39 | 10,216 (14.5) | 5489 (14.1) | 4727 (15.1) |
| 40 ≤ | 2607 (3.7) | 1364 (3.5) | 1243 (4.0) |
| Obstetric history missing n = 3589; 5.1% | |||
| Nulliparous | 22,665 (34.0) | 12,749 (34.5) | 9916 (33.4) |
| Multiparous, no previous cs | 35,423 (53.1) | 19,558 (52.8) | 15,865 (53.4) |
| Multiparous, with previous cs | 8646 (13.0) | 4700 (12.7) | 3946 (13.3) |
| Ethnicity missing n = 7843; 11.2% | |||
| White | 54,965 (88.0) | 31,199 (89.9) | 23,766 (85.5) |
| South Asian | 3244 (5.2) | 1622 (4.7) | 1622 (5.8) |
| Black | 1779 (2.8) | 692 (2.0) | 1087 (3.9) |
| Mixed | 1332 (2.1) | 661 (1.9) | 671 (2.4) |
| Other stated | 1160 (1.9) | 525 (1.5) | 635 (2.3) |
| Socioeconomic deprivation missing n = 3, < 0.01% | |||
| Least deprived = quintile 1 | 7373 (10.5) | 4268 (10.9) | 3105 (9.9) |
| Quintile 2 | 9765 (13.9) | 5528 (14.2) | 4237 (13.5) |
| Quintile 3 | 12,560 (17.9) | 6820 (17.5) | 5740 (18.4) |
| Quintile 4 | 16,522 (23.5) | 8492 (21.7) | 8030 (25.7) |
| Most deprived = quintile 5 | 24,100 (34.3) | 13,937 (35.7) | 10,163 (32.5) |
| Pregnancy risk factors missing n = 3589; 5.1% | |||
| Pre-existing diabetes | 1021 (1.5) | 502 (1.4) | 519 (1.7) |
| Pre-existing hypertensive conditions | 488 (0.7) | 250 (0.7) | 238 (0.8) |
| Gestational diabetes | 4294 (6.4) | 2263 (6.1) | 2031 (6.8) |
| Pre-/eclampsia | 1442 (2.2) | 779 (2.1) | 663 (2.2) |
| Highest level of pre-pregnancy contact | |||
| Community care contact | 53,098 (75.5) | 28,880 (74.0) | 24,218 (77.4) |
| CRT | 13,832 (19.7) | 8242 (21.1) | 5590 (17.9) |
| Hospital admission | 3393 (4.8) | 1925 (4.9) | 1468 (4.7) |
| Timing of most recent pre-pregnancy contact | |||
| > 5 years | 18,302 (26.0) | 10,219 (26.2) | 8083 (25.8) |
| 1–5 years | 34,672 (49.3) | 19,770 (50.6) | 14,902 (47.7) |
| < 1 years | 17,349 (24.7) | 9053 (23.2) | 8291 (26.5) |
Acute relapse
Of the 70,323 included women, 2862 (4.1%) had an acute care relapse (a psychiatric hospital or CRT contact) in the postnatal period (the year after birth). The risk of acute relapse varied according to regional CPMHT availability (Table 17). Of the 31,276 women who gave birth when a CPMHT available at the onset of pregnancy, 1117 (3.6%) had an acute relapse in the year after birth compared to 1745 (4.5%) of the 39,047 women who gave birth CPMHT when was not available (adjOR 0.77, 95% CI 0.64 to 0.92; p-value = 0.004).
| Primary outcomes | n (%) | No CPMHT at onset of pregnancy | CPMHT at onset of pregnancy | Unadjusted OR (95% CI) for CPMHT | p-value | Adjusted, with time trends only adjOR (95% CI) for CPMHT | p-value | Adjusted, with time trends and maternal characteristics.adjOR (95% CI) for CPMHT | p-value |
|---|---|---|---|---|---|---|---|---|---|
| Acute care contact (CRT/Hosp) | |||||||||
| Perinatal | 3932 (5.6) | 2310 (5.9) | 1622 (5.2) | 0.87 (0.74 to 1.02) | 0.095 | 0.87 (0.73 to 1.02) | 0.095 | 0.84 (0.71 to 0.99) | 0.041 |
| Postnatal | 2862 (4.1) | 1745 (4.5) | 1117 (3.6) | 0.79 (0.66 to 0.95) | 0.011 | 0.79 (0.66 to 0.94) | 0.009 | 0.77 (0.64 to 0.92) | 0.004 |
| Pregnancy | 1788 (2.5) | 1031 (2.6) | 757 (2.4) | 0.91 (0.75 to 1.11) | 0.373 | 0.92 (0.76 to 1.12) | 0.428 | 0.89 (0.73 to 1.08) | 0.248 |
| Any care contact (other/CRT/Hosp) | |||||||||
| Perinatala | 19,921 (28.3) | 10,033 (25.7) | 9888 (31.6) | 1.33 (1.22 to 1.47) | < 0.001 | 1.32 (1.20 to 1.46) | < 0.001 | 1.35 (1.23 to 1.49) | < 0.001 |
| Pregnancy | 14,555 (20.7) | 7001 (17.9) | 7554 (24.2) | 1.46 (1.31 to 1.63) | < 0.001 | 1.44 (1.29 to 1.62) | < 0.001 | 1.50 (1.34 to 1.68) | < 0.001 |
| Postnatal | 15,366 (21.9) | 7866 (20.1) | 7500 (24) | 1.25 (1.14 to 1.37) | < 0.001 | 1.24 (1.13 to 1.36) | < 0.001 | 1.23 (1.13 to 1.36) | < 0.001 |
There was no association between the regional availability of a CPMHT and acute relapse during pregnancy. Seven hundred and fifty-seven (2.4%) of 31,276 women with a CPMHT available had an acute relapse during pregnancy compared to 1013 (2.6%) of the 39,047 women who gave birth without CPMHT available (adjOR 0.89, 95% CI 0.73 to 1.08; p-value = 0.248).
Any specialist mental healthcare contact
Of the 70,323 included women, 19,921 (28.3%) had any type of specialist mental healthcare contact in the perinatal period (during pregnancy and in the year after birth). The percentage of women having any specialist mental healthcare contact in the perinatal period varied according to CPMHT availability (see Table 17). Of the 31,276 women who gave birth when a CPMHT was available, 9888 (31.6%) had a specialist mental healthcare contact in the perinatal period compared to 10,033 women (25.7%) of the women who gave birth when a CPMHT was not available (adjOR 1.35, 95% CI 1.23 to 1.49; p-value < 0.001). Higher percentages of women having any mental healthcare contact in regions with a CPMHT available were observed both during pregnancy and in the first year after birth.
Obstetric and neonatal outcomes
Table 18 shows that the overall risk of stillbirth and neonatal death in the first 7 days after birth was 0.4%, the risk of preterm birth 10.7%, and the risk of a birth of a SGA baby 6.8% in the 70,323 women with a pre-existing mental disorder. A maternal adverse outcome according to the English Maternal Morbidity Outcome Indicator (MMAOI) occurred in 1.1% of women. In the 70,096 women who had a liveborn baby, an adverse neonatal outcome according to the ENAOI occurred in 9.1%.
| n (%) | No CPMHT | CPMHT | Unadjusted OR (95% CI) for CPMHT | p-value | Adjusted, with time trends only adjOR (95% CI) for CPMHT | p-value | Adjusted, with time trends and maternal characteristics.adjOR (95% CI) for CPMHT | p-value | |
|---|---|---|---|---|---|---|---|---|---|
| Perinatal death | 316 (0.4) | 151 (0.4) | 165 (0.5) | 1.37 (1.10 to 1.68) | 0.003 | 1.31 (1.06 to 1.62) | 0.011 | 1.34 (1.09 to 1.66) | 0.006 |
| Preterm birth | 7508 (10.7) | 4341 (11.1) | 3167 (10.1) | 0.90 (0.79 to 1.04) | 0.146 | 0.87 (0.76 to 1.00) | 0.049 | 0.86 (0.74 to 0.99) | 0.032 |
| SGA | 4769 (6.8) | 2542 (6.6) | 2227 (7.2) | 1.10 (1.01 to 1.20) | 0.022 | 1.11 (1.02 to 1.20) | 0.021 | 1.10 (1.02 to 1.20) | 0.016 |
| ENAOI for live births | 6191 (9.1) | 3599 (9.5) | 2592 (8.6) | 0.90 (0.76 to 1.05) | 0.187 | 0.87 (0.74 to 1.03) | 0.109 | 0.86 (0.72 to 1.02) | 0.086 |
| MMAOI | 758 (1.1) | 418 (1.1) | 340 (1.1) | 1.01 (0.85 to 1.21) | 0.898 | 0.98 (0.82 to 1.18) | 0.862 | 0.93 (0.78 to 1.11) | 0.433 |
The risk of stillbirth and neonatal death was higher when a CPMHT was available at the onset of pregnancy. Of the 31,276 women who gave birth when a CPMHT was available, 165 (0.5%) had a stillbirth or a baby who died in the first 7 days after birth, compared to 151 (0.4%) of the 39,047 women who gave birth when a CPMHT was not available (adjOR 1.34, 95% CI 1.09 to 1.66; p-value = 0.006).
The risk of preterm birth was lower when a CPMHT was available (10.1%, compared to 11.1%, fully adjusted OR 0.86, 95% CI 0.74, 0.99; p-value = 0.032), whereas the risk of a SGA birth was increased (7.2%, compared to 6.6%, adjOR 1.10, 95% CI 1.02 to 1.20; p-value = 0.016). There were no statistical differences in the risk of adverse neonatal outcomes according to the ENAOI or maternal adverse outcomes according to the MMAOI between women with and without a CPMHT available at the onset of pregnancy.
Subgroup analyses
The first subgroups analyses included only women who had a pre-pregnancy psychiatric hospital admission or CRT, and the second subgroup analyses included only women who had a specialist mental healthcare contact in the year immediately before the pre-pregnancy. The pattern of results in both subgroup analyses were similar to those observed in the main analyses (see Report Supplementary Material 1; p-values for interaction always > 0.05).
Discussion
Our study of 70,000 with a pre-existing mental health disorder found that 9% of these women who gave birth in an English NHS hospital had a specialist (secondary) mental healthcare contact in the 10 years before the pregnancy, suggesting a pre-existing mental disorder. We found evidence supporting our hypotheses that the implementation of CPMHTs improves access to specialist (secondary) mental health services in the perinatal period and that in turn it reduces the risk of an acute relapse in women with pre-existing mental health disorders. This supports policies that ensure that comprehensive and high-quality PMH services are available for all women with pre-existing mental disorders in the perinatal period.16,19
The observation that the risks of stillbirth, neonatal death and the birth of a SGA baby were increased in regions where a CPMHT was available is a serious concern that needs further follow-up. The interpretation of this pattern of opposing associations is not straightforward and may be due to chance. However, based on this evidence, CPMHTs should ensure that they have close links with other health and social care professionals to minimise the risk that more intensive psychiatric support negatively affects the midwifery and obstetric support for women with severe mental disorders get during pregnancy and childbirth.
Strengths and limitations
Using national, linked data we assigned women to two groups according to the availability of a CPMHT in their region, rather than whether individual women did actually receive CMHPT support. This regional approach has two key implications. First, it eliminates the problem of bias by clinical indication because there is no direct link between the individual women’s characteristics and whether or not they receive support.94 However, regions with a population with higher baseline risks of adverse mental health or obstetric and neonatal outcomes may have implemented CPMHTs earlier and quicker which may not have been fully captured in our risk adjustment models. Second, the regional approach may have reduced the observed effect size of CPMHT perinatal mental health care, because even if women gave birth in a region where CPMHT support was available, not all women will have received it.
The observed effect size of CPMHTs may have been further reduced because we used a single date to distinguish whether or not a CPMHT was available in a particular region, without allowing an implementation period or a lag time for its effects to become apparent. However, Figure 8 demonstrates that only 70 regions changed from not having to having a CMPHT during the study period which limits the extent to which this underestimates the effect size.
As explained in this chapter, we did not have access to records of specialist mental healthcare episodes for a 4-month period (between December 2015 and March 2016).
Interpretation of the results and implications for practice
Our results suggest that in women with pre-existing mental disorders, the regional availability of CPMHTs providing tailored PMH support increased the use of mental health care during pregnancy and in the year after birth and that it reduced the risk of acute relapse in the year after birth.
However, the interpretation of the study’s pattern of obstetric results is less straightforward given that we found higher risk of stillbirth, neonatal death and birth of a SGA baby alongside a lower risk of preterm birth when a CPMHT was available. A possible explanation for the increases in obstetric and neonatal risks is that the primary focus of the CPMHTs to reduce psychiatric risks may ‘overshadow’ recognising and addressing modifiable behavioural and obstetric risk factors, such as motivating women to stop smoking, encouraging a healthy lifestyle and identifying other physical risk factors.95,96 Another unintended negative aspect of CPMHT involvement could be that care by CPMHTs during pregnancy highlights mental health needs, and may lead to experiencing discrimination when physical health care is needed.97
The increases in the risk of stillbirth and birth should not be overinterpreted at this stage, as it is possible that they may represent false-positive results.73 However, they do suggest that CPMHTs should work very closely with maternity and obstetric services, GPs in primary care, and generic specialist secondary care mental health service and communicate directly with health visitors and social workers.17