Parts of this chapter are reproduced with permission from Tassie et al.98 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.
Aim
The economic component of this study aimed to explore changes in patterns of service use and associated costs following implementation of CPMHTs in England, and to consider costs and outcomes in a cost–consequences framework.
We hypothesised that areas with CPMHTs would be associated with lower rates of acute care in the perinatal period, shorter duration of admissions, lower rates of Mental Health Act detentions, and lower costs than areas without CPMHTs.
Methods
Study design and population
As described in this chapter, the study used a longitudinal cohort design. The economic perspective was limited to secondary mental health services, and the time horizon was the perinatal period (from estimated start of pregnancy to 1-year post birth). The economic analyses used the same population as described in this chapter all women residing and giving birth in England between 1 April 2016 and 31 March 2018 with pre-existing mental illness, defined as having used secondary mental healthcare services in the 10 years before onset of pregnancy.
Intervention
As described in this chapter, the intervention was access versus no access to a CPMHT, determined by the woman’s place of residence at pregnancy start date.
Resource use measurement and valuation
The Mental Health Services Data Set (MHSDS) provided patient-level data on contacts with secondary mental health service use, including acute care contacts (psychiatric inpatient stays and CRT contacts) and other community mental healthcare contacts (‘community care’). Unit costs were in Great British pounds at 2018–9 prices. Inpatient costing was conducted within a mental health cluster framework which firstly involved combining the number of admitted days with unit costs for mental healthcare cluster assignment inpatient days99,100 provided in the NHS Reference Costs.101 Unit costs for CRTs and other community care contacts were taken from Curtis and Burns (2017)102 and inflated to 2018–9 prices.
Outcomes
Outcomes were those included in the clinical analysis Chapter 6 and in this chapter: perinatal mortality, preterm birth, SGA, ENAOI and EMMOI, plus we explored formal detentions under the Mental Health Act.
Analyses
Number and percentage of women using each service type (inpatient, CRT, community care) were reported as well as the mean and SD for number of contacts and costs of contacts. Generalised least squares (GLS) regression models were used to estimate the mean difference in mental health resource use costs associated with individuals in areas with and without CPMHTs. To account for the skewed nature of healthcare costs, non-parametric bootstrapping with 5000 estimates was used to generate the standard errors and 95% CIs around the cost differences.103 Models were estimated with random effects at the CCG level, and the final model included a binary exposure variable, a time variable and adjustments for sociodemographic characteristics (maternal age, ethnicity, socioeconomic deprivation), highest level of pre-pregnancy care contact (inpatient, CRT or community care), timing of most recent pre-pregnancy care contact (< 1 year, 1–5 years and > 5 years) and maternal risk factors (parity, pre-existing hypertension, pre-existing diabetes mellitus, pre-eclampsia and gestational diabetes).
Results
Complete economic data were available for 70,082 women (99.7% of the total cohort). Table 19 reports service use and cost summary statistics. In the perinatal period, those in CPMHT areas were less likely to use acute care services than those without CPMHTs (4.94% vs. 5.58%), but more likely to use community care (30.96% vs. 24.72%). When acute care was broken down into its two components, use of inpatient care was higher for the CPMHT group (1.43% vs. 1.13%), while use of CRTs was lower (4.41% vs. 5.23%). The CPMHT group had higher numbers of acute care contacts than the no CPMHT group (mean 1.11 vs. 0.87), longer inpatient stays (mean 0.85 vs. 0.55 days), fewer CRT contacts (mean 0.27 vs. 0.32) and higher community care contacts (mean 4.31 vs. 3.26). This translated into higher total costs over the perinatal period in the CPMHT group compared with no CPMHT (mean £651 vs. £414).
| Service | No CPMHT(n = 38,901) | CPMHT(n = 31,181) |
|---|---|---|
| Number using each service n (%) | ||
| Total acute care | 2171 (5.58) | 1540 (4.94) |
| Psychiatric admission | 441 (1.13) | 445 (1.43) |
| CRT | 2035 (5.23) | 1375 (4.41) |
| Community care | 9615 (24.72) | 9653 (30.96) |
| Number of days/contacts mean (SD); range | ||
| Total acute care | 0.87 (9.39); 0–461 | 1.11 (11.53); 0–446 |
| Psychiatric admitted days | 0.55 (8.49); 0–455 | 0.85 (10.62); 0–446 |
| CRT contacts | 0.32 (2.16); 0–100 | 0.27 (2.04); 0–107 |
| Community care contacts | 3.26 (10.75); 0–202 | 4.31 (12.14); 0–241 |
| Costs mean (SD); range | ||
| Total acute care | £282 (3615); 0–210,024 | £387 (4421); 0–177,702 |
| Psychiatric admission | £218 (3459); 0–208,813 | £332 (4245); 0–176,254 |
| CRT | £64 (437); 0–20,187 | £54 (412); 0–20,187 |
| Community care | £132 (434); 0–8153 | £174 (490); 0–9727 |
| Total | £414 (3807); 0–213,455 | £651 (4634); 0–182,506 |
The GLS model (Table 20) found costs to be significantly higher in the CPMHT group compared with no CPMHT for psychiatric inpatient stays (adjusted mean difference £90, 95% CI £18 to £163; p = 0.015), total acute care (adjusted mean difference £80, 95% CI £4 to £157; p = 0.040), community care (adjusted mean difference £22, 95% CI £10 to £34; p < 0.001) and total costs (adjusted mean difference £111, 95% CI £29 to £192; p = 0.008). Multiple imputation of missing data produced similar results (see https://link.springer.com/article/10.1186/s12913-024-10553-8).
| Service | Cost difference £ (SE) | 95% CI | p-value |
|---|---|---|---|
| Total acute care | 80 (39) | 4 to 157 | 0.040 |
| Psychiatric admissions | 90 (37) | 18 to 163 | 0.015 |
| CRT | −11 (6) | −22 to 0 | 0.058 |
| Community care | 22 (6) | 10 to 34 | 0.000 |
| Total | 111 (42) | 29 to 192 | 0.008 |
Maternal and neonatal outcomes are reported in full in this chapter. For formal detentions under the Mental Health Act, the odds of a detention for CPMHT versus no CPMHT was not statistically different (adjusted OR 0.926, 95% CI 0.760 to 1.130; p = 0.444). A summary of costs and consequences is provided in Table 21 which shows that that there is no clear pattern; total costs are significantly higher in the CPMHT group and outcomes vary, with some being poorer in CPMHT areas (perinatal mortality and SGA), some better (preterm birth) and the remainder not significantly different (EMMOI, ENAOI and detentions).
| Costs/outcomes | Summary of results |
|---|---|
| Total costs | CPMHT > no CPMHT |
| Perinatal mortality | CPMHT > no CPMHT |
| SGA | CPMHT > no CPMHT |
| Preterm birth | CPMHT < no CPMHT |
| ENAOI | CPMHT = no CPMHT |
| EMMOI | CPMHT = no CPMHT |
| Formal detentions | CPMHT = no CPMHT |
Discussion
The results of the economic analysis suggest that the implementation of CPMHTs has successfully supported increased access to secondary mental healthcare services, particularly community mental health care. However, no corresponding reduction in use or cost of psychiatric inpatient admissions was identified. These findings may reflect better identification and admission of more severely unwell women requiring intensive support, with better support being provided in the community for women less severely unwell. It may also reflect greater access in CPMHT regions to MBUs where admissions are typically longer than for traditional psychiatric wards.104 Length of admissions is influenced by many factors, but a key factor in PMH populations is whether women are together with or separated from their baby. In traditional psychiatric wards, there is likely to be a greater sense of urgency to discharge women so they can be returned to their baby.
Increased access to CPMHTs resulted in significantly higher costs and mixed results in terms of maternal and neonatal outcomes. However, CPMHTs are primarily focused on improving mental health outcomes, which this study was unable to examine, with the exception of formal detentions. Nor were we able to consider quality of life or measures preferred for economic evaluation, such as quality-adjusted life-years. We were also unable to measure any potential reduction in other costs that may have been impacted by access to CPMHTs, such as social care or non-mental health service costs or the longer-term impacts on costs and outcomes. These are important omissions and while regions with CPMHTs may incur higher costs in the perinatal period, the benefits in terms of potential increases in quality of life and potential reductions in broader societal costs may offset this increased cost.