Introduction and background
Aims and objectives
Method
Search
Databases | |
Medline, CINAHL, MIDIRS, DARE, EMBASE, OVID, PubMed, ProQuest, Informit Science Direct, Cochrane Library, NHSEED | |
Published between 1994 and 2018 | |
English language publications only | |
Article contained key search words or combined search terms: midwifery, midwife-led units, nurse-midwifery, birth centers, cost, cost-effectiveness, economic evaluation, economic outcomes, pregnancy risk classification, maternal outcomes, neonatal outcomes, clinical outcomes, maternity services | |
Primary research article or Systematic Review/Meta-analysis or Integrative Review | |
Economic analysis secondary to RCT accepted | |
Peer-Reviewed Journals | |
Population sample of childbearing women and/or their babies where risk classification profile defined and/or includes woman with high risk or complex pregnancy | |
Measurement of at least one economic outcome measure combined with clinical and/or other outcome measures, in midwifery care units or integrated midwifery continuity models that included antenatal, birthing and postnatal services, compared to other maternity service models | |
Economic perspective is funder/health service |
Inclusion criteria
Results
Evidence hierarchy level | Included studies |
---|---|
Level I: Systematic Review | Devane et al., 2010, Ryan et al., 2013, Sandall et al., 2016a |
Level II Randomised Controlled Trial with Economic Evaluation | Homer et al., 2001a,b, Kenny et al., 1994, Rowley et al., 1995, Tracy et al., 2013 |
Levels III and IV Quasi-experimental Cost Studies (cohort, cross-sectional, case control, non-randomised prospective, retrospective audit) | Gao et al., 2014, Jan et al., 2004 |
Econometric Studies – predictive cost, productivity, resource models using datasets | No studies relevant to complex needs |
Appraisal of studies
Aim of Study | Sample / Setting | Design | Major Cost Findings | Health Outcomes | Strengths / Limitations |
---|---|---|---|---|---|
1. Sandall et al. (2016) Compare effects of midwife-led continuity models with other models for childbearing women and their infants Primary outcomes antenatal, birth & immediate postpartum outcomes Secondary outcomes birth intervention, morbidity, some aspects of resource use & cost United Kingdom √ | Included: 15 RCTs 17 674 women (Canada, Ireland, Australia, UK) Excluded: 22 studies Only 6 of the 15 RCTs measured cost of model; only 4 of the 6 RCTs that measured costs included “mixed risk” pregnant women/high risk pregnancy: Kenny 1994 Rowley 1995 Homer 2001 Tracy 2013 | Systematic review Cochrane Pregnancy & Childbirth Group Trials Register + reference lists of retrieved articles. Selection criteria: published and unpublished trials, pregnant women randomly allocated to midwife-led continuity models of care or other models of care for pregnancy & birth Cost trend reported narratively as RCT cost method varied, e.g. cost analysis; CEA; or not stated | Trend to cost saving effect in midwife-led continuity Cost savings intrapartum care – all studies Antenatal: varied Postnatal: 1 study higher cost/ 1 study no difference Primary 0utcome in midwife – led models (RR) (CI) ↓ regional analgesia (0.85, 0.78 – 0.92) ↓instrumental birth (0.90, 0.83 – 0.97) ↓ pre-term <37 wk (0.76, 0.64 – 0.91) ↓ fetal loss <24 wk (0.84, 0.71 – 0.99) ↑spontaneous vaginal birth (1.05, 1.03 – 1.07) No difference CS or intact perineum Secondary 0utcome midwife – led models: ↓amniotomy; ↓ episiotomy; ↓ fetal loss <24 wks; No labour analgesia; longer labour (MD) 0.50 hrs, No difference for: fetal loss >24 wks; labour induction; A/N admission; A/N haemorrhage; augment labour; PPH; low birthweight; 5 min Apgar < 7; SCBU admission; initiate breastfeeding | Time horizon: RCT (cost included) 1994 – 2013 Women receiving midwife care less likely to have epidural, episiotomies, instrumental birth. Spontaneous vaginal birth rate increased. CS rate no difference. Women less likely to have pre-term birth, lower risk of losing babies < 24wks, More likely to be cared for in labour by a known midwife. No adverse effects compared with other models. Conclusion: most women should be offered midwife-led continuity of care BUT Evidence may not apply to women with serious pregnancy or health complications as these women were not specifically included in all studies / analysis for clinical effectiveness not stratified | Limited evidence CEA for women with complex pregnancy Combined results: low and mixed risk pregnant women 4 studies used different economic evaluation methods -: narrative report as cost assessment inconsistent Strong evidence cost improved in midwifery models for low risk with, reduced intervention + increased satisfaction. Mixed risk studies - ‘interpret with caution’ |
2. Ryan et al. (2013) Analysis of evidence on cost – effectiveness of midwife-led care compared with consultant –led care in UK settings. Estimate potential cost savings to accrue from expansion of midwife – led care in UK Used Section 3 CE of Devane et al. 2010 SR United Kingdom √ | Economic synthesis of 3 RCTs evaluated against guidelines for economic review Drummond and Jefferson (1996) 5796 women Hundley 1995 Young 1997 Begley 2009 Excluded: Flint 1989 (sub-group costing 49 of 1001 women only) | Systematic review 12 electronic databases for cost midwife led models: Cochrane Methodology Register NICE methods + multiple 1-way sensitivity analysis for economic synthesis of costs used 3 RCTs applied to 8 scenarios CE measure used Incremental Net Benefit (INB): expressed as Net Monetary Benefit (NMB) – £ value, and Net Health Benefit (NHB) – QALY, Quality adjusted life year gain | Mean cost saving £12.38 per woman midwife led (ML) care Expansion of ML care to 50% of all eligible women in UK projected aggregate £1.16 mil cost saving/yr Sensitivity analysis: cost change per woman varied from saving £253.38 (37.5 QALYs gained per year) to cost increase £108.12 dependent on assumptions with correspondent aggregate annual savings £23.75 million, or aggregate annual cost increase £10.13 million | Time horizon: RCT (cost included) 1995 – 2009 Three economic analyses used in synthesis of potential cost saving from increasing midwife-led services for eligible maternities. Issues identified around generalizability of findings. High rate of transfer from ML to medical-led care in studies demonstrates ‘risk’ assessment criteria unable to identify all women who will develop complications in pregnancy and labour | Rigorous health economic assessment measures: INB, NMB,QALYs Limited to UK system Excluded RCTs from Australia and other countries where no comparison with consultant-led model Mixed risk pregnancy profile; sub-group analysis show cost results consistent for groups as (RR) fetal loss and neonatal death overlap with 1.00 |
3. Devane, D. et al. (2010). Section 3: assessed CE of midwife-led care compared with consultant –led care. Estimated potential cost savings of expanding midwife-led care in UK (pp. 33–45) United Kingdom √ | Based on 3 of 4 RCTs See 2. above Hundley 1995 2844 women; Young 1997 1299 women; Begley 2009 1653 women | Systematic review see 2. above Sensitivity analysis x 3 based on 8 scenario SA 1: Systematically varying estimated cost savings SA 2: Systematically varying RR for overall fetal loss & neonatal death using low risk and ‘mixed risk’ cases SA 3: Systematically varying assumed uptake of ML service | As published in Ryan, Revill et al. 2013 | Time horizon: RCT (cost included) 1995 – 2009 Expanding midwife –led maternities show: Reduced rate of interventions in ML continuity of care, including: <AN hospitalization Reduced use of regional analgesia in birth, less episiotomy and instrumental delivery & greater numbers of women more likely to experience spontaneous vaginal birth BUT may not extrapolate to women with identified risk factors | Cochrane bias assessment tool used for trial internal validity Not generalisable, small number of studies CE varied with unit size, location and volume |
Aim of Study | Sample / Setting | Design / Method | Model used (link costs & health outcomes) |
---|---|---|---|
1. Gao, Y. et al. (2014). Compared CE two models, Midwifery Group Practice (MGP) against baseline cohort of Aboriginal mothers / infants. Clinical and cost analysis Australia √ | Regional hospital, Northern Territory MGP cohort: 7 communities MGP Women = 310 MGP Babies n = 315 (Sept 2009 – June 2011) Baseline cohort: 2 communities Baseline Women n = 412 Baseline babies n = 416 (Jan 2004- Dec 2006) All risk | Economic evaluation - retrospective records audit (Baseline Jan 2004-Dec2006) prospective data collection (MGP Sept 2009-June 2011) | Cost-consequences analysis: Australian dollars Measured/calculated direct costs per group Established comparative cost and changes post establishment MGP service from first antenatal appointment to 6 weeks postpartum for Aboriginal mothers and babies |
2. Tracy, S.K. et al. (2013). Assess efficacy, safety and cost of caseload midwifery versus standard hospital maternity care for women of mixed obstetric risk Dec 2008 -May 2011 Australia √ | Women of all pregnancy risk status (not stratified) Sample 1748 women 2 tertiary teaching hospital sites, 2 states, NSW / Queensland | 2 arm RCT Caseload care, Women with a named midwife n = 871 versus Women Standard Hospital Care n = 877 Intention to treat analyses | Cost- consequences analysis: Australian dollars Cost of care per woman based on DRG separation and direct and indirect costs for resource use collected from hospital financial system Primary & secondary clinical & cost outcomes Univariate logistic regression, OR 95% CIs and Pearson χ2 test; p values; non-parametric bootstrap percentile CIs infer significance of effects |
3. Jan S. et al. (2004). Holistic economic evaluation of an Aboriginal Community Controlled Midwifery Program in Western Sydney 1990-1996 Australia √ | Sample: 2 groups of Aboriginal women, Western Sydney birthing between Oct 1990 – Dec 1996, Nepean & Blacktown hospitals n = 834 Antenatal care at Daruk Aboriginal Community Controlled Program, or either hospital | Cost analyses estimated Direct Program costs and downstream savings. Retrospective case record audit | Cost analysis: Australian dollars Clinical and cost data linked from case record and NSW Midwives Data Collection 1991–1996 with hospital data linked with Australian National DRG cost weights; Medication: PBS (pharmaceutical benefits) Diagnostic tests: MBS (medicare benefits) Sensitivity analysis used to model uncertainty |
4. Homer C.S. et al. (2001). Assess clinical and cost difference – team community midwifery care -CMWC compared to control/ standard hospital care - SHC 1997-1998 Australia √ | Sample of women of mixed pregnancy risk n = 1089 CMW = 550 SHC = 539 One Australian public hospital State of NSW | RCT-Zelen Design Cost analysis: CMW vs SHC 2 teams each with 6 fulltime midwives provided care for 600 women/yr (25 births/mth/team) Calculated mean cost/woman for 9 components of maternity care | Cost analysis: Australian dollars Mean cost/woman/group - standard errors and 95% CI calculated using bootstrap technique Components of care and cost for resources used for each woman: antenatal clinic; antenatal admission; day assessment unit; labour and birth; hospital-based postnatal care; domiciliary postnatal care; and, admission of neonates to the special care nursery (SCN), on-call costs. Salaries and wages calculated at market prices Sensitivity analysis in 3 areas: Neonatal admission to SCN; Efficiency of AN clinics; Proportion of elective CS |
5. Rowley, M.J. et al. (1995). Examined cost/clinical differences for birth between 2 groups - Team Midwifery - 6 midwives vs routine hospital care Australia √ | Sample of women of mixed pregnancy risk n = 814 Discrete stratification of high risk = 275 women Team midwifery n= 405 Hospital care n = 409 One Australian public hospital State of NSW | RCT: 2 groups continuity team (midwives) vs routine care (hospital) Cost measured: Australian National Cost Weights for Diagnostic Related Groups (DRG) per birth / delivery Intention to treat | Cost-effectiveness: Australian dollars; direct costs Multiple outcomes measured. No single measure of effectiveness derived. Australian national cost weights for diagnosis-related groups (DRGs) applied to outcomes of women for whom complete results were avail. Performed retrospectively by clerk blinded to study - based on medical records, covered inpatient costs. Cost of intervention & comparative care estimated by analysing midwives' salaries. No discounting as time-period < one year. Costs and quantities not reported separately. No sensitivity analysis undertaken. No price dates given. |
6. Kenny, P. et al. (1994). Cost analyses: Team Midwifery Vs Standard hospital care. Included clinical outcomes Sept 1992 – July 1993 Australia √ | Sample n = 446 women Team Midwifery n =213 Standard Care n = 233 Westmead public hospital State of NSW | RCT 2 Arm Study Resource cost estimates: AN, birth, PN care Cost estimated where statistically significant difference in service use shown Included: direct costs, infrastructure, staff salaries - calculated for ‘low’ and ‘high’ risk women each group | Cost analysis (Drummond1987) Costs estimated based on resource use at AN, birth and PN (including domiciliary) stages of care separately Costs based on care delivered No sensitivity analysis undertaken. Costing assumptions: cost effective if resource costs of midwifery care shown to be less or equivalent to conventional care and health benefits of midwife care relative to conventional care are shown to be positive |
Study | Major Cost Findings | Health Outcomes | Strengths / Limitations |
---|---|---|---|
1. Gao, Y. et al. (2014) | Cost saving AUS $ 703 / mother-infant episode for MGP cohort was not statistically significant (p=0.566) MGP (midwifery model): ↓birth cost -$ 411, p=0.049 ↓SCN cost – $ 1 767, p=0.144 ↑ AN cost + $ 272, p<0.001 ↑PN cost + $ 277, p<0.001 ↑infant readmission costs + $ 476, p=0.05 ↑travel cost = $ 115, p=0.001 | Time horizon: Midwife cohort – all Aboriginal mothers who gave birth between Sept 2009 - June 2011 (and their infants) Baseline cohort – all Aboriginal mothers who gave birth between Jan 2004 – Dec 2006 (and their infants) Women who received midwife model had more antenatal care, more ultrasounds, were more likely to be admitted to hospital in antenatal period, had equivalent birth outcomes (i.e. mode of birth; pre-term birth; low birth weight) compared with baseline cohort. Babies in midwife model admitted to Special Care Nursery had significantly reduced length of stay | Mixed risk; small sample Cost assumptions used for economic analysis – expert opinion not primary data Missing data (3.7% – 24.5%); 51% all cases = missing data; Time trend confounding; Hostel costs & transport costs not included |
2. Tracy, S.K. et al. (2013). | Median cost saving of $ 566 AUS / woman with Caseload / named midwife | Time horizon: Dec 2008 – May 2011 Birth interventions reduced in midwifery model 30% > spontaneous onset of labor; ↓ analgesia; ↓elective caesarean; No significant difference for overall rate of caesarean between groups. Similar safe outcomes for mothers and babies between groups | Registered Trial: ACTRN12609000349246 All pregnancy risk status No stratification of risk profile Defined eligibility, inclusion/exclusion criteria Study sufficiently powered (80%) and Type 1 error 5% Sample bias challenged external validity Cross-overs – did not receive assigned model of care Non-masking of group allocation from clinicians |
3. Jan S. et al. (2004). | Net cost estimate AUS$1, 200 per client – calculated by subtracting cost savings to other centers Daruk Antenatal service saw 245 women for 339 pregnancies during study | Time horizon: Women birthing between Oct 1990 – Dec 1996 No significant difference in service birth weights or perinatal survival Daruk Antenatal care = Gestational age @ 1’st visit lower; mean number AN visits higher; attendance for AN tests better Women strongly positive toward midwife model for relationship, trust, accessibility, flexibility, information, empowerment and family-centered care | Mixed risk pregnancy Evaluation framework, both quant and qual methods Focused on antenatal care attendance and access; costs were broader than used in conventional economic analyses - included birth outcomes and antenatal attendance in a subsequent pregnancy Assumptions in sensitivity analyses / estimated downstream health costs |
4. Homer C.S. et al. (2001). | Mean cost/woman: CMWC A$2 579 vs SHC A$3 483 Excluding neonatal costs: CMWC A$1 504 (1449–1559; 95%CI) v SHC A$1 643 (1563–1729 95%CI) Mean cost saving 9 areas SHC – CMCW: Antenatal +28.84 Day Assessment Unit -5.42 Antenatal inpatient +38.74 On-call cost -21.81 Labour / birth +68.83 Hospital Postnatal care 43.85 Domiciliary care -11.06 Special Care Nursery +2801.28 Total/woman +904.09 | Time horizon: 1997 – 1998 (not specific) Caesarean rate: CMWM 13.3% vs SHC 17.8% (OR . 0.6, 95% CI 0.4±0.9, P = 0.02) No other significant differences were detected among women or babies for clinical outcomes or events during labour and birth between care models | Cost analysis alongside RCT; 10 000 bootstrap replications Mixed risk sample; Costs included resource use, clinician travel, neonate care; No equipment, capital or program development costs; No transfer rates; Caseload/midwife key to cost saving; Not possible to determine optimal caseload numbers; unclear if data analyzed by intention to treat |
5. Rowley, M.J. et al. (1995). | Mean cost ↓4.5% per birth: Team MW v Routine care A $3 324 vs A $3 475 | Time horizon: May 1991 – June 1992 Included first AN visit to 6 weeks after birth Team MW women: higher AN class attendance OR 1.73; 95% CI:1.23-2.42 ↓ birth interventions 36% vs 24% OR; 1.73 (1.28 – 2.34); p<0.001 ↓ pethidine use 0.32 (0.22 – 0.46) ↓ newborn resuscitation 0.59 (0.41 – 0.86) Maternal satisfaction with team care was greater on 3 elements: information giving; participation in decision-making, and relationships with caregivers. Less cost than routine care and fewer adverse maternal and neonatal outcomes | Cost study alongside RCT Included women of all pregnancy risk status Model was team midwifery care, not caseload continuity Costs based only on DRGs; i.e. top – down cost only / not detailed. Unable to compare with other economic evaluations |
6. Kenny, P. et al. (1994). | Team Midwifery vs Standard Care: Avg costs AN cost/woman High risk $ 427 vs $ 456 Low risk $ 135 vs $ 133 Average additional cost per birth / woman $ 4.21 vs $ 9.36 PN cost/woman: Hospital stay $ 356.64 vs $ 397.26 (earlier discharge) Domiciliary $45.45 vs $45.80 | Time horizon: Sept 1992 – July 1993 Significant differences: manipulative vaginal birth, episiotomy & perineal tears. Women in team midwife care reported higher levels of satisfaction over 3 periods of antenatal, birth and postnatal care with information, communication and midwife attitude and skill | RCT Level 1 evidence; All risk pregnancy included; Discrete costs: AN, birth and PN Robust, bottom-up costing; Team midwife model, not caseload; Low risk of bias, although blinding not stated; Loss to follow up - 19 in TM vs 22 in SHC |