Homebirth vs birth centre vs standard labour ward.

Four studies described cost analysis or economic analysis of homebirth, birth centres and labour wards for low-risk women. [14, 19, 21, 29] These varied considerably in terms of the sample size and methods. The three larger studies concluded that home birth was less costly and the fourth, which was much smaller, found no difference in cost. Schroeder et al (2012) compared birth at home, alongside midwifery units (AMU), free-standing midwifery units (FSU) and standard obstetric units (OU). They found that homebirth conferred the greatest benefit overall, compared with birth centres and labour wards. This involved a saving of £565 when women gave birth at home compared to the hospital obstetric unit and savings of £195 and £170 when compared to freestanding birth centres and alongside birth centres respectively.

In another large study of 11,592 women who gave birth at home and 11,788 in hospital In the United States, Anderson and Anderson (1999) compared mortality and caesarean section rates and costs of homebirth and hospital birth. The authors concluded that homebirth was less expensive at US$1711 compared to birth in hospital estimated at US$5382. Janssen et al (2015) in British Columbia linked data from six administrative databases from 2001 to 2004 of 9864 women who planned to give birth at home under the care of a “regulated” midwife, or in a hospital under the care of a midwife or a physician. All women met the criteria for home birth eligibility. This study demonstrated significant cost saving between birth settings, home birth being the least costly at CAN$2275, followed by planned hospital birth with a midwife CAN$4613 and hospital birth with a physician CAN$4816 [19]. The smallest study, a prospective non-randomised cost analysis of 418 women conducted in the Netherlands, found no significant difference in the cost between the alternative birth settings. [14]

Birth centres vs standard labour wards.

The remaining studies examined cost in birth centres and standard labour wards. [23, 28, 30, 33] The majority of these found a cost saving in providing birth centre care [23, 30–33], one study reported no difference in cost [28] and a study in Scotland found a net increase in cost per woman however, this study included the costs of introducing this new model of care and birth setting. [34] Again, the sample sizes and methods varied considerably. Differences included the type of alternative facility, namely alongside or free-standing birth centres, and the recruitment of the women was largely based on vaguely reported criteria.

Toohill (2012) in a small Australian cohort study of 102 women compared the costs and outcomes of women receiving Midwifery Group Practice (MGP) care, who gave birth in a birth centre, with women receiving ‘standard care’ by midwives or GPs, who gave birth in a standard labour ward. When the costs were compared overall, there was a significant reduction in the cost of giving birth in the birth centre under MGP care compared to standard care at AU$4722 and AU$5614 respectively (p<0.001). Further differences in costs between the settings were found in studies by Bernitz (2012) in Norway, and Howell (2014) in the USA. Bernitz studied 1100 women in a 4-year RCT, calculated costs per woman (using a top-down and bottom-up approach), and found women who had care in an alongside midwifery unit (AMU) were less costly to care for than those in the standard obstetric unit (SCU) (€1672 vs €1950 p<0.001) in the same hospital complex. An RCT in Ireland also found the cost of care of women in alongside Midwifery-Led Units (MLU) was €182 lower than in the hospital setting led by consultant doctors. [32]

The cost-analysis by Howell (2014) involved 43,859 participants, 872 of whom gave birth in one birth centre and 42,987 who gave birth in a ‘usual care’ setting who were matched on observable variables on birth certificate data by propensity score modelling. This study found a 16% reduction in costs for every pregnancy conducted within a birth centre; a saving of US$1163 per birth. [30]

In Canada, Reinharz (2000) in a control-matched cohort study measured the cost effectiveness of 2000 women in seven pilot projects and found differing results across all settings. These projects examined the processes and outcomes of women who chose to give birth in midwife-led birth centres compared to women giving birth under the care of midwives or physicians in a hospital setting. The findings varied in cost and effectiveness, with overall conclusions suggesting that midwives are less costly in three of the seven birth centres and in the remaining four, cost overlaps suggested no real cost difference. Sensitivity analyses revealed that the upper limit of the midwifery groups’ costs resulted in a marginally lower cost than the physician costs. [28]

The Scottish RCT by Hundley et al (1995) with 2844 participants examined intrapartum costs of introducing a midwifery-led birth centre compared to the standard consultant-led labour ward. Nine sensitivity analyses of differing factors were tested and the result indicated an additional cost of 10.5% for the introduction of the midwifery-led service (with a range of a 2.5% saving to 11% additional costs). The birth centre was newly established and the results of the trial showed that the setup costs of the birth centre and the higher level of staff employed to staff the new unit factored significantly in the overall cost thus making it a slightly more costly option. This was the only study that showed the costs associated with an alternative birth setting were higher than the standard hospital setting.

Using a decision analytic model, Stone and Walker (1995) measured the cost-effectiveness of birth centre birth compared to hospital birth for women with matched low risk profiles (11,814 birth centre births and 2256 hospital births). The analysis concluded that it was less costly to give birth in a birth centre (US$3385) compared to a hospital labour ward (US$4673). Effectiveness results favoured the birth centre for this low risk cohort, calculating that only when the transfer rate from the birth centre to the hospital exceeded 62%, would the cost-effectiveness reverse.

Cost-effectiveness analyses.

Three of the cost effectiveness analyses in this review were published since 2012, two of which examined the cost of providing alternative birth settings at one health service or campus [23, 31] alongside an RCT [31] and a cohort study [21, 23]. All three studies took the perspective of the health care provider. Two studies used administrative data to report costs either with a ‘Cost Per Patient’ (CPP) measure [31] based on the hospital’s activity-based costing system, or direct hospital costs derived from patient records that were then aligned with Australian-Refined Diagnostic-Related Groups (AR-DRGs). [23] Toohill (2012) also derived the costs of primary health care episodes (funded federally in Australia) through a diary that was given to the participants at the start of the trial as well as the woman’s own Pregnancy Handheld Record (PHHR).

Schroeder et al. (2012) performed a cost-effectiveness study from the perspective of the National Health Service (NHS) and calculated the incremental cost and incremental effects of births planned to be in the different birth settings and were expressed as mean cost per woman in each birth setting and mean net benefit. Unit cost estimation combined top-down/ bottom-up costing methods and resources costed were listed and augmented by an expert advisory group involving focus groups of midwives. This data was obtained through the questionnaires and datasheets associated with the larger Birthplace in England Study. Bootstrapping and sensitivity analyses were used to address uncertainty around ratios and variation in costs.

The three remaining cost-effectiveness studies were published between the years 1995 and 2000. Stone and Walker (1995) and Anderson and Anderson (1999) examined the “National Birth Centre Study”. The National Birth Centre Study (1989) studied 11, 814 low-risk women admitted for labour and delivery at 84 free-standing birth centres in the United States. They were followed though their labour, birth, transfers and for at least four weeks after birth[36] A decision analytic model was devised by Stone and Walker which allowed all logical and ‘chance’ events to be considered in an analysis. Direct costs were limited to “hotel costs” (inpatient accommodation costs) and antenatal care costs by a Certified Nurse Midwife (CNM), CNM fees for labour and postpartum care, medical costs for caesarean delivery and ambulance costs for transfers. Indirect costs (for example fixed equipment and clinician education costs) included the cost of the equipment needed to open the birth centre and the annual building lease. This data was obtained from field interviews with financial managers of the obstetric department and the birth centre provided patient charge data. Local ambulance companies also reported costs of transfers.

Anderson and Anderson (1999) took the perspective of the cost to Medicaid and reported charges by CNMs from 29 US states through a survey mailed out to every CNM known to the American College of Nurse-Midwives (ACNM). The Medicaid Program in the USA provides free or low-cost medical benefits to low-income people who have no medical insurance or have inadequate medical insurance. These charges were compared to obstetrician charges from published work [37] and the Health Insurance Association of America. No resources or accommodation charges were reported in this study and the charges were obtained at different time periods, however this was adjusted for by calculating projected cost increases due to inflation.

In Canada, an observational study by Reinhartz et al (2000) used data collected from perinatal medical records, client postnatal questionnaires and physician billing information from the Regie de l’assurance du Quebec (RAMQ), the Quebec health insurance board. The costs were considered from four perspectives in total—the Health Ministry, the regional board, the client and her family. Other costs were obtained from non-physician professional associations to value visits made by women, for example a chiropractor or dietician and unit costs for drugs were calculated on generic prices where available.

Cost analyses/studies.

Hundley et al (1995) and Kenny et al (2105) also performed an economic analysis alongside an RCT. Kenny et al (2015) estimated costs associated with medical and midwifery staff for antenatal, intrapartum and postnatal care as well as investigations and interventions and hospitalisation costs. The financial data were collected by an experienced hospital accountant and costs relating to resource use estimated and verified with financial and services managers. Midwifery managers provided advice on the various pathways women in the service may take throughout their pregnancy. Hundley (1995) performed the cost analysis from the perspective of the hospital by costing the settings separately and classifying the costs into four main groups: staff costs; consumables; capital costs; and overheads. The main method of data collection was by questionnaire to the midwife managing the birth and the women on discharge from each facility. Medical record review also elicited information regarding consumables and interventions such as monitoring, pharmaceuticals, operative delivery and vaginal examination.

In the Netherlands Hendrix et al (2009) used the Dutch manual for costing in healthcare, which is a methodological reference for performing cost studies. This six-step reference included regarding the scope of the research, the choice of cost categories (this study measured the health sector costs and non-health care costs), resources to be included, the volume of resources, the valuation of the resources, and the calculation of the unit costs. Hendrix et al took the societal perspective and collected data identifying health care costs such as contact with health care professionals, medications, medical interventions, analgesia and hospitalisation. Non health care costs were regarded as costs to the woman and family and were collected using cost diaries, questionnaires given to the women and birth registration forms. [14]

The second study in this review from the perspective of Medicaid is by Howell et al (2014). This study estimated costs based on a previous study that used propensity score reweighting and instrumental variable analysis. Major components of the cost of care were estimated for the antenatal, intrapartum and postnatal care and these included physician costs (from Medicaid fees), midwife costs (from an algorithm that calculated a proportion of the Medicaid and Medicare physician fee) and Maternal and infant hospital costs were calculated from the National Inpatient Sample and the National Perinatal Information Centre for costs incurred in the Neonatal Intensive Care Unit (NICU). Howell et al. (2014) went on to report that the most costly component of maternity care was the maternity facility cost regardless of the model of care.

The results varied among the evaluations largely due to the differing costing methods, the data collection, the measurement and valuation of the resources identified. The authors concluded in eight of the eleven evaluations that it was less costly to give birth in a setting alternate to a standard labour ward. [19, 21, 23, 29, 31–33] Two evaluations resulted in no significant difference in the cost of the alternative setting [14, 28] and one showed an increased cost; however this evaluation included the cost of the introduction of the birth centre facility. [34]

Home birth.

There is high quality evidence that midwifery-led care in pregnancy, labour and birth is safe, with comparative or improved outcomes for women at low risk of complications. [41–44] The comparative cost of birth settings is more difficult to find. How costs are measured and estimated adds to the difficulty of drawing firm conclusions regarding where the costs lie and how to make a model of care and birth setting affordable for the health system and acceptable for the women who wish to access alternative pregnancy care. These limitations included the variety of approaches to measuring and indeed valuing resource use, data sources from which this information was obtained, and the methodologies used.

Homebirth costs included midwifery time and some consumables whereas the charges in the hospital group included accommodation for the women as well as the fixed costs of facility use and variable costs including staff time and consumables. Broadly speaking, most studies included in this review reported lower costs associated with giving birth at home or in birth centres. Given the ethical implications of performing an RCT to measure the differences in cost and outcomes of place of birth, the cohort study by Schroeder et al (2012) provided the most methodologically robust results. Prospective data collection in real time allowed the data collectors to document detailed use of resources, staff time and consumables, a method that could provide high quality data in future detailed cost analyses.

Birth centres.

The difference in costs for women attending birth centres were associated with fewer interventions and procedures [21, 31] shorter length of stay or lower accommodation costs and facility overheads. [21, 32, 33]

Cost savings can be found by moving low-risk childbearing women away from hospital settings where there are established fixed costs to consider, as well as the availability of the convenient use of technology and interventions that are motivated by increased throughput and medical practices. Logically, home births require fewer resources and do not incur accommodation costs. [45] The intervention rates and use of technology are typically lower in birth centre settings, interestingly, more so in freestanding birth centres than alongside birth centres. [21]

Standard labour wards.

Undeniably, standard labour wards carry the highest fixed costs and staff costs. Standard labour wards are necessary but the level of care is not required for all women. Many low risk women give birth spontaneously, in Australia around 50 percent [7], and probably don’t require the level of resources available in the labour ward setting. This means there is a high capacity to shift care outside the high level care settings without compromising outcomes for mothers and babies. [7, 41] Economies of scale become less applicable for standard labour wards and more favourable for birth centres and homebirth as the numbers increase in out-of-hospital settings [46], further supporting the evidence for providing similar services for women at low risk of complications.