What happens when a private hospital comes to town?
The impact of the ‘public’ to ‘private’ hospital shift on regional birthing outcomes
Article Outline
- Summary
- Introduction and background
- Research aims
- Research questions
- Method
- Results
- Discussion
- Conflict of interest
- Acknowledgements
- References
- Copyright
Summary
Purpose
To examine the regional impact of a shift from public to private hospital care on birthing outcomes.
Procedures
A retrospective regional cohort study analysed the birth outcomes for 20,826 live singleton births of gestation ≥37 weeks, within one regional area in New South Wales between 1 January 1997 and 31 December 2003. Rates of intervention for induction of labour (IOL), epidural pain relief and operative mode of birth were established and analysed according to hospital type. A cascade model was then constructed for total births by hospital type.
Findings
Regional birthing outcomes were significantly affected by a shift from public to private hospital care. The introduction of a new private hospital birth facility in the region studied, led to 90% of all privately insured births within the region shifting to the private hospital. During the period 1997–2003, overall regional rates for IOL increased from 38 to 45%, epidural use in labour increased from 10.4 to 21.1% and the caesarean section rate increased from 14.1 to 24.75%.
Principal conclusions
The introduction of a new private hospital birthing facility into the regional health area studied and the shift from public to private hospital birth had a profound impact on the overall birthing experiences of women in the region. This suggests that private hospital services are not a direct substitute for public hospital birthing services. The cascade effect was present for women regardless of risk category and more pronounced in the private hospital. Women who are privately insured require better information to assist them in choosing their birthing environment, rather than assuming that they are simply buying a comparable product through private insurance.
Keywords: Childbirth, Intervention rates, Pregnancy outcomes, Health insurance
Introduction and background
From 1 July 1997, the Australian Federal Government introduced a series of major private health insurance incentives, including a 30% rebate on premiums in 1999 and Lifetime Health Cover (LHC) from July 2000, with the aim of reversing the long-term downward trend in private health insurance coverage.1, 2 These incentives were effective initially (Fig. 1), although the rate has been slowly decreasing since.3 Implicit in these incentives was that Australians could expect health outcomes within private hospitals to be at least equivalent to public hospitals in terms of safety and quality.1 Previous research indicates that outcomes may in fact be less than optimal, compared with evidence-based recommendations, for a range of interventions during birth.1
Figure 1.
Percent of population covered by private hospital insurance (New South Wales) (Private Health Insurance Administration Council (PHIAC)11).
Women using private health insurance for pregnancy may be unaware that, in doing so, they are more likely to experience interventions such as caesarean section (CS), epidural anaesthetic, instrumental birth (forceps and vacuum) and episiotomy.1, 4, 5, 6 This is consistent with the suggestion that public and private hospitals are not actually ‘perfect’ substitutes for one another.7 In explaining this further, Stephen Duckett suggests that clinical management of patients may reflect a different culture within private hospitals influenced by specialisation in elective surgery, for example, [7, p. 745]7. It could be suggested that women with private health insurance, birthing in private hospitals, do not have the same opportunities to experience spontaneous labour resulting in normal vaginal birth, when compared to women attending public hospitals for birth. Introduction of new private birthing facilities could in effect be limiting choices for women who are privately insured.
Throughout Australia, there has been a general increase in activity within the private hospital sector between 1994 and 2004 [7, p. 746]7. Health insurance changes have meant that in general terms patient care episodes have shifted from the public to private hospitals. The extent to which this has occurred in obstetrics is speculative and will be the focus of empirical evaluation within this paper. Empirical evidence is needed to illustrate the extent of change brought about by policies to increase private health insurance and flow-on effects created by greater use of private hospital services for birth.
Research aims
This study aims to examine the regional impact on birthing outcomes related to the introduction of private hospital birthing services within one identified regional area. The area selected was characterised by the opening of a new private hospital birthing unit within a private hospital facility from the beginning of 1998. Immediately prior to the opening of this birth facility, privately insured women overwhelmingly gave birth at the larger of the two major public hospital maternity units in the area. Previous research has indicated that interventions and outcomes for privately insured women birthing in public hospitals tend to fall in between levels experienced by publicly-insured women and privately insured women birthing in private hospitals.1, 6 This suggests that any transfer of births from the public to the private hospital sector will affect rates of intervention and birth outcomes in the region studied in addition to effects attributable to increases in private insurance coverage.
The primary objective of this study is to estimate the overall effects on the regional pattern of birth interventions and outcomes due to the opening of the private hospital birthing facility. The analysis has been conducted in two stages. In this descriptive paper, we document trends in birth interventions which have occurred during the period 1997–2003, focussing on regional effects brought about by differences between the public and private sectors.
A paper to follow will examine the extent to which differences observed are mitigated by differences in obstetric risk profiles between public and privately insured women in the region.
Research questions
What was the regional impact of the new private hospital birthing unit on:
Method
Data
Following approval from NSW Health for the release of data for epidemiological analysis de-identified unit record data were provided from the NSW Midwives Data Collection, which covers all NSW births as recorded by midwives in NSW hospitals, augmented by additional data available to NSW Health regarding insurance status. Ethics approval was received from University of Wollongong, Human Research Ethics Committee (HE06/014).
This study utilised unit record data from the NSW Midwives Data Collection, covering all births within the selected area health service over the period 1997–2003. The eligible study population (n
=20,826) comprised all live singleton infants at term (≥37 weeks gestation) in the nominated area in New South Wales between January 1 1997 and 31 December 2003 (92% of all births). For some of the analyses it was necessary to exclude the year 1997 due to changes in the definitions used by the MDC for induction of labour and some maternal conditions and obstetric complications. When 1997 was excluded the eligible study population was 17,939 births. It should also be noted that insurance status was not available for 2003.
Birth outcomes data includes labour (e.g. spontaneous, induction, augmentation) and pain relief (e.g. nitrous oxide, pethidine, epidural), mode of birth (e.g. normal vaginal, instrumental vaginal, caesarean section). The SPSS for windows release 11.5 software was used for all analyses of this dataset.
Results
Births by insurance status
The shift in births from the public hospital to the private hospital in this region between 1997 and 2003 was profound (Fig. 2). Privately insured births in public hospitals in the region declined from 29.0% of all births in 1997 to only 3.9% in 2002. By 2003 the private hospital accounted for 33.3% of all births in the region. Publicly insured births fell from over 73% in 1999 to 63.5% in 2002. This decline was particularly pronounced in 2001 (from 72.3% in 2000 to 67.2%), presumably due to the effects of LHC. Thus, the new private hospital unit had gone from zero to one-third of total market share in only 5 years (1998–2003), and captured almost 90% of the privately insured market. In 2002, the private hospital was responsible for 969 of the 1085 privately insured eligible births, or 89.3%. Anecdotal evidence suggests that this was at least partly due to an obstetric policy that privately insured women use the new private facility. Conversely, over the study period eligible births in the region's two public hospital units declined by over one-third, from 2965 in 1997 to 1967 in 2003.
Figure 2.
Eligible births by hospital type, 1997–2003 (percent). Public Ins: public insurance; Priv/pub: privately insured in a public hospital; Private Hosp: privately insured in a private hospital; Public Hosp: public hospital.
Trends in obstetric interventions and birth outcomes 1997–2003
Despite the oft-used argument that women in private hospitals are at greater risk because they are older, there is no clinical evidence that their objective risk profile is greater than women birthing in public hospitals. The MDC defined pre-existing medical and current obstetric conditions to include essential hypertension, diabetes, gestational diabetes, pregnancy induced hypertension and antepartum haemorrhage. Privately insured women are generally found to enjoy lower levels of measured maternal and obstetric conditions presumably due to the positive correlations observed between education, income, private insurance and health status [7, p. 63-8; 8, p. 24]7, 8. Table 1 compares women birthing in the public versus private hospital in terms of pre-existing medical conditions, obstetric complications, age, parity and previous CS history. Privately insured women were older, on average, and more likely to have experienced previous CS.
Table 1. Factors associated with obstetric risk by type of hospital, eligible births 1998–2003 (percent) (n=17,939)
| Risk factor | Type of hospital | |
|---|---|---|
| Public (n=13,330) | Private (n=4609) | |
| Maternal condition | 1.2 | 0.5 |
| Obstetric complication | 11.0 | 11.8 |
| Age 35+ | 11.8 | 20.4 |
| Primiparous | 39.5 | 39.9 |
| Previous CS | 8.7 | 13.2 |
| Smoking during pregnancy | 26.8 | 4.8 |
However, when objective conditions and complications, as defined above, are considered, there is little difference between private and public women. Women birthing in the private hospital were much less likely to have a pre-existing condition but slightly more likely to be reported as having an obstetric complication, with little difference in overall parity profiles. Smoking during pregnancy, which constitutes a risk factor in its own right and may also serve as an indicator of low education, income and socioeconomic status, was well over five times more prevalent among public hospital women (26.8% versus 4.8%). In total, differences in risk profiles would appear to be unlikely to explain large variations of the types documented below in birth interventions or outcomes. As noted above, a paper to follow will examine this issue in greater detail.
Elective caesarean section (CS)
Elective CS rates were consistently two to three times higher in the private hospital (Fig. 3). Although elective CS increased only modestly in both sectors over the period 1997–2003, the overall regional increase was from 7.5 to 12%. This appears to be due to the shift of births into the private sector. Hence, the observed increases in elective CS rates appear to have been driven by a combination of shifts away from public care (with greater midwifery inputs) towards obstetrician-led modes of care, plus a general trend over time towards elective CS observable across the spectrum of childbirth care services. Differences between private and public women in previous CS profiles may also explain part of the differences observed between the two hospitals.
Figure 3.
Elective caesarean section by hospital type, 1997–2003 (percent).
Management of labour (induction of labour)
For women who experienced labour, many experienced an artificially induced labour rather than spontaneous labour. Induction of labour (IOL) is recommended for situations where the benefits in terms of maternal and neonatal outcome are thought to outweigh the risks associated with the method of induction. For the purpose of analysis, induction of labour was defined as labour that is artificially induced by pharmacological means (oxytocin and/or prostaglandins).9
The overall IOL rate for eligible labouring women in the region increased from an already substantial 38.0% in 1998 (due to definitional changes, comparable IOL figures for 1997 are not available) to almost 45% by 2003, despite only modest increases overall in public hospitals (Fig. 4). This appears to be due to both a shift to the private sector for birthing and to increasing IOL rates in the private hospital (up from 46.3% in 1998 to almost 58% by 2003).
Figure 4.
Induction of labour by hospital type, 1998–2003 (percent of women having labour).
Management of labour (epidural)
Total epidural anaesthetic rates more than doubled from 10.4% in 1997 to 21.1% in 2003 (Fig. 5). Private hospital rates increased from 15.5% in 1999 to 26.1% in 2000 and to 32.8% by 2003. Public hospital epidural rates increased more modestly, from 10.4% (1997) to 15.8% (2003). This suggests that the overall increases were driven primarily by the shift from public to private modes of care.
Figure 5.
Epidural anaesthetic by hospital type, 1997–2003 (percent of labouring women).
Birth outcomes for labour
Mode of birth as an outcome for labour differed according to hospital type. Overall emergency CS rates almost doubled from 1997 to 2003, from 6.6 to 12.7% (see Fig. 6). There was, however, little difference between the two hospital types on this dimension. The growth in emergency CS was apparently mainly due to trends over time in both sectors. This is not to imply that CS in total was equally common in both hospital settings. Taking elective CS into account, total CS rates were both substantially higher and grew more rapidly in the private sector, reaching a high of 31.9% of all eligible births in 2003, versus 21.3% in the public sector.
Figure 6.
Emergency CS by hospital type, 1997–2003 (percent of labouring women).
Similarly, instrumental birth rates were consistently higher in the private sector, so that the incidence of operative birth (emergency CS or instrumental birth) was much higher at the private hospital (Fig. 7). Operative birth rates grew from 22.5% in 1997 to 32.8% in 2003 at the public hospitals, but from 45.0% in 1998 to 53.7% in 2003 at the private hospital. In other words, women labouring at the private hospital had less than a one in two chance of experiencing a normal vaginal birth at the private hospital, but better than two chances in three at the public units.
Figure 7.
Operative birth by hospital type, 1997–2003 (percent).
Cascade effects
Cascade of intervention has been raised in previous research6, 10 and can best be described as a ‘flow-on’ effect from one intervention to another. In this study the flow-on effect from induction of labour, to epidural pain relief and operative mode of birth was modelled. A cascade effect was demonstrated for this region according to the interventions of both IOL and epidural, for women birthing in both public and private hospitals (Figure 8, Figure 9). The ‘cascade’ models further demonstrate that privately insured women have higher rates of all interventions (elective CS, IOL and epidural) and higher rates of operative birth.
Figure 8.
Cascade effects 1998–2003 public hospital. NVB: normal vaginal birth; Instr.: instrumental birth (forceps or vacuum); EmCS: emergency CS. *Missing data for mode of birth for one woman in this category.
Figure 9.
Cascade effects 1998–2003 private hospital. NVB: normal vaginal birth; Instr.: instrumental birth (forceps or vacuum); EmCS: emergency CS. *Missing data for mode of birth for one woman in this category.
Table 2, which summarises cascade effects modelled in Figure 8, Figure 9, demonstrates that women birthing at the private hospital were much less likely to experience normal vaginal birth. Only 34.5% of women who had both IOL and epidural in the private hospital went on to experience a normal vaginal birth. Even in the public hospital, once epidural was combined with IOL, the normal vaginal birth rate fell to 41.2%. This is compared with 91.0% of the women experiencing normal vaginal birth when they had neither IOL nor epidural during labour.
Table 2. Normal vaginal birth rates by level of intervention in labour 1998–2003 (%) (excludes elective CS)
| All eligible births ≥37 weeks | ||
|---|---|---|
| Public hospital | Private hospital | |
| No IOL or epidural | 91.0 | 81.6 |
| IOL only | 82.0 | 76.9 |
| Epidural only | 46.3 | 34.1 |
| IOL+epidural | 41.2 | 34.5 |
Figure 8, Figure 9 also demonstrate that IOL seems to be associated with epidural, which in turn is associated with operative birth. Private hospital women were almost three times as likely to have an epidural if they had IOL (OR=2.89), whilst the odds ratio was 5.20 for public hospital women. The main importance of this is that, again, women who had epidural were 3.23 (private hospital) to 5.16 (public hospitals) times more likely to experience operative birth. IOL also elevates the probability of operative birth, but not to the same extent. Note that, overall, the incidence of operative birth among labouring women varied from 65.5% for private hospital women who had IOL and epidural down to only 9.0% for public hospital women who had neither intervention.
Discussion
Public to private health insurance shift, further enhanced by public to private hospital shift, appears to have had a pronounced impact on regional birthing outcomes. When a new private hospital birthing service was introduced into the region studied, it was remarkably successful in capturing market share (almost 90% of regional privately insured, and one-third of all, singleton term births by 2002). This very success makes it more difficult to assess the impact of the private hospital on regional patterns of birth interventions and birthing outcomes. Use of the private hospital has become virtually synonymous with having private insurance, making it extremely difficult to disentangle the effects of changes in health insurance status from those due to the opening of the private hospital maternity unit.
The success of the private hospital in attracting consumers of privately insured birthing services away from the public hospital sector also raises a number of policy issues. Firstly, there has been, since changes in private health insurance policy led to an increase in insurance coverage and increased use of private hospital services, a tendency to regard such increases as being beneficial to the hospital sector as a whole, by “taking pressure off the public sector”. This has certainly occurred in terms of birthing in the region studied. As the private hospital maternity unit grew, births in public hospitals fell by approximately one-third. Yet, given that the evidence seems to suggest that the main effect of this switch may have been to lead to large increases in more expensive birthing interventions which would otherwise not have occurred, one may question whether providers of maternity services in public hospitals should in fact be dismayed by their substantial loss of market share. Rather than applauding a policy change which has the effect of sending women they would otherwise be caring for to a private hospital environment where they will have to face much higher probabilities of experiencing CS, IOL, epidural, instrumental birth and so on, perhaps public hospitals should be seeking to more aggressively market their own services and birthing philosophies, and win back this market from the private hospital.
A related question might be to what extent the resources released to other public hospital services by the dramatic fall in birthing services provided have been wisely re-deployed, or indeed re-deployed at all. To what extent has the public hospital reduced its demand for maternity service inputs (beds, midwives, other types of labour, etc.)? Queues for elective surgery are, in this region as elsewhere, alleged to be a perennial problem for public hospitals. Is there any tangible evidence that the loss of one-third of the public hospitals responsibilities for providing regional birth services has had a positive impact on reducing these “queues”, for example? There seems to be little information available on such questions at the local level.
Thirdly, it appears that the large increase in private health insurance coverage that occurred in 2000 once LHC and the 30% insurance rebate were both in place may have been temporary, as coverage has quickly resumed its long-term downward trend (Fig. 1). Having achieved 90% market share, how will the private hospital react in future if private insurance coverage continues this decline? Will it be content to absorb the diminution in numbers of patients and revenues implied by such declines, or will it react in order to preserve revenues and profits? Does this imply even higher levels of birth interventions in future, or perhaps lobbying of governments to provide even higher subsidies for private health insurance?
In this study, two implications of the “Cascade effect” analysis are striking. Most importantly, the link between epidural pain relief and birthmode complications seems incontrovertible. Epidural pain relief, apart from constituting an additional cost to birth10 is associated with clinical risks in its own right, dramatically increasing the likelihood of both emergency CS and instrumental birth, which are procedures with both clinical and resource usage implications. Even at the region's public hospitals, women who had epidural went on to experience a 61.4% rate of operative birth, compared to just 11.9% of those who did not have epidural. For the private hospital, these rates were 65.6% versus 20.3%. Given that there are alternative pain relief measures, one may question whether it is appropriate that so many women are being given this intervention. Yet by 2003, almost one-third of women labouring at the private hospital had this intervention, which had grown from around 15% in 1998. Even at the public hospitals, there was a trend toward epidural, the incidence of which increased from 10.4 to 15.8% over the period of the study.
One reason may be the growth of IOL as the mode of initiating labour, due to the clear link between IOL and epidural. Yet 58% of private hospital labours were induced by 2003, and around 45% of public hospital labours as well. IOL appears to, by itself, increase the incidence of operative birth somewhat, but the more important effect on birthmode is via its effect on epidural rates. From 1998 to 2003, 31.7% of public hospital women who had IOL also had epidural, compared with just 6.1% of women who did not experience IOL. The equivalent figures for the private hospital were 38.5 and 13.3%. Perhaps it is time for a major overhaul of clinical policy relating to the use of these two birth interventions?
The results of this study also emphasise the need to re-examine the subsidised use of specialist services for ‘healthy pregnancy’. At the very least we need to inform women about probable birth outcomes when choosing models of pregnancy care because the choice has potentially far-reaching implications for their health beyond the immediate postnatal period. There is also both clinical and financial justification for private sector services (including private insurers) to consider incorporating midwifery models of care as part of providing choices for women who are privately insured. In addition, both the private sector and those in the public sector responsible for allocating taxpayer's funds to healthcare should re-consider the assumption that specialist obstetric care is a superior product, when evidence suggests that it may be merely a more expensive product.10
Implications for future research
This study has documented both that a new private hospital birthing service has been extremely successful in attracting custom and market share in the current environment of large-scale subsidies and other incentives for holding private health insurance, and that this seems to have led to significant increases in the medicalisation of birth in the region studied. However, it is possible to argue that the pattern of interventions and birthmode outcomes documented can be justified, either on the basis of changes over time and/or differences in risk profile between private and public women, or by improvements in maternal and/or neonatal outcomes. A follow-up paper will address these issues. The second phase of this study will also use the more restrictive subset of ‘low-risk’ births to explore questions relating to additional outcomes measures, including neonatal outcomes (e.g. Apgar scores, admission to NNICU), to assess whether it is plausible that improvements in these types of outcomes could justify the levels of increase observed for interventions.
This study contributes to the body of evidence on the striking relationship between the use of epidural pain relief and operative birth outcomes.10 Given that there are many alternative forms of pain relief available to labouring mothers, and that the MDC, for example, provides detailed information on the usage of these various products, an urgent need appears to exist for further research into relationships between pain relief and birth outcomes in the context of the Australian hospital birthing environment.
Conflict of interest
The authors verify that, within their knowledge, the publication of this paper will not constitute any conflict of interest.
Acknowledgements
We wish to acknowledge the NSW Health Department in preparation and provision of the data for this project and clearly state that the author(s) opinions, as expressed in this paper, do not necessarily reflect those of NSW Health. Funding support was received from the University of Wollongong for the preparation of this manuscript.
References
- . Impact of private health insurance on obstetric outcomes in NSW hospitals. Austr Health Rev. 2004;27(1):27–38
- . Policy change and private health insurance: did the cheapest policy do the trick?. Austr Health Rev. 2002;25(6):33–41
- . The effects of increase private health insurance: a review of the evidence. Austr Health Rev. 2004;28(1):119–138
- . Episiotomy in NSW hospitals 1993–1996: towards understanding variations between public and private hospitals. Austr Health Rev. 1999;22(1):18–32
- . Perineal outcomes in NSW public and private hospitals: analysing recent trends. Austr J Midwifery. 2002;15(2):5–10
- . Rates for obstetric intervention among private and public patients in Australia. Population based descriptive study. BMJ. 2000;321:137–141
- . The Australian health care system. 2nd ed.. Oxford: Oxford University Press; 2004;
- . Health economics. 3rd ed.. Boston: Addison Wesley; 2003;
- . Labor induction rate variation in upstate New York: what is the difference?. Birth. 2003;30:168–174
- . Costing the cascade: estimating the cost of increase obstetric intervention in childbirth using population data. BJOG. 2003;110:717–724
- Private Health Insurance Administration Council (PHIAC). Coverage of Hospital Insurance Tables Offered by Registered Health Benefits Organisations by State; 2003. http://www.phiac.gov.au/statistics/membershipcoverage/hosquar.htm. Accessed 26th September 2006.
PII: S1871-5192(07)00017-0
doi:10.1016/j.wombi.2007.02.001
© 2007 Australian College of Midwives. Published by Elsevier Inc. All rights reserved.
