What do maternity services produce? An exploration of potential output measures to assess the eciency and productivity of maternity services in Australia.

Background: In maternity services, as in other areas of healthcare, increasing emphasis is placed on improving “eciency” or “productivity”. The rst step in any eciency and productivity analysis is the selection of relevant input and output measures. Within healthcare quantifying what is produced (outputs) can be dicult. The aim of this paper is to identify potential output measures that reect the principles of woman-centred care and that can be used in an assessment of the eciency and productivity of maternity services in Australia. Methods: This paper will survey available perinatal and maternal datasets in Australia to identify potential output measures; map identied output variables against the principles of woman-centred care outlined in Australia’s national maternity strategy; and based on this data, create a preliminary composite outcome measure for use in assessing the eciency and productivity of Australian maternity services. Results: The identied composite measure consists of labour and birth outcomes indicators where data is available from the National Perinatal Data Collection. The composite measure makes it very clear that there are signicant gaps in Australia’s maternity data collections with regard to measuring how well a maternity service is performing against the values of respect, choice and access. Conclusions: Adoption in Australia of the collection of woman-reported maternity outcomes would substantially strengthen Australia’s national maternity data collections and provide a more holistic view of pregnancy and childbirth in Australia beyond traditional measure of maternal and neonate morbidity and mortality. These terms


Background
In maternity services, as in other areas of healthcare, increasing emphasis is placed on improving "e ciency" or "productivity" (1,2). These terms are sometimes seen as synonymous with cost-cutting, and those on the frontline of delivering care may feel that the terms are used as a means of facilitating the reduction of resources with little concern for how that effects the quality of care (3,4). Such scenarios are actually not congruent with what e ciency and productivity relate to. E ciency and productivity measurement allows comparison of the relative performance of a given set of entities (for example hospitals) that produce the same or similar goods and/or services (for example, maternity care).
Formally, productivity is de ned as the ratio of outputs to inputs and can be represented by a production frontier, with input(s) on the (x) axis and output(s) on the (y) axis. The production frontier represents the maximum output that could be produced from each input level given current technology. Firms operate either on that frontier, if they are technically e cient or beneath the frontier if they are not technically e cient. If a rm is beneath the frontier, this indicates that they could be producing more outputs then they currently are. Technical e ciency is de ned as the production of the maximum amount of output from a given amount of input (or alternatively the production of a given output with minimum input quantities) given current technology. Allocative e ciency is similar to technical e ciency but places a cost on inputs or outputs. Allocative e ciency is de ned as the input mix that minimizes cost, given input prices or when the output mix maximizes revenue, given output prices. Together, technical and allocative e ciency comprise overall economic e ciency (5). E ciency and productivity measurement places equal emphasis on inputs (or costs) and outputs (what is actually produced). If costs are reduced and what is produced also simultaneously declines, productivity is not increased, and e ciency is unlikely to be reached.
In any e ciency and productivity analysis the selection of relevant input and output measures is an essential rst step. However, within healthcare capturing what is produced can be di cult. Output measures should re ect the function and key activities of a given industry and allow comparison of both the quantity and quality of output (5). For the health industry, the most relevant output variable would be one that measures the health gains of individual patients who seek treatment (6,7). However, there is often limited data available on individual patient outcomes. Many e ciency and productivity studies therefore utilise proxy measures of health outcomes, such as number of patients treated or length of stay.
The limitations of proxy measures of health outcomes, such as number of patients or length of stay, is further pronounced when considering the performance of maternity services and the importance of woman-centred care. Woman-centred care promotes the principles of choice, control, continuity of caregiver and self-determination (8,9). It is increasingly being incorporated in Australia and other jurisdictions as the foundation of the provision of safe and effective maternity care (10, 11) Woman-centred care recognizes that a 'successful' birthing experience is de ned by more than the delivery of a healthy baby and the physical safety of the mother. The selection of output measures to assess the e ciency and productivity of maternity services should therefore also move beyond simple measures of maternal and neonate morbidity and mortality and indicators of clinical activity, and towards those variables that capture the entirety of the birthing experience.
The aim of this paper is to identify potential output measures that re ect the principles of woman-centred care and that can be included in an assessment of the e ciency and productivity of maternity services in Australia. This paper will survey available perinatal and maternal datasets in Australia to identify potential output measures; map identi ed output variables against the principles of woman-centred care outlined in Australia's national maternity strategy Woman-centred care: Strategic Directions for Australian Maternity Services; and based on this data, create a preliminary composite outcome measure for use in assessing the e ciency and productivity of Australian maternity services.
The Australian national strategy, Woman-centred care: Strategic Directions for Australian Maternity Services, outlines a means to support the delivery of maternity services for women from conception until 12 months after pregnancy or birth. The Strategy outlines four values -safety, respect, choice and access -which underpin twelve principles for woman-centred maternity care that apply to all health professionals providing maternity services (12). An additional le outlines these twelve principles (see Additional le 1).
This national strategy provides a useful framework for considering potential output variables for assessment of the e ciency and productivity of maternity services in Australia. Ideally, any e ciency and productivity analysis would incorporate output variables that correlate with and indicate how well a maternity service is delivering care in accordance with these twelve principles (12).

Data scoping
A search was conducted for datasets available within Australia that related to maternal health care. A Google search engine (Chrome) was used to search for and identify relevant datasets. The following keywords and phrases were included in the search: Maternal; Maternity; Perinatal; Pregnancy; Childbirth; Data; Collection; Indicators Data mapping Identi ed datasets relating to maternal health care were reviewed. The contents of each dataset was mapped to the four values of safety, respect, choice and access outlined in Australia's national maternity strategy. From this mapping exercise a preliminary composite outcome measure was constructed for use in assessing the e ciency and productivity of Australian maternity services.

Results
Existing maternity and perinatal data sets in Australia National Perinatal Data Collection and National Core Maternity Indicators The National Core Maternity Indicators (NCMIs) provide information on measures of clinical activity and outcomes in relation to maternity care across Australia. The purpose of the indicators is to establish baseline data to monitor and evaluate maternity care in Australia and enable continuous improvement in care. The NCMIs are clinical indicators of maternity care, where a clinical indicator is de ned as a measure of the clinical management and outcome of care and is based on evidence that con rms the underlying causal relationship between a particular process or intervention and health outcome (13). The NCMIs are constructed from data items from the Australian Institute of Health and Welfare (AIHW) National Perinatal Data Collection (NPDC), a national populationbased collection that provides information on the pregnancy and childbirth of mothers, and the characteristics and outcomes of their babies. The NPDC captures all births in Australia in hospitals, birth centres and the community (14).
Tables 1 and 2 map the NCMIs and NPDC data item against the four values of the national strategy -safety, respect, choice and access. a. without instruments to assist the birth b. assisted with instruments X X General anaesthetic for women giving birth by caesarean section X Women having their second birth vaginally whose first birth was by caesarean section X X

Birth Outcome Indicators
Apgar score of less than 7 at 5 minutes for births at or after term X Small babies among births at or after 40 weeks gestation X Third and fourth degree tears: a. for all vaginal first births b. for all vaginal births X *Rather than the whole population, these indicators are measured only for 'selected women'. This is women whose characteristics indicate they have a lower risk of birth complications and therefore provide a better indication of what are expected outcomes in 'standard' cases.
Selected women are aged between 20-34 years; gave birth between 37-41 completed weeks of gestation; had a singleton baby who presented in the vertex (head down) position (17) It can be seen that the NCMI and NPDC data items relate predominantly to issues of safety and largely neglect those of respect, choice and access. Nevertheless, many of the NCMIs have some utility when considering the e ciency and productivity of maternity services.
Labour and Birth Indicators capture a number of common interventions in delivery. Although interventions in delivery are often required to ensure the safety of mother and baby, Australia is known to have a high rate of potentially unnecessary Caesarean sections, induction and episiotomy (15,16). This can be seen as symptomatic of the medicalisation of the birthing experience and in the context of woman-centred care there is a clear impetus to eliminate unnecessary birth interventions. These labour and birth indicators are therefore highly relevant to include as output variables in e ciency and productivity measurement, as they can provide some indication as to how well a maternity service is providing care that adheres to the national strategy values of safety, respect, choice and access. Maternity services with a similar casemix should exhibit a similar rate of birth intervention. Differences in intervention rates could there indicate a high rate of unnecessary birth intervention in a given service and a deviation from the values and principles of the national strategy.
Birth Outcome Indicators capture important information regarding the physical health of mother and baby following labour and delivery. A baby's Apgar score assesses the clinical status of a baby immediately following childbirth. Third and fourth degree tears are classi ed as severe trauma to the perineum and can occur spontaneously or as a result of obstetric intervention during vaginal birth. Birthweight is a key indicator of a baby's health and is used as measure of health and wellbeing of the mother in pregnancy, as well as an indication of a baby's chance of survival, health, development and wellbeing (17). The physical health of mother and baby is obviously central to the provision of safe and effective maternity care and these indicators are therefore logical choices for inclusion as output variables in an e ciency and productivity analysis.
The NCMI Antenatal Period Indicators capture maternal behaviours that have been shown to be associated with child health outcomes (13). Smoking during pregnancy is a risk factor for many adverse outcomes in pregnancy, including pre-term birth, placental complications and perinatal death of the baby.
Antenatal care in the rst trimester is associated with better maternal health in pregnancy, fewer interventions in late pregnancy and positive child health outcomes (13).
Indicators such as these are more appropriately included in an e ciency and productivity analysis as measures of input, rather than output measures. This is because they in uence the complexity of care a woman may receive and therefore the resources consumed in delivering this care. E.g. women who smoke or do not receive antenatal care are at higher risk of a number of adverse outcomes in pregnancy, requiring higher complexity and more resource intensive maternity care. Indicators like these relating to maternal behaviours or characteristics are also a re ection of the casemix of any given maternity service. In an e ciency and productivity analysis, input measures can be 'cost adjusted' to account for the casemix of a given hospital, which can help distinguish between those maternity services who consume more inputs because they treat more complex patients and those who consume more inputs as a result of technical or allocative ine ciencies. Tables 1 and 2 indicate NCMI and NPDC items that are more appropriately included in an e ciency and productivity analysis as a cost adjustment measure.
The NCMI and NPDC collect a number of data items that can usefully be incorporated as output variables in an assessment of the e ciency and productivity of maternity services. However, this data is largely quantitative in nature and provides little insight into women's lived mental and emotional experience of birth. Indeed, the national strategy itself recognizes that the maturity and effectiveness of existing administrative data sets can be improved to include the collection of woman-reported outcomes, wellbeing and experiences, e.g. using patient-reported experience and outcome measures. Other jurisdictions in Australia also have in place surveys to measure patient experience. Western Australia (22) and South Australia (23) employ randomized surveys to collect and measure data regarding patient experience, but none of these surveys relate speci cally to consumers of maternity care. Victoria also employs randomized surveys to measure patient experience, but also includes specialized questionnaires for maternity clients. However, none of these surveys are as comprehensive or as widely reported as the New South Wales or Queensland surveys. Implementation of the New South Wales and/or Queensland Maternity Care survey methodology across Australia would go some way to ensuring that woman-reported outcomes of labour and delivery are represented in national data collections. It would also be of signi cant utility as a source of woman-reported outcomes for inclusion as output variables in any assessment of the e ciency and productivity of maternity services in Australia.

International Consortium for Health Outcomes Measurement Pregnancy and Childbirth Standard Set
The International Consortium for Health Outcomes Measurement (ICHOM) is a not-for-pro t organization that was established to promote and facilitate the global uptake of value-based health care. Value-based health care is a theoretical framework that places patients at the centre of care. It de nes value as the ratio of outcomes of care divided by the cost of achieving those outcomes, where outcomes are de ned as relevant end results of care from the perspective of the patient. To facilitate the implementation of value-based care, ICHOM works with international Working Groups of clinicians, researchers and patients to de ne standardized outcome measure sets (Standard Sets) for evaluating value in speci c condition areas (24).
ICHOM has developed a Pregnancy and Childbirth Standard Set that identi es 24 outcome measures to evaluate care during pregnancy and up to 6 months postpartum. Speci c outcome measures are grouped across four domains: Patient satisfaction with care; survival; morbidity; and patient-reported health and well-being. The Standard Set also includes a list of case-mix factors to allow comparison of outcomes across various patient populations. Table 3 shows the Pregnancy and Childbirth Standard Set outcomes mapped to four values outlined in Australia's national maternity strategy, and identi es those outcomes that are more appropriately included in an e ciency and productivity analysis as cost adjustments.
of the e ciency and productivity of maternity services, allowing for the inclusion of output variables that directly relate to woman's experience of pregnancy and childbirth.
Constructing an output measure based on current data Output measures for assessing e ciency and productivity of maternity services in Australia should ideally re ect the values and principles of woman-centred care. Australia's national maternity strategy clearly articulates Australia's vision for the provision of woman-centred maternity care based on the values of safety, respect, choice and access. This strategy therefore provides a useful framework to consider potential output measures for assessment of the e ciency and productivity of maternity services in Australia. Ideally, any e ciency and productivity analysis would incorporate output measures that correlate with and indicate how well a maternity service is delivering care in accordance with these values. They should also be able to be applied nationally, using data that is collected and accessible in every state and territory.
With these principles in mind, we have constructed a composite output measure that can be used in assessing the e ciency and productivity of maternity services in Australia. This composite measure consists of antenatal, labour and birth and birth outcomes indicators where data is available from the National Perinatal Data Collection. These measures are shown in Table 4 and are mapped against the four values of Australia's national maternity strategy: safety; choice; respect; access.

Discussion
The composite outcome measure put forward by this paper is constructed from antenatal, labour and birth and birth outcomes indicators where data is available from the National Perinatal Data Collection, mapped against the four values of Australia's national maternity strategy: safety; choice; respect; access. The value in constructing a composite output measure is that it captures multiple outcomes that are representative of the provision of woman-centred care -it represents the birth experience that should be attainable for the majority of women. Whilst recognizing that for some women attainment of these outcome measures would be inappropriate or unachievable, composite outcome measures nevertheless provide a useful measure of the overall performance of a maternity service in delivering 'best practice' women-centred care across multiple dimensions. However, as is evident from Table 4, the individual indicators in this composite outcome measure map largely to the values of safety and very little to the values of respect, choice or access. The current Australian national maternity data collections are largely quantitative in nature and provide little insight into women's lived mental and emotional experience of birth. At a state level, only two states routinely deliver and report woman-reported maternity experiences. Within Australia there is therefore no national data available that provides any speci c indication of how well maternity services are performing in delivering woman-centred maternity care.

Conclusion
This composite measure developed in this paper makes it very clear that there are signi cant gaps in Australia's maternity data collections with regard to measuring how well a maternity service is performing against the values of respect, choice and access. Adoption in Australia of the collection of womanreported maternity outcomes would substantially strengthen Australia's national maternity data collections and provide a more holistic view of pregnancy and childbirth in Australia beyond traditional measure of maternal and neonate morbidity and mortality. Widespread adoption of across Australia of maternity care surveys, such as those conducted in New South Wales or the implementation of the ICHOM Pregnancy and Childbirth Standard Set would go some way to providing a more comprehensive and nuanced assessment of how well Australian maternity services are performing in the delivery of women-centred care.