Cost, cost-consequence and cost-effectiveness evaluation of a practice change intervention to increase routine provision of antenatal care addressing maternal alcohol consumption

Alcohol consumption during pregnancy is associated with adverse obstetric and infant outcomes that can have lifelong social and economic consequences [1, 2]. Specifically, alcohol consumption during pregnancy increases the risk of miscarriage, still birth and Fetal Alcohol Spectrum Disorder (FASD) which is the most common preventable cause of intellectual impairment in the western world [2]. Despite this, the 2019 Australian National Drug Strategy Household survey reports that 55% of Australian women consumed any alcohol during pregnancy, and 14.5% continued to drink once they knew they were pregnant [3]. No safe level of alcohol exposure has been established and Australian national alcohol guidelines recommend women abstain from drinking alcohol whilst pregnant, trying to become pregnant, or breastfeeding [4].

Public maternity services are important settings for the provision of antenatal care to a large proportion of pregnant women [5, 6]. International [7] and Australian [8, 9] antenatal clinical practice guidelines recommend that during initial and subsequent antenatal appointments all pregnant women: have their alcohol consumption assessed; be advised that it is safest not to consume alcohol during pregnancy and of the potential risks of consumption; and be offered referral for additional alcohol treatment services if required [3]. Despite such guideline recommendations, assessment and care for antenatal alcohol consumption in public maternity services is sub-optimal [10, 11]. For example, in Canada approximately only half of surveyed health professionals reported providing advice to pregnant women regarding the consumption of alcohol [12]. In the United Kingdom two thirds of women reported receiving such advice from a midwife [13]. In a recent Australian survey less than two thirds of pregnant women reported that they received an assessment of their alcohol consumption and just over one third received advice and referral appropriate to their level of alcohol consumption at their initial antenatal visit [11]. Less than 10% of women received recommended care at subsequent antenatal visits [11]. Practice change strategies have been demonstrated to be effective in increasing the provision of evidence-based care in various clinical settings. Such strategies include educational meetings, local opinion leaders providing expert opinion, audit and feedback and electronic prompt and reminder systems [14‐17]. No studies have reported the effectiveness of such strategies in improving the provision of care addressing maternal alcohol consumption by maternity services. Whilst effective, the delivery of these practice change strategies should be considered against their resource requirements [18].

Despite the increasing use of economic evaluation in health services research, its application to the assessment of the cost and cost effectiveness of practice change intervention strategies is limited [18] [19, 20]. A 2019 systematic review by Roberts et al. identified 30 studies that included implementation or improvement as part of an economic evaluation [20]. Of those, 14 were implementation studies and the most common focus was on implementation strategies of new care pathways or novel services [20]. Of these, seven included a cost-effectiveness analysis (CEA), of which two were conducted in the Australian setting, one was a web-based intervention [21] implementation designed to improve the management of minor head injury in emergency departments [22]. Of the 14 implementation studies identified in this review, only six included specific implementation costs, such as those associated with staff training and education, the impacts of new processes on patient and carer costs and the cost of developing new processes [20]. Another recent systematic review of economic evaluations and cost analyses of guideline implementation strategies identified 235 implementation studies, of which only 10% provided information about implementation costs, with none providing detailed cost information [23]. Furthermore, this review identified 63 studies (27%) that reported an economic evaluation, however, overall the methodological quality was poor and very few included conclusions on the effectiveness or efficiency of implementing the guideline into practice [23]. Similarly, a systematic review of economic evaluations of antenatal nutrition and alcohol interventions and their implementation identified 12 studies, ten addressing nutrition intervention effectiveness and two addressing alcohol interventions [24]. The review found that although the alcohol interventions were cost-effective or cost saving, the cost and cost-effectiveness of the intervention and its associated practice change interventions were not reported, and have not been reported previously [24]. Failure to identify, measure and value explicit costs associated with implementation risks underestimation of the investment required to change practice. This has been identified as a missed opportunity to develop evidence about the importance of fixed and recurring costs associated with practice change interventions [20].

The importance of economic evaluation in the context of implementation science and how these analyses can be most efficiently incorporated into decision-making process has been recognised [19]. To address this evidence gap, a trial-based economic evaluation was conducted of a practice change implementation intervention that aimed to improve delivery of guideline recommended antenatal care addressing alcohol consumption by women during pregnancy. The aims of the economic evaluation were to:

The secondary aims for the economic evaluation were to report each of the aims by sector.

This economic evaluation outlined the costs, consequences, and cost-effectiveness associated with the first published randomised controlled trial to examine the effectiveness and efficiency of delivering a practice change intervention to improve the implementation of guideline recommended care for maternal alcohol consumption. The intervention was effective at improving reported receipt of all guideline elements of care. The incremental cost of intervention delivery was calculated to be $326,774. Labour (staff time) was the main cost driver and variation in delivery costs across sectors was associated with design differences in labour time and travel costs for each sector.

The average cost over the trial period of delivering the practice change intervention per eligible clinician was $993, and $591 per woman who received all guideline elements. The cost of delivering clinician training was upfront and is not expected to require additional investment beyond the trial horizon. Clinician training modules were developed during the intervention and made permanently available online for staff to access as part of their ongoing clinician training. Further, in an effort to sustain the intervention effect size, implementation strategies were deliberately designed to affect permanent and scalable change within the health system, e.g. the inclusion of specific modules into the state-based e-maternity patient record management system, and performance measures were embedded into the health system. Therefore, these implementation strategies are expected to sustain the effect size post-intervention. As such, the average cost per woman is expected to reduce markedly with continued delivery of this new model of care over time (e.g. beyond the trial time horizon), with increased women receiving antenatal care and with scaling up to involve other maternity services/clinicians. Similarly, with respect to potential ongoing costs (beyond the trial time horizon) associated with this intervention and maintaining the intervention effect size beyond the trial time horizon.

Direct comparison between the results of this and previously reported studies of the cost and cost effectiveness of practice change interventions is not possible given the significant differences in the design and methods of the various studies. Broad comparisons with studies included in previous systematic reviews were also limited as the costs associated with the implementation of the practice change strategies is not often considered [20, 24, 37]. The review by Roberts et al. [20] identified one modelled CEA of a universal alcohol screening and brief intervention program in primary care in England [38]. This analysis compared the health and social care costs verses health benefits and found screening patients for alcohol consumption upon registration with a family doctor would steadily capture up to 96% of the population over a 10-year programme [38]. This study showed that alcohol screening and the provision of brief advice, led by practice nurses, provided cost savings to the health care system of £120m over 30 years [38]. Similar to the methods and findings of the present study, the cost of the intervention was estimated using activity costs and identified that resourcing needs for this intervention would be highest in the early years of the program due to the volume of new patients being screened, and would decrease over time [38]. This study did not include the cost of implementing this model of care into routine practice or addressing barriers to care delivery by relevant health service providers. In comparison, the practice change strategies used in the current study were designed to be embedded into the health care system to maintain the new model of care beyond the intervention delivery and follow-up period. As such, it is plausible that the demonstrated change in practice could be sustained beyond the trial time horizon and the average cost per woman screened would decrease as more women attend the antenatal service. Future studies could include longer intervention follow-up periods to capture the number of women receiving care over a longer time period, as well as the health and societal benefits for mothers and their infants. Further research is also required to determine the cost of delivering the intervention at scale and whether economies of scale can be achieved in its delivery.

Implementation costs are recognised as an understudied aspect of implementation science [39, 40]. Saldana et al. proposed that one of the reasons implementation costs are not routinely examined is the lack of standard measurement [39]. More recently, Cidav et al. proposed a pragmatic approach to systematically estimating detailed resource use and costs of implementation strategies that combine time-driven activity-based costing with a leading implementation science framework [41] to guide specification and reporting of implementation strategies [33]. One of the key strengths of this study was the use of this time-driven activity based micro-costing to map implementation processes with actions, actors and strategies. This method provides transparent, granular cost estimation and allowed for a cost comparison of the different implementation strategies. It has been proposed that there could be value to using standardised methods for estimating implementation costs as it could allow decision makers responsible for determining the viability and feasibility of adopting new practices to benefit from the ability to generalise across settings [39, 42]. The transparent reporting of activity costs using a published time-driven activity-based costing method is intended to inform researchers and decision makers how specific components of an implementation intervention influence the total cost.

Another strength of this economic evaluation is the identification, measurement and valuation of development costs. That is, once the practice change intervention has been delivered, there are non-recoverable intervention components that remain within each sector. For example, the cost of developing and producing local clinical guidelines, the cost of developing educational materials and the cost of developing site-specific systems level monitoring and accountability measures. Saldana et al. argue development costs are an important consideration in decision making as policy makers must decide ex ante whether to invest in a new model of care and unrecoverable costs must be considered amongst the future benefits [39]. Transparent reporting of costs associated with the practice change intervention is intended to inform decision makers of when, during the practice change process, different costs and benefits can be expected [19].

It was not possible to calculate an ICER for the cost per additional service user (pregnant woman) who received all guideline elements of care. The study was designed and powered to measure effectiveness at the health sector level, thus precluding this ICER calculation. Given this limitation, the average cost per additional service user (pregnant woman) exposed to the recommended treatment was calculated. Another, limitation of the study was that the ICER was calculated on the assumption that the intervention was wholly additional to usual care, that is, no usual practice activity was displaced as a result of the intervention. Identifying, measuring and valuing the cost of usual care was beyond the scope of the trial and hence a limitation of the analysis. The trial time horizon was 35-months, and as such only upfront and short-term costs to health service providers were included. Similarly, the potential impact of increased referrals to drug and alcohol services and the longer-term benefits of alcohol-harm reduction to society at large, was not captured in this trial and is a noted limitation. Future economic evaluations should endeavour to include these components. The range of costs associated with sustaining changes in provider behaviour and maternal alcohol consumption is complex [2], and beyond the scope of the analyses.

The practice change intervention was effective at improving women’s reported receipt of all guideline elements of care. The incremental cost of delivering the intervention was calculated to be $326,774. To our knowledge, no similar studies have been published in the literature. The economic evaluation provides information for decision and policy makers regarding the cost, cost-consequence and cost-effectiveness of delivering a practice change intervention to support the introduction of a model of care for addressing alcohol consumption by pregnant women. Given the positive trial findings, further research is required to assess sustainability and determine the cost of delivering the intervention at scale and whether economies of scale can be achieved.