This review of the literature on economic evaluations of smoking cessation and relapse prevention programs among pregnant women reveals a dearth of studies on the subject and provides justification for further research support in this critical public health area. Although none of the studies were performed entirely in accordance with Panel recommendations or BMJ guidelines (e.g., none employed incremental CEA or CUA), numerous studies included certain aspects of these guidelines and provided useful research findings on the value of prenatal smoking cessation to women, infants, providers, and society at large. Thus, portions of each study may be used as an example for future analyses. It is also worth noting that, since most of the studies we reviewed were published before the development of the guidelines for conducting and reporting economic evaluations, one would expect some divergence between recommended guidelines and study results, further bolstering efforts of the mid-1990s to standardize economic evaluations to reduce uncertainty among researchers, reviewers, and journal editors about methods and reporting practices. Standards established by the Panel, the Cochrane Group, the BMJ, and others are likely to improve the quality and consistency of future economic evaluations in health and medicine.
Thus, while differences in the design, reporting, and description of economic evaluation and in model assumptions, data definition and estimation, discount rates and perspectives, to a certain extent limit our ability methodologically to draw head-to-head conclusions about the most effective and cost-effective strategies, together these studies offer distinct insights into the value of reducing smoking among pregnant women.
For example, despite the fact that reporting practices of economic analyses varied widely, most studies in this field have employed CBA and have found favorable cost–benefit ratios, even when maternal benefits are excluded. Moreover, estimates of program costs appear to be based on similar assumptions and to be similar. It is unclear why CBA has been the method of choice [24
], and it is worth noting that the types of CBA performed in the studies reviewed here are not traditional (e.g., valuing a life-year saved in monetary terms) but are more akin to cost-saving analysis in that they estimate savings in health-care expenditures resulting from not smoking. It appears that the literature has focused on these types of studies because of the interest and ease of estimating hospital costs (primarily costs of neonatal intensive care) associated with not quitting smoking during pregnancy.
The studies that aimed to estimate the break-even cost of smoking cessation programs demonstrated that such programs pay for themselves because, by and large, they save more than they cost. Many studies did not, however, adopt the societal perspective, nor did they include all relevant costs (e.g., training costs), suggesting these research findings are conservative estimates of potential savings. Although it is difficult to say whether such programs produce incremental health gains (at the margin) and to assess the marginal cost of those gains (because few studies used comparable outcome measures [health gains measured in life-years saved or QALYs saved]), resource savings as an outcome is relevant for analyzing the financial investment of prevention programs.
In summary, prenatal smoking cessation offers both health and economic benefits for women, infants, providers, and society. When women quit smoking by the first trimester of their pregnancy, for example, their infants are likely to have the same body weight as infants of nonsmokers and they significantly reduce the risk of intrauterine growth retardation. Prenatal smoking cessation programs are also relatively inexpensive on average, with costs in some cases as low as $25 per person for brief counseling. Combining costs and benefits reveals highly favorable benefit–cost ratios (of up to $6 saved for every dollar invested), significant cost savings (of up to $1,000,000), and modest costs for health gain, ranging from $100 per percent that quit to $4000 per LBW birth prevented to $63,000 per perinatal death prevented to $11,000 per life-year saved and $210,000 per SIDS death averted. By any measure, such programs compare favorably to more than 80% of clinical preventive services that are not cost-saving; in some cases, the cost per life-year saved of these services can range from $165,000 to $450,000. These findings could be considered by state public health departments, Medicaid agencies, employers, and health insurance plans (including managed care) to provide smoking cessation benefits and work with providers on screening, counseling, and behavioral interventions for tobacco use.
A final note on methodology, while such limitations hinder comparisons among programs, they also represent an opportunity for future analyses to better adhere to sound and consistent methodologies as recommended by the Panel and the BMJ. A lack of economic studies is common in health and medicine, however, especially in the area of community prevention services [25
]. Thus, even the limited number of economic studies on smoking cessation among pregnant women, although not surprising, signal the value of such studies.
In some respects, however, these results are somewhat surprising, given the general trend toward more systematic reporting in the evaluation of tobacco-related interventions [26
]. There has also been a longstanding emphasis in health economics toward standard reporting of study assumptions and basic cost and health outcomes [4
Our analysis has several limitations. First, it is restricted to a specific intervention type (smoking cessation and relapse prevention) and therefore does not include the economic evaluation of other types of programs that aim to reduce the health risks to mothers and infants during and after pregnancy. Second, as with any review that uses key words in a literature search, it may have missed some relevant studies. Third, an in-depth assessment of the merits of all clinical and economic assumptions and research methods was not made. As noted above, the considerable heterogeneity among study types, methods, main assumptions, interventions, and outcomes precluded a formal meta-analysis and the use of strict quality criteria relating solely to cost-utility analyses. Therefore, we included all studies that met our basic inclusion criteria and derived no weights to assess the significance or quality of any individual study. The lack of comparability across studies limits our ability to determine specific public policy implications.
Despite these challenges, efforts to sort out some of the finer methodological challenges in both the cost and effectiveness elements of CEA will eventually lead to more uniformity in reporting. For example, efforts to enumerate cost categories and determine and measure health benefits (in terms of life-years gained) will eventually improve the generalizability and comparability of the results of CEA. In combination with established standards, such efforts will help in the development of an analytic framework for making future studies more comparable.
In conclusion, ideal studies of economic evaluations of smoking cessation and relapse prevention programs for pregnant women would prospectively apply standardized methods noted herein. Characteristics one would be looking for in such a study would include an economic evaluation planned prospectively alongside a randomized clinical trial in which all inputs consumed in the interventions would be measured and valued alongside the clinical trial to enhance the reliability and validity of intervention costs. Costs collected would include those necessary to reproduce the intervention in a nonresearch setting and such inputs would likely include time spent with clients for intervention delivery and follow-up and materials. The cost analysis would be extended to the societal perspective by including cost savings for neonatal intensive care, chronic medical conditions, and acute conditions during the first year of life and cost savings for maternal health care (cardiovascular and lung diseases). The primary outcome measures would be quit and relapse rates measured during the trial and extended to the societal perspective by converting such rates into life-years saved and QALYs saved. Future benefits would be discounted at a 3% annual rate as recommended by the Panel on Cost-Effectiveness in Health and Medicine. CEAs employing these data would then be performed from the societal perspective to estimate incremental cost-effectiveness ratios, expressed as net resource cost per life-year gained or QALY gained. Sensitivity analysis performed by varying important parameters singly, and in combination, through clinically meaningful ranges would examine the robustness of the ratio estimates. We are publishing one such study that meets these criteria [27
]. The studies reviewed in this article might therefore be seen as useful embryonic elements of a more systematic framework for conducting meaningful and useable analyses of smoking cessation and relapse prevention programs for pregnant women.