Cost-effectiveness and quality of life of a diet intervention postpartum: 2-year results from a randomized controlled trial

Overweight and obesity are growing health problems globally, affecting more than half of the adult population today [1]. Along with the increased risk of adverse health effects and all-cause mortality, obesity has a strong negative impact on health-related quality of life (QOL), which includes the individual’s perception of physical, mental, and social wellbeing [2]. Previous research has reported that nearly all aspects of QOL are adversely affected by elevated body mass index (BMI), and that women with excess weight have lower QOL compared to men of corresponding BMI [3, 4]. In addition, obesity contributes to increased societal costs through both direct health care costs and indirect costs. The latter is a result of decreased years of disability-free life, increased mortality before retirement, early retirement, disability pensions, and reduced productivity [5, 6].

Among women, pregnancy has been identified as an important risk factor for the development and exacerbation of overweight and obesity [7]. This is mainly explained by excessive gestational weight gain and subsequent postpartum weight retention, which increase the risk of complications during succeeding pregnancies [8] and influence long-term maternal health [9, 10]. However, the postpartum period may also spark motivation for lifestyle changes to lose the extra weight gained during pregnancy. Facilitators that converge in this period include increased energy requirement during lactation [11], motivation to return to pre-pregnancy weight [12], desire to serve as a parental role model [13], and an established contact with health care professionals. Also, in Sweden, women can benefit from parental leave until the child is 18 months old old. In addition to the reduced risk of maternal metabolic disease and future pregnancy complications [14], postpartum weight loss may also have an immediate impact on QOL and health care costs [15, 16]. Importantly, increased QOL is a highly relevant patient-centered outcome and an essential component in cost-effectiveness analyses. However, data on the long-term effect of postpartum lifestyle interventions on QOL and cost-effectiveness are missing, especially in real world settings. This information is critical to guide politicians and financers involved in decision-making processes about resource allocation.

We have recently conducted an effectiveness trial to evaluate whether a 12-wk diet intervention can produce weight loss among postpartum women with overweight and obesity within a primary health care setting in Sweden. The results showed that women randomized to diet intervention achieved a greater weight loss after 12 wk. (6.1 vs 1.6 kg, p < 0.001) and 1 y (10.0 vs 4.3 kg, p = 0.004) compared to the control group [17]. When women with a new pregnancy between 1 and 2 y were excluded, an effect emerged also at 2 y (8.2 vs 4.6 kg, p = 0.038) [18]. In this report, we evaluate the cost-effectiveness of the diet intervention and explore changes in QOL, as compared to a control group, in postpartum women with overweight/obesity within the context of primary health care in Sweden.

We set out to evaluate the impact of a primary health care-based diet intervention on QOL and cost-effectiveness in a postpartum weight loss trial. We found that the D-group improved their QOL more at 12 wk. and 1 y, and had a greater increase for the dimensions general health and mental health, and the mental component summary score, than did the C-group. Medium effect sizes were reached according to Cohen’s classification [31]. Furthermore, the cost per QALY was 7889 USD (SF-6D), 2367 USD (EQ-5D-3 L) and 1704 USD (EQ-VAS). With a willingness to pay 50,000 USD for a QALY, the likelihood of cost-effectiveness was 0.77–1.00. To the best of our knowledge, this is the first effectiveness trial to demonstrate that a diet intervention postpartum performed in real life settings improved QOL and was cost-effective. This is in addition to the favorable effects previously reported on anthropometric outcomes [17].

Previous research among non-pregnant individuals with obesity has reported that weight loss is associated with an increase in physical health but not in mental health [32]. In the current trial, there was no difference between study groups in the physical component summary score; however, a between-group difference was observed for the mental component summary score at 12 wk. and 1 y. This finding could be related to fewer physical limitations (excluding recovery from childbearing) in this young female study population (mean age 32.2 y) with relatively low BMI in the context of weight loss trials (pre-pregnancy BMI 28.4 kg/m², hence within the overweight range). Furthermore, although recovery from childbearing and being on maternal leave are plausible reasons for the improved QOL within both study groups, the explanation the greater improvement in the D-group is probably multifactorial. For example, the increase in QOL in the D-group could be related to the weight loss per se, and/or to the feeling of being able to control one’s lifestyle and weight. In addition, success in adhering to the new dietary regime, reaching pre-pregnancy weight and the design of the diet intervention may also have played a role. Nevertheless, between 1 and 2 y, there was a general decrease in QOL for both groups, in line with results from other lifestyle interventions [33, 34].

In the present trial, the decline in QOL and mental health from 1 to 2 y could be related to regaining weight between 1 and 2 y, indicating decreased compliance with the diet regime. The decrease in QOL could also be associated with returning to work or studies, as most women in Sweden stay at home from between 1 to 2 years. Hence, the transition to “normal life” may have impacted their QOL and also made it more challenging to maintain lifestyle habits established during parental leave. They were thus facing new barriers to weight management. Novel strategies and additional support might be needed to maintain weight loss and QOL in working life as compared to achieving improvements during parental leave. These strategies could include knowledge about stress management and the addition of physical activity, which has been shown to increase both physical and mental aspects of QOL [16, 35]. In sum, our results indicate that diet interventions postpartum should be supplemented with further efforts related to returning to working life.

The estimated costs per gained QALY (1704–7889 USD) for the diet intervention can be considered low. There are, however, no formal thresholds for what is considered good cost-effectiveness. What can be prioritized on a restricted budget depends on a number of factors besides cost-effectiveness, including affordability, budget impact, fairness, and feasibility. WHO argues that a threshold should simply be seen as an indication of poor, good, or very good value for money [36]. Despite this, there is ongoing discussion of what cost-effectiveness ratios are usually accepted. Cost-effectiveness ratios of 50,000–100,000 USD in USA [29] and 32,000–50,000 USD in UK [29] have often been accepted by stakeholders in Sweden and other Western countries. Thus, the cost-effectiveness ratios of the current trial can be considered low in relation to what Western countries are willing to pay for a QALY, and the likelihood of cost-effectiveness is high. Although there are no general recommendations for the threshold for the likelihood of cost-effectiveness for a change in routine care, there are arguments that it should be close to 0.50 [37]. Sensitivity analyses, in which higher costs and a lower gain in QALY were modeled, did not change the assumption of cost-effectiveness. This strengthens the conclusion that the diet intervention can be considered cost-effective.

The diet intervention in the LEVA in Real Life trial was developed and evaluated in a previous efficacy trial, the LEVA trial [38], conducted by our research group. In that trial, the diet intervention produced a weight loss of 9% after 12 wk. that was maintained at 10% after 1 y. In the related cost-utility analysis [39], the diet intervention was found to be cost-effective, with a cost per gained QALY of 8643–9785 USD and a likelihood of cost-effectiveness of 0.87–0.93. Compared to the LEVA trial, the present diet intervention was less costly (incremental cost 225 USD vs 303 USD). In the LEVA in Real Life trial, the diet intervention was somewhat cheaper but had approximately similar effects. However, it is important to note that it was performed in an ordinary health care setting. As far as we know, there are no other cost-effectiveness studies of diet intervention for postpartum women with overweight or obesity. In addition, diet interventions in other patient groups confirm that diet interventions for individuals with overweight or obesity are in general cost-effective [40].

This trial has several strengths. First, the diet intervention was delivered and evaluated within ordinary primary health care in Sweden; thus, the effect and costs should be representative of those involved in running the diet treatment in routine care. Nevertheless, future research should examine the effect of postpartum diet treatment in countries with less generous parental leave policies. Second, QOL can be viewed from different dimensions. In Sweden, SF-36 is the most established QOL-instrument, and has the advantage of providing information on multiple dimensions of QOL. Specific changes in the different dimensions of QOL represent an interesting additional aspect beyond what is provided by regular health economic QOL instruments that show QOL in relation to only a single dimension. Third, there is always a degree of uncertainty when estimating QALY based on a QOL instrument. It can be asked whether the questionnaires capture the actual change of QOL, whether the preference valuation is made by a relevant group of individuals and which of the methods of preference valuation (time trade-off, standard gamble, or rating scale) is used. We used three different QOL instruments in which the preferences were evaluated by a general population as well as by the affected patients. We also used all three valuation methods. Thus we should have reduced the uncertainty involved in estimating QALY. The conclusion is the same regardless of which QOL instrument was used. Further confirmation is provided by the fact that similar results were found in the previous LEVA trial, where the same diet intervention was evaluated. Finally, the cost-effectiveness analysis is based on real life data, thereby minimizing the need for assumptions.

The present report also has some limitations. Although attrition was low compared to other weight loss [41, 42] and postpartum [7, 43] trials, the women who dropped out during the trial may differ systematically from those who remained. Therefore, to reduce the risk of overestimating the effect, several intention-to-treat analyses were performed using different imputation strategies [44]. All models generated the same result, strengthening and confirming the conclusions from this trial. Note, however, that due to the explorative approach used to examine the dimensions of QOL, p-values should be interpreted with caution.

A diet intervention that produced clinically relevant postpartum weight loss also resulted in increase in QOL and was cost-effective. Maternity leave usually entails changes in everyday life, and so this period provides a unique opportunity for health care to offer lifestyle treatment and promote postpartum weight loss through dietary changes.