Economic evaluation of stuttering treatment in preschool children: The RESTART-study

Highlights

• Differences in effects and costs between the LP and RESTART-DCM treatment were small.

• Cost–effectiveness ratios were favorable for the LP compared to RESTART-DCM treatment.

• The LP is a good alternative to RESTART-DCM treatment in Dutch primary care.

Abstract

Purpose

The purpose of this study was to evaluate the incremental cost–effectiveness and cost–utility of the Lidcombe Program (LP) compared with treatment based on the Demands and Capacities Model (RESTART-DCM) for preschool children who stutter.

Method

A cost–effectiveness and cost–utility analysis were carried out alongside a Randomized Clinical Trial (the RESTART-study). In total, 199 children in 20 speech clinics participated. Outcome measures included the number needed to treat, based on the percentage of children who did not stutter at 18 months, and Health-related Quality of Life (EQ-VAS and HUI3) at 3, 6, 12 and 18 months. Health-related Quality of Life scores were used to calculate quality adjusted life years (V-QALYs for the EQ-VAS and U-QALYs for the HUI3). Direct and indirect costs were measured by cost questionnaires. Missing data were multiply imputed. Percentages of children who did not stutter in both groups were compared by a chi-square test. Between-group differences in mean QALYs and costs, as well as cost effectiveness and cost–utility ratios, were evaluated by applying bootstrapping techniques.

Results

After 18 months, health outcomes were slightly better in the LP group, although only the difference in V-QALYs was statistical significant (0.018; 95% CI: 0.008 to 0.027) with a small effect size (Cohen's d = 0.17). Mean costs for the LP group were significantly higher compared to the RESTART-DCM group (€3199 versus €3032), again with a small effect size (Cohen's d = 0.14). The incremental cost–effectiveness ratio was €3360 for one additional child who did not stutter with the LP, and the estimated cost–utility ratios were €10,413 (extra cost per extra V-QALY) and €18,617 (extra cost per extra U-QALY). The results indicated a high probability that the LP is cost-effective compared to RESTART-DCM treatment given a threshold for willingness-to-pay of €20,000 per QALY.

Conclusions

Differences in effects and costs between the LP and RESTART-DCM treatment were small. Cost–effectiveness and cost–utility ratios were in favor of the LP. The LP is considered a good alternative to RESTART-DCM treatment in Dutch primary care.

Introduction

Persistent stuttering can lead to a decreased health related quality of life (HRQOL) in the psychological, emotional and social domains of functioning (Craig et al., 2009, Koedoot et al., 2011, Menzies et al., 2009), as well as to substantial health care costs (Blumgart, Tran, & Craig, 2010). To prevent stuttering becoming persistent, treatment is best initiated in the preschool years. Treatment should preferably result in a high percentage of children recovering from stuttering at acceptable societal costs. Information on costs and effects of available stuttering treatments could help clinicians and policy makers in decisions on therapy choice and reimbursement. Although the last decade has shown an increasing number of studies into the efficacy of stuttering therapy in preschool children, there is a lack of evidence on the cost–effectiveness of available treatments.

Two widely applied treatment approaches for preschool children who stutter are treatment based on the Demands and Capacities Model (Starkweather and Gottwald, 1990, Starkweather, 2002) and the Lidcombe Program (LP; Onslow, Menzies, & Packman, 2001). In the Netherlands, children are commonly treated according to the former approach. Currently, about 10% of Dutch speech- and language pathologists (SLPs) working in private practices are also trained in the LP. The LP is supported by a larger body of evidence than any other treatment (Nye et al., 2013), but a head-to-head comparison against other types of treatment is presently unavailable. However, it is known that the LP requires a relatively long maintenance phase after fluent speech has been attained. The LP is therefore expected to be more costly than DCM based treatment. The average number of treatment sessions for DCM based treatment has been estimated at 12 sessions (Starkweather & Gottwald, 1990), while the LP requires almost double. This raises the question whether the presumably higher treatment cost of the LP is compensated by a greater proportion of recovered children, fewer relapses, and better individual speech outcomes, as suggested by Onslow et al. (2001).

An economic evaluation can provide insight into the costs and effects of a new health care intervention compared to usual care. All types of economic evaluations assess costs, but health consequences can be measured in different ways (Drummond, Sculpher, Torrance, O’Brien, & Stoddart, 2005). The most common forms of economic evaluation are cost–effectiveness analysis (CEA) and cost–utility analysis (CUA). In a CEA, the health consequences are expressed in terms of natural units (i.e., survival or a desired clinical outcome like recovery), while in a CUA the effects are valued in terms of generic measures of health, such as quality adjusted life years (QALYs; Drummond et al., 2005). The comparison of costs and effects of a new intervention with usual care results in an incremental cost–effectiveness or cost–utility ratio. This metric can be used to judge whether the additional effects are large enough to justify the extra costs. To get the most benefit from resources available to society and, accordingly, to guide implementation and reimbursement decisions, an economic evaluation should be conducted from a societal perspective. This implies that all costs and health benefits are included, regardless of to whom costs are related to or who receives the benefits (Drummond et al., 2005). In the field of speech and language pathology economic evaluations are scarce, but crucial to provide a basis for decisions on implementation and reimbursement of therapies (e.g., Robey, 2004). The aim of the present study was to determine the incremental cost–effectiveness and cost–utility of the LP compared to DCM based treatment.