Introduction
Achilles tendon rupture is a common injury and the incidence has been shown to be increasing over the past few decades [
14,
16,
20,
23,
24]. While the previous studies have compared whether surgical or non-surgical treatment is the most beneficial in terms of function and outcome, no significant differences have been observed, except for the risk of re-rupture [
9,
18,
31,
32]. The debate with regard to the best treatment for Achilles tendon rupture remains ongoing.
A cost-minimisation analysis of the management of acute Achilles tendon rupture has demonstrated that the overall costs of surgical management are higher than those of non-surgical management [
29]. However, to the best of our knowledge, surgical versus non-surgical treatments of acute Achilles tendon rupture have not as yet been put in the context of a cost-effectiveness analysis where the differences in cost are compared with the difference in quality-adjusted life years (QALYs), which can show the cost of each QALY gained by a given treatment.
Cost-effectiveness analysis is being used increasingly to inform decision makers with regard to setting priorities in healthcare. Comparing and ranking treatments based on the cost per gained QALY (the lower, the better) can indicate how to maximise patient health benefits given limited healthcare budgets [
17]. QALY is a health outcome metric that combines health-related quality of life (HRQoL) and “quantity” of life (life length). One QALY can be viewed as 1 year lived in the best possible health state. The HRQoL used to calculate QALYs is (typically) based on patients’ self-assessed valuations of different health states and often referred to as a preference-based measurement [
28]. Different types of preference-based instrument are used to measure the preference-based HRQoL score. These instruments could be condition-specific, but they are commonly generic, i.e., suitable in theory for all kinds of healthcare treatment, and include the EQ-5D, the six-dimensional health state short form [
4] and the Health Utilities Index [
15], for example. There is no consensus on which preference-based measurement should be used in cost-effectiveness analyses, although the EQ-5D has become increasingly recognised [
5,
30].
The Achilles Tendon Total Rupture Score (ATRS) is a primary patient-reported outcome measurement related to symptoms and physical activity after treating total Achilles tendon rupture. The score is reported to have high reliability, validity, and sensitivity [
6,
13,
25]. However, it lacks a preference-based score, i.e., how do patients weight the importance of the different items. As a result, it is not possible to use the ATRS directly to calculate QALYs to assess treatments in cost-effectiveness analyses [
25]. This problem has been encountered multiple times in clinical studies [
5], where a non-preference-based measurement has been the only suitable health measurement available for the condition in question. To solve this problem, a method known as mapping is being used more and more frequently [
5,
7,
21,
22]. Mapping investigates the statistical relationship between a non-preference-based measurement and a preference-based measurement, producing an algorithm (“map”) to be used in the calculation of a preference-based HRQoL score. To make this feasible, the method requires a data set of the source measurement (e.g., ATRS) and the target measurement (e.g., EQ-5D) that have been administered alongside each other to the same patients in the relevant clinical trial [
5,
30].
If a statistical association between the ATRS and the EQ-5D can be established, i.e., allowing the ATRS to be directly applicable for cost/QALY analyses, it will be valuable in the assessment of treatment for total Achilles tendon rupture. It was hypothesized that a statistical association between the ATRS and the EQ-5D could be established with mapping as an approach. The purpose of this study was to develop an algorithm to convert the ATRS to the EQ-5D by mapping.
Discussion
The most important finding of the present study was a model for predicting the EQ-5D score from the ATRS. The high
R2 (0.57) indicates a high goodness of fit, even though the model only demonstrated a correlation in three of ten items. In particular, the impact of pain and daily activity was proven to be of specific interest. Brazier et al. [
5] reviewed 30 mapping studies for various condition-specific health states, with a total 119 different mapping models. They reported an
R2 of 0.17 for one of the poorer fitting models and an
R2 of 0.51 for the better model when mapping condition-specific measurements onto generic measurements. This suggests a higher level of fitness in this model compared with most other available mapping models.
With its high validity, reliability, and sensitivity, the ATRS is the only patient-reported measurement for the outcome of an acute Achilles tendon rupture [
25]. One of the main strengths of mapping the ATRS to a preference-based measurement is that it can be used for cost-effectiveness analysis. There is a little agreement about the most appropriate preference-based measurement for this purpose. Moreover, different preference-based measurements are not guaranteed to generate the same values for the same sample of patients [
5]. However, with the extensive use of the EQ-5D and as the most commonly selected target measurement in mapping studies, the EQ-5D has also been the chosen target measurement in the present study.
Although mapping is gaining in popularity, the validity of this method has not been fully addressed. Round and Hawton [
27] raised questions about the validity of mapping, arguing that translation from one score to another does not mean that the same health preference is being measured. There are a number of fundamental concerns about mapping, the first of which is the differing sensitivity between the instruments. Generic instruments are designed to measure general health aspects but are insensitive to small health changes. In contrast, condition-specific instruments are inadequate for measuring general health but are sensitive to changes specifically related to the condition of interest. The second is the degree of conceptual overlap between the dimensions measured by the instruments. The less overlap between the dimensions, the weaker the mapping function, and vice versa. Regardless of the degree of overlap, the loss of information associated with dimensions in either of the involved instruments is difficult to avoid when mapping is done. The potential consequences of poor validity are overestimating/underestimating utility values.
Considering these concerns, there is no doubt that mapping to generate utility values is only second best to using preference-based measurements in the first place. However, given that many clinical studies are missing, or were unable to incorporate a preference-based measurement (e.g., not suitable for the relevant condition) and interest in performing QALY-based economic evaluations with clinical studies is growing, mapping is an increasingly used as an alternative solution.
Ideally, the current mapping algorithm may potentially play an essential role in the assessment of Achilles tendon rupture treatment. As already mentioned, the treatment for acute Achilles tendon rupture is either surgical repair or non-surgical treatment. No significant differences in terms of symptoms, function, or result have been shown in the previous studies [
9,
18,
31,
32]. There is a reduced risk of re-rupture with a surgical repair (3.1–3.5%) in comparison with non-surgical (12.6–13%) treatment [
3,
19], but the downside to this is a higher rate of complications such as infections and adhesions [
1]. As the benefits are comparable, there is still no consensus on the best treatment for acute Achilles tendon rupture. However, the current mapping algorithm makes it feasible for researchers and medical practitioners to estimate a utility value for QALY calculation in clinical studies or settings in patients with an acute Achilles tendon rupture where the ATRS is being administered. As a cornerstone of economic analysis, the QALY enables the measurement of economic benefits between healthcare interventions, while incorporating the impact on quality and quantity of life. Given that there are two comparable treatment options for Achilles tendon rupture, using QALYs to measure benefits in cost-effectiveness analyses for both may provide important input in clinical practice, as well as in political decision-making. As a result, the mapping algorithm developed in this study is potentially valuable when assessing the treatment of acute Achilles tendon ruptures.
It should be noted that the mapping algorithm presented in this study will only be applicable for fairly healthy patients with an EQ-5D of 0.47 as the lowest possible score. This is expected, as the analysis is performed on a sample with a high EQ-5D score. It remains to be determined whether the algorithm is applicable to patients with a poorer health state, i.e., by repeating the experiment on a sample with lower EQ-5D scores.
Conclusions
Utility values are best obtained directly using preference-based measurements, while deriving them with mapping is an alternative solution in clinical trials where only non-preference-based measurements are available. In this study, a mapping algorithm between the ATRS and the EQ-5D was developed, thus providing a way to perform QALY-based cost-effectiveness analyses of acute Achilles tendon rupture treatment.