Result card

  • ECO5: What is the incremental cost-effectiveness ratio?
English

What is the incremental cost-effectiveness ratio?

Authors: Suvi Mäklin, Taru Haula, Kristian Lampe, Jaana Leipälä, Ulla Saalasti-Koskinen

Internal reviewers: Aurora Llanos-Mendez, Sergio Marquez, Urs Brügger, Mark Pletscher, Iñaki Imaz, Sonia García Pérez

Acknowledgments: Information Specialist Jaana Isojärvi (FINOHTA/THL, Finland)

Systematic literature review

An evidence table from the systematic review described in the domain methodology was used to compile this answer. The incremental cost-effectiveness ratios (ICERs) reported in economic analyses are extremely difficult to compare in an exact manner, since they are heavily dependent on e.g. the healthcare setting, the methods of the original analyses, the timeframe and changes in currency rates over time. Therefore, no exact threshold levels were used, but instead a descriptive analysis of the data was performed, considering the numbers in euros and using foreign currency rates during the first week of May 2012. Detailed results of each study are available in the evidence table in Appendix ECO-2 to this result card.

In some cases the authors listed more than one ICER value reflecting, for example, changing assumptions about follow-up times. In this result card the values reflecting the longest available time horizon were used whenever available.

Modelling

Decision-analytic modelling was used to estimate the ICERs. A previously constructed model combining a decision-tree and a Markov model was used {2}, with some modifications {3}. The structure of the model is presented in the domain methodology. The analysis was done from the healthcare provider’s perspective using a time horizon of the expected remaining life time of the 65- year-old population. Base case analysis was done using Finnish data and data from the literature. The analysis was conducted using TreeAge Pro HealthCare software (version 2011, TreeAge Software Inc.).

Systematic literature review

Twenty-six studies identified in the literature review contained either a cost-effectiveness analysis (CEA) or cost-utility analysis (CUA) of population-based ultrasound screening for AAA compared with no screening. Of these, 19 studies were published less than 10 years ago (2003 or later).

The time horizon in the included studies was typically expected lifetime or 20 years or more. Only a few studies employed a time horizon of less than 10 years. Two included studies actually reported the same study, but one had a time horizon of 7 years and, the other, 10 years {10, 35}.

A clear majority of CEA or CUA analyses included men either as a single cohort (typically 64 or 65 years of age) or as a group between 64 and 75 years of age. Somewhat younger (from 50 years onwards) or older men were included in 6 studies {20, 33, 36-39}. Women were included in three studies only (Mäklin 2011, Wanhainen 2006, Russell 1990). All but three studies originated from the Nordic countries, the UK or North America. The three other studies were from Italy {28} and the Netherlands {25, 36}, one of them partly from Norway.

The ICER varied considerably across studies as expected since the analyses had been made in different healthcare systems, during different times and using different time horizons. The lowest ICER was found in a Danish study published in 2010 {26}, with ICERs of 157 € per LYG and 179 € per QALY. The highest ICER was likewise found in a Danish study published in 2009 {2}, with an ICER of 43 485 €/QALY (see comment section below).

In 22 of the 26 studies the cost per LYG or per QALY was lower than approximately  – and in several of them clearly less. This was the case for 17 of the 19 studies published less than 10 years ago (2003 or later).

In four studies the cost was clearly higher than approximately 10 000 €, expressed either as €/LYG or €/QALY {2, 35, 38, 40}. These results, however, may be less reliable (see comment section below). Five out of seven studies published more than 10 years ago (2002 or earlier) indicated an ICER at the very low or very high end of the range.

The four studies that observed higher ICERs than the remaining studies, originated from Canada {38}, Denmark {2}, and the UK {35, 40}. More recent studies indicating a lower ICER also originated from all these countries.

The most recent of the three studies, which included women as well as men, concluded that screening women may also be cost-effective, but in the absence of information on women the analysis was done mainly with information concerning men, and hence this result should be interpreted with caution {3}. The second study,  which was the only study considering women only, concluded that screening women may be cost-effective and that women should be included in future evaluations of AAA screening {32}. The third study concluded that a screening programme should include both men and women {33}.

Four systematic literature reviews were also identified, published between 2007 and 2010 and covering literature up to June 2008 at the latest{34, 41-43}. One of them concluded that AAA screening will probably gain additional life years and quality of life for men aged 65 years or older at acceptable extra costs {41}. Another concluded that AAA screening seems to be both effective and cost-effective, but that economic evaluations do not always take into account peri- and postoperative mortality {42}. The third review was more critical, concluding that most of the existing health economic evaluations have employed optimistic assumptions and included too few sensitivity analyses, and hence further cost-effectiveness analysis of AAA screening is recommended (Ehlers 2008). This review also included other types of screening for AAA apart from population screening and hence its focus was somewhat different. Four out of five of its authors were also authors of the Danish study (Ehlers 2009) that found the highest ICER for AAA screening. The fourth review concluded that the cost-effectiveness of AAA screening may be acceptable, but needs further expert analysis, and that the evidence to demonstrate a benefit in women is insufficient {43}.

In conclusion, the currently available evidence from 26 CEA or CUA analyses supports the cost-effectiveness of this technology in many European settings, at least in the UK, the Scandinavian countries, the Netherlands and Italy. A more specific assessment can only be done in specific settings using local data and values.

Detailed characteristics of the studies are available as Appendix ECO-2.

Modelling

Base case analysis

Table 6 summarises the results of the base case cost-effectiveness analysis. The ICER of population-based AAA screening in Finland would be 8433€ per LYG, when ultrasound screening is offered once to 65-year-old men. The ICER for women is lower at 7198€/LYG, due both to, lower incremental costs and lower incremental effectiveness of AAA screening compared with no population-based AAA screening. The results for women should be interpreted with caution as they are heavily based on data for men, and on assumptions that the natural course of AAA is similar in men and women.

Table 6. The results of the base case analysis of cost-effectiveness of population-based AAA screening compared with no population-based screening in Finland. The analysis was run for men and women separately.

  

Costs

(€)

Incremental costs (€)

Effectiveness

(life years)

Incremental effectiveness

Incremental cost-effectiveness ratio

Men

No AAA screening

350

 

11.52425

  
 

AAA screening

579

229

11.55136

0.02711

8433 €/LYG

Women

No AAA screening

99

 

15.67382

  
 

AAA screening

192

93

15.68668

0.012857

7198 €/LYG

Sensitivity analysis

The robustness of the results was tested using probabilistic sensitivity analysis, varying most of the parameter values simultaneously. The results of the probabilistic sensitivity analyses for both sexes are shown as cost-effectiveness acceptability curves in Figures 3 and 4. The cost-effectiveness acceptability curves represent the probability that intervention is cost-effective at different threshold values. In the base case analysis with a willingness-to-pay threshold of 24 000 €/LYG, population-based AAA screening for men is cost-effective with a probability of 95%. For women, the corresponding threshold value is 10 300€/LYG.

 106.ECO Fig 3

 Figure 3. The cost-effectiveness acceptability curve, AAA screening for men (probabilistic sensitivity analysis).

 106.ECO Fig 4

Figure 4. The cost-effectiveness acceptability curve, AAA screening for women (probabilistic sensitivity analysis).

Also, a one-way sensitivity analysis was undertaken for some key parameters (Table 7) in order to gauge the effect of these on the results. Varying the attendance rate at the screening had only a minor effect on the ICER. Variation of the discount rate had a stronger effect on the results –yet the ICER stayed at what could be considered to be an acceptable level.

Table 7. The results of one-way sensitivity analyses, incremental cost-effectiveness ratio of population-based AAA screening compared with no population-based AAA screening, €/LYG.

 

Men

Women

Base case

8433

7198

Discount rate 0%

6104

4080

Discount rate 5%

12 216

8855

Compliance with screening 70%

8728

7407

Compliance with screening 85%

7980

7185


The 26 studies identified in the review represent a variety of healthcare settings, periods and methodological approaches. Also the target population varies to some extent in terms of age. Consequently, great care should be taken when extrapolating results from these studies to specific settings.

Overall there seems to be a reasonably uniform body of evidence indicating that population screening for AAA using a single ultrasound investigation entails an ICER of less than approximately 10 000 €. The number may also be considerably lower, even less than 200 € (e.g. {26}).

Those four studies in which the ICER was estimated to be the highest all contain features that make their applicability in other settings questionable in present-day Europe. The Canadian study {38} and one of the UK studies {40} are from the 1990s and hence their results may be outdated. One should also notice that this early UK study reported two different ICERs depending on the aneurysm size {40}. The other UK study is more recent {35}, but its time horizon is only 7 years. Its authors conclude that it compares well with other screening programmes and that they expect its lifetime cost-effectiveness to be highly favourable. The 10-year results of the same study were reported in another paper included in this review and they were in line with the majority results {10}. The authors of the Danish study concluded that AAA screening would be unlikely to be cost-effective {2}. This study has been heavily criticised for its methods {44-46}. Four out of five authors of the Danish study were also authors of the most critical literature review {34}.

Most of the studies that we have referenced above in detail are studies that reported upper and lower limits of the range of findings. Three key studies were not explicitly referenced but belong to the analysis {47–49}.

The four systematic reviews we found were not uniform in their conclusions {34, 41-43}. Our current review, however, contains seven studies that have been published after the time limits of the literature searches in the four reviews.

It should also be noted that the current evidence does not contain justification for excluding women as a target group for AAA screening. The evidence concerning women is less extensive though and would probably benefit from further research.

Critical
Partially
Mäklin S et al. Result Card ECO5 In: Mäklin S et al. Costs and economic evaluation In: Jefferson T, Vicari N, Frønsdal K [eds.]. Abdominal Aorta Aneurysm Screening [Core HTA], Agenzia nationale per i servizi sanitari regionali (age.na.s), Italy; 2013. [cited 28 May 2023]. Available from: http://corehta.info/ViewCover.aspx?id=106

References