Author: Xuanqian Xie

The economic analysis of diagnostic/screening tests is relatively complex. I would like to present an example of using microvolt T-wave alternans (MTWA) to select high risks patients for implantable cardioverter defibrillators (ICD) (Filion et al., 2009). In the future, I will extend the discuss of cost effectiveness analysis of diagnostic tests to more complex conditions, such as 2 tests versus 1 test, optimized sequences of 2 or more diagnostic tests, different decision rules (i.e. disconjunctive positivity criterion or conjunctive positivity criterion), and so on.

When we try to model an economic evaluation of a single diagnostic/screening test (i.e. test vs. no test), the first question is what would happen without this test. Commonly, we assume that all patients with a suspected disease are treated or none of them are treated. Using a diagnostic test, the patients were divided into 4 groups, true positive(TP), false positive(FP), true negative(TN) and false negative(FN). The patients with positive (both true and false) test results would be treated and the patients with negative results would not receive the treatment. Then, we can model the clinical consequences following TP, FP, TN and FN. Sutton et al. discussed this issue in their methodology article (Sutton et al, 2008). (Dr. Sutton published a couple of great papers in meta-analysis and economic analysis. I have learned a lot from his publications.) But, frankly speaking, due to the paucity of evidence, usually it is not easy to obtain the precise estimate of the clinical consequences of wrong treatment following the FP results.

In my example of screening test, we assumed the MTWA can help to identify high risks patients, and ICD for the selected high risk patients would be more cost-effective than to treat all. Briefly, patients with positive results (high risk patients) in MTWA tests were treated by ICD, and negative results (low risk patients) were received the medical treatment only. Risk ratio (95%CI) of non-negative vs. negative MTWA test was 2.6 (1.4, 5.8). Without MTWA, we assumed that all patients received ICD, or all patients received the medical treatment. We applied the Markov model for each treatment strategy. See Figure 1.

Figure 1: The cost utility analysis model

See the Table below for the main results. MTWA marginally improves the cost-effectiveness of ICDs for primary prevention in patients with severe left-ventricular dysfunction (ICER: 108.9 K $CAD/QALY or 81.3 K $CAD /LYs versus 121.8 K $CAD /QALY or 91.1 K $CAD /LYs).

More details of assumptions of the model, data inputs and analysis can be found in our publication (Filion et al., 2009).

**Reference: **

Filion KB, **Xie X**, van der Avoort CJ, Dendukuri N, Brophy JM. Microvolt T-wave alternans and the selective use of implantable cardioverter defibrillators for primary prevention: a cost-effectiveness study. *Int J Technol Assess Health Care* 2009; 25(2):151-60

Sutton AJ, Cooper NJ, Goodacre S, Stevenson M. Integration of meta-analysis and economic decision modeling for evaluating diagnostic tests. *Med Decis Making* 2008; 28(5):650-667.

### Health Economic Evaluation

### Meta-analysis

### Statistics in Medicine

### Statistical graph and poster

### SAS /R /Winbugs Code

### Contact me

E-mail: shawn.xieq@gmail.com