Developing cost-effectiveness models (CEM) for cost-effectiveness analysis is an important tool for establishing the value of a new drug and gaining market access. Such analysis allows us to examine the ranges for cost-benefit ratio of a new drug and compare it to the treatment results with existing analogues. As a rule, CEM is developed based on a number of assumptions that may be inaccurate. Therefore, a robust cost-effectiveness analysis is impossible without a sensitivity analysis (SA).

What is sensitivity analysis?

Sensitivity analysis assesses the impact of changes in the initial parameters of the model on the outcome of interest. The sensitivity analysis technique consists of changing selected parameters within certain limits, provided that the other parameters remain unchanged. The greater the range of variation of the parameters, at which the final result remains a positive value, the more robust the model results are. Sensitivity analysis also allows you to identify the most critical variables that are most likely to affect the feasibility and effectiveness of the project.

Sensitivity analysis is used to evaluate simulation models. This study allows you to determine the impact of fluctuations in input variables' values on the model's output characteristics. It is carried out when it is necessary to establish at what variation of input data the validity of the main conclusions made on the basis of simulation results is preserved. The ease of conducting Sensitivity analysis in simulation is one of the advantages of this method.

How sensitivity analysis is useful

Recalculating results under alternative assumptions to determine the effect of a variable in SA allows:

1. Verify the robustness of the simulation model or system results.
2. Better understand the relationships between input and output variables in the system or model.
3. Reduce uncertainties by identifying model inputs that cause significant uncertainty in the outputs and therefore, should be the focus for improved reliability through further research.
4. Detect errors in the model (e.g., unexpected relationships between inputs and outputs).
5. Simplify the model by correcting inputs that do not affect the outputs, or by identifying and removing redundant parts of the model structure.
6. Improve communication between model developers and decision-makers.
7. Detect the range of input factor values for which the model output is maximum or minimum or satisfies some optimal criterion.
8. In the case of a model calibration with many parameters, primary sensitivity checking can simplify the calibration step by focusing on the sensitive parameters.

Thus, the consequence of SA is the development of a better model.

Sensitivity analysis in cost-effectiveness modelling

The SA of a cost-effectiveness model allows us to determine the project's resilience to possible changes in:

• the economic situation in general (changes in the rate of discount, increase in the costs);
• internal indicators of the project (changes in sales volumes, drug cost, etc.).

In other words, SA of economic efficiency allows for determining the stability of the project in relation to the fluctuations of financial market conditions and possible changes in the macroeconomic environment.

Sensitivity analysis is conducted in several stages:

1. Selection of key parameters of the project (ICER, NPV, IRR, FV, etc.), changes in which will significantly affect project flows and factors that affect their values (revenue, cost, wages, taxes, etc.).
2. Calculation of key parameters at basic values of factors.
3. Consistent change of factors and calculation of key parameters under new conditions.
4. Checking sensitivity of selected parameters with a probability of deviations of the first type (probability that the factor will change, i.e., become more, less, or remain within the target) and the second type (if the factor is still below the target level, then with a probability of 60% the deviation will not exceed 10%).
5. Determination of the most sensitive to these changes of key parameters and factors that have the greatest impact.
6. Comparison of the sensitivity of the output for each input.

The wider the range of parameters in which performance indicators remain within acceptable values, the higher the safety margin of the project, and the better it is protected from fluctuations of various factors that affect the project results.

Sensitivity analysis and cost-effectiveness modelling in healthcare

Digital cost-effectiveness model development is a popular service with drug and medical device manufacturers. Analysis using this tool allows you to:

Easily present evidence and change cost-effectiveness model parameters in real-time;
obtain a tool to help in negotiations with local or national government officials.

Individuals who make decisions about accessing the market for new medications or using new treatment protocols especially appreciate the additional information confirming CEA's reliability and validity. For this reason, Digital Health Outcomes recommends that sensitivity analysis should not be neglected in the economic models.