Longexe

Measurement

What should be exactly stated when the primary outcome measure is explained?

An explanation about every used method and her comparative methods should be given.

Report why a method was used and when a comparative method was taken into consideration.

Common practices are methods like the QALY'S or also "willingness to pay". If intermediate results are used as final results discuss their trustworthiness.

What is to do when the result of health benefits have been valued details?

Should the result of a cost utility analysis be in healthy years, then QALY's were used to

determine the results. The used data was collected by several programs simultaneously.

Going into detail with either the methods of QALY or with the calculation of the healthy

years, he will stumble over the concepts “Victims” and “loss”. In this context the words

"Standard Gamble" and, "Time trade off" should be familiar. This is a reminder of a possible loss, profit or better state.

Further mention which sources were used to achieve the data, e.g. patient data, members of the general public or health care professionals.

What is with the indirect benefits when there changes productivity?

Should indirect benefits be considered or not? If yes, they should be reported separately and the relevance to the study must be given. Now the method "willingness to pay" should be taken into consideration again. Not only the willingness itself, but the liquidity of the patients too is important. When using monetary-values, regard the different social states. This can be a question with moral aspects, has the willingness of a rich person a bigger priority compared to a poor person? In general, this willingness-to-pay is the main factor, the others are less important.

return

short

---

Modelling

What is a model?
There are several definitions for the term "model". Those definitions have in common, that a model may not be a perfect copy but a simplified representation of reality. The extent of this simplification depends on the needs of those, who want to use the model.

Why Modelling?
In most cases economic data, that were collected in clinical trials, cannot only be used for a fully informed decision. This can also not available or they are not transferable. Often date dervided from different databases. Thereby is the modelling of target size an essential component of evaluation. In addition, modelling techniques enable an evaluation to expand beyond what was observed.

What demands should be meet on a modelling?
The main requirement is that the model is explicit and clear. To ensure this, the following factors are essential:

• full transparency with clearly described and well-founded model inputs and assumptions

• sufficient depth for adequate representation of the modeled diseases with its associated cost and the corresponding measures

• flexibility to calculate multiple scenarios with varying constellations and adoption settings

• possibility of determining the uncertainty in the predicted cost

• use of relevant data

Further are the following points necessary:

• The progress of clinical outcomes should be extrapolated about that what was observed in a trial.

• Final outcomes should be transformed from intermediate measures.

• The relation between inputs and outputs should be examined in a production function model to estimate or apportion resource use.

• The evidence from trials should be used to reflect what might happen in a different clinical setting or population.

Some types of models:

decision trees

discrete event simulation

Markov model

transmission model

agent based model

Monte-Carlo-Simulation

epidemiological model

return

short

---

Forms of Evaluation

Different forms of economic evaluation

Study type	Measurement of benefits	Question posed
Cost minimisation analysis	Benefits found to be equivalent	Which is the most efficient way of achieving a given goal (or objective)?
Cost effectiveness analysis	Natural units (eg life years gained)	What is the most efficient way of spending a given budget?
Cost-utility analysis	Healthy years (eg quality adjusted life years, healthy years equivalents)	See cost effectiveness analysis
Cost-benefit analysis	Monetary terms	Should a given goal (or objective) be pursued to a greater or lesser extent?

---

return

short

Handling Uncertainty

Uncertainty is a side effect if a sample is used. The intention is not to negate the effect, that is not possible. The intention is to scan the effect. To figure out the influence of uncertainty on the results a sensitivity analysis should take place. By varying variables a range of results is shown. Depend on the target figures it could be continuativ to choose variables uncertain in the way that they are extreme values (e.g. the highest possible price, the longest time to wait).

One way to approach uncertain data (e.g. about long term effects of a therapy) is working with estimations for the future.(best backed up or confirmed by follow-up-studies)

At the end it is essential to report about uncertainty for handling it. As it is not possibble to cancel the effect, the effect should be reported well.

return

short

---

Discounting Data

Should the data be discounted?

In Germany discounting has a fixed rate of 5%. It is necessary to add the results of this discounting rate in german publications, whether you actually work with it or not.

Provide sensitivity analyses with the discount rates of 0%, 3%, 7% and 10% additionally. The ratio which is used mainly should be justified, e.g. using a 0% discount rate to avoid falsification of results by preventive methods is the most common reason.

What else is there to discounting?

Further adding the undiscounted and inflation-cleared data is advisable, so that a reader has the possibilty to recalculate the results with other discount rates.

discounting of non-monetary values is common practice nowadays and meant to display further influences like rising income, age and pure time preference on the relative value of current and future health states.

Should the perspective be given, out of which is discounted?

The perspective out of which is discounted should be given so the reader has the possibility to judge the arguments in a more objective way.

What about the influences of inflation?

If inflationary data is used, mention why, the inflation rate itself should be taken from either 'Bundesgesundheitsministerium -> Statistik zu KV -> Finanzierung', , or at least from the 'Statistisches Bundesamt' .

return

short