Have you already performed health economic analyses? Then you will definitely recognise the following.

To what level is it relevant to include costs in the analysis? For quite a few parameters, it is clear that they can have a decisive influence on your analysis and in your decision-making process. However, sometimes the effort you have to put into looking up data is not reflected in a meaningful impact on the results.

Of course, as a scientist you strive for perfection, that should always be your basic attitude. But as a consultant we also strive for an optimal project related cost/benefit towards our partners. We will advise our clients if there is no point in conducting a survey among doctors if the expected impact of the parameter to be estimated is virtually nil.

A well-known example is the costs related to side effects in oncological treatments. Most treatments generate side effects. A rule of thumb is that only ‘treatment related adverse events’ (TEAEs) of ‘grade 3’ or higher and occurring in at least 5 % of the study population are included in the base case analysis. It goes without saying that this is not always possible. A rare side effect that can lead to very high costs (e.g., amputation and/or revalidation) of course needs to be included in the analysis.

But if a side effect occurs in more than 5 % of your sample and the associated cost is very low (e.g., fatigue), then the likely cost-impact of this parameter is probably so low that there is no point in consulting several experts for this. An estimate may in that case be sufficient. After all, your analysis includes an uncertainty analysis as well.

Additional questions may arise when addressing side effects:

1.  Do you include these in each cycle of your model or as a one-off? In quite a few cases, our advice here is to include the costs in the model once. To this end, you can make an average cost estimation per treatment and include it in your cost structure once.
2.  And at what stage do you include it, at the start, at the end, …? In general, you may want to include it at the start of treatment. By using this approach, the discounting effect has no impact and you calculate at full cost as in ‘real-life’ clinical practice.
3.  But what to do if you need to check with Key Opinion Leaders but don’t have the financial resources to question doctors? Literature can often offer a solution. Of course, as always, be careful to compare to a similar setting. There is little point in using costs from other continents and linking that to a European study. In addition, you need to align the found cost data and correct for inflation and possibly implement a currency conversion. Also note that the data from different sources can vary greatly as the authors probably used different methodologies to arrive at their costs (registers / surveys / estimates / top-down costs / …).
4.  If you do have data from a (European, or even country specific) survey, this clearly is preferable. There however may still be things missing. Well, dare to work with the average cost you obtained by means of your survey as a point of reference. Conversely, of course, a proposal can also be worked out first and then validated. It’s up to you!

As mentioned, the financial impact of an adverse event is normally limited. Dare to be critical and always discuss the options that lie ahead with the team. Our advice is to focus on the main ‘drivers’ of your analysis. You will see that this is rarely side-effect driven (unless, of course, the main advantage your innovation offers is that there are significantly fewer side effects).

Anyway, good luck with your analysis!

If you have any questions after reading this blog, do not hesitate to contact us.

contact

See you soon,

Your Hebias team