Startup Investing: Throw a snowball

This article is how I came up with a medtech startup idea, which is based on my previous article about filtering ideas Startup Investing: How to build a snowball.

How I got an idea

Half a year ago I read the high interesting book "Risk Savvy" by Gigerenzer. The book is about decision making under uncertainty and risk. A couple of chapters are about medical decision making for doctors and patients. Basically Gigerenzer promotes the idea of using natural frequencies and fact boxes to communicate risks of a therapy, drugs and illnesses.

Natural Frequencies describe probabilities as absolute numbers. Saying "5 out of 100 people get ill" is better than saying "5% of people get ill". The main advantage of saying "5 out of 100" is, that its easier to visualize. A second advantage is that it explicitly says on what the probability refers to and prohibits evil relative probability.

Relative probability is mathematically correct, but really evil, if not communicated correctly. 'Newer pills [...] double the chances of blood clots' is evil usage of relative probabilities. The same relative probability ('double the chances') expressed as natural frequencies is the following. First generation (contraception) pills cause blood clots in 12 out of 10,000* women. Second generation of (contraception) pills cause blood clots in 22 out of 10,000* women. The probability of having blood clots is 0.12%* (1st gen) to 0.22%* (2nd gen) or an absolute increase of 0.1%* or in relative probabilities 83%*. You see, just saying 'double the chances of blood clots' misrepresents the real information! That is why relative probabilities are evil or at least have to be used very, very cautiously.

On the sad side, most journalist are not cautious and misrepresent this life-depending fact wrong! Because of this article above (and similar others), many women did not use the pill any more. Thereby getting pregnant and causing more(!) abortions. As abortions are way more life threatening than using the 2nd gen pill, it is fact that many more women suffered physically and psychologically or even died because of this misrepresentation of facts. A chill went up my spine: because information about 'chances of blood clots' got misrepresented, many women suffered and some even died! Got anyone arrested because of not doing their journalism job? NO!

Fact boxes are a way to represent (medical) information in a transparent and compact way [1]. The following fact box represents the pill case* above:

                                    1st gen pill       2nd gen pill
Women treated               10,000              10,000
Blood clots occured              12                     22

The comparison between 1st and 2nd generation pill represented as fact box is obvious and easily understood. Lets make a Keynes twist to the fact box ("It is better to be roughly right than precisely wrong."), because 10,000 is a large, hardly countable and imaginable number:

                                  1st gen pill       2nd gen pill
Women treated              1,000                   1,000
Blood clots occured             1                          2

Now its obvious that the risk increase of taking the 2nd gen pill is there but in absolute terms small. In other words 'double the chances of blood clots' was transformed into 'an increase from 1 to 2 cases of blood clots per 1,000 women treated'. I do not think masses of women will stop taking their pill, risking an abortion, based on this representation of facts.

MedTech Startup

The idea of using fact boxes to represent medical information is in my opinion a very valuable service provided to society. Right now I am in the third step in my process and I have to do small capital investments (server) and large time investments (website). Here is my first website in German: diagnose.orolle.net

My primary competitor is informulary., which already provides fact boxes to the American market (FDA). Its founded by the inventors of the fact box. My target market is Europe. In Europe the drug admission, languages and health care systems is quite different, so that American and European markets are quiet distinct from each other. I do not think head-to-head competition is likely in short- and mid-term even if the product is exactly the same.

At the moment the business model is quiet straight forward. Create a fact box once and sell the fact box multiple times. The cost structure of creating the fact box is primarily manual labour (time-depended) of a specialized pharma-graduate. The online selling per unit has near 0-costs. The drug admission documents are free of cost, while access to medical journal publications have to be paid for (in some European Countries medical publications are open and free of charge).

The customers might be patients and health-insurer (consumer) or doctors, hospitals and health-insurer (professional). Maybe some subscription model is more favourable than pay per download? Maybe some customization and reselling model is also more interesting to my customers? Maybe some kind of reviewer & education services can be provided to consumers and professionals? The acceptance of the user interface and service has to be verified, so that it is easily accepted by my customers. All these questions have to be tested in step 3 and is yet to be done.

The primary market barrier is a trusted customer relationships. After a trusted and proven relationship is established, the switch barrier is relatively high, as the cost on my customers P&L statement is relativity small compared to other cost positions and the use and feel of my service should create some switching cost which have to be overcome by my competitor. In my opinion there are only a few paying customers: doctors & doctor's office, hospitals and health insurers. In my opinion going after the health insurer first should be most promising.

Sources & Comments
[1] Schwartz, Lisa M., Steven Woloshin, and H. Gilbert Welch. "Using a drug facts box to communicate drug benefits and harms: two randomized trials." Annals of Internal Medicine 150.8 (2009): 516-527

* Estimated numbers based on: http://www.dailymail.co.uk/health/article-3098335/Newer-types-Pill-taken-million-women-double-chance-blood-clots-compared-older-brands-contraceptive.html