Data Network Effects in Healthcare
This article describes the application of network effect theory to the healthcare field. Healthcare startups that have just recently formed often struggle to obtain the amount of data that they need in order to feed machine learning algorithms with the goal of improving patient outcomes. One example of this is Mediktor, a symptom-checker startup whose product was built on an algorithm that analyzed patient symptoms and recorded outcomes. The accuracy of the symptom checker improved with each new user that input data because the algorithm had more data in order to make itself “smarter.”
The article defines a network effect cycle that results in more an more people using the healthcare product as more users join: as users join, more data is collected, allowing algorithms to get smarter, which in turn results in a better product that even more people want to use. Thus, the more people who use the product, the more appealing the product becomes. This is the exact definition of a network effect.
It is likely that healthcare startups need to get past an initial threshold of users in order for their product to be successful. We found in class that if the “lower equilibrium point” is surpassed, then the amount of users will converge to a much higher equilibrium, corresponding to a huge success for the company. However, if this point is not reached, the amount of users will eventually converge to zero because the product would not be worth enough for people to want to continue using it. The article claims that startups often use public databases in order to get over this initial “hump” and teach their algorithms enough that the product becomes useful to many new customers. It would be interesting to study the exact locations of the lower equilibrium points for different companies and try to correlate a company’s success with whether or not this initial equilibrium point was surpassed.