Common pitfalls of digital health startups and how to avoid them (2023) | Hacker News Reader

Common pitfalls of digital health startups and how to avoid them (2023)

bettychang.xyz

76 pointsgone3510 comments

Wrote Hacking Healthcare for O'Reilly, former hospital executive, built and operated hundred of hospitals and multi-practice groups. Created the open source ClearHealth/HealthCloud EMR and WebVista.

I advise a lot of health related startups. I think the list is OK but the two most important things I have seen come up over and over again are 'have a show pony' and KISS.

"Have a show pony" means you need to have a doctor associated with your business, idea, etc that is high profile and a public advocate for it. Doesn't necessarily mean as an employee or consultant, etc, that can get complex because of corporate practice of medicine regulations but advisory boards and research partners are often good fits. This lets you be taken seriously by what are extremely risk averse institutions in most cases, especially when dealing with "two guys and a hard drive" startups.

KISS is the old adage, keep it simple stupid, but really in healthcare the overwhelming high priority problems are insanely simple people issues. Getting people to wash their hands, getting patients to show up at the right time and the right place, having an organized and non-duplicative flow of operating information. A tremendous amount of healthcare is done via spreadsheet, think at that level. Complex solutions can be profitable but almost never endure and can be extremely hard to sell.

I think a lot can be gleaned about the general dysfuction of healthcare as an industry from this previous HN thread: https://news.ycombinator.com/item?id=39186252

They’ve forgotten the classic error of engaging only with client at the management and IT and ignoring clinical requirements until UAT, and a clinician revolt ends in project failure, adverse media coverage, millions of wasted dollars, and ongoing patient harm from unsatisfied requirements and persistent legacy systems.

Also needs an asterisk about not relying on your customer’s BAU resource to deliver projects, particularly those with large integration components, and one about not buying a business for its product as a replacement for yours which is EoL, then immediately trying to push it on your customers as an “upgrade” when you’ve lost half the acquisition’s engineering talent and you don’t understand the product.

One important thing which isn't mentioned in the article is that you need at least 1m EUR/USD to become a company with a medical device. So you better have a plan to to raise such money, spend and recoup it with revenues.

For most hardware devices that's correct. But software as an FDA Class I medical device can be done cheaper than that. Under US regulations, pure software with no hardware component can be considered a "medical device" depending on functionality.

In my past company we looked at EU MDR class 2a certification, we established it would cost around 100k€/a to get and maintain all necessary certifications. Essentially hiring a full-time resource to do all documentation, plus costs due to slower work cycles, plus certification reviews, plus external consulting as needed.

Ultimately we closed shop before we starting going down that route, so I cannot tell if that number was correct.

Can you go into a little detail on the expenses that 1m needs to cover?

Quality Management setup and certification

Clinical study

Technical file / DHF creation

FDA Submission (510k/DeNovo) or CE marking (via Notified Body)

Depends a lot on your product, your target market(s), number of supported languages, number of employees, and most of all: the risk class of the product.

QMS can be around 15-20k

Study: not necessarily needed if you do not need new evidence. Then you can write a literature review, which is A LOT cheaper

TechDoc: yeah, takes time

CE marking: depends a lot on the notified body. Here's a slightly outdated list and apparently, you can pay anything between 15k€ to 500k€: https://openregulatory.com/notified-bodies/

i'm an executive for a large medical device commercialization consultancy. we do it all from pure r&d, development/engineering, software/firmware (including SaMD), quality/regulatory, commercial strategy, preclinical, precision medicine, manufacturing, etc.

what i see most often from "digital health startups" is 1) a severe lack of knowledge of the device development process (generally due to immaturity) 2) under appreciation of the importance of a true regulatory strategy 3) poor product generational planning (boil the ocean).

this generally results in startups not understanding why development costs are what they are.

(2023)

Reply on news.ycombinator.com