Wow.

That's pretty amazing. You're accusing me of active malice for explaining the (entirely reasonable) thought process behind the government's legal system for regulation that pharma must comply with to operate.

I think when you move from the realm of anecdote to dealing with statistics, products with 100 million users, and historical cases (Therac-25), you gain a perspective that regulation is probably the simplest and most effective mechanism for keeping deaths at a minimum while allowing new products to reach the market.

You're actively ignoring the point and just being offended which is natural because it was put extremely bluntly. But you've just restated that the current state of affairs is "entirely reasonable" when the comment you are replying to explains why it's actively harmful. You've provided no argument for why it's reasonable, just taken offense and waved around cases of catastrophic failure.

I don't understand what's unreasonble about:

1) companies have a profit motive and they want to design and sell products

2) historically, we have seen that companies acting with profit motives often have incentives to ignore problems with their products and even actively suppress them

3) the government wants to incentivize companies to make new products that help people

4) but they also don't want large numbers of people to die, or even small numbers of people, due to errors of any kind.

5) we have ample evidence that when unregulated, errors occur and people die. Statistically across large numbers of drugs and medical treatments, we have enough data to make strong, but not perfect statements about which processes minimize negative problems while maximizing availability of product to sick people. In some cases, people die because executives explicitly chose to cut corners. With regulation, we can catch those executives and put them in jail.

6) in general, the one thing that comes out of all of this is that extensive record keeping is required for investigators to deal with problems when they occur. Most of regulation is just record keeping- what did we do, what happened when we did it- and change control. Literally, my job was to ensure that firmware updates didn't change the observable properties of a system that was making a drug that saved people. Yes, when a firmware of a machine that makes drugs is updated, and it has a bug, people can die.

7) given about a hundred years of experience in this area, enough people who work on "reliability" (like SRE in tech) people have developed a set of techniques, called regulation, which attempts to minimize the number of deaths, while still allowing products to reach the market. Like SRE, it's a combination of hard-won experience and a lot of intelligence, along with not enough monitoring data.

8) I also feel pain and unhappiness that pharma doesn't move faster. It wasn't until I worked daily on product development and implementing regulation that I began to appreciate that this is a truly complicated problem with a lot of variables, incentives and players.

9) Even giving my best good faith effort, I simply cannot credit people who give anecdotal evidence about friends with diabetes. There are a number of reasons for this, but basically, if you do enough large-scale decision making, you'll find that anecdotes introduce bias that can lead to poor decision making. I see problems, also, in the DIY community where people sort of assume the best possible results; in science, after you run enough experiments, you get to see: it's easy to fool yourself and think what you've done is good, but it really has to percolate through the system and get evaluated and reproduced by many people before you can trust it.

If you don't think the above logic chain is reasonable, that's OK. Many people aren't happy with how slowly pharma and medical DX industries and FDA work. But, you should basically understand that the entire goal here is to ensure good products make it to the market without "too many unnecessary deaths".

The worst part of all of this is that I dedicated many years of my life to improving the speed at which drugs could be discovered and brought to patients, only to learn that the primary reasons we can't, are political, structural, systemic, and incentivized. If somebody had a better solution than what the current market has, I can point funding at them.

In some ways, I think the FDA is a victim of its own success. If I'm not mistaken, it was created in part because of the 100+ deaths around Elixir sulfanilamide. But since the FDA has helped prevent things like that from recurring, we underestimate the potential dangers of under-regulating.

Not that the FDA couldn't be more efficient (possibly considerably more so), but it's clear that most products that go into production work, and that can't be discounted.

https://en.wikipedia.org/wiki/Elixir_sulfanilamide

Context: I am a type 1 diabetic.

As far as doctors are concerned I am an exemplary patient who has shown blood glucose controls as good as they come (A1C of less than 6.5 for more than a decade now for those that know what it means).

I regularly hear from "concerned" people how they know someone else who's blood control isn't as good and they should do better, etc..., etc...

I have personally used some of these DIY products and so I want to be very clear; the difference between "somewhat poorly controlled blood sugar" and just one really bad mistake in "tightly/ best-controlled blood sugar" is this

1) Poorly controlled - in a decade I will start to have some additional health problems (vision impairment/neuropathy/etc)

2) Tightly controlled w/dosing mistake - Coma/Brain Dead/Dead

There is a good reason to test this equipment and take a conservative approach.

Physical activity level and how much active or "on board" insulin you need when you are being active vs inactive can vary by 10x. As an example of this; before I do cardio workouts I intentionally do a carb load and let my blood sugar raise before starting to counter the effect the exercise will have on my blood sugar. Despite this on two occasions during running I have experienced double drops (super fast blood sugar drop to unsafe levels) to levels so low I started to black out.

Nightscout/OpenAps/Dexcom/Medtronic, it doesn't matter open source or closed source none of these devices can handle these scenarios so at the end of the day it is still on me to intervene and make decisions based on what I am going to do. Until something better than pump systems come along (and these "artificial pancreas" systems as they are calling them to day are still just pump systems; I will have to intervene no matter how sophisticated the algorithms.

In summary; just because you care about someone with diabetes, do not presume you know what is best for us.

OpenAPS doesn't mean you never have to intervene, nor does it claim to. It does mean that the interventions needed are less frequent and more predictable. Ie, in the case of exercise--if it's only a small drop, shutting off basal is enough to counteract it, so it'll do that, and you don't have to do anything. If it's a larger drop, you need to eat carbs to counter it--and it'll send you a notification telling you how many carbs you need to eat, and this notification will typically come quite a bit earlier than the low-glucose alarm the CGMS will eventually send.

Yep; I'm aware of all of that.

I'm glad that the open source community is getting involved and I want to see these things help move the industry forward.

My comments were to the parent who presumes

1. To know what is best for diabetics 2. The device is inherently less risky than current options

Nightscout absolutely would not get through FDA approvals today if for nothing other than catastrophic failures. I love the features and think it's way more robust than what is provided by Dexcom, but at the end of the day, I switched back to Dexcom's receiver and app because it doesn't crash.

OpenAPS is great, but requires a level of research and effort than many patients aren't capable of. Then on top of this OpenAPS has this nice disclaimer as an opener, "By proceeding to use these tools or any openaps repositories, you agree to abide by the copyright agreement and release any contributors from any liability."

Now I understand why that statement is there, but when someone dies using this "product" that hasn't been approved by the FDA; well let's just say I hope all parties involved have really good umbrella policies, because I'm not too sure that's going to hold up in a court of law.