gforge

The pipe is one of those things that once you've tried it you never want to go back.

The pipe came to R a few years back and I would argue that it is one of the best features. I admit that data munging is a very stepwise process but I find myself frequently missing it in Ts when I code. It makes the code denser as you use fewer intermediate variables while the code id still easy to debug.

The package.json should be able to actively ignore vulnerability id's. As id's disappear with audits the npm audit could just remove those, eg a "prune"

I love the new tabs, I found the old ones clumsy and ugly. Design is unfortunately something that is constantly a moving target. If Mozilla stops updating they will surely die - for most users the technical details don't matter, what matters is that their friends go, wow what is the browser you're using? And thus, redesign is just a way of staying alive...

We have our own dataset that we're annotating.

There is though a lot of good research coming out of the US but from my understanding they can be rather expensive. Sweden has probably some of the best hip arthroplasty outcomes but the cost of these is roughly 5-10 times lower than in the states. I think the system is somewhat stacked against you, if doctors (also human beings) spend 80-100 hours/week during their residency, they expect a financial reward afterwards or no-one would agree to that.

Orthopedics is also somewhat of an odd medical specialty with lots of incentive issues and a scarcity of good clinical trials. In a study from Sydney they found that among the 50% of the surgeries that had been evaluated with randomized clinical trials (RCT), only in 50% of the cases the trials actually supported the surgery. It is though much more difficult to randomize people to surgery, doing the blue or red pill is much easier.

I'm pretty sure also that a big part of the problem is us doctors failing to identify if a patient is part of the long tail or not, i.e. does he/she not fall into a particular study's inclusion criteria or not? Hopefully we will be able to shortcut this problem with deep learning tools but as most self-driving-car enthusiast know - there is a big difference driving around in the parking lot from the open streets.

I'm an orthopedic surgeon with over two decades of coding experience and the last six years I've been working on developing deep learning models for muskuloskeletal images. Despite being in a rather narrow field, only focusing on trauma radiographs, the complexity of the task is hard to overstate. We have piloted our software with our radiologists at the hospital, but the amount of features you need so that the clinicians actually feel that there is any benefit, are huge. At the moment we're training 480+ different labels and I still think we're not even half way for trauma radiographs.

Yes, especially if you have images or other things except text. It surprisingly bad when you want a constant layout between the two.

While I knew the concepts of log I first truly understood it when I started looking at everything as log2. In hindsight I feel like a moron, obviously it is about doubling and halving risks is what it's about (I'm an MD so everything is a risk ratio in my field).

I remember a colegue that was presenting results during their defense of a thesis where one risk ratio 0,45 and the other 2,1. I asked which effect was biggest and they automatically replied 2,1. I'm pretty sure that 80%+ of my colegues would make the same mistake. Everyone understands double/half - we should try to teach people this as the word log is just too intimidating for so many.

Please stop viewing health care as one single entity. It is equal to saying that all IT is the same. It is a huge beast that quite often lacks margins.

The pipe operator has was introduced to R a few years ago. It is one of those things I miss the most in js.

Debugging is an issue though. I commonly find myself splitting up the pipes into section in order to inspect the results and find exceptions. Breakpoints and stepping through code is probably the features I most often struggle with.

Yeah, I remember some spray version when I was a kid. The 10+ is how long I think it has been present at my particular hospital.

The interesting thing is that it hasn't really caught on. We don't have any financial pressure or other external factors, it is really convenient, yet I'm pretty sure that only 10-20% of the doctors use it.

Probably fine. We often have blisters where there is an intact surface protecting the new skin. The big difference is that an intact blister is sterile by nature while most other wounds will always have some contamination.

Probably, but it is probably not that cost motivated. Also, you quickly learn how to avoid that particular problem.

MD here: perhaps as an alternative to outpatient wounds but we've been using glue for 10+ years now. It works when you have a clean wound but often you need to reconstruct messy wounds and then glue doesn't get the job done. You need to be able to pull the skin and with glue you will just end up gluing your glove onto the wound (yup, personal experience).

In the operating theater I don't really see it happening. We've been discussing using glue there but then you need an additional layer of sutures which usually take longer to do (intracutaneous). Anything that takes time in the threatre is usually a deal breaker.

There are probably use cases outside my field but even for burns I'm uncertain if it is a good fit. Most wounds are not sterile and encapsulating bacteria under a layer of fiber seems like a recipe for pushing the bacteria deeper and generating a more severe infection.

Yes, there have been numerous people pointing out this in this thread. I think that UX design is viewed as frosting on the cake instead of the thing that actually makes it work. Unfortunately I think one of the issues is that the SDLC is broken.

For some reason we keep getting features that few of us actually requested but that often make sense in that they are life saving, that is if people actually used them. I have a ton of indicators flashing all-over the place and after a while your brain simply adapts and ignores them. I doubt that anyone has actually done any analysis to see if these tools actually save lives.

I've been involved in some projects when I've spent 30 minutes with UX people and then never heard back from them. When the final product arrives it is full of annoying flaws, e.g. the latest software I was involved in the work flow is: 1. click on "open software" from the patient chart 2. a new window opens where I manually have to click on "log in" (note, I'm not adding a password) 3. the UI goes into authentication 4. the UI goes back to the login screen 5. the UI asks me a question that I _always_ answer the same thing to 6. I actually get to the data I was interested in

The most annoying part is number 4 - if I accidentally click on "log in" everything breaks and I have to close the window and redo everything from start.

Most of these UX issues fall under the category - murder by a thousand cuts. It is mentally tiring to be faced with all these issues but none of them is like open heart surgery where one mistake can be fatal. What happens is that doctors/nurses become tired and don't put the effort into the patient interaction where the actual value in healthcare is created. I'm pretty sure that bad UX in the end kills many more patients through this than malfunctioning critical hardware.

I truly wish that people in this forum that create health-care software think of UX and actually use their software more than just "I clicked through and everything works". Your bosses will probably never ask for UX features because they have a deadline, but you are in charge of designing and during that process it is easy to fix many of these annoyances (e.g. just disable the log in button).

I am an orthopedic surgeon during the day and at night I write code. I just don't believe that spending a few hours watching surgery actually helps. You need to really work with the software, a solve daily problems and be aware of what problems are trivial software fixes.

For instance, when I sign lab results I get to see the references for twenty year old subjects. It would be much more helpful to see the patients last lab result (in 60+% of the cases this is available).

Another example is in the ER where we have a list of all the patients. There is a column for my name so that I know which cases I am handling, unfortunately that column is too the far right and only visible after resizing the other columns. This setting is also not saved so each time I open the list I have to resize over and over again. If you could move that column to the left would be great, or even just bold font my patients' name.

The major problem that I see is that we have:

- huge monolithic software that make any change a living nightmare

- all changes are treated as we were doing open heart surgery. Most stuff we do is trivial but having a standard that high makes it almost impossible to deploy updates - especially the minor changes that really would change things.

- 80% of users complain about the software but very few understand the underlying problem.

- much of our software is spec driven - I can sign my lab results and I can see the patients in the ER, no one bothered to add the last tweaks to actually make it easy.

Have you considered asking the authors of the typedefs to contribute to that section? My guess they would be pretty honored to be able to help out directly with such a high profile package.

I like types (I've use both Flow and Typescript) in general, but I can also admit that they can cause more type specific coding that is unrelated to the actual problem that I want to solve. It's somewhat of a rabbit hole though, it is hard to go back once you've started. It would therefore be great if there was something more than the test cases provided at flow-typed when trying to to learn redux typing.

I just encountered Flow type issues after the 85 release. I know the type defs are in flow-typed but it would be nice if you could also provide some guidance. There are just so many parameters for connect that it just blows my mind when I try to find the source error.

Cool, odd that I as a Swedish researcher haven't heard of this... Maybe they ment Switzerland - people have major difficulties discerning the two.

Tack, ha en bra dag!

I'm a little late to the game but I would like to point out that this doesn't apply to infected implants and some other special orthopaedic infections where you just as with TB need to treat for a long subclinical period. Otherwise I think it is an excellent article that many of my colleagues should read. I've actually also always wondered where this myth comes from.

That's what I tell my patients, seems to work

A common problem that we have is that there is no such thing as the perfect drug. Any drug that is stronger than placebo will inevitably have side-effects (note that even placebo has side effects). It is always a cost-benefit and this is a tricky thing to fully convey to patients - what does 10% risk of side-effect mean? For the patient that gets the side-effect they usually have it a 100%.

I wish that we doctors were immune to the industry's influence but unfortunately we also sometimes fall prey to the cunning tricks. I think the OxyContin scandal says it all http://theweek.com/articles/541564/how-american-opiate-epide... Mind-blowing that doctors actually believed that there is a free lunch - a morphine product without addiction...

There is hope though, here in Sweden doctors during the 90:s were frequently taken to nice dinners and other events in the name of drug companies representative costs. In early 00 this basically disappeared and today we get a sandwich during lunch break in order to hear about the new thing that they want to sell us. I think the awareness of the "no free lunch" concept (see http://www.nofreelunch.org/) is much higher today than 20 years ago when I started med-school.

One cost that you have to be prepared for is that the system becomes "communistic": if we prescribe any drug that is outside the recommended set of drugs our clinic has to cover the costs. The adoption of new drugs is generally slower (although not for important drugs e.g. cancer treatments) and we are no longer allowed to make local deals with suppliers, everything has to go through centralized deals (even if we can get a better deal). In my mind it's a small price to pay if it keeps our patients safe and we get to spend the money on our health-care where it matters, but I think this system would never pass the American legislative system...

I could not agree more. Our bodies have had millions of years of handling disease without drugs, we should tap into that resource. The problem I face daily is that many patients come to me and say they don't have the time for physiotherapy and want surgery. They often add that their neighbor had great experience with surgery for the exact same illness. It's actually amazing how illnesses cluster...

It is hard to convince people of taking the slow route. This is a similar to convincing people to write unit tests for their code, we understand that it is the right thing to do but we always have a deadline. Some threads here report that they started with physiotherapy after years of other solutions, I would love to know how to change years into months.

There is an inertia but I would say that it is often of benefit to the patients. The problem is that I as a surgeon love to operate (note that we don't even have a monetary incentive) and I truly want to help my patients. This makes me biased and the inertia has saved me quite a few times for starting performing new cool procedures.

The biggest problem is that we often have a big problem with discontinuing treatment s, see my post on clavicle fractures http://gforge.se/2017/08/clavicle-gate/ I am not a cardiologist so I don't know the full details on the statins, but I would be careful not to skip them. Cardiovascular mortality has decreased hugely the last decades and this one of the major treatments - they must be doing something right.

Also be careful of flat-earth arguments, there are plenty of people who want to disprove medicine - some of us are researchers but others are just writing blog posts after finding a single article contradicting a treatment. It is incredibly hard to do medical research, it takes years to recruit patients and collect data. You always think that there is a ton of patients with this condition and then you find that the inclusion and exclusion criteria that you believed to be so brilliant (others had forgotten about them in their studies) actually get in your way of being able to finish the study. If this wasn't enough, once you are done with disproving the treatment, you will find that the person that a few years ago popularized it is no longer doing it... They would never do that, now they're doing treatment B, a completely different thing. It's a game of whack-a-mole.

I agree, there are serious issues with many of the drug treatments and psychiatry has some dark history - but so does orthopaedics (e.g. the pseudotumors associated with hip resurfacing). Even if anti-depressants are no more effective than placebo, I still think a knife may be an even worse alternative and I've seen patients spiraling down into an abyss of pain after repeated failed surgeries. My hope is that if we are able to correctly identify the cause we can start getting somewhere with the treatment.

This said, there are indications that there is a subset of cases where the pain actually is triggered by the spine. The SweSpine register has some interesting numbers that show the results after surgery over time (http://4s.nu/pdf/Report_2014_Swespine_Engl_ver_141204.pdf page 37, DDD). There is also the very fascinating study by Arlbert et al (https://www.ncbi.nlm.nih.gov/pubmed/23404353) that show that antibiotics may help. As always LBP is most likely a mix of diagnoses where we need to identify the correct subgroups before we even start to test treatments.

Lovely article, as an orthopaedic surgeon I can tell you that most of us know this. Treating chronic pain with surgery is generally a bad idea, I spend a large share of my time explaining this. Some patients are grateful for this, but I think that many go and try to find "a real doctor".

An interesting development is that there is generally an increased acceptance for psychiatric diagnoses. More are accepting that just as the pancreas may stop delivering insulin, the brain may fail to uphold the proper seretonin levels. Unfortunately it is not as simple as that, fixing seretonin or any other substance is like pouring oil over an engine and hope that it reaches the target. Still, I'm sure that we will see a huge change as true targeted treatments are no longer just sci-fi. Getting people to the right specialist is until then a good start.