Hacker News: dwelch

New comment by dwelch in "The madness of reduced medical diagnostics"

dwelch — Fri, 20 May 2022 13:39:51 +0000

Agreed. I think the article and parent comment seem to be from the perspective of someone who has never had to, say, read or interpret a chest CT. Or had to tell someone that they have an indeterminate module that may be cancer but may be nothing.

The article also seems to imply that their is one entity or person who controls all aspects of diagnosis (hence the meme at the top), from taking a history and doing a physical exam to perfectly understanding the intricacies of all additional diagnostic testing. Of course, your primary care doctor may not understand exactly what percentage of people in a low (or high) risk cohort with this lung nodule with certain specific imaging features turn out to have cancer. Hopefully the radiologist does—but they didn’t talk with the patient to assess their risk tolerance and data-collection preferences before saying “consider further evaluation with tissue sampling” in that report.

At the end of the day, respecting patient autonomy matters. It sounds like whoever wrote this article would lean towards asking their doctor to collect as much information as possible. Others wouldn’t. In a perfect world, the patient will be well informed by their doctor, be well-counseled on the risks and benefits of a test / intervention, and will be capable of understanding things like statistics well enough to come to an informed decision that works for them. In reality, that often does not happen for a number of reasons.

New comment by dwelch in "California tells coronavirus-positive medical workers to stay on the job"

dwelch — Thu, 13 Jan 2022 13:18:51 +0000

I can count on two hands the number of nurses at my one hospital who were fired (or forced to resign, or whatever corporate language you’d like to put on it) last year because they refused to get vaccinated and had recently had COVID within the past few months.

New comment by dwelch in "“Please don't add any of my stuff to this project”"

dwelch — Mon, 14 Jun 2021 17:50:14 +0000

I agree. This is kind of bizarre behavior from someone who is such an important contributor to Home Assistant, a large scale open source platform. I can’t imagine most of the leaders in Home Assistant would share this attitude, but maybe I’m wrong.

New comment by dwelch in "Portable MRI promises to provide immediate diagnosis in virtually any setting"

dwelch — Wed, 09 Sep 2020 20:09:59 +0000

Imaging in general has an important role in screening for certain diseases, and in some cases is already more important than physical examination. This mainly depends on how "well" physical examinations can detect an illness, the characteristics of a particular imaging study (cost, radiation exposure, etc), and the characteristics of a disease (prevalence in a population, potential benefit of early detection/treatment). So in short, for MRI to replace physical exam and/or lab testing in screening for a particular disease, the cost needs to come down or the benefit should outweigh the cost.

Prostate cancer screening in general is no longer officially recommended by the AAFP as a rule [1], but if a patient decides to opt-in to screening, the "assay" (PSA) is preferred over a digital rectal examination, which has been shown to be increasingly useless in the detection of prostate cancer. MRI doesn't currently have a role in prostate cancer screening, but Prostate MRI has an increasingly critical role in diagnosis.

CT colonography is currently a viable option for colon cancer screening instead of colonoscopy, though colonoscopy is still considered the gold standard.

Chest CT is currently recommended to screen for lung cancer in patient's that meet certain criteria (age, smoking history).

Mammography is obviously the workhorse of screening for breast cancer, rather than physical examination. But MRI is recommended as a screening mechanism for certain patient populations with very high risk of breast cancer (e.g. certain BRCA mutations). The small population, high prevalence, and benefit of early detection combine to make this a viable option.

[1] https://www.aafp.org/family-physician/patient-care/clinical-...

New comment by dwelch in "Breast cancer detection in mammography using deep learning approach"

dwelch — Tue, 31 Dec 2019 01:52:12 +0000

While I think I know what you’re getting at, there are of course examples of “detecting pathology early and then treating it” preventing morbidity, mortality, and saving money. This is the basis of all (good) public health cancer screening practices. Think of a screening colonoscopy identifying a local pre-invasive adenocarcinoma and cutting it out (cure).

New comment by dwelch in "Don’t buy the MacBook Pros even on sale, in my opinion"

dwelch — Wed, 25 Apr 2018 20:29:41 +0000

I do clean installs on my Late 2013 13" MBP annually with each major OS update, and my performance is still very good.

New comment by dwelch in "Launch HN: Medumo (YC W18) prepares patients for surgeries"

dwelch — Fri, 09 Feb 2018 03:16:57 +0000

As a medical student with an interest in software, this is great to see. A great solution to a costly problem that is likely underestimated by most. As I spend more time "on the ground" in the healthcare system, I see more of these practical problems that have very real consequences, with outdated pseudo-solutions. I'm excited by the prospect of emerging solutions like this.

New comment by dwelch in "As opioid overdoses exact a higher price, communities ponder who should be saved"

dwelch — Sun, 16 Jul 2017 15:22:00 +0000

Furthermore, I'd imagine the cost of the naloxone alone is a small percentage of the total cost of the emergency responders + acute care given in life-threatening overdose. The article doesn't discuss a recurring theme in medicine: acute care is really, really expensive. It'd be far more cost-effective to reduce the need for acute care by providing access to addiction treatment before things get to this point. Easier said than done I suppose.