Live data from Hacker News

What would a “good” WebMD look like?

blog.tjcx.me

111–120 of 323 posts

Re: What would a “good” WebMD look like?

#111
post #15

A better question might be: Why is the private sector responsible for providing accurate health information? As this article shows, the incentives for people running medical websites and the people reading them are not aligned. I'd say the UK NHS website and symptoms/medications pages hit the nail on the head - https://www.nhs.uk/conditions/ . It has no advertising and provides short, easily readable and actionable i…

That's a nice page, and I'll definitely bookmark that one. But I don't see most of the public using something like that over WebMD. WebMD is complete trash (I even block it in my Kagi search preferences), but it succeeds over sites like the NHS one because of blog SEO. WebMD is really a glorified blog with posts about every condition you might get paranoid over. I don't know how precisely they achieve this, but they'…

I wonder if UK users see the NHS site further up the rankings?

I'd be curious to know if people are putting in 'NHS' as a search term as well.

Re: What would a “good” WebMD look like?

#112
I worked for a company that tried to do exactly what this article is proposing, I was responsible for parsing the data from publications into exactly this sort of table.

The primary reason is that it is very hard to come up with a schema that even 5% of papers would adhere to. The vast majority of this knowledge is phrased as natural language.

There are databases that track compounds and the publications related to them, but those papers again are natural language and cannot be readily converted to tabular data. Our first basic approach involved POS tagging and then trying to associate proper nouns with numeric values. Again the issue became how do you interpret a sentence like "may lead to sudden death" as a symptom? Something like "may lead to symptom X in Y% of respondents" is a nightmare to consistently parse without heavy ML running over huge datasets of just text.

In the end we wound up having to shut down concluding that until papers are released with not only arbitrary XML tables/results we were not equipped to handle the task. And even worse, what if our models didn't interpret things correctly and a consumer got {symptom:"sudden death", chance:0%} and insisted on that compound for their indication only to later realize the paper stated "in lab setting 0% of animals didn't experience sudden death after being administered X after diabetes diagnosis". Paying a hundred students to work around the clock couldn't get the volume processed accurately for months, let alone getting a second army of validators to confirm each entry.

Re: What would a “good” WebMD look like?

#113

Earlier quoted context omitted.

So there's a private option that's terrible (WebMD) and a private option that's terrific (Merck Manuals). And it's the terrible one that ranks well at Google. Sounds like a Google problem and apparently only a Google problem.

That's where I continue to want to see what results in a search engine would look like that heavily punishes presence of advertisements in a result. All the SEO spam pages are ad-driven, so cutting out anything following that incentive should result in removal of all pages that follow that terrible SEO spam pattern that ruins search results.

Punishment/vengeance is a popular idea around here, but you have to also remember that a search engine is supposed to bring you the most relevant results.

Filtering out, say, Stack Overflow or Reddit because it has ads doesn't help you when it answers your question and is perhaps the only thing on the internet that truly does.

People seem to think there's this ad-less replica of the internet, sitting right behind our ad-riddled internet, where everything they want exists for free, it's just hidden. In reality, the websites making money are the ones providing the vast majority of things people are looking for.

Use https://search.marginalia.nu if you want to severely punish ads.

Re: What would a “good” WebMD look like?

#114

Earlier quoted context omitted.

That's a nice page, and I'll definitely bookmark that one. But I don't see most of the public using something like that over WebMD. WebMD is complete trash (I even block it in my Kagi search preferences), but it succeeds over sites like the NHS one because of blog SEO. WebMD is really a glorified blog with posts about every condition you might get paranoid over. I don't know how precisely they achieve this, but they'…

I wonder if UK users see the NHS site further up the rankings? I'd be curious to know if people are putting in 'NHS' as a search term as well.

One anecdote from a UK internet user. I just tried googling

long covid

and the top result was NHS. Not sure if it is for other nationalities but they're definitely getting good ranking in the UK

edit: I also tried flu and meningitis and NHS is top for both of those queries too.

Re: What would a “good” WebMD look like?

#115

Earlier quoted context omitted.

So there's a private option that's terrible (WebMD) and a private option that's terrific (Merck Manuals). And it's the terrible one that ranks well at Google. Sounds like a Google problem and apparently only a Google problem.

That's where I continue to want to see what results in a search engine would look like that heavily punishes presence of advertisements in a result. All the SEO spam pages are ad-driven, so cutting out anything following that incentive should result in removal of all pages that follow that terrible SEO spam pattern that ruins search results.

Imagine a world where the biggest search engine made its money from advertising. In that kind of a world, wouldn’t the search engine primarily be incentivized to show you the results pages with the most advertising, regardless of the quality of the content?

Re: What would a “good” WebMD look like?

#116
post #94

Earlier quoted context omitted.

It does. Google treats gov.uk and nhs.uk sites as authoritative sources and ranks accordingly. We typically pay no consideration to SEO on non-campaign sites. It doesn't stop ads from sneaking in above the results though. There's a whole industry of shady companies 'hijacking' and charging extra for free/cheap government services.

What does "non-campaign" mean in this context?

The sites that sit on gov.uk are usually either:

* transactional services - those that let a person do a task, like applying for a driving licence or registering a trademark

* information services - a collection of guidance on a subject, like maximum working hours or data protection responsibilities

There are some that don't really fall into either category and are there to advertise something. They aim to be informational and inspirational. Some examples:

* https://helpforhouseholds.campaign.gov.uk/

* https://www.apprenticeships.gov.uk/

* https://national-security.campaign.gov.uk/

* https://skillsforlife.campaign.gov.uk/

The reason SEO is more important on these is that there's often competition. For example, if I search for 'teach in uk' the Get Into Teaching campaign site will be competing with lots of teacher training colleges and companies, job sites, etc.

Re: What would a “good” WebMD look like?

#117

Earlier quoted context omitted.

> Sounds like a Google problem and apparently only a Google problem. I want something like webrings to become A Thing again. A user curated search engine. And the users doing the curated need to be vetted. I don't know if this is even possible, but I get tired of having to come to HN to get a human recommendation that is miles better than the algorithmic crap from the current search engines.

> And the users doing the curated need to be vetted. It comes back to the age-old question: Who vets the vetters?

Unsolvable since it's a network of fallible humans we attempt to topologically score.

You can make decent attempts, such as academic peer review. Even this system perpetuates its own problems (beta amyloid) and has perverse incentives (publish or perish, falsified results), though.

Semantic web had some good ideas about this. Networks of signed FOAF data attached to articles and posts. You could form a side graph of trust information that you could revoke at any time.

Re: What would a “good” WebMD look like?

#118
post #15

A better question might be: Why is the private sector responsible for providing accurate health information? As this article shows, the incentives for people running medical websites and the people reading them are not aligned. I'd say the UK NHS website and symptoms/medications pages hit the nail on the head - https://www.nhs.uk/conditions/ . It has no advertising and provides short, easily readable and actionable i…

If the “government” becomes the only source of truth, that source of truth is necessarily political. The oft-mentioned NHS — do they not have an incentive to lower health care expenditures? Would providing medical information to the public that prescribes a suggested course of action incentivize the course of action that is least expensive to the government but not necessarily the best treatment? Is there any potential for a conflict of interest there? Of course there is. If some condition is best treated with an expensive procedure, would the NHS be incentivized to recommend a cheaper, but less effective approach? Of course they would. The NHS routinely denies or delays life-changing surgeries over cost despite such denials not being in the best interests of the patient. (Try getting an NHS knee replacement.) There are treatments NHS won’t recommend because a person is too old. Not because they can’t handle the procedure but because the expense isn’t worth the amount of lifespan remaining in the patient.

Medical information provided by governments is often centered around what’s best for “populations” and not necessarily the individual. Covid is a prime example: masking kids was bad for the kids but allegedly “good” for the population. Damaging to the individual for the alleged benefit of the so-called greater good. Despite all the kid masking “science” turning out to be garbage. Covid vax for kids is another example: the government right now in the US promotes vaccinating even toddlers despite their individual risk of serious illness being statistically zero.

HIV is also something that the government really got it wrong in the 1980s. https://www.aier.org/article/fauci-was-duplicitous-on-the-ai...

Covid information is another example: the government was wrong on multiple occasions throughout the pandemic. Yet when “official sources” are the only approved source of information, the public gets mislead. The marketplace of ideas is a real idea and it should be embraced. We need more private sector competition for information, not less.

If WebMD is bad, the solution isn’t government, it’s a competitor.

Re: What would a “good” WebMD look like?

#119
post #15

A better question might be: Why is the private sector responsible for providing accurate health information? As this article shows, the incentives for people running medical websites and the people reading them are not aligned. I'd say the UK NHS website and symptoms/medications pages hit the nail on the head - https://www.nhs.uk/conditions/ . It has no advertising and provides short, easily readable and actionable i…

Check out the National Library of Medicine’s MedlinePlus resource for an equivalent to the NHS’s consumer-oriented health info portal: https://medlineplus.gov/

It’s a fabulous and under-utilized resource!

Re: What would a “good” WebMD look like?

#120

I worked for a company that tried to do exactly what this article is proposing, I was responsible for parsing the data from publications into exactly this sort of table. The primary reason is that it is very hard to come up with a schema that even 5% of papers would adhere to. The vast majority of this knowledge is phrased as natural language. There are databases that track compounds and the publications related to t…

In the end the only functional parts that worked algorithmicly are exactly those featured in the GlacierMD demo. What trials are running related to this indication, what compounds are being tested for the indication and what other indications are related.

That's the easy part, it's effectively a word association game, TF-IDF did this job admirably, scoring proper nouns by their uniqueness and then associating them with one another and searching for publications with similar words as the requested indication. Effectively a medical word cloud for each indication and compound. Parsing them into symptoms is the first nightmare, the second is numeric values associated with those symptoms and paper results.

There is a very good reason the demo only has one indication and a handful of symptoms, it's being done manually and then at best showing publications related to the words encountered.

It's not a matter of cost, although the author is all but doomed if they want to cover more than a few indications, it's a matter of not forcing publicly funded health publications to use an electronically parseable Format despite the simplicity of them being able to parse their paper by definition.

See the standards XKCD, the issue is getting many different academics and departments to agree on a set of schema to include alongside their publications. PubMed at least tries with their XML dumps but even those are inconsistent at best and non-syntactically interpretable at worst. The Japanese compound tracker is great to learn about a specific compound and their indications but stops there.

Post reply on HN