[COGlory]

>“We definitely do see a slight change compared to the total population, we do see a slight pullback in overall basket,” John Furner, the chief executive officer of Walmart’s sprawling US operation, said in an interview Wednesday. “Just less units, slightly less calories.”

This implies they're looking at how people on GLP-1 analogs are behaving relative to the general population....but how do they know who is on the drugs?

[kevinmchugh]

They do have pharmacies. I'm sure they have a lot of people who do all their shopping there.

[dymk]

That’s a big hipaa violation.

[mc32]

No, but it's probably a HIPAA violation.

[kevinmchugh]

They also probably have a pretty good model of what weight loss looks like absent semaglutide. The semaglutide patients probably have a different set of purchasing behaviors from dieters and LAP band surgery patients.

[whycome]

Not if just looking at correlation.

[dymk]

Incorrect

[voisin]

You can’t anonymize the data first?

[dymk]

Nope, it’s not like A/B testing a button color on an ad; it’s medical history

[alistairSH]

It’s HIPAA.

[curiousllama]

Could totally be creepy. But could also totally be a survey. Those are still probably pretty cheap to run on something like Walmart+ delivery customers.

[electric_mayhem]

You don’t think Walmart’s pharmacy data can be correlated to the same customer’s shopping habits?

[dymk]

No, it can’t

[I_Am_Nous]

>debit card ending in 6969 purchased Ozempic at pharmacy >debit card ending in 6969 purchased $200 of food, 5% less than they did last month

They don't have to know anything about the specific person to correlate data they likely already have in their POS system.

[dymk]

HIPAA violation

[deadlydose]

No, I don't think it would be. The last four digits of a CC purchase at a pharmacy won't contain PHI. It will just say "pharmacy." But if you have some data to back that claim up, I'd love to read about it.

[kelipso]

They probably have records for each individual (not just last four digits of CC); I don't think it's hipaa violation as long as the records don't contain personally identifying information.

[electric_mayhem]

Come on now.

Can’t? Really?

Maaaybe it’s against some law or privacy policy or mandatory annual training.

But do you honestly believe companies follow laws and policies if they think they can get away with not?

And even if you can ignore that corporations are regularly -publicly- wrist-slapped for failings in those areas and still believe they are virtuous, privacy-respecting, law-abiding entities (rofl) … are you ready to argue that no executive or other employee ever, (knowingly or unknowingly) uses data to run a calculation or check a theory against published policy?

The only thing that surprises me about the above scenarios is there’s a human alive who would believe their improbabl3a let alone, as “can’t” would imply, impossible.

[dymk]

I don't think you've worked in the medical industry, or you'd know just how big of a deal HIPAA is.

[Judgmentality]

I've never worked in the medical industry but I know many people who have, who basically told me HIPAA violations are extremely common and only enforced for a fraction of violations that actually occur. My ex used to work in medical insurance (for a very, very big company) and estimated that maybe 3% of HIPAA violations are actually enforced. I used to think HIPAA was a huge deal until she told me story after story of violations that were ignored.

I think HIPAA is the sort of thing where if you hear about it then it's taken seriously, but the overwhelming number of violations are just ignored and you never hear about them. I'd like to be wrong but unfortunately that's the information I've been fed by people more knowledgeable than me.

[Judgmentality]

Walmart also has a pharmacy. I'm not an expert on the privacy laws here, but I'd be surprised if they couldn't use that information (even if it's a legally grey area).

[philip1209]

They can probably reasonably infer income, and use that as proxy when doing cohort analysis.

[dymk]

Surveys