Craig Monsen

150 posts

Craig Monsen

Craig Monsen

@cmonsen

MD and CMIO at Atrius Health. Mostly read/tweet thoughts on #valuebasedcare, #providersatisfaction, #digitalhealth, #interoperability.

Katılım Nisan 2009
181 Takip Edilen126 Takipçiler
Craig Monsen retweetledi
The CARIN Alliance
The CARIN Alliance@carinalliance·
While we appreciate and applaud the @linakhanFTC @FTC statement today that reminds health apps they are subject to the @FTC Breach Notification Rule (ftc.gov/system/files/d…), it's important to note we have already gone MUCH farther to protect health data outside of HIPAA. 🧵
English
1
3
5
0
Michael L. Barnett
Michael L. Barnett@ml_barnett·
Sometimes I remember the tiny printer that spit out EKG strips for all of the telemetry events on the cardiology unit. It was horribly inaccurate, always printing out useless stuff. To handle this, there was a hole in the table so the printer strips went directly into the trash.
English
2
3
26
0
Craig Monsen
Craig Monsen@cmonsen·
Seems like #aducanumab is “fake it til you make it” in drug dev. On the one hand, it seems like a long shot that this will help patients today measurably. On the other hand, does it help Alzheimer’s research cross a chasm out of limited public r&d funds? nejm.org/doi/full/10.10…
English
0
0
0
0
Craig Monsen retweetledi
Isaac Kohane
Isaac Kohane@zakkohane·
How robust will ur favorite AI application be if ICD coding shifts, if reimbursement patterns change care, if response to a dangerous virus changes, or new drug introduced? Just a few examples of data set shifts that doctors muddle through daily that systematically upend ML apps.
Sam Finlayson@IAmSamFin

Letter: "The Clinician and Dataset Shift in Artificial Intelligence" out in @NEJM w/ @_asubbaswamy , @suchisaria, @kdpsinghlab, @zittrain, @zakkohane, et al Link: nejm.org/doi/full/10.10… Full version of Table (w/ details+citations): sgfin.github.io/assets/dataset… Brief thread: 1/n

English
1
7
36
0
Craig Monsen retweetledi
Ben Moscovitch
Ben Moscovitch@benmoscovitch·
The Physician Fee Schedule proposed rule is out❗️ A key change: public health measures in Promoting Interoperability CMS proposes making electronic case reporting & immunization information system use required (both have been critical in COVID response) public-inspection.federalregister.gov/2021-14973.pdf
Ben Moscovitch tweet mediaBen Moscovitch tweet mediaBen Moscovitch tweet media
English
1
6
17
0
Craig Monsen
Craig Monsen@cmonsen·
Looking good for healthcare prices this year. Bounceback inflation of pandemic affected businesses may mae healthcare start to look affordable. 👇
Council of Economic Advisers Archived@WHCEA46Archive

Inflation as measured by CPI increased at a 5.4% rate year-over-year last month and 0.9% month-over-month. Core inflation—without food/energy—rose 4.5% year-over-year and 0.9% month-over-month. A large part of the increase is due to cars and pandemic-affected services. 1/

English
0
0
1
0
Craig Monsen
Craig Monsen@cmonsen·
@dp_oneill Good points. My hypothesis would be that though death is a relatively infrequent cause of attrition it drives the MLR down subsequently owing to high cost of end of life care. Survival bias I think would overwhelm the marginal cost associated with the entire cohort aging 1 year.
English
1
0
1
0
Dan O'Neill
Dan O'Neill@dp_oneill·
@cmonsen Not sure we can assume that mortality is the main driver of attrition, nor that it produces a healthier cohort over time. Also seems plausible that attrition is primarily rooted in plan/provider switching, and surviving cohort members also get older over the period, no?
English
1
0
4
0
Craig Monsen
Craig Monsen@cmonsen·
Am I the only one uncertain about interpreting an MLR for a cohort without describing the methodology for handling right censoring? Deaths lead to a more healthy remaining cohort over time. (One Medical S4 for Iora) #page89" target="_blank" rel="nofollow noopener">investor.onemedical.com/static-files/f…)
Craig Monsen tweet media
English
2
0
1
0
Craig Monsen
Craig Monsen@cmonsen·
Still a little unclear what’s under the hood here, but seems as though if this were nothing more than an “easy button” for new players participating in Carequality framework-based data exchange, well, I think that would be a great addition to the marketplace.
English
0
0
1
0
Karandeep Singh
Karandeep Singh@kdpsinghlab·
ML researcher: One of the biggest limitations of linear models is that you can *only* use them to learn linear relationships. Statistician: I have the same problem with linear algebra. ML researcher: 👀 🤔 Statistician: 👁 👁 ML: I’m wrong, aren’t I… Stats: Yes.
English
7
18
207
0
S. Trent Rosenbloom
S. Trent Rosenbloom@trentrosenbloom·
Re: preferences about Immediate online result release versus waiting until your provider can discuss: it’s Friday and your biopsy result shows cancer, but the doc won’t be able to review with you until Tuesday. You don’t know the result. How do you want to get your diagnosis?
Wisconsin, USA 🇺🇸 English
6
5
5
0
Craig Monsen
Craig Monsen@cmonsen·
@GauravSingalMD Before concluding that, may want to apply some adjustment for feedback received on platform that is popularized on timeliness (ie twitter). I think fair to describe RW feedback has been more mixed for certain results enriched for “could be cancer” (eg mammo).
English
0
0
2
0
Gaurav Singal, MD
Gaurav Singal, MD@GauravSingalMD·
This thread has really opened my eyes. Almost every doc in this thread wishes they could break bad news to the pt personally (with genuine pt interest at heart). Almost every pt explains docs aren’t as good at that as we think we are, and they’d rather just get the results.
Justin Maykel@justinmaykel

Called a patient today to tell them their colonoscopy biopsies showed cancer. Patient already knew—got an automated alert earlier in the day and read the results on the online medical record. At home, alone. This new law granting access to test results is a BIG problem.

English
5
8
52
0
Michael L. Barnett
Michael L. Barnett@ml_barnett·
I've had multiple patient inquiries directly citing this thread. The evidence is promising BUT - We can't practice medicine for unapproved approaches via tweet thread! Given your expertise, many take your word seriously. IMO a direct clinical recommendation is stepping too far.
Dr. Angela Rasmussen@angie_rasmussen

If you are in the US and got J&J, I encourage you to talk to your provider about whether this is something you should think about. If you live in a community with overall low vaccination I’d suggest you strongly consider doing so.

English
4
0
10
0

Keşfet

@FTC @pewtrusts @ONC_HealthIT @ml_barnett @dp_oneill @kdpsinghlab @trentrosenbloom @GauravSingalMD