Patients were at a x1.56-x4.30 greater risk of experiencing an #opioid#overdose and had $86.37-$366.06 higher health care expenditures in buprenorphine treatment gap months relative to treated months. https://t.co/d28d1sCNAI #Research
Check out this thread from @BradleyDStein on our recent study! We mined a huge prescription database spanning multiple years to understand dynamics around bup prescribing.
The disheartening punchline is that many start and then rapidly stop.
Sustaining prescribing is vital!
The Call-For-Papers for the next round of @HlthEconSeminar is open! And new for this round of EHEC, we will have some slots open for graduate students to present their own work.
Submit by July 14th at 5pm. And please share with your colleagues/students!
https://t.co/1gQPo5g4Sn
ATE, ATT, ATO... how do you choose? Different methods target different estimands, yielding effects with different interpretations. How do these estimands differ, and which one is right for you?
@Lizstuartdc and I explore that in our new article: https://t.co/Ldn1jXLlbi >>
Thanks for organizing a great session, Ashley & thanks @ana_c_moura for the thoughtful comments on Amanda Abraham, @EmilyCLawler, & I’s paper! In case you missed it, we find NPs are driving increases in buprenorphine availability following CARA in full NP SOP states in Part D.
Come check out my poster at Academy Health regarding the effect of Vitamin D supplementation on COVID-19! It is a large observational study, and has very promising results. I will be around from 12:00-1:00PM to answer any questions. #ARM2021 https://t.co/f5vI5upGWp
🚨🚨Did you miss this amazing talk with Andrew Goodman-Bacon @agoodmanbacon and Pedro Sant'Anna @pedrohcgs? Do you want to watch it again? Or make your students watch it? You are in luck!! The recording is available at https://t.co/Rq3RCy2m2w and
https://t.co/I2eROAXRKC
@causalinf "Our estimates of group-time average treatment effects run in 3.0 seconds on....
a laptop with a 2.80-GHz Intel i5 processor with 8GB of RAM"
3 seconds flat on a machine that has basically the same hardware as a modern-day toaster...
Further, Callaway and Sant'Anna (2020) provide a number of different weighting approaches for the aggregation, including those that allow one to evaluate treatment effect heterogeneity and treatment effects after a certain period of time following treatment
Summary:
TWFE suffers with staggered tx adoption by downweighting effect from Tx groups with shorter exposures. Solution: Use ATTs that compare each treated group against pre-treatment periods of other treated groups, aggregate ATTs using sample size weights, and bootstrap
This was an outstanding workshop by @agoodmanbacon
and @pedrohcgs in today's @HlthEconSeminar. Great practical advice on how to manage TWFE (and when not to). If you missed it, happily it's being archived. My advice: grab a notepad and take a (another) look.
Should clarify that the bigger problem with TWFE that was mentioned in the talk and described in Goodman-Bacon (2019) is that negative weights can arise when treatment effects are changing over time. Hence, tx can be positive for all subjects but the TWFE will be negative.
New paper shows 18 yo are more likely than 17 yo to be prescribed opioids for pain in the ER (despite being clinically similar). This bump in prescribing rates for “adults” is associated with a jump in opioid related harms.
Decision “noise” in opioid prescribing is harmful