Read on Twitter
When I read 
headline,
"Evidence generation above all"
https://rockhealth.com/rock-weekly/evidence-generation-above-all/
I had > a few thoughts.
1. No shit. I was baffled when I left academia 3 yrs ago & remain flummoxed that health tech is so late to the game w/this. It doesn't have 2 be this way...
#thread
headline,"Evidence generation above all"
https://rockhealth.com/rock-weekly/evidence-generation-above-all/
I had > a few thoughts.
1. No shit. I was baffled when I left academia 3 yrs ago & remain flummoxed that health tech is so late to the game w/this. It doesn't have 2 be this way...
#thread
The smoke & mirrors is painful b/c people/payors pay for digital health b/c it delivers on, wait for it... CLINICAL outcomes (setting Goop & snake oil aside). Using "clinical outcomes" as a catch all 2 include cost savings 4 re-hospitalizations, 
med adherence,...
re-hospitalizations, med adherence,...
... reduced morbidity, mortality & prevalence of chronic disease, & increased QOL (among other things), it strikes me as completely obvious 
that evidence gen should be a pre-requisite for success - defining success as valuation, market share, capital raise.
that evidence gen should be a pre-requisite for success - defining success as valuation, market share, capital raise.
Now before you roll ur eyes & protest "but evidence gen is hard!" let's talk about the following:
2. There's loads of ways to generate evidence. Sure, decentralized clinical trials r sort of new but they're not that new. And pragmatic trials? they've been around awhile...
2. There's loads of ways to generate evidence. Sure, decentralized clinical trials r sort of new but they're not that new. And pragmatic trials? they've been around awhile...
& then there's multiphase optimization trials, factorial experiments, all kinds of cohort designs, Hill's criteria - I could go on but u get the picture.
METHODS MATTER - & science has a lot of methods 4 evidence gen!
Ex. Check out these ppl:
https://www.methodology.psu.edu/
METHODS MATTER - & science has a lot of methods 4 evidence gen!
Ex. Check out these ppl:
https://www.methodology.psu.edu/
4. Why aren't these methods routinely used 4 evidence gen?
- B/c scientists r not employed in enough #s or in the right positions in health tech cos
- B/c health tech was born from tech & tech tech ways were sloppily applied 2 health
- B/c, 4 fuck's sake, #whitepapers
- B/c scientists r not employed in enough #s or in the right positions in health tech cos
- B/c health tech was born from tech & tech tech ways were sloppily applied 2 health
- B/c, 4 fuck's sake, #whitepapers
5. What is mind-numbingly frustrating is that scientists have been raising the alarm on lack of evidence as problematic for YRS.
Here is but a sample:
https://www.psychologytoday.com/us/blog/digital-mental-health/202003/there-s-app-is-it-the-right-one
https://www.nature.com/articles/s41746-019-0093-1
https://pubmed.ncbi.nlm.nih.gov/29254914/
https://www.nature.com/articles/s41746-019-0111-3 https://pubmed.ncbi.nlm.nih.gov/24139770/
Here is but a sample:
https://www.psychologytoday.com/us/blog/digital-mental-health/202003/there-s-app-is-it-the-right-one
https://www.nature.com/articles/s41746-019-0093-1
https://pubmed.ncbi.nlm.nih.gov/29254914/
https://www.nature.com/articles/s41746-019-0111-3 https://pubmed.ncbi.nlm.nih.gov/24139770/
6. The kool aid drinking has got to stop. Well, go ahead & drink it if u must, but u will lose market share.
Health tech is getting increasingly competitive. Want to gain & retain ur B2B employer contracts? Better deliver on outcomes - REAL outcomes. Not cherry picked BS...
Health tech is getting increasingly competitive. Want to gain & retain ur B2B employer contracts? Better deliver on outcomes - REAL outcomes. Not cherry picked BS...
... The thing about health is that it is measurable. Even latent constructs are (e.g., my depressive symptoms). Also, humans take time to change, change is non-linear, & pop health is different from personalized health.
Sound complicated? It is. So is making drugs.
Sound complicated? It is. So is making drugs.
7. Want to win in health tech?
- Learn what effective engagement looks like for ur customers/end users:
https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2768347
- Model individual differences over time & acknowledge that pop means generally suck for personalizing products & services:
https://csteps.asu.edu/asu-engineer-makes-mobile-health-apps-more-effective
- Learn what effective engagement looks like for ur customers/end users:
https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2768347
- Model individual differences over time & acknowledge that pop means generally suck for personalizing products & services:
https://csteps.asu.edu/asu-engineer-makes-mobile-health-apps-more-effective
8. Hire scientists. We are as important as the engineers, UX, & all ur standard product people. I wrote this awhile ago but in case u missed it, here's more about how beh science is needed in health tech product design. Most tech is aiming for beh chng. https://mdisrupt.com/blog/healthtech/behavioral-science-building-health-product/
9. Experiments r amazing. Causal evidence is THE gold standard.
There r many more exp designs beyond the RCT. It depends on what ur doing (i.e., standard of evidence req varies) but u have options.
And always remember https://en.wikipedia.org/wiki/Correlation_does_not_imply_causation
There r many more exp designs beyond the RCT. It depends on what ur doing (i.e., standard of evidence req varies) but u have options.
And always remember https://en.wikipedia.org/wiki/Correlation_does_not_imply_causation
10. No, Big Data data science is not the panacea.
Don't believe me? Here r some reads:
https://bmcmedicine.biomedcentral.com/articles/10.1186/s12916-018-1122-7
https://www.jmir.org/2019/4/e12286/
http://veekaybee.github.io/2019/02/13/data-science-is-different/ @vboykis
Don't believe me? Here r some reads:
https://bmcmedicine.biomedcentral.com/articles/10.1186/s12916-018-1122-7
https://www.jmir.org/2019/4/e12286/
http://veekaybee.github.io/2019/02/13/data-science-is-different/ @vboykis
In closing...
Delivering on the promise of health tech demands that experts have a seat at many tables - in the C Suite, in strategy & corp dev, in product dev, as managers & ICs. Not 2 mention as VCs, KOLs, & partners.
I'm an optimist but woman, we have a LOOOONGG way 2 go.
Delivering on the promise of health tech demands that experts have a seat at many tables - in the C Suite, in strategy & corp dev, in product dev, as managers & ICs. Not 2 mention as VCs, KOLs, & partners.
I'm an optimist but woman, we have a LOOOONGG way 2 go.
