Propeller Health

Short Summary of Intervention:

“The Leading Mobile Platform for Respiratory Health Management,” according to the company website, Propeller Health helps patients manage their asthma by using an electronic sensor to help patients record their inhaler use and track their medications.

Links to Materials Being Reviewed

This abstract of these interim findings for a study financed by the California Healthcare Foundation, which is also an investor in Propeller Health, also links to others.

Other treatments include Forbes’ David Shaywitz Two Examples Highlight Challenges of Demonstrating the Benefit of Digital Health Technology/ and The Health Care Blog’s Meet Propeller Health, Digital Health’s Poster Child for Invalid Savings Reporting (Al Lewis and Vik Khanna)

Summary of key figures and outcomes:

• 40 people studied for four months
• No change in Asthma Control Test Scores
• 25% less rescue drug use
• 62% reduction in inpatient utilization
• $688/asthmatic in savings from reduced ER and Inpatient usage (vs. $282 “improvement” for control group), before the costs of the equipment is added back in
• 67% of asthmatics presenting in ER admitted to the hospital (vs. 20% for country as a whole)

Questions for Propeller Health:

Why shouldn’t a disinterested observer question the linkage between California Healthcare Foundation financing the study and also investing in your company?

ANS: Refused to answer

Doesn’t the large treatment effect violate Dr. John Ioannidis’ conclusion that large treatment effects are usually wrong, especially because both groups received the same treatment?

ANS: Refused to answer

Wouldn’t one expect that such large improvements in avoided hospitalizations be caused by equally large improvements in the official Asthma Control Test Scores, instead of no improvement in those test scores?

ANS: Refused to answer

Should the Asthma Control Test Score algorithm be revisited due to its failure to predict this massive improvement?

ANS: Refused to answer

In disease management, typically a reduction in inpatient utilization is achieved by an improvement in adherence generating more drug use.   How were you able to achieve the opposite?

ANS: Refused to answer

Why did you write “Patients on Propeller save more than $600/year” on your website when the net reduction in spending (vs. the control group) was only $406?

ANS: Refused to answer

How do you reconcile your claim of $688 in savings per asthmatic in ER and inpatient utilization with the fact that commercial health plans only spend about $400 per asthmatic on asthma-coded ER and inpatient utilization?

ANS: Refused to answer

Your principal investor calls the $206 reduction in costs for the control group an “improvement” in costs. How could it be called an improvement when there was no intervention in the control group and therefore any change is due to regression to the mean and the Hawthorne effect? Wouldn’t it be more ethical to acknowledge that?

ANS: Refused to answer

Why did you decide to overlook the 67% inpatient admission rate of patients who present at the ER, and why is it several times the expected rate of 20% or less?

ANS: Refused to answer

Your own principal investigator, Rajan K. Merchant, is on record as saying that this data shouldn’t have been released by “Proppeller” because the study isn’t complete yet. “Anything prior is pure speculation,” he observed. How do you respond to your principal investigator’s opinion that you should not have released this data?

ANS: Refused to answer

Responses, addenda, apologies, retractions, changes etc.

Addendum: Email exchange

Email from Propeller:

The trial [is] still underway and we look forward to the principal investigators submitting the findings for peer review later this year.

Our Response:

Thank you.  So then another question is: why did you release this information now if indeed the trial is still underway and the principal investigator says the information shouldn’t have been released?   (Note:  It is OK to apologize if it was a mistake.  There is a spot on the website where we will note apologies.)

ANS: Refused to answer or apologize