The Journal of the American Medical Association recently released the results of a study that looked at the effect of technology and behavioral interventions on patient outcomes following a heart attack.
The researchers found no significant difference between the medication adherence or clinical outcomes of those in the control group vs those that were given a combination of technologies and incentives.
According to one of the study’s authors, Dr. David Asch, executive director of Penn Medicine’s Center for Health Care Innovation, “It was a surprise. We went into this study thinking that it would be effective and it wasn’t”. Fellow author, Dr. Kevin Volpp echoed that same sentiment: “What we found was a little bit surprising and a little bit disappointing”.
The study was conducted at the University of Pennsylvania Health System (UPHS) over a span of 12 months. There were 1,509 patients involved in the study; all hospital inpatients who had experienced a heart attack and had been hospitalized between 1 and 180 days. The average age of the study group was 61 and they were all insured with five carriers who had Medicare fee-for-service arrangements with UPHS. All of the patients had been prescribed at least two daily medications (statin, aspirin, beta-blocker or antiplatelet).
The control group of 506 patients was given the standard post-discharge medication instructions and treatment. The remaining 1,003 patients were given additional tools to help them post-discharge:
- A smart pill bottle that tracked whether or not it had been opened at the prescribed intervals
- Participation in a daily lottery with a 1 in 5 chance of winning $5 and a 1 in 100 chance of winning $50 each day medications were taken as prescribed
- An option to enlist a friend or family member to receive notifications if the participant failed to use the smart pill bottle twice in any 3-day span
- Access to social work resources
- A hospital-based advisor to answer questions and reinforce medication adherence over the phone
On the surface, the failure of this level of support and intervention is disheartening for anyone developing medication adherence technology or involved with helping a loved one recover from being hospitalized. However, if you listen to the post-study podcast or spend time looking at how the incentives/technology was administered to the study group, important clues emerge as to why this failure may not mean abandoning hope for technology-based interventions.
First, only 878 of the 1,003 patients activated their smart pill-bottles and only 70% of that same group fully participated in the incentives and technology. This indicates that the lack of adherence may not have anything to do with technology when its working, but rather that there is a challenge to get patients using that technology in the first place.
Second, the fear of another heart attack may have been enough of an incentive to keep patients on their medication regimens. Put another way, perhaps the control group already had sufficient incentive to follow their prescriptions and thus technology would have little impact.
Third, and perhaps most significantly, an average of 41 days elapsed between the time the patients were discharged from the hospital and the time they were activated on the incentive/intervention program. This delay was attributed to the delay in the insurance process. According to Dr. Volpp:
If we had been able to engage these patients earlier, for example. If this had been a hospital-based intervention and this could have started at the time of discharge [rather than weeks later], then we would have had a greater opportunity to influence these patients and change the course of their care
I personally found this last point by Dr. Volpp fascinating. This study may have inadvertently shown that the timeliness of implementing post-discharge behavior and technology incentives matters as much as the types being implemented. 41 days after discharge is a long time – almost a month and a half.
Consider this example. Say you get caught for speeding and as part of the ticket-writing process the police officer activated a reminder system in your car that (a) warned you when you were 5 miles over the limit and (b) sent a message to your significant other whenever you receive two such warnings in the same day. From personal experience I can tell you that the week after I got a speeding ticket, I followed every posted speed limit. Why? Because the trauma of getting caught was still fresh in my mind.
Now imagine the same scenario but instead of immediately activating the warning system it took until 41 days after getting your ticket. By the time the system was in place you would have already fallen back into old habits and assured yourself that you were “fine” driving the way you were.
It would be interesting to see if analysis of this study’s data revealed a correlation between the length of time before incentive implementation and adherence. Even if it doesn’t, this study holds a cautionary tale for anyone in HealthIT – timeliness of implementation may matter as much as your solution itself.