The American College of Cardiology’s (ACC) Scientific Session and Expo, held March 16-18 in New Orleans, kicked off Saturday morning with the initial results of the much-anticipated Apple Heart Study. For those who keep pace with the fast-moving tech world, I’m sure it felt like an eternity since the study had been announced. But by healthcare standards, the study went quickly and accomplished something unthinkable: It enrolled a breathtaking 419,297 individuals in eight months using a novel, entirely virtual study design.
The Apple Heart Study was designed as a prospective, single-arm, open-label study with the objective of helping primary care providers, general cardiologists and electrophysiologists figure out what they should do when a patient walks into their office because their Apple Watch sent them a notification of an irregular heartbeat. Do they ignore it because the technology isn’t clinical-grade? Order follow-up testing? Or even base a diagnosis on it?
While that objective and the intent to detect longer, more dangerous atrial fibrillation episodes (instead of every A-fib episode) changes the lens in which the results should be interpreted, the study as a whole confirms the clinical promise of wearable technologies. Far from flooding exam rooms with anxious patients, only 0.5% of enrolled individuals were sent notifications of an irregular heartbeat, and of those who completed a follow-up ECG, 34% ultimately were found to have atrial fibrillation. By means of comparison, incidence in the general population is estimated to be 2% for individuals under 65 years old, and 9% for those over 65 years. When they compared episodes of A-fib while the individual was wearing both the watch and the ECG patch, the watch notifications correlated with a positive A-fib patch reading 84% of the time. While the trial didn’t rely on the newest Apple Watch’s built-in ECG monitor, as more consumers use that new feature it will no doubt drive patient awareness, education and likely demand for confirmatory testing such as MCOTs and Zio patches.
In some ways, the study methodology was just as interesting as the results. The Apple Heart Study was an entirely virtual study, with patients recruited, screened, enrolled and managed via technology applications. While the results of the study were mixed, there are some key takeaways for those medtech companies that want to learn from Apple’s virtual, partner-based approach. Here are four:
- Patient-reported data should be taken with a grain of salt. In the Apple Heart Study, enrollees needed to be 22 years of age or older and a U.S. resident, in addition to having both an iPhone 5s or newer and an Apple Watch (without the ECG feature). Exclusion criteria included a prior diagnosis of atrial fibrillation or flutter, or the current use of anticoagulation medication. Interestingly, of the 2,161 individuals who received a notification of irregular heartbeats during the study, 161 fessed up that they actually knew they had A-fib prior to enrolling. While in this case that stat brought a chuckle from the audience at the ACC session, it demonstrates one of the clear downsides of such methodologies, especially when the technology being tested is consumer-facing and surrounded by this much hype. Some of this risk could be mitigated by linking electronic health records or claims data for the enrolled patient, or using the self-reported enrollment data as simply a first step in the process of identifying patients.
- Incorporating virtual elements could reduce cost and enrollment time frames. One of the consistent challenges many studies face today is the ability to reliably identify and recruit patients who meet a sometimes very specific set of criteria. Virtual screening, followed up with confirmation of study eligibility by a healthcare professional, could be a way to cast a wider net without relying solely on patient-reported data. In studies such as COAPT, which took almost six years to enroll because it was looking for such a specific type of patient, additional scale could have worked to the team’s advantage: If they’d had the ability to short-list prescreened candidates, perhaps it could have gone faster, saving time and money.
- It’s important to be mindful of how technology access could impact the composition of the study. Given the “grassroots” nature of virtual enrollment, the study population has the possibility of being more diverse from a geographic, ethnic and potentially socioeconomic standpoint than many randomized, controlled trials today. In the case of the Apple Heart Study, enrollment was restricted to those consumers who owned multiple, relatively expensive Apple products. The “buy-in” biases the sample toward those who can afford to pay for these products and are motivated enough to voluntarily seek out the study. That said, the promise of virtual enrollment is scale, limited only by the types of devices required for the study. If anyone with a phone or web access could be considered for enrollment, the study would be accessible to a wider group, potentially countering some of the bias we see in trials today where ethnic minorities and women are oftentimes underrepresented.
- Collaboration is key. With more technology solutions bridging the gap between high-tech and medical devices, each company will bring its own strengths and weaknesses to the table. By partnering with Stanford—along with American Well for telehealth study visits and BioTelemetry for the ECG comparator—Apple was able to bring the study to fruition quickly and lend credibility to the results. With the follow-on Heartline study planned with Johnson & Johnson, Apple will seek to prove the impact of its technology on overall health outcomes, albeit using a more traditional study design.
While many studies will continue to require physical interaction between HCPs and enrolled patients, this study does offer a glimpse into ways that future trials could reduce cost, accelerate enrollment, and become more relevant to the diverse patient populations that many HCPs see in their practices today. Virtual study elements offer significant upsides but also present risks that need to be hedged to protect the integrity of the study.