Wearable Data Security and App Efficacy Standards

May 7, 2015, Neurogaming Conference, San Francisco—Tracey Tien from the LA Times moderated this panel that addressed the intertwined issues of security and efficacy. Panelists were Ariel Garten from Interaxon, Eddie Martucci from Akili Interactive, Jason Moehringer from PsyberGuide, John Reppas from Neurotechnology Industry Organization (NIO), and Cory Lathan from Anthrotronix.

Is neurogaming unique with respect to privacy?
Reppas responded that the area is not special for privacy and security, a both converge to be like healthcare.
Lathan agreed that the data are important, but not a distinct category.

Exporting data?
Garten noted that existing data are just state and emotion data. Even with increased resolution of the new sensors and better algorithms, we cannot detect thoughts. The data are more like disease markers, but we need standards and better infrastructure. The existing standards for privacy differ from training and fundamental brain functions. People should own their own data, something current market standards don’t cover.
Lathan added that we can do better, as in the medical analogy for data and security. Analysis requires larger arrays of subjects, so data have to be merged and cleaned for aggregation.

Efficacy standards?
Moehringer noted that such standards do not exist for consumer-level apps. This areas is still emerging and the FDA ( Food and Drug Administration) is trying to work on the issues. There are alternative paths to market that allow the borderline apps at the studies level to get exposure, then they are moved to the FDA as they mature.
Garten retorted that the FDA has set a very high bar, requiring scientific validation through extensive testing.
Lathan suggested apps try to fit into one of three tiers: fun and interest, not a disease claim, and FDA. The first has very low standards, the second doesn’t exist in the current workings, and the third is difficult. The industry needs to define and promulgate useful standards. Insurers are wiling to pay for some neurogaming interventions, mostly in employee perks to record compliance data.
Reppas commented t ht patients, disease advocacy groups, and physicians are waiting for solutions. The payers, insurers, are the de facto guarantee bodies and seem willing to pay, so the industry needs market makers.
Martucci added that the FDA is working on some type of class enforcement discussions for things like mobile apps.

When to seek FDA proof?
Reppas offered to label as not a device, but this raises the question of how to define similar functions.
Martucci considered safety as one key attribute. In a similar way to other diagnostic and analysis areas try to set up as a “do your own analysis” like 23 and me. Call the technology computerized cognitive technology. In use for concussions and PTSD, define the tools as non-diagnostic, instead call them monitoring and tracking tools.
Lathan noted that the FDA will become involved if the device is intended to be used for safety and clinical decisions. But the government always overstep its bounds. Another issue is the label description of functionality and how the device is used in practice. It is important to limit intended use on labels and collateral materials.
Reppas declared that the government has to learn from consumer generations. FDA wants full reproducibility of claims and results. The organization is trying to change for diagnostic instruments to reduce the risk of wrong results.
Lathan commented that both consumer and clinical groups need the same measurements. The problem is if a device or process is used for a different or additional area.
Garten claimed that full FDA approval is not that hard if you do clinical studies at a large research organization like the Mayo Clinic, etc. who are trustworthy and credible organizations.

Efficacy regulations?
Moehringer responded that it is hard in the tier 2 area. The differences between medical and consumer end users is important. The end user wants curation, and validation and wants these functions in a forum to be able to compare claims, etc.
Martucci noted that a trusted curator is supplemented by analysts like SharpBrain who do consumer research.

Are third parties enough?
Moehringer declared that if one standard or the market exists, the flakes will wash out.
Lathan called for better judgment, but acknowledged that it is hard to develop a good check list for scientific and medical evaluations and the quality of those evaluations.
Garten commented that the costs of the standards partially depend on the ability to prove cyber trust. The good actors can claim some trust, but trust is quickly eroded by the bad actors.
Reppas added that Apple released their health kit to enable collection of mobile measurements on an aggregate basis. They are partnering with Sage Biometrics as the broker for the data.

Sharing of databases and brainwaves?
Garten reiterated that everyone owns individual data. The user can make some portion accessible to accredited third parties.

Trials for non-disease like better perception?
Moehringer called for a way to prove effective use of models, since it is so hard to develop standards and metrics for individuals as compared to masses of subjects.
Reppas commented that the trials have to be within a clinical framework, making it hard to do studies. The categories of data are the last to exist.
Lathan advocated separating medical and preventive classes. It is ok to change a designation to a general wellness class.
Martucci opined that the various types of data is problematic. In pediatrics, you can get similar data for other indications.

Big pharma is always helping physicians develop off label uses?
Reppas any MD can do experiments in good faith. They can do anything due to the low barriers in managing patients. Some tracking capabilities are available and are fairly simple to use. This aspect of care delivery is a good channel but can slow adoption if not written up in a major journal.
Martucci enjoined that clients tend towards prevention, regulations, etc. and can check if best practices are available. the problem is that most physicians don’t check for guidelines.
Lathan added that one issue is there is no definitive test for many problems, just good feasibility or high likelihood.

Size needed for trials?
Moehringer claimed that trial size can vary greatly causing a distraction. Standards play a central role in helping define sample sized for various targets.
Martucci stated that the FDA is actually capable of negotiating and will help in defining the critical data and evidence to submit.
Reppas wondered if the market exists for some products. It is understood that the MDs and payers have a high threshold for acceptance and adoption. The approval processes may change for therapeutic to include full clinical trials.
Garten noted that the trials need to include sham neurofeedback as a part of the trials to prevent observer effects and prove independently that the interventions and algorithms are effective.
Lathan compared the FDA and MS responses and suggested that the averages for a study need to include the control group and that the size of the studies should target the doctors’ need for similar data.