Kaiser Permanente Ventures puts on a series of healthcare-related presentations. Last week’s focus was a check up on the applications of digital health technology to mental health treatment. This coincided with a national focus on mental health after 2 prominent figures committed suicide. But importantly, the presentation focused away from those specific incidents. Instead, it concentrated on topics that face us every day – the high frequency of mental health issues generally that will come home to roost in every 1 out of every 5 people, and the stigma attached to mental health problems that may keep those affected from seeking treatment.
Dr. Mordecai, KP’s National Leader for Mental Health and Wellness, raised several intriguing considerations for developing digital health technologies for mental health, including some that got me thinking about their impact on startups and IP protection.
The first points raised by Dr. Mordecai addressed the type of care mental health apps could address.
Replicate or facilitate?
A replicative approach, as it sounds, replicates an existing treatment and envisions replacing a clinician with an app. A facilitative approach would design an app solution that is used in conjunction with existing treatment, keeping the clinician in the mix. For instance, the patient tracks moods and events on the app, and the clinician can review this data, as well as check in by chat as needed with the patient.
From a development perspective, Dr. Mordecai put his enthusiasm on the facilitative apps. He encouraged developers to take into account all stakeholders when designing the apps, particularly that both patients and clinicians would find the app easy to use and helpful when integrated into existing modes of treatment.
From an IP perspective, these approaches each offer their challenges. First, in the US, there is the subject matter eligibility aspect. Abstract ideas and methods that simply move into software what can be done by thinking and pen to paper don’t qualify for patent protection. Think about what the app will do, and how this differ from the existing face-to-face treatments when looking at how to protect innovations in this space.
Another aspect to consider in developing IP protection is the multiple persons or entities that participate in carrying out the steps of what will be a method claim in a patent. Enforcing patent claims against potential infringers becomes a challenge when a method claims requires more than one person to carry out the steps. For example, if the method requires the app, input from the patient and a response from a clinician, this would involve 3 parties (including the app provider) in the mix. Who of the potential infringers controls all of the steps? Likely no one – and therein lies the dilemma from a legal perspective.
Predictive versus reactive
Predictive technology collects and sorts through data to make predictions about mental health behaviors. The example provided by Dr. Mordecai was stratifying the risk of out-patients who had been previously treated for an incident and were at risk of future imminent self-harm. The overall risk for the full set of outpatients could be broken into lower risk and high-risk subsets based on certain predictive indicators. This allows the clinicians to focus on the higher risk patients and/or check in with this subgroup more frequently than the others. On the other side of the equation are the reactive approaches. Reactive technology, like it sounds, reacts to a set of circumstances already present. In the above example, a reactive approach might focus on a patient that has already experienced self-harm and now identified, is provided treatment.
The predictive side taps into the big data and machine learning trends. It can harness the power of the analytics to be proactive in treatment and act preventatively rather than waiting for a crisis to occur before intervention.
From an IP perspective, the power of the predictive apps is in their algorithm and the data sets they employ. In many cases, an algorithm is protected as a trade secret and there are pros and cons of this approach. On the one hand, trade secrets don’t offer protection from someone independently arriving at the same algorithm, unless they’ve stolen your information to get there. On the other hand, patenting means putting your algorithm out there “in print” for everyone to see. This leaves the app open to design-arounds that avoid any infringement but deliver a similar result.
The data sets are another matter. Access to data is very powerful and often start-ups don’t have their own healthcare data. Large organizations such as Kaiser Permanente (KP) are rich sources of data. But as Dr. Mordecai cautioned, larger organizations may not be the best “first stop” for startups developing mental health apps. KP, for example, takes on apps when they are more generally more mature. They look for prior testing in a smaller setting, and scalability since their membership is large and testing can involve a rollout to thousands if not millions of patient-members. As I have addressed in a previous blog , who owns the data – the patient, the healthcare provider – is another ball of wax.
BIG BROTHER or big brother?
The consideration of data collection and use leads into the third point raised by Dr. Mordecai: Is the software watching us like capitalized BIG BROTHER or lending a helping hand, the lower case big brother? and which do we want for mental health or is it a blend?
Create it and they will come? Maybe or maybe not.
Also interesting to me from Dr. Mordecai’s discussion was his take on the potential adoption of software solutions for mental health. From KP’s perspective as a healthcare provider, it looks for 3 features in apps: (1) initial uptake; (2) engagement for a meaningful length of time; and (3) contribution to clinically-meaningful outcomes. The first of these is relatively easy to measure – do the KP members download and open the app?
The second parameter is more subjective, what is a meaningful length of time? This is dependent on the objectives of the app and whether it is to treat a short-term or chronic condition. For example, if the app provides education materials about a condition, perhaps only a few engagements are necessary to provide impact. If the app is a therapy for an eating disorder, more frequent and repeated use by a patient may be required for the app to be meaningful in treatment.
The last parameter is arguably the most important and yet also the most difficult to assess. What is clinically-meaningful? It could range from curative solutions to more process-related improvements, like increased patient attendance at follow-up in-person therapy sessions. A recent dilemma highlighted for apps is how to conduct a properly controlled trial, and whether there should be a sham app or just no app used as the control (see e.g., here for an example)
This third parameter is also of interest from an IP protection standpoint. If the software provides tangible and measurable results, it offers a strategy for protecting efficacious solutions. While pure functional claiming can be difficult, a hybrid of method steps with defined measurable outputs could provide differentiation for patentability over prior art, as well as for a competitive edge in the marketplace.