The time is right to explore how sensor-enabled digital adherence monitoring systems can benefit patients, sponsors, and payers. Non-adherence to prescribed medications is a key issue facing our healthcare industry, costing the US healthcare economy between $100-300 billion annually, representing between 3-10% of total healthcare costs.1,2 Further, non-adherence also decreases our ability to capture dose-response relationships and safety profiles during clinical development.3 The continuing serious financial burden of non-adherence to the healthcare economy, combined with the risks of an inadequate understanding of dose-response and dose-safety profiles has created a landscape that requires our full attention.
This week I gave a lecture on “Developing Wireless Adherence Solutions at the Interface of Technology and Pharmaceutical Science” during the “Essential tools for improving patient adherence” session at the FIP International Congress of Pharmacy and Pharmaceutical Science in Buenos Aires, Argentina. The FIP Congress is focused on providing clinical pharmacists and pharmaceutical scientists around the world with current best practices and cutting edge approaches to improving patient care. My lecture was based on previous industry development efforts I was involved in to submit the first digital medicine new drug application (NDA) to the United States Food and Drug Administration, which not only combined cutting edge technology with an active pharmaceutical, but also employed a unique ‘reverse’ application of pharmacometrics to guide the development of this complex drug product.
While mobile applications,4 ingestible sensors,5 and new technologies are becoming more prevalent in the adherence space, there are few efforts destined for regulatory submission that introduce these technologies in combination with active pharmaceuticals. To my knowledge, the aforementioned joint submission between Otsuka and Proteus is the only one of its kind:6 The project combined a 510(k) approved ingestible/wearable sensor with an FDA approved anti-psychotic drug to provide objective adherence information to the patient’s care team (Fig. 1).
Ideally, such systems will improve adherence in participating patients. Although this may seem an intuitive assumption, proving this claim is not trivial. Clinical outcomes, such as event free survival or relapse rates, could be used to make separate label claims and demonstrate efficacy of the system; however, these metrics are separate from claims regarding adherence that imply improved maintenance of prescribed dosing regimen. Generating sufficient evidence that one product ‘improves’ adherence over standard oral therapy requires comparing adherence-related observations for both treatments—one must prove that those observations are superior in the treatment arm. The only such adherence-related observation for such a claim would be plasma (blood) time-concentration data given the absence of objective dosing information in an active control arm. This suggests that a pharmacometric endpoint would need to be provided to obtain such a label claim. How could one apply pharmacometric principles to demonstrate that these systems improve adherence? Reverse them.
Traditional pharmacometrics is the process of building a quantitative model that is not only capable of capturing the pharmacokinetic/pharmacodynamics profile of a given compound over time, but also the variability associated with those profiles. After this has been completed (assuming there is confidence with the model parameter values and variability), one can fix those values and use that information in a separate data set to partially tease out deviations of that variability in observed versus expected exposures. The development and application of this approach is described in the work that my colleagues and I conducted7 to determine whether or not a successful development path could be paved for an “improves adherence” label claim for our digital medicine system.
Our evidence suggested that rigorously quantifying ‘improved adherence’ over standard of care oral therapy was unlikely for multiple reasons: the modest natural variability in the compound; the lack of correlation with clinical screening criteria to identify patients with lower plasma exposures than expected; and higher measured exposure values than expected (around 65%). One of the main drawbacks of this scenario is that patients in the standard of care oral therapy arm can overdose or severely misrepresent the time of their last dose and nothing can correct for this. In contrast, a more objective control arm, such as the Medication Event Monitoring System ([MEMS], Aprex Corp., Fremont, Calif.)—electronic cap technology—could be used; however, in this case, the study would need to run a long time to minimize any potential ‘white coat’ type effects where patients become—and may stay—more adherent simply because they are participating in a study. This effect can be exacerbated if the control arm patients also have technology-enabled dosing—which would not be a desirable outcome for the study. This is a bit like trying to maintain uncertainty in Schrodinger’s cat box without closing the lid; it negates the principle. This application of ‘reverse’ pharmacometrics is one of the few published applications of pharmacometrics used to inform a development pathway without actually generating any PK tables—it suggests that there may be other opportunities for pharmacometricians to engage our clinical colleagues (and vice versa!) to create more informed development pathways as our products become more complicated.
 Institute, N.E.H., Thinking Outside the Pillbox: A System-wide Approach to Improving Patient Medication Adherence for Chronic Disease. 2009.
 Iuga, A.O. and M.J. McGuire, Adherence and health care costs. Risk Manag Healthc Policy, 2014. 7: p. 35-44.
 Smith, D.L., Patient Nonadherence in Clinical Trials Could There Be a Link to Postmarketing Patient Safety. Therapeutic Innovation & Regulatory Science, 2012. 46(1): p. 27-34.
 Dayer, L., et al., Smartphone medication adherence apps: potential benefits to patients and providers. J Am Pharm Assoc (2003), 2013. 53(2): p. 172-81.
 Hafezi, H., et al., An ingestible sensor for measuring medication adherence. IEEE Trans Biomed Eng, 2015. 62(1): p. 99-109.
 US FDA Accepts First Digital Medicine New Drug Application for Otsuka and Proteus Digital Health. 2015 6 September, 2016]; Available from: http://www.proteus.com/press-releases/u-s-fda-accepts-first-digital-medicine-new-drug-application-for-otsuka-and-proteus-digital-health/.
 Knights, J. and S. Rohatagi, Development and application of an aggregate adherence metric derived from population pharmacokinetics to inform clinical trial enrichment. J Pharmacokinet Pharmacodyn, 2015. 42(3): p. 263-73.
All information presented derive from public source materials.
To learn how modeling and simulation was used to facilitate regulatory approval for a long- acting injectable antipsychotic drug, read this case study.