TABLE 2 Framework for the Ethical Evaluation of DOT and mHealth Interventions for TB Treatment Adherence, by Deceasing Accuracy of Adherence Detection
InterventionCriteria for Ethical Evaluation
(1) Monitoring Technology(2) Feature(s) to Enhance Patient Adherence(1) Accuracy(2) Stigmatization and Intrusiveness(3) The Use of Incentives(4) Balance of Individual and Public Good
Direct monitoring technology (metabolite testing) (DM)Follow-up by health care workers when non-adherent

SMS reminders to take medication throughout treatment

Reward incentives (e.g., in-kind goods, or reductions in insurance contributions)a conditional on good adherence

Penalty incentives (e.g., insurance surcharges) when non-adherent
Most accurate: medication ingestion and metabolization is necessary for adherence detectionLeast risk: urinalysis can be used with greater privacy and does not enable surveillance; respect for patient agency and autonomyGeneral risk of coercion with penalty incentives

Incentive size: incentives that are too small may fail to address all relevant adherence barriers; incentives that are too large may have greater risk of coercion

Incentive type: guaranteed rewards may have greater risk of coercion than lotteries and may be more likely to “crowd-out” intrinsic patient motivation

Incentive frequency: less frequent incentives may increase patients’ susceptibility to present-biased preferences
General risk of violating patient freedom and privacy when using individual adherence data to develop a picture of population-level adherence or to assist in contact tracing
Direct monitoring technology (embedded sensors) (DE)Most accurate: medication ingestion is necessary for adherence detection, but patients could induce vomitingMost risk: if visible, wearable hub is potentially stigmatizing; even if invisible, may enable location surveillance and identification of “problem” patients; restricted patient agency and autonomy
Video observation of therapy (VOT)Fairly accurate: swallowing is observed, but patients can feign ingestion or, where a human observer is required, collude to create false reportLeast risk: can be used with greater privacy and does not enable surveillance; respect for patient agency and autonomy
Direct observation of therapy (DOT)Most risk: association of patient with health care worker can be stigmatizing; restricted patient agency and autonomy
Indirect monitoring technology (patient- and device-facilitated) (IP and ID)Least accurate: swallowing not observed; patient can place false call or remove cap without taking medicationLeast risk: can be used with greater privacy and does not enable surveillance; respect for patient agency and autonomy
Recommended interventionFollow-up by health care workers when non-adherent

SMS reminders to take medication throughout treatment

Reward incentives (e.g., in-kind goods, or reductions in insurance contributions) conditional on good adherence
Maximize accuracy by minimizing opportunity for patient deception and adherence overreportingMinimize stigmatization and intrusiveness to preserve patient agency and promote autonomyUse reward incentives, but minimize risk of coercion by using 2-way SMS or video conferencing between patients and providers

Reward value should be carefully tailored to local social and economic context (smaller value to address patient factor barriers; larger value to address non-patient factor barriers)

Daily/weekly lottery
Develop population-level picture of adherence to more efficiently use resources, learn about best practices and regions where improvement is needed, and ensure timely drug restocking

Strive for anonymity, thus promoting public good while minimizing restriction of individual freedom and privacy
  • a For additional examples, see CDC, 2012.38