1. Jessica E. Haberer, MD, MS 11 October 2017
Does the monitoring of medication packaging tell us what we need to know about TB treatment adherence?
Jessica E. Haberer, MD, MS
11 October 2017

2. Financial interests No conflicts to declare
Grant funding: US National Institutes of Health, Bill and Melinda Gates Foundation, USAID
Consulting: Merck, Natera
No financial relationships with the technologies discussed in this presentation

3. Outline Basic principles of TB treatment and adherence
Directly observed and self-administered therapy
Need for novel adherence-informed approaches to TB treatment
Features of “smart” medication packaging
Electronic adherence monitors
99DOTS (cell phone-based monitoring)
Next steps
Summary/conclusions

4. Basic principles of TB treatment and adherence
TB is nearly always curable if patients are treated with appropriate, effective, uninterrupted therapy (Nahid, CID 2016; Int J Tuberc Lung Dis 1999)
Directly observed therapy (DOT) in theory allows for:
Dose-by-dose monitoring
Delivery of tailored interventions as needed for those not presenting for treatment
Phone calls, in person visits
Counseling, structural assistance

5. Challenges with DOT
DOT does not improve outcomes over self-administered therapy (SAT) (Pasipanodya, CID 2013; Cochrane, 2015)
Significant cost to healthcare systems and significant burden to individuals receiving treatment (Tanimura, Eur Resp J, 2014; Cochrane, 2006)
Implementation low in many settings in practice
China: 20% of patients are managed with conventional DOT; over 50% self-administer (Hou, IJTLD 2012)
14% followed with DOT; 28% lost to follow-up (Lei, Infect Dis Poverty 2016)
The field is moving away from DOT (WHO 2014; McLaren, BMC ID, 2016)

6. Is SAT adequate?
(WHO Global TB Report, 2016)

7. Extent of relapse and drug resistance
Difficult to know how much of incident infection is new versus relapse
Study from southern India (Thomas, IJTLD 2005):
503 cured patients followed for 18 months
62 (12%) relapsed
Odds for relapse >2x for patients with irregular versus regular adherence
Drug resistance is a serious problem and likely underestimated due to lack of testing (Toczek, IJTLD 2012)
 ~480,000 people developed multidrug-resistant TB in 2015 (WHO Global TB Report)

8. (WHO Global TB Report, 2016)

9. WHO call for enhanced approaches
As treatment supervision alone is not likely to be sufficient to ensure good treatment outcomes, additional treatment adherence interventions need to be provided
When patients receive treatment adherence interventions in conjunction with DOT or SAT, the treatment outcomes are significantly improved compared to DOT or SAT alone
(WHO guidelines, 2017 update)

10. WHO recommendations
(Compendium of WHO guidelines: End TB Strategy)

11. Adherence-informed approaches to TB treatment

12. Adherence and relapse in TB clinical trials
Historically, post-treatment TB relapse has been low in clinical trials (<5%)
However, data from recent trials highlight substantial differences in relapse rates based on medication adherence
Patients taking HRZE with <90% adherence had 5.6 times increased hazard ratio for TB recurrence in a meta-analysis of the OFLOTUB, REMox, and Rifaquin trials
Months of follow-up
Source: TB ReFLECT Consortium, unpublished data
(with permission from Dr. Rada Savic)

13. What do we need to know about adherence?
3 stages of adherence: initiation, execution, and persistence
Initiation
Execution
Persistence

14. What information is currently available?
Self-report
Sensitivity to non-adherence
Commonly an overestimate due to social desirability and recall bias
Pharmacy refill
Maximum predicted adherence
Gives blunt information
Pill counts
Sensitive to non-adherence
Pill dumping is common
Better measures are needed to understand all phases of adherence

15. Medication packaging approach to adherence monitoring
Examples
Electronic adherence monitors
99DOTs
Common features
Dose-by-dose monitoring
Delivery of tailored interventions as needed for those not presenting for treatment
Phone calls, in person visits
Counseling, structural assistance
Same as DOT

16. Medication packaging approach to adherence monitoring
Unlike DOT
Lack of direct observation (e.g., by a caregiver)
Patient-centric and low burden
Use of packaging as a monitor is passive and requires minimal effort
Avoids daily transport to clinic
Additional options/features
Reminders
Education to avoid dosing confusion

17. Electronic adherence monitors
MEMS in use >25 years in >700 studies
Real-time devices in use ~10 years with >40 studies
(Permission given for photo)

18. Rationale for electronic adherence monitors
Automated system
Outreach monitoring (e.g., SMS) may require high levels of clinic engagement to know who needs support; e.g., ~30% in Wel-Tel (Lester, Lancet 2010)
Follow-up limited to those without evidence of dosing
Adherence challenges
Technical failures
Device non-use
Enables knowledge of the complete denominator of patients (not just those engaging with active monitoring strategies)
Automated pairing of monitoring with interventions
Clinic engagement

19. Medication monitor in India
MDR Boxes
DSTB Boxes
Pediatric Boxes
© Bill & Melinda Gates Foundation

20. Education through the medication monitor
DSTB
MDR-TB
PEDIATRIC DS-TB
© Bill & Melinda Gates Foundation

21. Evidence base for electronic monitoring
Review of 14 medical conditions in US, including hypertension, hyperlipidemia, and diabetes (Checchi, JAMA 2014)
37 peer-reviewed studies through 2014 (32 RCT)
Common characteristics:
Recorded dosing events and stored records of adherence
Audiovisual reminders to cue dosing
Digital displays
Real-time monitoring
Feedback on adherence performance
Increase in mean adherence by 3-34%
Percent “adherent” ranged from -8 to 50%

22. Evidence base for electronic monitoring*
Review of RCTs for SAT (Demonceau, Drugs 2013)
79 studies ( ) involving 23 diseases
Studies with EM-feedback type had the largest mean improvement in adherence measures
Average difference of 20% in adherence outcomes between patients receiving EM-feedback versus control; almost 2x the average difference among patients receiving other interventions
© Bill & Melinda Gates Foundation | 22

23. ART adherence improves with real-time monitoring + follow-up
112 individuals taking ART in rural Uganda
Monitoring with MEMS, followed immediately by Wisepill + follow-up of adherence gaps >48 hours
Adherence immediately increased and was sustained >6 months
11% increase
(Haberer, AIDS, 2017)

24. Counseling linked to electronic monitoring improves adherence to ART
120 individuals in Nanning, China
3 months monitoring, then randomized to device-triggered SMS reminders + enhanced counseling
Among suboptimal adherers, RR 2.4 for >95% adherence
(Sabin, JAIDS, 2015)

25. 40-50% improvement with medication monitor compared to standard of care in TB treatment
China’s National Tuberculosis Control Program
4,500 participants
Rural and urban settings
“Poor adherence” = 20+% missed doses
Limitations
Frequent, minor technical problems
Intensive case management likely underutilized
(Liu, PLoS Med, 2016)

26. Validation of electronic adherence monitoring
Smart packaging is a proxy for medication ingestion
Inaccuracies may be present
Device non-use (“pocket doses”)
Curiosity events
Pharmacokinetic modeling data show good agreement between concentrations predicted by electronic monitoring data and actual concentrations (Vrijens, JCP 2005)
Numerous studies show high degree of correlation between electronic monitoring and viral suppression (Musinguzi, AIDS Behav 2017; Haberer, JAIDS 2015)

27. Electronic adherence monitoring significantly associated with HIV viral suppression
95 adults and children in rural Uganda
Electronic monitoring and weekly IVR/SMS self-report over 6 months
Loss of viral suppression associated with electronic monitoring (p=0.02), but not IVR/SMS self-report (p=0.54)
(Haberer, AIDS 2013)

28. Validity of electronic adherence monitoring
Partners PrEP Study (phase 3 RCT)
Adherence sub-study of 1,147 individuals taking tenofovir-based PrEP with MEMS
Electronic monitoring performed best of all adherence measures (AROC ~0.7) compared to plasma tenofovir
Plasma has limitations as comparison standard
(Musinguzi, AIDS 2016)

29. Validity studies for TB
Electronic monitoring showed 99.5% sensitivity and 95.5% specificity versus urine Rifampicin among 432 patients in China (Huan, Chin J Antituberculosis 2012)
Validity evaluation of among 625 patients using 99DOTS compared to random urine Isoniazid ongoing (National Institute for Research in Tuberculosis, India)

30. Implementation currently ongoing
China
Ongoing health outcomes-oriented cluster RCT using the evriMED500
3,800 participants
ISRCTN
Planned distribution of 63,000 devices by end of 2017; provinces at a time through national system
India
250 patients with MDRTB using the evriMED1000
Planned distribution of 2,000 by year end

31. Additional trials using electronic monitors
A5300B/I2003B: Protecting Households On Exposure to Newly Diagnosed Index Multidrug-Resistant Tuberculosis Patients (PHOENIx)
Target enrollment of 3,452 household contacts
12 countries, 27 sites
evriMED1000
Evaluation of the Effect of 3HP vs Periodic 3HP vs 6H in HIV-Positive Individuals (WHIP3TB)
Target 4,000 participants
evriMED500

32. Cell phone-based monitoring: 99DOTS for self-reported dosing

33. 99DOTS toll-free phone number

34. 99DOTS toll-free phone number
Combination of Caller ID and numbers called shows that doses are in patient’s hands.

35. Notification of new patients
99DOTS adherence data
Notification of new patients
Alerts to field staff

36. 99DOTS implementation in India
Over 50,000 Patients
Scale-Up planned for 5 states with 250,000 patients
FDCs and 99DOTS envelopes to state drug store
FDCs in 99DOTS envelopes and shipped to TB centers
Patients issued 28 days of 99DOTS-wrapped meds
Patients registered on eNikshay (all phones)
TBHV notified of new assigned patients via SMS
Patients receive daily SMS reminders
Patients take meds, call random toll free number. Response “Thank you”
Patient calls populate adherence dashboard.
TBHV receives twice weekly SMS with “prioritized” patients
TBHV contacts/visit “prioritized” patients for counseling
Adherence dashboard used for enhanced counseling
REFILL
© Bill & Melinda Gates Foundation

37. 99DOTS and medication monitor ICT environment
Both technologies integrated/inter-operable with E-Nikshay and 99DOTS platforms
© Bill & Melinda Gates Foundation

38. Acceptability
Acceptability assessed within the RCT of electronic monitoring in China
High user performance, acceptability, and satisfaction among both TB patients and medical staff (Liu, Int J Environ Res Public Health 2017)
Desire for more platform development and training optimization
Acceptability study ongoing for 99DOTS (National Institute for Research in Tuberculosis, India)
Additional studies needed in diverse settings

39. Acceptability for real-time monitoring with ART
Monitors associated with habit formation
“I know that every night at 10:00 pm, I am supposed to be opening that device. It blinks every time I open it. So I am now used to it blinking every day because I take my medicines every day.”
Monitoring seen as caring
“I do not mind because I think it is good for me; I can take my medicine well knowing that I am being watched. When someone is waiting to see if you are taking your medicine, it means that they care and you should also have
the responsibility to take your medicine well.”

40. Cost
evriMED
Device costs (assuming 1 re-use)
~$5 per patient for “basic” version
~$10 per patient for “real time” version
$1/month for real-time data and SIM costs
99DOTS
~$5 per 6-month treatment course (all inclusive)
6 month DS-TB medication cost approximately $20 per patient

41. Next steps More data is needed on the use of this “smart” packaging
WHO recommendations are based on a relatively small evidence base
Need more studies on feasibility, acceptability, accuracy, impact, and scalability
Integration with other healthcare delivery improvements
No silver bullet: one approach won’t fix everything; one approach won’t work for everyone
Need “layers” of interventions
More data is needed on which adherence patterns are most critical; we know this in HIV, but not well in TB

42. Threshold for triggering interventions
Short window in HIV; what is it for TB?
479 individuals in rural Uganda
Real-time adherence monitors used to assess new viral rebound
587 adherence interruptions observed
13 instances of viral rebound detected in the field
OR 1.25 for each day beyond a 48 hour threshold (p=0.007)
(Haberer, JAIDS 2015)

43. Summary/conclusions
“Smart” medication packaging presents a novel paradigm for patient- centered care delivery
Electronic adherence monitors and 99DOTS both enable day-to-day monitoring that allow for differentiated support while limiting facility- based care
Evidence-base is growing for acceptability, feasibility, and impact on adherence and clinical outcomes

44. “New technologies in the diagnosis and treatment of TB are going to be key if we are to accelerate progress towards eliminating TB”
~José Luis Castro, Executive Director of the Union

45. Acknowledgments Bruce Thomas, Arcady Group
Bill and Melinda Gates Foundation
Wisepill Technologies
Ramnath Subbaraman
Discussions with my co-presentors

46. Thank you for your attention!