Transcript No Slide Title

ELECTRONIC MONITORING TO ASSESS ADHERENCE AND VALIDATE ALTERNATIVE ADHERENCE MEASURES
IN TUBERCULOSIS PATIENTS ON COMMUNITY-BASED DIRECTLY OBSERVED TREATMENT
Jossy van den Boogaard
1,2,
Ramsey Lyimo 2, Martin Boeree 1, Gibson Kibiki 2, Rob Aarnoutse
1
1Radboud
University Nijmegen Medical Centre, The Netherlands;
2Kilimanjaro Christian Medical Centre, Tanzania
ABSTRACT
BACKGROUND
Problem statement: Community-based Directly Observed Treatment (DOT) can be an effective strategy to improve adherence to tuberculosis
(TB) treatment in settings where facility-based DOT is causing overburdened healthcare facilities. However, the strategy may lead to irregular
drug intake in practice. This is difficult to prove in the absence of a simple and valid adherence measure.
Objectives: We assessed adherence rates of TB patients on community-based DOT by using the Medication Event Monitoring System (MEMS
bottles), and we used MEMS as a reference standard to determine the validity of alternative adherence measures.
Design and setting: This was a longitudinal study among outpatients attending four public TB clinics in Tanzania’s Kilimanjaro Region. The
Tanzanian TB programme allows patients to choose between facility- and community-based DOT. Patients on community-based DOT have to
select a treatment supporter (usually a relative or spouse) who should provide DOT in the home setting.
Study population: Adult TB patients who presented with newly diagnosed TB between February and May 2010 and who had chosen for
community-based DOT were eligible to participate. Fifty patients were enrolled; 37 completed treatment, six died, three defaulted and four
dropped out.
Outcome measures: MEMS data was used to calculate adherence rates by dividing the number of days on which at least one bottle opening was
registered by the total number of monitored days, multiplied by 100%. Adherence rate cut-off values of 100% and 95% were used to differentiate
between adherence and non-adherence, and to determine the validity and accuracy of the test measures. The test measures included a urine test
for isoniazid, urine colour test for rifampicin, Morisky scale, Brief Medication Questionnaire (BMQ), adapted version of the AIDS Clinical Trials
Group (ACTG) adherence questionnaire, pill counts and clinic attendance for medication refills.
Results: Adherence rates ranged from 50.0 to 100% (median 98.4%) in all patients, and from 89.3 to 100% (median 98.4%) in the patients who
completed treatment. In the latter group, 70% of patients were less than 100% adherent and 19% less than 95%. The ACTG questionnaire and
urine colour test had the highest sensitivity (70-100%) but lowest specificity (20-37%) for detecting non-adherence, and the Morisky scale and
clinic attendance the highest specificity (80-100%) but lowest sensitivity (14-35%). The sensitivity of the routinely used combination of pill
counts and clinic attendance improved when the ACTG questionnaire was added.
Conclusions: The high adherence rates suggest that the Tanzanian model of community-based DOT can be an effective strategy to prevent nonadherence. Studies in patient populations with a wider range of adherence rates are needed to confirm the validity of (combinations of) adherence
measures that are feasible for use in resource-limited settings.
METHODS (2)
OBJECTIVES
•Active TB is the second leading cause of death from infectious
diseases worldwide, causing 1.6 million deaths every year
• Non-adherence to TB treatment is a major barrier to global TB
control
• WHO’s solution of Directly Observed Treatment (facility-based DOT)
is causing overburdened healthcare facilities worldwide
• Involvement of community members in the provision of DOT is an
alternative, but adherence under community-based DOT is unknown
and this strategy may need monitoring by adherence measures
• Medication Event Monitoring System (MEMS) bottles can be used to
validate other simple and affordable adherence measures
•To assess adherence rates of TB patients on community-based DOT
by direct and indirect adherence measures
•To determine the validity of other locally feasible adherence
measures with MEMS as the reference standard
When applied
Indirect
MEMS
Continuously
Drug refill visits
Every week in first 2
months;
Every 2 weeks in last 4
months
Pill counts
Every week / every two
weeks
Morisky scale
Week 4, 8. 12
Brief Medication Questionnaire (BMQ)
Week 4, 8, 12
METHODS (1)
• Longitudinal study among 50 adult TB outpatients on communitybased DOT in Tanzania’s Kilimanjaro Region
• Adherence to once-daily TB drugs was assessed throughout 6
months TB treatment
• Patients filled a questionnaire about MEMS use at completion of
treatment. Pocket dosing (taking out medication for later use) as
identified by this questionnaire was excluded from analysis
Funding: This study was financially supported by KNCV Tuberculosis Foundation (the Netherlands).
Measure
Adapted AIDS Clinical Trials Group (ACTG) questionnaire Week 16
Direct
Urine colour test for rifampicin
Week 4, 8, 12, 16
Urine test for isoniazid
Week 4, 8, 12, 16
•MEMS adherence rate cut-off values of 100% and 95% were used to
differentiate between adherence and non-adherence and to calculate
sensitivity, specificity and accuracy of the test measures
RESULTS (1)
RESULTS (2)
RESULTS (3)
•50 patients were enrolled:
Figure 1. Non-adherence to TB treatment of patients who completed treatment (n=37), assessed by the different
test measures
Table 1. Sensitivity, specificity and accuracy of adherence measures to differentiate between adherent (= at least 95%
adherent according to MEMS) and non-adherent patients
62% male,
mean age 42 yrs,
76% completed primary education only,
4% HIV-positive; 27% of HIV-positives on antiretroviral treatment
Sensitivity
Specificity
Accuracy
Drug refill visits
14
83
70
Pill counts
29
63
57
Morisky scale
14
90
76
BMQ
57
55
56
•37 completed TB treatment, 6 died, 3 defaulted and 4 dropped out of the study
ACTG questionnaire
100
37
49
•MEMS adherence rates:
Urine colour test
86
27
38
INH urine test
43
60
57
Refill visits + pill count
29
57
51
Refill visits + pill count + ACTG
100
20
35
Refill visits + pill count + urine colour test
86
23
35
Refill visits + pill count + BMQ
71
38
44
Refill visits + pill count + INH urine test
71
37
43
Refill visits + pill count + Morisky scale
29
57
51
•7/50 patients had no formal treatment supporter despite being on community-based DOT
In all patients, mean MEMS adherence was 96.3%, SD 7.7%
Among all patients, 70% were <100% adherent, 21% were < 95% adherent and 2% were <80%
No difference in MEMS adherence was found between those with and without a treatment
supporter (p=0.79, independent -samples T-test)
RESULTS (4)
CONCLUSIONS
•The combination of pill counts and refill visits that is used in routine practice had moderate
sensitivity and specificity
• This is the first study in which MEMS was used to assess adherence in community-based DOT during the full 6-months TB
treatment
• Sensitivity of routine combination improved when an adherence questionnaire was added
• High adherence rates were observed in a Tanzanian population, not confirming the concern that patients on community-based DOT
are prone to non-adherence. However, community-based DOT was turned into self-administered treatment by some patients
• Adherence questionnaires have the additional advantage that they provide information about perceived
facilitators of and barriers to adherence
• Supplementing pill counts and clinic attendance with a third measure such as the ACTG questionnaire, helps to identify potentially
non-adherent patients who could benefit from tailored adherence-promoting interventions
RESEARCH AND POLICY IMPLICATIONS
STUDY LIMITATIONS
• Adherence rates in this study could have been biased due to study participation. However, other studies
suggest that the ‘interventional effect’ by MEMS ceases when used for months.
• Community-based DOT seems to be an effective alternative in settings where facility-based DOT is causing overburdened
healthcare facilities
• Additional research in larger patient populations is needed to confirm this finding
• The small sample size does not allow for firm conclusions with regard to the effectiveness of communitybased DOT
• Due to the generally high adherence rates found in this study, only adherence rate cut-off values of 95%
and 100% could be applied
• The combination of pill counts, clinic attendance and an adherence questionnaire can be a useful tool to monitor adherence of
patients on community-based DOT
• Additional research is needed to confirm the sensitivity and specificity of this and other combinations of adherence measures that
are feasible to use in resource limited settings
This study was financially supported by the ‘Netherlands-African partnership for capacity development and clinical interventions against poverty-related diseases’ (NACCAP/EDCTP)
and by the KNCV Tuberculosis Foundation , The Netherlands