Nathan Geffen, TAC

In the October 2009 issue of HTB South, we described a pilot study by Atkins and colleagues in which TB patients self-administered treatment. [1] At the World Lung Conference, a poster by MSF examined self-administered treatment at a site in Homa Bay, Kenya. [2] This is in contrast to direct observation of treatment recommended by the World Health Organisation (WHO).

This cross-sectional study consisted of a survey of 279 potentially eligible patients taken from the TB register in November 2008 and June 2009. Of these, 67 (24%) did not participate because five never started treatment, 20 defaulted before the survey, 11 were dead, four were hospitalised, 13 refused to consent to the home visits necessary to carry out the survey and 14 were not found. Patients received education on TB treatment. They had to collect their medicines weekly during the intensive phase and monthly during the continuation phase.

The survey consisted of four measurements: an interviewer-administered questionaire, visual analogue scale in which the patient estimated adherence in the last month, pill counts and an isoniazid urine strip test. For those who participated the median age was 35 years, 46% were female, 79% had pulmonary TB and 69% were HIV-positive. Of those who were HIV-positive, 73% were on ART.

The interviewer-administered questionaire asked questions about pill intake over the past four days. If patients responded that they had taken all their pills, their adherence was described as exact. If they took 75% of their pills, their adherence was described as satisfactory, else their adherence was described as unsatisfactory. 95% were exact, 3% satisfactory and 2% unsatisfactory.

On the Visual Analogue Scale assessment, patients were asked: “How much of your prescribed TB medications have you taken in the last month?” 93% answered 90-100%. 7% answered 80-90% and only one patient answered less than 80%.

Overall, 98% of patients tested positive for isoniazid. Pill counts results were described as exact if 100% of pills were taken, satisfactory if at least 80% of pills were taken and unsatisfactory otherwise. 84% were exact, 10% satisfactory and 6% unsatisfactory. However pill count data was unavailable for 64 patients.

There was no significant difference observed between the intensive and continuation phases for any of the four measures of adherence nor were there significant differences between HIV-positive and HIV-negative people, nor between people on ARVs and people not on ARVs.

Reasons given for non-adherence varied with 17 patients giving 36 reasons, the most common being “ran out of pills” (8 patients), followed by “away from home” (7 patients), followed by, “forgot to take medication” (6 patients).

The Kappa coefficient measures agreement between qualitative assessments. A Kappa < 0.4 is moderate or poor agreement, 0.4 – < 0.75 is fair to good agreement and ≥ 0.75 is excellent agreement. The only combination of assessments that had fair to good agreement was INH positivity and the questionaire (Kappa = 0.43). All other binary comparisons of the assessments had poor to moderate agreement.

The authors pointed out that the INH test was expensive and needed a cold chain and that pill counts had poor reliability and inconsistent data. Therefore they state that in routine settings, the questionaire and visual analogue scale should be the preferred adherence measurement tools. They also state that adherence measurements should not be based on just one tool.

The authors concluded that adherence was high among surveyed patients but acknowledged that the study was limited by several factors including that only patients who consented to home visits were surveyed, the adherence visit was sometimes too close to the last visit and patients who defaulted before the survey were not included.

comment

At the time of the conference, data was not yet available to adequately compare patient outcomes and adherence. Hopefully the researchers will publish this soon.

While this study has some serious limitations–particularly that it does not examine patients who were not adherent before the survey started–observational data is accumulating that self-administered treatment for TB can compete with the directly observed treatment model promoted by the WHO. A randomised open-label trial to compare these methods is warranted. Such a trial should compare health outcomes and adherence, but also perceptions of patient dignity, travel costs incurred by patients and cost to the health system.

References

1. Geffen N. 2009. Overview of TB-related studies at IAS.

http://i-base.info/htb/5919

2. Huerga H et al. 2009. Adherence to self-administrated tuberculosis treatment in Homa Bay, Kenya. 40th World Lung Conference, Cancun, December 2009.