Table 4. Multivariate association between treatment interruption pattern characteristics for each 6 month treatment period and poor treatment outcomes (default, failure, death) among MDR TB patients with at least one missed dose during the course of treatment (n = 542).

	0–6 months		6–12 months		12–18 months		18–24 months
Characteristic*	RR† (95% CI)	p	RR† (95% CI)	p	RR† (95% CI)	p	RR† (95% CI)	p
Interruption Pattern
Short, regular	reference		reference		reference		reference
Short, sporadic	–		3.32 (0.9–12.0)	0.07	2.54 (1.1–5.7)	0.03	0.77 (0.2–3.9)	0.75
Long, regular	–		–	–	–	–	1.0 (1.0–1.0)	1.0
Long, sporadic	1.10 (0.8–1.6)	0.61	4.37 (1.2–15.8)	0.03	3.38 (1.6–7.1)	0.001	2.78 (0.6–13.1)	0.20
Missing >10% of Treatment Doses	1.55 (1.1–2.4)	0.05	1.97 (1.2–3.4)	0.01	2.35 (1.3–4.3)	0.005	1.26 (0.5–3.0)	0.60

RR, rate ratio; CI, confidence interval.

^*Each variable represents characteristic for the 6 month treatment period (e.g., 0–6 months).

^†All RRs adjusted for age, number of treatment interruptions, variability of time on treatment, and whether they were underweight (BMI<18), or received kanamycin/amikacin or clarithromycin during the treatment period. RRs presented are adjusted for all other variables in the model. Dashes (–) for RR and 95% CI indicate there were not enough observations in the interruption category for that time period to derive an estimate.