Measuring adherence to antiretroviral treatment in resource-poor settings: The clinical validity of key indicators

Table 3 Correlations among dispensing, self-report, and attendance-based adherence measures at time of clinic visit closest to 180 days after treatment initiation.

Correlation Significance N (bold = p < 0.05)	Coverage >80%	Coverage >95%	Coverage gap >30 days	Average self-report adherence *	Any self-report adhe-rence less than perfect*	% of visits on or before scheduled day**	% of visits within 3 days of scheduled day**	% of visits before last dispensing finished
% of days covered	0.335 < .0001 488	0.744 < .0001 488	-0.321 < .0001 488	-0.186 .0004 358	-0.185 .0004 368	0.324 < .0001 458	0.357 < .0001 458	0.560 < .0001 488
Coverage >80% &		0.388 < .0001 488	-0.734 < .0001 488	-0.002 .9762 358	0.010 .8474 368	0.186 < .0001 458	0.305 < .0001 458	0.237 < .0001 488
Coverage >95% &			-0.374 < .0001 488	-0.178 .0007 358	-0.162 .0018 368	0.264 < .0001 458	0.376 < .0001 458	0.399 < .0001 488
Coverage gap >30 days &				-0.003 .9480 358	-0.010 0.8471 368	-0.120 .0104 458	-0.222 < .0001 458	-0.206 < .0001 488
Average self-report adherence					0.983 < .0001 358	-0.136 .0104 357	-0.160 .0024 357	0.072 0.1725 358
Any self-report adherence less than perfect &						-0.141 .0069 367	-0.164 .0016 367	-0.044 0.3961 368
% visits on/before scheduled day							0.630 < .0001 458	0.360 < .0001 458
% visits within 3 days of scheduled day								0.367 < .0001 458

* Self-reports (n = 368) coded 1 = perfect, 2 = good, 3 = poor; excludes n = 120 cases from Ethiopia where self-reports were scored good/fair vs. poor
** n = 458 for appointment keeping variables due to missing scheduled visit dates
& Kendall tau rather than Spearman rank-order correlations are reported when these variables are correlated with each other

ISSN: 1472-6963