Which adherence measure - self-report, clinician recorded or pharmacy refill - is best able to predict detectable viral load in a public ART programme without routine plasma viral load monitoring?

被引：30

作者：

Mekuria, Legese A. ^{[1
,2
,3
,4
]}

Prins, Jan M.

Yalew, Alemayehu W. ^{[3
]}

Sprangers, Mirjam A. G. ^{[1
]}

Nieuwkerk, Pythia T. ^{[1
]}

机构：

[1] Univ Amsterdam, Acad Med Ctr, Dept Med Psychol, Amsterdam, Netherlands

[2] Erasmus Univ, Med Ctr, Dept Epidemiol, Netherlands Inst Hlth Sci, Rotterdam, Netherlands

[3] Univ Addis Ababa, Sch Publ Hlth, Addis Ababa, Ethiopia

[4] Univ Amsterdam, Div Infect Dis, Dept Internal Med, Acad Med Ctr, Amsterdam, Netherlands

来源：

TROPICAL MEDICINE & INTERNATIONAL HEALTH | 2016年 / 21卷 / 07期

关键词：

adherence; detectable viral load; HIV/AIDS; combination antiretroviral therapy; resource-limited setting; ACTIVE ANTIRETROVIRAL THERAPY; TREATMENT RESPONSE; INHIBITOR THERAPY; HIV-1; INFECTION; VALIDITY; OUTCOMES;

D O I：

10.1111/tmi.12709

中图分类号：

R1 [预防医学、卫生学];

学科分类号：

100235 [预防医学];

摘要：

ObjectiveCombination antiretroviral therapy (cART) suppresses viral replication to an undetectable level if a sufficiently high level of adherence is achieved. We investigated which adherence measurement best distinguishes between patients with and without detectable viral load in a public ART programme without routine plasma viral load monitoring. MethodWe randomly selected 870 patients who started cART between May 2009 and April 2012 in 10 healthcare facilities in Addis Ababa, Ethiopia. Six hundred and sixty-four (76.3%) patients who were retained in HIV care and were receiving cART for at least 6months were included and 642 had their plasma HIV-1 RNA concentration measured. Patients' adherence to cART was assessed according to self-report, clinician recorded and pharmacy refill measures. Multivariate logistic regression model was fitted to identify the predictors of detectable viremia. Model accuracy was evaluated by computing the area under the receiver operating characteristic (ROC) curve. ResultA total of 9.2% and 5.5% of the 642 patients had a detectable viral load of 40 and 400RNAcopies/ml, respectively. In the multivariate analyses, younger age, lower CD4 cell count at cART initiation, being illiterate and widowed, and each of the adherence measures were significantly and independently predictive of having 400RNAcopies/ml. The ROC curve showed that these variables altogether had a likelihood of more than 80% to distinguish patients with a plasma viral load of 400RNAcopies/ml from those without. ConclusionAdherence to cART was remarkably high. Self-report, clinician recorded and pharmacy refill non-adherence were all significantly predictive of detectable viremia. The choice for one of these methods to detect non-adherence and predict a detectable viral load can therefore be based on what is most practical in a particular setting.

引用

页码：856 / 869

页数：14

共 48 条

[1]

[Anonymous], 2003, Adherence to Long-Term Therapies: Evidence for action

[2]

[Anonymous], 2015, Interpreting diagnostic tests

[3]

[Anonymous], ART SCAL ETH SUCC CH

[4]

[Anonymous], ABB REAL TIM TM HIV