Www.ias2013.org Kuala Lumpur, Malaysia, 30 June - 3 July 2013 Factors related to non-adherence to antiretroviral (ART) drugs among adult ART clients attending.

www.ias2013.org Kuala Lumpur, Malaysia , 30 June - 3 July 2013

Factors related to non-adherence to antiretroviral (ART) drugs among adult ART clients attending 18 facilities in Tanzania, Uganda and Zambia

Julie Denison, FHI360Olivier Koole, Institute of Tropical Medicine


Study Team

• FHI 360, USA• Muhimbili University of Health and Allied

Sciences, Tanzania• Infectious Diseases Institute, Makerere University

Medical School, Uganda• Tropical Diseases Research Centre, Zambia• Institute of Tropical Medicine, Belgium • Massachusetts General Hospital, USA • Centers for Disease Control and Prevention, USA

• Gideon Kwesigabo• Fred Wabwire-Mangen• Modest Mulenga• Julie Denison• Sharon Tsui• Cindy Geary• Meng Wang• Ya Diul Mukadi• Leine Stuart• Eric Van Praag• Kwasi Torpey• Joris Menten• Robert Colebunders• Olivier Koole• David Bangsberg• Andrew Auld• Simon Agolory• Seymour Williams• Jonathan Kaplan• Aaron Zee


Context

• Massive ART scale-up: 8 million people on ART

• High levels of ART adherence required for viral suppression and good clinical outcomes

• Adherence levels higher in sub-Saharan Africa– Mills et al pooled estimate: 55% N. America; 77%

SSA (JAMA 2006)– Nachega et al (Curr Opin HIV/AIDS 2010)


Objectives

Primary To characterize the current level of adherence

among ART clinic patients across multiple program settings

SecondaryTo identify important factors associated with

ART adherence, including both individual risk factors and program characteristics


Study Population and Sites

• Study Population– 18+ years at ART

initiation at study site– Initiated 3 ARVs at least

6 months prior to data collection

• Study sites– 3 countries – 6 sites per country,

purposively selected

UGANDA

TANZANIA

ZAMBIA


Methods

Adherence

Self-reports: 3-day, 30-day, 30-day VAS, Case Index

Provider estimates: 30-day

Pharmacy data: medication possession ratio (mpr)

Known risk factors

Stigma, depression, alcohol use, social support, disclosure, wealth index

Pre ART CD4 count, WHO Stage, ART regimen, pill burden, number of years on ART

Cross-sectional design: 250 patients per site systematically selected April-August 2011


Methods

• Viral load testing conducted at six sites

• June to July 2011, Health Care Manager questionnaire conducted at 18 sites – Rural/Urban– Level and type of health facility– Community ART dispensing– ART initiation requirements– Stockouts– Type of clinic staff (i.e. Lay counselors)


Analysis

1. Adherence levels described with cutoff points determined by ROC curve analysis using HIV viral load (VL) < />=1,000 copies/ml

2. Optimal adherence measures selected based on degree of association with viral load failure

3. Multivariate regression analysis, adjusted for site-level clustering, assessed associations between non-adherence and individual and program factors


Patient AccountingNumber of ART clients screened 6,825Number of ineligible ART clients 1,854 Initiated ART <6 months before study 1,529 Less than age 18 729 Did not speak a study language 783Number of eligible ART clients 4,971 Number who did not consent 482 Number who consented and are in the analysis 4,489

Country Number of sites Number of participants

Tanzania 7 1498

Uganda 6 1495

Zambia 6 1496


Selected Participant CharacteristicsCharacteristic 4,489

n (%)

Age (year), mean (SD)

40.9 (9.4)

Sex, Female 3,045 (67.8)Years on ART < 2 1,102 (24.6) 2-5 2,249 (50.1) >5 1,138 (25.4)Pre-ART CD4 <=250 3,020 (67.3) >250 567 (12.6) Missing 902 (20.1)

Characteristic 4,489n (%)

Current ART regimend4T(30),3TC,NVP 731 (16.3)TDF,3TC/FTC, EFV 699 (15.6)AZT,3TC,EFV 773 (17.2)AZT,3TC,NVP 1,427 (31.8)Other 684 (15.2)Missing 175 (3.4)


Selected Program Characteristics

Characteristic nTotal number of clinics

19

Setting Rural/peri-urban 6 Urban 13Number of adults on ART < 2000 8 2000-4000 7 > 4000 4

Characteristic nLevel of health facility National referral 4 Provincial/Regional 4 District 6 Primary/ community based

5

Type of health facility Government 10 Mission 5 Non-religious NGO 4


1. Levels of Incomplete Adherence

• 9.9% (148/1,496) had VL >1000 copies/ml

• VL failure ranged from 7.2%- 17.2% by study site


1. Levels of Incomplete Adherence

Adherence measure

ROC Cutoff

n/N % IncompleteAdherence

3-day 93% 340/4397 7.7

30-day 99% 611/4387 13.9 30-day VAS

99% 2401/4450 54.0

Case Index 11 751/4473 16.8Health provider 98% 2617/4415 59.3Pharmacy MPR 90% 1634/4070 40.1


2. Selection of Adherence Measures: Associations with Viral Load Failure

Adherence measure

Category VL Failure >1,000 copies/ml

OR (95% CI)

Self-Report

3-day <93% 1.40 (0.80,2.46)

> 93% 30-day <99% 1.13 (0.71,1.78) >99%

30-day VAS <99% 0.86 (0.57,1.30)>99%

Case Index <11 1.30 (0.84, 2.01)>11


2. Selection of Adherence Measures: Associations with Viral Load Failure

Adherence measure Category VL Failure >1,000

copies/mlOR (95% CI)

Provider Report <98% 1.57(1.02,2.41)

>98%

Pharmacy MPR <90% 1.48 (1.02,2.15)

>90%


3. Model Building: BivariateIncomplete Adherence Pharmacy MPRIndividual characteristicsSignificant and selected (p<0.1) Non –significant and not

selected (p>0.1)- Female sex- Younger than 35 years- On ART between 2-5 years- Greater internalized stigma- Positive screen for depression- Taking four or more pills daily- Missing pre-ART CD4 count

- Marital Status- Social support- CAGE alcohol abuse- DHS Wealth Index- Cost to clinic- Time to clinic- Disclosure- Pre-ART WHO-stage


3. Model Building: BivariateIncomplete Adherence Pharmacy MPR

None of the Program characteristics were significant at p<0.2

- Level of health facility- Type of health facility- Dispensing of ARVs at community level- Buddy needed for ART initiation- Lay counselor provides adherence counseling- ART stock out (past six months)- Location (urban/rural)


3. Model Building: MultivariableIncomplete Adherence Pharmacy MPR

Adjusted Odds Ratios (95% CI)

Age >35 0.85 (0.73-0.99)

Years on ART <2 1

2-5 1.21 (1.02-1.43)

5+ 0.88 (0.72-1.10)

Stigma High 1.26 (1.06-1.50)

Pill Burden 4+ daily 1.35 (1.12-1.62)

Pre-ART CD4 Missing 1.37 (1.15-1.63)


Conclusions

• Wide variability in adherence levels

• Pharmacy MPR and provider estimates associated with VL failure

• Interventions needed to– Reach younger ART clients– Support long-term adherence– Address internalized stigma


Special thanks to our funders, partners, study clinics and ART clients

Www.ias2013.org Kuala Lumpur, Malaysia, 30 June - 3 July 2013 Factors related to non-adherence to antiretroviral (ART) drugs among adult ART clients attending.

Documents

kuala lumpur

art regimen

art clinic patients

non adherence

antiretroviral art drugs

adult art clients

art crosssectional design

current level of adherence