AIDS and Behavior

, Volume 15, Issue 7, pp 1397–1409

Once Daily Dosing Improves Adherence to Antiretroviral Therapy

Authors

    • Division of Infectious DiseaseUniversity Health Network
    • Dalla Lana School of Public HealthUniversity of Toronto
  • Maggie Li
    • Division of Infectious DiseaseUniversity Health Network
  • Sharon Walmsley
    • Division of Infectious DiseaseUniversity Health Network
    • Departments of Medicine and Health PolicyManagement and Evaluation, University of Toronto
  • Curtis Cooper
    • The University of Ottawa, The Ottawa Hospital Division of Infectious Diseases
  • Sandra Blitz
    • Division of Infectious DiseaseUniversity Health Network
  • Ahmed M. Bayoumi
    • Departments of Medicine and Health PolicyManagement and Evaluation, University of Toronto
    • Centre for Research on Inner City Health, The Keenan Research Centre in the Li Ka Shing Knowledge Institute and Division of General Internal MedicineSt. Michael’s Hospital
  • Sean Rourke
    • Departments of Medicine and Health PolicyManagement and Evaluation, University of Toronto
    • Centre for Research on Inner City Health, The Keenan Research Centre in the Li Ka Shing Knowledge Institute and Division of General Internal MedicineSt. Michael’s Hospital
    • Ontario HIV Treatment Network
    • Department of PsychiatryUniversity of Toronto
  • Sergio Rueda
    • Centre for Research on Inner City Health, The Keenan Research Centre in the Li Ka Shing Knowledge Institute and Division of General Internal MedicineSt. Michael’s Hospital
    • Ontario HIV Treatment Network
  • Anita Rachlis
    • Departments of Medicine and Health PolicyManagement and Evaluation, University of Toronto
    • Division of Infectious DiseasesSunnybrook Health Sciences Centre
  • Nicole Mittmann
    • Health Outcomes and Pharmaco Economics (HOPE) Research CentreSunnybrook Health Sciences Centre
  • Marek Smieja
    • Department of Pathology and Molecular MedicineMcMaster University
  • Evan Collins
    • Division of Infectious DiseaseUniversity Health Network
    • Department of PsychiatryUniversity of Toronto
  • Mona R. Loutfy
    • Departments of Medicine and Health PolicyManagement and Evaluation, University of Toronto
    • Maple Leaf Medical Clinic
    • Women’s College Research InstituteWomen’s College Hospital
Original Paper

DOI: 10.1007/s10461-010-9818-5

Cite this article as:
Raboud, J., Li, M., Walmsley, S. et al. AIDS Behav (2011) 15: 1397. doi:10.1007/s10461-010-9818-5

Abstract

We studied the association of once-daily dosing with self-reported adherence among participants of the Ontario Cohort Study who were currently taking ART and who had completed a 90-min interviewer-administered questionnaire. Suboptimal adherence was defined as missing ≥1 dose of ART in the 4 days prior to the interview. Participants (n = 779) were 85% male, 69% men having sex with men, 67% white, median age 48 years (IQR 42–54), median years of ART 9 (IQR 5–13) and median CD4 count 463 cells/mm3 (IQR 320–638). Fifteen percent of participants reported suboptimal adherence in the 4 days prior to the interview. In a multivariable logistic regression model, participants on once daily regimens were half as likely to miss a dose during the 4 days prior to the interview. Other independent correlates of suboptimal adherence were younger age, lower positive social interaction and increased frequency of consuming > 6 alcoholic drinks on one occasion.

Keywords

AdherenceHIVAntiretroviral therapyOnce-daily dosing

Introduction

Although the introduction of combination antiretroviral therapy (ART) has led to dramatic reductions in disease-specific morbidity and mortality, several factors limit its universal success. Despite our advances in the use of combination ART, even in clinical trials 10–20% of participants fail to achieve maximal viral suppression and much of these relates to poor adherence. A high level of adherence (>95%) has previously been shown to be required to maintain full virologic suppression, maximize clinical response and prevent the development of drug resistance [14] but consistent adherence to ART regimens remains a challenge for many patients. While this degree of adherence may not be necessary with newer regimens [5, 6], physicians are still advised to encourage their patients to adhere as closely as possible to their ART regimens [7].

Factors that been shown to affect antiretroviral adherence include age [810], HIV risk factor [11], general health status [12], living with children [13], patients’ beliefs about their need for ART and their concerns about potential adverse effects of ART [14, 15], frequency and severity of symptoms [8, 16, 17], stressful life events [18], depression[1923], mistrust of the medical system [24], HIV stigma [25], and health literacy independent of education [26]. Further, adherence can vary over the course of HIV infection [27].

In the last decade, efforts have been made to improve antiretroviral adherence by reducing the number of pills and the frequency of administration of ART. This has been possible due of improved understanding of the pharmacokinetics of many of the agents, the use of different formulations with higher doses, and the use of coformulations. Adherence has been shown to be associated with pill count [28] and drug class [29] and has been inconsistently associated with once-daily dosing in controlled trial settings [3036]. There is limited assessment of adherence to current antiretroviral drugs and regimens outside of controlled trial settings [29, 37]. Furthermore, the interaction between antiretroviral dosing and demographic and social factors on adherence has yet to be investigated. We assessed the effects of dosing of antiretroviral regimens and demographic, clinical and social factors on adherence to antiretroviral medications in a large observational cohort study.

Methods

Study Population and Design

We studied participants enrolled in the Ontario HIV Treatment Network Cohort Study (OCS). The OCS was initiated in 2007 and included both new enrollees as well as participants who had previously participated in the HIV Ontario Observational Database (HOOD) and consented to continue their enrolment. The OCS is a voluntary clinic-based cohort study that currently enrolls at 11 active HIV care sites in Ontario.

Since October 2007, OCS participants have completed an annual interviewer-administered questionnaire. Participants at seven sites completed a 15 min “core” questionnaire, while participants at four sites in Toronto completed a 90 min “extended” questionnaire, which collected extensive socio-behavioral and demographic information including data on antiretroviral adherence.

At most sites, data were extracted from medical charts every 6 months and included specific laboratory data such as HIV diagnosis date, CD4 counts and viral load measurements, and clinical information including antiretroviral medications, diagnostic codes, adverse events and hospitalizations. At sites which had established databases for patient care, data for consenting participants were downloaded directly into the OCS.

The current analysis included all participants on ART who had completed the extended questionnaire at enrolment.

Outcome Measures

Adherence was assessed by self-report with the AIDS Clinical Trials Group (ACTG) Adherence Questionnaire. Suboptimal adherence in this analysis was defined as having missed any doses of antiretroviral medications during the 4 days prior to the interview. Secondary measures of adherence included the number of days participants missed taking all of their doses during the 4 days prior to the interview, whether or not antiretroviral medications were missed during the past weekend, the last time the participant missed any medications, whether or not the antiretroviral medications required special instructions and the frequency of following the special instructions.

Exposure Variables

The exposure variable of primary interest was the frequency of dosing of the antiretroviral regimen, dichotomized into categories of once daily vs more than once daily. This classification was based on the antiretroviral medications participants were taking at the time of the interview. If any single antiretroviral medication in the regimen was taken more frequently than once daily, the regimen was classified as more than once daily.

Demographic covariates of interest included age, gender, HIV risk factor, race, region of birth, immigration status, education, employment, income and housing. Clinical variables included general health, years of HIV infection, viral load, CD4 count, years of ARV therapy and protease inhibitor (PI)- vs. non-nucleoside reverse transcriptase inhibitor (NNRTI)-based ART regimens. The number of HIV-related symptoms the participant experienced in the past 4 weeks was tabulated in addition to the number of symptoms that “bothered the patient a lot” using the ACTG Symptom Distress Questionnaire [38]. “Binge” drinking was defined as the consumption of six or more alcoholic drinks on one occasion.

A number of psychosocial instruments were incorporated into the OCS questionnaire. The MOS Social support survey measures emotional/informational support, tangible support, positive social interaction (Appendix A) and affectionate support [39]. Scores were transformed to a 0–100 scale with the following formula: 100 × (observed score − minimum possible score)/(maximum possible score − minimum possible score). The Centre for Epidemiologic Studies Depression Scale (CES-D) [40] is a survey of 20 depressive symptoms, for which the frequency of each symptom during the past week was ranked for a total score from 0 to 60. A total CES-D score of 16 or greater is indicative that the participant potentially has clinical depression. The Brief COPE is 28-item instrument [41] which is a subset of the COPE inventory, measuring adaptive coping (active coping, planning, emotional support, instrumental support, positive reframing, acceptance, humour and religion) and maladaptive coping (venting, denial, substance use, behavioural disengagement, self-distraction and self-blame) [42]. The Pearlin Mastery scale is a seven item scale measuring the amount of control participants feel they have over various situations [43]. A modified version of the Berger HIV Stigma survey is a 16 item survey of stigma associated with HIV disease [44]. The National Population Health Survey Stress Questionnaire is a 34-item survey of stress measuring recent life events (10 questions), chronic strains (17 questions) and childhood adversities (7 questions) [9].

Statistical Methods

Baseline demographic, clinical and social characteristics were summarized using medians and interquartile ranges (IQR) for continuous variables and frequencies and percentages for categorical variables for the entire cohort. Characteristics were compared between participants who had missed at least one dose of any antiretroviral drug in the past 4 days and those who had not missed any doses in the past 4 days using chi square tests, Fisher’s exact test or Cochran-Armitage test for trend as appropriate for categorical variables and Wilcoxon rank sum tests for continuous variables. Characteristics were also compared between participants on once-daily dosing to those on more than once-daily dosing using chi square tests or Fisher’s exact test as appropriate for categorical variables and Wilcoxon rank sum tests for continuous variables.

Logistic regression models were used to determine the odds ratio of suboptimal adherence during the 4 days prior to the interview associated with once-daily dosing and to identify other independent correlates of suboptimal adherence. Covariates which were significant in univariate models with a significance level <0.10 were considered as candidates for inclusion in the multivariable logistic regression model.

Results

Description of Study Population

2127 participants in the OCS completed baseline interviews between October 2007 and May 2009. Of these, 913 completed the 90 min “extended” questionnaire which contained the questions regarding antiretroviral adherence. Seven hundred and eighty-three participants were taking ART at the time of the interview. Four participants who responded “don’t know” to the adherence question were excluded, for a final analysis sample size of 779.

The demographic, clinical and social characteristics of the study population are summarized in Table 1. The study population was 85% male, with a median age of 48 (IQR 42–54), 69% men having sex with men, 70% Caucasian and 63% Canadian born. Sixty-eight percent of participants had some post secondary education, 43% were employed and the median household income was $40,000–60,000 CDN. Ninety-seven percent reported living in a house, condominium or apartment, 3% lived in a room in house, motel, boarding house, group home or supportive housing and less than one percent reported being homeless. Using the SF-36, 46% of the participants reported very good or excellent health but 38% satisfied screening criteria for possible clinical depression according to the CES-D. Alcohol use was moderate and substance use within the past 6 months was reported by 16% of participants. The numbers of recent stressful life events, ongoing problems and early childhood adversities were indicative of moderate chronic stress [9].
Table 1

Baseline characteristics of study cohort by adherence in past 4 days

Variables

Total n = 779

Adherenta in past 4 days n = 664

Non-adherent in past 4 days n = 115

Test statistic

P value

Demographics

 Dosing of ARV regimen

  Once daily

233 (30%)

197 (32%)

26 (23%)

3.14

0.08

  More than once daily

516 (70%)

427 (66%)

86 (77%)

  

 Gender

  Male

658 (85%)

566 (85%)

92 (80%)

2.16

0.14

 Age group

  <30 years

21 (3%)

11 (2%)

10 (9%)

3.28

0.001

  30–40 years

115 (15%)

96 (14%)

19 (17%)

  

  40–50 years

311 (40%)

264 (40%)

47 (41%)

  

  >50 years

332 (43%)

293 (44%)

39 (34%)

  

 Risk factor

  Men having sex with men

536 (69%)

469 (71%)

67 (58%)

6.98

<0.01

  Heterosexual

74 (10%)

61 (9%)

13 (11%)

0.51

0.47

  Injection drug use

49 (6%)

40 (6%)

9 (8%)

0.54

0.46

  Blood Product

22 (3%)

17 (3%)

5 (4%)

1.14

0.29

  Endemic country of origin

115 (15%)

91 (14%)

24 (21%)

3.99

0.05

 Region of birth

  Canada/North America

488 (63%)

424 (64%)

64 (56%)

8.20

0.08

  Caribbean/South, Central America

99 (13%)

83 (13%)

16 (14%)

  

  Africa

86 (11%)

65 (10%)

21 (18%)

  

  Europe

68 (9%)

59 (9%)

9 (8%)

  

  Asia and Middle East

37 (5%)

33 (5%)

4 (3%)

  

 Immigrant within 10 years of first visit

59 (8%)

49 (7%)

10 (9%)

0.24

0.62

 Race

  White

521 (70%)

459 (72%)

62 (57%)

18.56

0.002

  Black

125 (17%)

100 (16%)

25 (23%)

  

  South/East Asian

42 (6%)

36 (6%)

6 (6%)

  

  Aboriginal

14 (2%)

8 (1%)

6 (6%)

  

  Latin American

35 (5%)

29 (5%)

6 (6%)

  

  Arab/West Asian

8 (1%)

5 (<1%)

3 (3%)

  

Social factors

 Living condition

  House/condo/apartment

753 (97%)

645 (97%)

108 (94%)

3.81

0.16

  Room

20 (3%)

14 (2%)

6 (5%)

  

  Homeless

6 (<1%)

5 (<1%)

1 (<1%)

  

 Live with children

87 (11%)

69 (10%)

18 (16%)

2.71

0.10

 Education

  High School and lower

251 (32%)

202 (30%)

49 (43%)

3.20

<.01

  Some/completed college

298 (38%)

253 (38%)

45 (39%)

  

  University and higher

230 (30%)

209 (31%)

21 (18%)

  

 Working for pay

339 (43%)

289 (44%)

50 (43%)

0.0001

0.99

 Household gross yearly income

  <$40,000

290 (39%)

241 (38%)

49 (46%)

1.36

0.17

  $40,000–$60,000

184 (25%)

160 (25%)

24 (22%)

  

  $60,000–$80,000

85 (11%)

69 (11%)

16 (15%)

  

  >$80,000

187 (25%)

166 (26%)

21 (19%)

  

Alcohol use

 Frequency of alcohol consumption

  Never

211 (27%)

181 (27%)

30 (26%)

−0.51

0.61

  Monthly or less

182 (23%)

152 (23%)

30 (26%)

  

  2–4 times a month

181 (23%)

159 (24%)

22 (19%)

  

  2–3 times a week

103 (13%)

90 (14%)

13 (11%)

  

  4 or more times a week

102 (13%)

82 (12%)

20 (17%)

  

 Frequency of having 6+ drinks

  Never

543 (70%)

477 (72%)

66 (57%)

−3.43

<.001

  Less than monthly

140 (18%)

116 (17%)

24 (21%)

  

  Monthly

61 (8%)

44 (7%)

17 (15%)

  

  Weekly

26 (3%)

20 (3%)

6 (5%)

  

  Daily or almost daily

8 (1%)

6 (<1%)

2 (2%)

  

 With alcohol consequence

73 (9%)

55 (8%)

18 (16%)

6.26

0.01

Drug useb

 Substance use in last 6 months

128 (16%)

106 (16%)

22 (19%)

0.71

0.40

 Cocaine use in the last 6 months

64 (8%)

52 (8%)

12 (10%)

0.87

0.35

 Crack use in the last 6 months

35 (4%)

25 (4%)

10 (9%)

5.55

0.02

 Club drug use in the last 6 months

51 (7%)

43 (6%)

8 (7%)

0.04

0.85

 Methamphetamine use in the last 6 months

33 (4%)

28 (4%)

5 (4%)

0.004

0.95

Symptom distress

 Number of symptoms

8 (4, 11)

8 (4, 11)

9 (5, 13)

6.42

0.01

 Number of bothersome symptoms

3 (1, 6)

3 (1, 6)

5 (1, 8)

6.44

0.01

General health

 Excellent/very good

354 (46%)

310 (47%)

44 (38%)

−2.34

0.02

 Good

254 (33%)

218 (33%)

36 (31%)

  

 Fair/poor

170 (22%)

135 (20%)

35 (30%)

  

Depression

 No. of moderate depressive symptoms

3 (1, 6)

3 (1, 6)

4 (1, 8)

3.63

0.06

 Total CES-D score

12 (6, 21)

11 (6, 20)

15 (7, 26)

4.85

0.03

 Potential clinical depression

297 (38%)

245 (36%)

52 (45%)

2.87

0.09

Social support

 Emotional Support Score

72 (47, 88)

72 (50, 91)

63 (28, 78)

12.52

<.001

 Tangible Support Score

69 (31, 100)

75 (38, 100)

56 (25, 75)

7.88

<.01

 Affectionate Support Score

75 (50, 100)

75 (59, 100)

67 (33, 92)

8.32

<.01

 Positive Social Interaction Score

75 (50, 100)

75 (50, 100)

56 (25, 81)

17.26

<.0001

 Total Social Support Score

68 (45, 88)

71 (46, 89)

59 (32, 78)

13.74

<.001

Stigma

 Stigma Score (16–80)

48 (40, 57)

48 (39, 56)

49 (42, 57)

1.91

0.17

Stress

 Number of recent life events

0 (0, 1)

0 (0, 1)

0 (0, 2)

2.41

0.12

 Number of ongoing problems

3 (1, 5)

2 (1, 4)

3 (2, 6)

12.64

<.001

 Number of early childhood adversities

1 (0, 2)

1 (0, 2)

2 (1, 3)

5.75

0.02

 Total Number of Stressors

5 (2, 8)

4 (2, 7)

6 (3, 9)

15.92

<.0001

Mastery

 Total Mastery Score

20 (17, 21)

20 (17, 21)

20 (16, 21)

0.49

0.49

Coping

 Adaptive Coping Score (0–48)

29 (22, 35)

29 (23, 35)

27 (20, 34)

2.50

0.11

 Maladaptive Coping Score (0–36)

9 (5, 13)

8 (5, 13)

10 (7, 16)

12.94

<.001

Clinical

 Years since HIV positive test

12 (7 , 18)

12 (7, 18)

13 (8, 17)

0.31

0.58

 CD4 count (cells/mm3)

463 (320, 638)

461 (331, 634)

467 (288, 639)

0.13

0.72

 Viral load (<50 copies/ml)

673 (86%)

584 (88%)

89 (77%)

9.29

<0.01

 Years of ARV therapy

9 (5, 13)

9 (4, 13)

10 (6, 13)

0.16

0.69

 Year of ARV initiation

1998 (1995, 2003)

1999 (1994, 2004)

1998 (1995, 2003)

0.54

0.46

 First regimen

152 (20%)

133 (20%)

19 (17%)

0.77

0.38

 Second regimen or greater

627 (80%)

531 (80%)

96 (83%)

 ARV with special instructions

502 (66%)

425 (65%)

77 (69%)

0.64

0.42

 Current Regimen

  NNRTI based

333 (43%)

293 (44%)

40 (35%)

3.50

0.06

  PI based

502 (64%)

420 (63%)

82 (71%)

2.73

0.10

Results are reported as frequency and percent, or median and interquartile range

ARV antiretroviral, NNRTI non nucleoside reverse transcriptase inhibitor, PI protease inhibitor

aAdherence and non-adherence in the past 4 days is a binary response to the question, “Have you missed any of your doses of antiretroviral medications over the past 4 days?”

bDoes not include alcohol use

Summary of Cohort’s Adherence

Of the 779 participants, 115 (15%) had missed at least one dose of their ART regimen in the 4 days prior to the interview. Adherence varied from site to site: the proportions of individuals missing a dose in the past 4 days at each of the four sites were 11, 14, 19 and 21% (test statistic = 8.92, P = 0.03). The other adherence measures are summarized in Table 2.
Table 2

Summary of adherence measures

Question

Response categories

N (%)

Have you missed any of your doses of antiretroviral medications over past 4 days?

No

664 (85%)

Yes

115 (15%)

Don’t know

4 (1%)

If you missed doses during the past 4 days, on how many days have you missed taking all of your doses?

None

52 (45%)

One day

44 (38%)

Two days

9 (8%)

Three days

2 (2%)

Four days

5 (4%)

How closely did you follow your specific antiretroviral schedule during the past 4 days?

Never

13 (2%)

Some of the time

22 (3%)

About half of the time

14 (2%)

Most of the time

196 (25%)

All of the time

537 (69%)

When was the last time you missed any of your medications?

Never skip medications

235 (30%)

More than 3 months ago

143 (18%)

1–3 months ago

119 (15%)

2–4 weeks ago

95 (12%)

1–2 weeks ago

60 (8%)

Within the past week

129 (16%)

Don’t know

2 (0%)

Did you miss any of your anti-HIV medications last weekend?

No

702 (90%)

Yes

76 (10%)

Don’t know

5 (1%)

Do any of your anti-HIV medications have special instructions such as “take with food” or “on an empty stomach” or “with plenty of fluids”?

No

260 (33%)

Yes

504 (64%)

Don’t know

19 (2%)

If you have special instructions, how often did you follow those special instructions over the last 4 days?

Never

14 (3%)

Some of the Time

25 (5%)

About half of the time

10 (2%)

Most of the time

105 (21%)

All of the time

349 (69%)

Correlates of Adherence

Characteristics of participants who had missed at least one dose in the four days prior to the interview were compared to those of participants who did not (Table 1). In a univariate logistic regression model, there was a trend towards participants on once-daily dosing being less likely to miss a dose during the past four days (OR = 0.66, 95% CI (0.41, 1.05), Wald chi square = 3.11, P = 0.08) (Table 3). Covariates which were associated with an increased likelihood of missing at least one dose in the past 4 days were age <30 years, an African birthplace, Black or Aboriginal race, binge drinking more than once a month, number of symptoms of depression, number of ongoing problems, number of early childhood adversities, total number of stressors and a higher maladaptive coping score. An HIV risk factor of men having sex with men, university education and higher levels of all types of social support (emotional, tangible, affectionate and positive social interaction) were associated with decreased likelihoods of suboptimal adherence (Table 3).
Table 3

Univariate logistic regression models with outcome of non-adherence in past 4 days

Variables

Odds Ratio

(95% CI)

Test Statistic

P value

Once daily therapy (vs > once daily)

0.66

(0.41, 1.05)

3.11

0.08

Male vs Female

0.69

(0.41, 1.14)

2.15

0.14

Age

    

 <30

6.84

(2.73, 17.14)

16.80

<.0001

 30–40

1.49

(0.82, 2.70)

1.71

0.19

 40–50

1.34

(0.85, 2.11)

1.56

0.21

 >50

1

1

  

HIV risk factora

 Men having sex with men

0.56

(0.38, 0.85)

7.63

<0.01

 Heterosexual contact

1.71

(1.13, 2.60)

6.36

0.01

 Injection drug use

1.23

(0.58, 2.59)

0.29

0.59

 Receipt of blood products

1.62

(0.76, 3.49)

1.54

0.21

 Other

2.92

(0.53, 16.13)

1.51

0.22

Region of birth

 Canada/North America

1

1

  

 Africa

2.14

(1.23, 3.74)

7.15

<0.01

 Caribbean, South or Central America

1.28

(0.70, 2.32)

0.65

0.42

 Asia or Middle East

0.80

(0.28, 2.34)

0.16

0.69

 Europe

1.01

(0.48, 2.14)

0.00

0.98

Immigrated within 10 years of questionnaire

1.20

(0.59, 2.43)

 

0.62

Race

 White

1

1

  

 Black

1.85

(1.11, 3.09)

5.55

0.02

 South/East Asian

1.23

(0.50, 3.05)

0.21

0.65

 Aboriginal

5.55

(1.86, 16.54)

9.48

<0.01

 Latin American

1.53

(0.61, 3.84)

0.83

0.36

 Arab/West Asian

4.44

(1.04, 19.05)

4.03

0.04

Education

 High school

1

1

  

 Some/completed college

0.73

(0.47, 1.14)

1.87

0.17

 University and higher

0.41

(0.24, 0.72)

9.99

<0.01

Frequency of 6 alcoholic drinks on one occasion

 Never

1

1

  

 Less than monthly

1.50

(0.90, 2.49)

2.40

0.12

 Monthly and more often

2.58

(1.53, 4.36)

12.57

<0.001

Consequence of Alcohol use

2.05

(1.16, 3.65)

6.05

0.01

No. of moderate depressive symptoms

1.07

(1.02, 1.12)

7.48

<0.01

General Health

 Very good or excellent

1

1

  

 Good

1.16

(0.72, 1.87)

0.39

0.53

 Fair or poor

1.83

(1.12, 2.97)

5.86

0.02

Social Supportb

 Emotional Support Score

0.88

(0.82, 0.94)

14.38

0.0001

 Tangible Support Score

0.93

(0.88, 0.98)

6.88

0.01

 Affectionate Support Score

0.92

(0.87, 0.97)

8.08

<0.01

 Positive Social Interaction Score

0.87

(0.82, 0.93)

18.39

<0.0001

 Transformed Total Social Score

0.87

(0.81, 0.93)

14.93

0.0001

Stress

 Number of recent life events

1.18

(0.98, 1.43)

3.11

0.08

 Number of ongoing problems

1.17

(1.08, 1.26)

14.27

<0.001

 Number of early childhood adversities

1.16

(1.03, 1.32)

5.77

0.02

 Total Number of stressors

1.11

(1.05, 1.17)

15.09

<0.001

Maladaptive Coping Score (0–36)

1.06

(1.03, 1.09)

13.52

<0.001

Clinical characteristics

 PI based therapy

1.43

(0.93, 2.21)

2.63

0.11

 NNRTI based therapy

0.67

(0.44, 1.02)

3.56

0.06

 Years since HIV positive test

1.01

(0.98, 1.04)

0.32

0.57

 Years of antiretroviral therapy

1.01

(0.97, 1.04)

0.18

0.67

* Non-adherence in the past 4 days is defined as answer yet to the question: “Have you missed any of your doses of antiretroviral medications over the past 4 days?”

aNot mutually exclusive

bPer 10 unit increase on a 0–100 scale

In a multivariable logistic regression model, once daily dosing was found to be independently associated with a decreased likelihood of missing at least one dose in the past 4 days (OR = 0.47, 95% CI (0.28, 0.78), Wald chi square = 8.33, P = 0.004) after adjusting for positive social interaction score, age, and frequency of binge drinking (Table 4). While several social support and mental health measures were associated with suboptimal adherence in the univariate analyses including the number of ongoing problems, total number of stressors, depression score, and maladaptive coping score, the positive social interaction score had the strongest association with suboptimal adherence in the multivariable model.
Table 4

Multivariable logistic regression with outcome of non-adherence in the past 4 days

Covariate

Odds ratio

95% CI

Test statistic

P value

Age group

 <30

9.26

(3.35, 25.57)

18.45

<.0001

 30–40

1.74

(0.92, 3.29)

2.85

0.09

 40–50

1.37

(0.85, 2.22)

1.66

0.20

 >50

1

   

Positive social interaction scorea

0.86

(0.81, 0.93)

17.44

<0.0001

Frequency of >6 alcoholic drinks on one occasion

 Never

1

(0.73, 2.19)

0.70

0.40

 Less than monthly

1.27

(1.44, 4.42)

10.41

0.001

 Monthly and more

2.52

   

Once daily dosing

0.47

(0.28, 0.78)

8.33

0.004

aPer 10 unit increase on a scale from 0 to 100

Table 5 compares characteristics of participants by frequency of dosing. Participants on regimens with once-daily dosing were more likely to be younger, to have immigrated within the past 10 years, to have used substances within the past 6 months, to consume more than six alcoholic drinks on one occasion more frequently, to have higher levels of stigma, more ongoing problems and a higher total number of stressors. In multivariable logistic regression models, age was the only variable which was independently associated with once-daily dosing (data not shown).
Table 5

Characteristics at baseline by frequency of dosing

Characteristics

More than once daily dosing

Once daily dosing

Test statistic

P value

Gender

 Male

438 (85%)

188 (84%)

0.07

0.80

Age group

 <30 years

7 (1%)

14 (6%)

6.03

<.0001

 30–40 years

61 (12%)

47 (21%)

  

 40–50 years

197 (38%)

97 (43%)

  

 >50 years

251 (49%)

65 (29%)

  

HIV risk factor

 Men having sex with men

368 (71%)

150 (67%)

1.22

0.27

 Heterosexual

290 (56%)

123 (55%)

0.07

0.79

 Injection drug use

30 (6%)

17 (8%)

0.86

0.35

 Blood product recipient

31 (6%)

8 (4%)

1.82

0.18

Immigrated within 10 years

29 (6%)

24 (11%)

6.18

0.01

Living condition

  

0.43

0.81

 House/condo/apartment

501 (97%)

217 (97%)

  

 Room

12 (2%)

4 (2%)

  

 Homeless

3 (1%)

2 (1%)

  

Working for pay

215 (42%)

110 (49%)

3.71

0.05

Education

  

1.16

0.24

 High school or less

172 (33%)

67 (30%)

  

 Some/completed college

198 (38%)

84 (38%)

  

 Some/completed university

146 (28%)

72 (32%)

  

With Alcohol Consequence

48 (9%)

22 (10%)

0.60

0.81

Substance use within 6 months

73 (14%)

46 (21%)

4.84

0.03

ARV with special instructions

330 (66%)

150 (68%)

0.37

0.54

Potential clinical depression

201 (39%)

83 (37%)

0.20

0.66

Household gross yearly income

  

−0.54

0.59

 <$40,000

198 (40%)

70 (33%)

  

 $40,000–$60,000

117 (24%)

62 (29%)

  

 $60,000–$80,000

49 (10%)

33 (15%)

  

 >$80,000

132 (27%)

49 (23%)

  

Frequency of alcohol consumption

  

−0.50

0.62

 Never

148 (29%)

53 (24%)

  

 Monthly or less

115 (22%)

55 (25%)

  

 2–4 times per month

118 (23%)

58 (26%)

  

 2–3 times per week

65 (13%)

30 (13%)

  

 4 or more times per week

70 (14%)

27 (12%)

  

Frequency of having >6 drinks

  

−2.47

0.01

 Never

367 (71%)

143 (64%)

  

 Less than monthly

95 (18%)

42 (19%)

  

 Monthly or more

53 (11%)

38 (17%)

  

General health

  

1.92

0.05

 Excellent/very good

220 (43%)

116 (52%)

  

 Good

179 (35%)

63 (28%)

  

 Fair/poor

116 (22%)

44 (20%)

  

Social support

 Emotional support score

72 (46–88)

75 (47–91)

0.41

0.52

 Tangible support score

69 (38–94)

69 (25–100)

0.03

0.86

 Affectionate support score

75 (50–100)

75 (50–100)

0.03

0.85

 Positive social interaction score

75 (50–100)

75 (50–100)

0.82

0.82

 Total social support score

70 (46–88)

70 (43–89)

0.04

0.84

Stigma

 Stigma score

48 (40–56)

50 (43–58)

6.98

<.01

Stress

 Number of recent life events

0 (0–1)

0 (0–1)

3.98

0.05

 Number of ongoing problems

2 (1–4)

3 (2–5)

5.54

0.02

 Number of early childhood adversities

1 (0–2)

1 (0–2)

0.04

0.83

 Total number of stressors

4 (2–7)

5 (3–8)

4.43

0.04

Mastery

 Total Mastery score

20 (17–21)

20 (17–21)

0.07

0.79

Discussion

In our study population of HIV positive individuals in clinical care, we found high levels of suboptimal adherence according to various definitions of adherence. While only 15% of study participants reported missing an antiretroviral dose during the 4 days prior to the interview, 70% reported ever missing a dose, with 54% of those participants missing a dose within the past 4 weeks. Of individuals who reported special instructions for taking their ART, only 69% followed the instructions all of the time. These levels of suboptimal adherence are similar to those reported in other studies.

Our study confirms research demonstrating the association between once daily dosing of ART and improved adherence in the cohort [30] and clinical trial settings [3134]. We found that even among a cohort of well-housed and well-educated individuals who were willing and able to complete a 90 min questionnaire, participants taking their ART once daily were about half as likely to miss a dose in the prior 4 days as those taking their ART twice a day or more. While the frequency of dosing only tended towards significance in univariate models, it was significant in multivariable models because individuals who were younger or more likely to consume >6 alcoholic drinks on one occasion more than once a month were more likely to have used once daily regimens.

Despite the small number of participants less than 30 years of age, age was highly predictive of suboptimal adherence. Other studies have also shown younger age to be associated with poorer adherence to antiretroviral therapy [810]. This may be because older patients have more experience with HIV disease and have recognized the benefits of therapy while younger patients view HIV as a chronic disease with many options for therapy and are thus less careful about preserving the effects of their medications.

Several studies have shown the association between alcohol use and poor adherence. In our study, binge drinking was associated with suboptimal adherence. Other studies have shown that the association of alcohol and non-adherence may be a mediator to the association of life stressors or other social maladaptive factors and non-adherence as a method of coping. This theory is supported in our study as the effect of alcohol on adherence is reduced in the multivariable analysis relative to the univariate analysis after adjusting for social support. Other studies have found that substance use is associated with non-adherence. In our study, we did not find this association, possibly due to the fact that individuals using illicit drugs who were able to participate in a study with a 90 min questionnaire might not be representative of other substance users. Further, the effect of substance use on adherence varied according to the substance used. In our study, the use of crack cocaine was associated with suboptimal adherence in univariate models but the use of cocaine in other forms, club drugs or methamphetamines were not.

Novel findings of our study include the demonstration that participants at higher risk of non-adherence were more likely to be prescribed once-daily dosing. It is important to adjust for participant characteristics in observational studies to control for confounding by indication, in which physicians selectively prescribe once daily regimens to patients at higher risk of non-adherence. While the trends in the odds of suboptimal adherence associated with once daily therapy were not statistically significant by age group, the possibility of differential impacts of once daily therapy according to patient characteristics is worth investigating in other research studies.

It should be noted that a missed dose in a once-a-day regimen results in a 24 h period of lost coverage which may lead to low drug levels whereas a missed dose in a twice-a-day or three-times-a-day regimen results in a shorter period without medication.

Significant strengths of this research are the detailed information on psychosocial measures for the participants and the diverse nature of our study population which includes individuals from multiple risk groups, a variety of ethnic backgrounds and immigrants from countries with high HIV prevalence rates.

There are several limitations of our study. There was the potential for misclassification of once daily regimens as non-once daily due to the retrospective collection of antiretroviral prescription data from medical charts. The cross-sectional nature of the study precluded us from determining a temporal or causal relationship between many exposure variables and adherence. Our results may not be generalizable to all HIV positive individuals since our study population consisted of relatively healthy, well-educated, well-housed individuals who sought care in a large urban centre and volunteered to do a 90-min questionnaire. Small numbers of homeless individuals and Aboriginal peoples prevented us from studying these groups in more detail.

Our findings confirm that one strategy for improving adherence is the use of once daily dosing of ART which could be of benefit to these higher risk populations. Furthermore, our findings underscore the importance of considering demographic and social factors when considering starting or managing ART. Patients that are younger, have less social support and who have a tendency for binge drinking may require additional monitoring of ART adherence and could benefit from adherence counseling and management prior to the start of ART.

Acknowledgments

The OHTN Cohort Study (Principal Investigator, Dr. Sean B. Rourke) is supported by the AIDS Bureau—Ontario Ministry of Health and Long-Term Care. Data collection sites and members of the Scientific Steering Committee include: Drs. Irving Salit and Janet Raboud (Toronto General Hospital), Dr. Ahmed Bayoumi (St. Michael’s Hospital), Drs. Mona Loutfy, Graham Smith, Tony Antoniou and Fred Crouzat (Maple Leaf Medical Clinic), Dr. Anita Rachlis, Dr. Nicole Mittmann (Sunnybrook Health Sciences Centre), Dr. Wendy Wobeser (Kingston General Hospital), Dr. John Cairney (McMaster University and Centre for Addiction and Mental Health), Dr. Liviana Calzavara (University of Toronto), Dr. Curtis Cooper (University of Ottawa), Dr. Marek Smieja (McMaster University, Hamilton), Dr. Ken Logue (St Clair Medical Associates), Dr. Don Kilby (University of Ottawa Health Services), Dr. Anurag Markenday (St. Joseph’s Health Care, London), Dr. Roger Sandre (Sudbury Regional Hospital and Dr. Jeff Cohen (Windsor Clinic).

We gratefully acknowledge all of the people living with HIV who volunteer to participate in the OHTN Cohort Study and the work and support of the inaugural OCS Governance Committee: Darien Taylor (Chair), Dr. Evan Collins, Dr. Greg Robinson, Shari Margolese, Patrick Cupido, Tony Di Pede, Rick Kennedy, Michael Hamilton, Ken King, Brian Finch, Lori Stoltz, Dr. Ahmed Bayoumi, Dr. Clemon George, and Dr. Curtis Cooper. We thank all the interviewers, data collectors, research associates and coordinators, nurses and physicians who provide support for data collection and extraction. The authors wish to thank the OHTN staff and their teams for data management and IT support (Mark Fisher, Director, Data Systems) and OCS management and coordination (Virginia Waring, Project Manager, OCS). The viral load data in the OCS was supplemented through a linkage with the viral load database of the Ontario Agency for Health Protection and Promotion.

Six investigators are also the recipients of salary support from the Ontario HIV Treatment Network (JR, SW, CC), the Canadian Institutes of Health Research (MRL, MS, SR).

Conflicts of Interest

There are no conflicts of interest related to this paper and project.

Copyright information

© Springer Science+Business Media, LLC 2010