Intervention Review

You have free access to this content

Telephone delivered interventions for reducing morbidity and mortality in people with HIV infection

  1. Sarah Gentry1,
  2. Michelle HMMT van-Velthoven2,
  3. Lorainne Tudor Car3,
  4. Josip Car2,*

Editorial Group: Cochrane HIV/AIDS Group

Published Online: 31 MAY 2013

Assessed as up-to-date: 3 APR 2013

DOI: 10.1002/14651858.CD009189.pub2


How to Cite

Gentry S, van-Velthoven MHMMT, Tudor Car L, Car J. Telephone delivered interventions for reducing morbidity and mortality in people with HIV infection. Cochrane Database of Systematic Reviews 2013, Issue 5. Art. No.: CD009189. DOI: 10.1002/14651858.CD009189.pub2.

Author Information

  1. 1

    Peninsula College of Medicine and Dentistry, Exeter, UK

  2. 2

    Imperial College London, Global eHealth Unit, Department of Primary Care and Public Health, School of Public Health, London, UK

  3. 3

    University of Split, Split, Croatia

*Josip Car, Global eHealth Unit, Department of Primary Care and Public Health, School of Public Health, Imperial College London, St. Dunstans Road, Hammersmith, London, W6 8RP, UK. josip.car@imperial.ac.uk.

Publication History

  1. Publication Status: New
  2. Published Online: 31 MAY 2013

SEARCH

 
Characteristics of included studies [ordered by study ID]
Collier 2005

MethodsRandomised controlled trial conducted over a 96 week period.


ParticipantsParticipants (n=282) with advanced HIV infection (CD4 count less than or equal to 200 cells/mm3 or >80,000 HIV RNA copies/mL of plasma at screening) who were initiating antiretroviral therapy (no previous use of lamivudine, nonnucleoside reverse transcriptase inhibitors, or protease inhibitors) were enrolled in this adherence substudy of a treatment trial. Participants were recruited from 30 sites in the United States of America, Puerto Rico and Italy, 48% identified as white, 32% black and 15% Hispanic and 19% of participants were women. There were no significant baseline differences between the intervention and control groups.


InterventionsParticipants were randomised to a usual support measures group (n=140) or a telephone intervention group (n=142). Participants in the usual care support group received their study site's usual adherence support measures. These may have included 35 minutes of in-person counselling, written material or informal telephone calls. Those in the telephone intervention group received their site's usual support measures plus up to 16 scripted telephone calls from study site members (mostly nurses) over the 96 week study period. Calls focused on each subject's medication taking behaviour and barriers to adherence, and developed individual strategies to increase adherence. Social support and assistance with medical side effects were also provided.


OutcomesOutcomes were virologic failure and self-reported adherence.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNo mention of sequence generation in the study.

Allocation concealment (selection bias)Unclear riskNo mention of allocation concealment in the study.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskBlinding of participants and personnel was unlikely to be feasible in this study, but we do not think this will have affected the results.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo mention of blinding to outcome assessment in the study.

Incomplete outcome data (attrition bias)
All outcomes
Low riskOf the 282 enrollees, 239 (85%) completed the trial, 35 (12%) discontinued prematurely, six (2%) died, and two (1%) discontinued for other reasons. The reasons for not completing the trial were similar between groups.

Selective reporting (reporting bias)Unclear riskProtocol of the study was unavailable.

Other biasHigh riskWe judged the risk of contamination to be high. The same study site staff members were counselling participants in both groups. Subjects who were anticipated to have problems with adherence at all study sites received counselling. It is possible that the approach used for the usual support measures group may have been influenced by the additional support measures used for the calls group if the study site staff members inadvertently used counselling strategies they learned from the standardised script. According to a study site survey, 67% of study sites reported that they provided written materials to study participants as part of their usual adherence support measures. In addition, 41% of study sites reported making at least one telephone call to selected participants (who were judged by study site staff members as being at high risk for low adherence) as part of their usual support measures; telephone calls were made to a minority of participants, the number of telephone calls per subject was limited, and the content was not standardized.

Cox 2006

MethodsA randomised controlled trial of 12 months duration.


ParticipantsHIV positive persons (n=61) in the United States of America. In the intervention group 70.5% were black, 23.5% white, 3% Hispanic, and 3% mixed Hispanic and African American. In the control group 81% were black and 19% white. 50% of the intervention group, and 46% of the control group, were female.


InterventionsParticipants were randomised to an intervention group (n=34) or a control group (n=27). The control group received monthly visits until their viral load was <50, including physician visit, CD4, VL and pharmacist adherence review. The intervention group received the control group intervention plus telephone calls from a pharmacist at day four, two weeks, then monthly for six months.


OutcomesOutcomes were: adherence; change in CD4 lymphocyte count; and virologic response.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNo details of random sequence generation in the study.

Allocation concealment (selection bias)Unclear riskNo mention of allocation concealment in the study.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskBlinding of participants and personnel was unlikely to be feasible in this study, but we do not think this will have affected the results.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo mention of blinding to outcome analysis.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskIn the intervention group, one participant withdrew and five were lost to follow up. In the control group, one participant withdrew and seven were lost to follow up. Reasons for these were not mentioned in the study.

Selective reporting (reporting bias)Unclear riskProtocol was unavailable to us.

Heckman 2007

MethodsA randomised controlled trial.


ParticipantsRural (defined as residence in a community of 50,000 persons or fewer that was at least 20 miles from a city of 100,000 or more) persons (n=299) with a self-reported diagnosis of HIV in the United States. The majority (75%) of participants were white. 30% were women. No inclusion or exclusion criteria related to psychological functioning were employed.

Study quote: 'There were no differences between the treatment conditions at pre-intervention with the exception of education. Post-hoc comparisons revealed that Information Support Group Intervention participants completed fewer years of education than Coping Improvement Group Intervention participants.'


InterventionsParticipants were assigned to a telephone-delivered eight session information support group (n=84), a telephone-delivered eight session coping improvement intervention group (n=108), or a usual care control group (n=107).

Both intervention conditions had separate groups for men who had sex with men (MSM), heterosexual men and women. There were six to eight participants per group, and the interventions were conducted using teleconference technology. The control group received no active intervention, but were able to access services provided by their AIDS service organisation.


OutcomesOutcomes were: BDI score; Symptom Checklist 90-Revisied score; HIV-Related Life-Stressor Burden Scale score; Provision of Social Relations Scale score; The Barriers to Care Scale score; Functional Assessment of HIV Infection Inventory score; Coping Self-Efficacy Scale score; and demographic characteristics.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNo information on random sequence generation in the study.

Allocation concealment (selection bias)Unclear riskNo mention of allocation concealment in the study.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskBlinding of participants and personnel was unlikely to be feasible in this study, but we do not think this will have affected the results.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo mention of blinding of outcome assessment.

Incomplete outcome data (attrition bias)
All outcomes
Low risk73% (n=78) of the usual care condition, 68% (n=57) of the Information Support Group condition, and 82% (n=88) of the Coping Improvement Group condition completed eight month follow-up assessment.

Selective reporting (reporting bias)Unclear riskProtocol was unavailable to us.

Kalichman 2011

MethodsA randomised controlled trial.


ParticipantsParticipants were 26 men and 14 women recruited from infectious diseases clinics in Atlanta, Georgia, receiving and less than 95% adherent to anti-retroviral therapy. The majority (92.5%) were black, with the remainder (7.5%) being white. 35% of participants were women.

Study quote: 'Initial comparisons between conditions on all baseline characteristics and adherence did not indicate any significant differences between conditions, indicating that the randomisation scheme achieved balanced conditions.'


InterventionsParticipants were randomised to telephone-delivered behavioural self-regulation counselling (n=21) or a usual care control group (n=19).

Study quote: 'Those in the telephone intervention group received a single office session followed by four biweekly cell phone counselling sessions that were grounded in behavioural self-management model of medication adherence using data from phone-based unannounced pill counts to provide feedback-guided adherence strategies. The control condition received usual care and matched office and cell phone/pill count contacts.'


OutcomesOutcomes were: ART adherence; medication adherence self-efficacy; adherence behavioural strategies; process measures.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskStudy quote: 'Allocation was accomplished by using a computer generated simple randomization scheme. Randomization was not breached throughout the trial.'

Allocation concealment (selection bias)Low riskStudy quote: 'Recruitment, screening, and office-based assessment staff remained blinded to condition throughout the study.'

Blinding of participants and personnel (performance bias)
All outcomes
Low riskBlinding of participants and personnel was unlikely to be feasible in this study, but we do not think this will have affected the results.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskStudy quote: 'office-based assessment staff remained blinded to condition throughout the study and the adherence counsellor never conducted outcome assessments.'

Incomplete outcome data (attrition bias)
All outcomes
Low riskOne control group participant was lost to follow up.

Selective reporting (reporting bias)Unclear riskProtocol was unavailable to us.

Other biasLow risk

Lovejoy 2011

MethodsRandomised controlled trial conducted between December 2009 and March 2010.


ParticipantsSelf-reported HIV-positive people (n=100) over 45 years of age (mean age 54) who participated in risky sexual behaviour (had engaged in one or more occasions of unprotected anal or vaginal intercourse in the past 3 months). Participants were recruited from five metropolitan areas: New York City, Atlanta, Philadelphia, Cincinnati and Columbus The majority were black (87%), and the remainder white (13%). 46% were female. Participants did not differ on any baseline demographic or clinical variables across the three study conditions.


InterventionsParticipants were randomised to four sessions of telephone delivered motivational interviewing (n= 38), one session of telephone delivered motivational interviewing (n=39) or a control group (n=23) in which they received no motivational interviewing sessions.

Participants in the four session MI group received four sessions of MI, in which a therapist aimed to explore the participants’ relationship dynamics, increase readiness to engage in condom-protected sex, discuss and overcome barriers to condom use and improve clients’ confidence to engage in condom-protected sex. Exactly what was included depended on the patients’ readiness to change. Participants in the one-session MI group received the same intervention as the four session group, with the exception of treatment dose (one session as opposed to four sessions). Participants in the control group did not receive any active intervention, but were encouraged to seek support from their local AIDS service organisations, HIV-related and other support groups, information websites, and any other resources available to them. No formal referrals to these services were made.


OutcomesOutcomes were: the number of self-reported non-condom protected anal and vaginal sex acts in the last three months; and readiness to engage in condom protected behaviours (assessed by asking participants to endorse one of five statements which most represented their willingness to change).


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskStudy quote: 'Prior to initiating the study, off-site personnel utilized a random numbers table and weighted simple random assignment to generate the allocation sequence, yielding a 2 (1-session MI): 2 (4-session MI): 1 (control) ratio.'

Allocation concealment (selection bias)Low riskStudy quote: 'Immediately following the baseline interview, participants were assigned to one of three parallel study conditions using concealment of allocation procedures. Sealed, consecutively numbered, opaque envelopes prepared by off-site personnel contained the randomly allocation conditions.'

Blinding of participants and personnel (performance bias)
All outcomes
Low riskStudy quote: 'All research personnel were blind to participant condition prior to, and during, the baseline interview.' Blinding of participants was unlikely to be feasible in this study, but we do not think this will have affected the results.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskStudy quote: 'Interviewers were blind to participant condition when conducting all interviews.'

Incomplete outcome data (attrition bias)
All outcomes
Low riskTwo participants in the four session MI group, three participants in the one session MI group, and zero participants in the control group did not complete six month follow-up. Reasons were: unable to locate (n=2); dropped out of study (n=1); deceased (n=1); and too sick (n=1).

Selective reporting (reporting bias)Unclear riskProtocol was unavailable to us.

Lucy 1994

MethodsRandomised controlled trial of 16 weeks duration.


ParticipantsEnglish-speaking men (n=17) aged 25-68 and in the early stages of HIV. They were well educated (college=5, graduate school=12), largely employed (full time=10, part time=2, unemployed=3, student=1, retired=1). 71% were white, 18% African American and 12% Latino. There were significant differences in sexual orientation between the intervention and the control group. In the intervention group one participant was bisexual and eight homosexual. In the control group, three participants were heterosexual and five were homosexual.


InterventionsParticipants were randomised to a telephone intervention group (n=9) or a control group (n=8). Those in the intervention group received weekly structured telephone calls for 16 weeks. These incorporated psychosocial support; ongoing monitoring of health, stress, mood and interpersonal satisfaction; information and education regarding HIV infection and AIDS; and early referral to other services where appropriate.The control group was a wait-list no intervention control group.


OutcomesOutcome was psychiatric distress, measured using the General Severity Index of the BSI.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNo details about random sequence generation.

Allocation concealment (selection bias)Unclear riskNo mention of allocation concealment in the study.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskBlinding of participants and personnel was unlikely to be feasible in this study, but we do not think this will have affected the results.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo mention of blinding of outcome assessment in the study.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNo mention of incomplete outcome data in the study.

Selective reporting (reporting bias)Unclear riskProtocol was not available to us.

Other biasHigh riskHigh risk of bias for baseline differences. There were significant differences in sexual orientation between the intervention and the control group. In the intervention group one participant was bisexual and eight homosexual. In the control group, three participants were heterosexual and five were homosexual.

Ransom 2008

MethodsA pilot randomised controlled trial carried out between February 2006 and March 2007.


ParticipantsSelf-reported HIV (n=79) positive persons living an urban area (definition) and with depression (met diagnostic criteria for a depression-spectrum disorder according to a telephone-administered mood module of the Primary Care Evaluation of Mental Disorders). The majority of participants were white (77%), followed by African American (10%), Latino (9%), Hispanic (3%) and Native American (1%). Just 16% of participants were women. There were no significant baseline differences between the intervention and control groups.


InterventionsParticipants were randomly allocated to an intervention group (n=41) or a usual care control group (n=38). The intervention group received a six-session, telephone-delivered interpersonal psychotherapy intervention plus usual care. Those in the usual care control group received no active therapeutic intervention, but had access to services provided by their AIDS service organisation (for example support groups).


OutcomesOutcomes were BDI-II score; Outcomes Questionnaire score; Provision of Social Relations Scale score; and the University of California Los Angeles Lonliness Scale score.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskStudy quote: 'randomly assigned'. No further detail.

Allocation concealment (selection bias)Unclear riskNo mention of allocation concealment in the study.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskBlinding of participants and personnel was unlikely to be feasible in this study, but we do not think this will have affected the results.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo mention of blinding of outcome assessment in the study.

Incomplete outcome data (attrition bias)
All outcomes
High riskStudy quote: 'Participants in the teletherapy group were marginally more likely to discontinue study involvement (n=10) than those in the control group (n=3), although this association was not significant.'

Study quote: 'Participants with missing postintervention data (n=13) were retained for final outcome analyses by using a last-observation-carried-forward approach.'

Selective reporting (reporting bias)Unclear riskThe only information provided on the University of California Los Angeles (UCLA) Lonliness Scale and Provision of Social Relations Scale (PSRS) results is: 'No hypothesized condition x time interactions were found for the UCLA Loneliness Scale or the PSRS.'

Reynolds 2008

MethodsA randomised controlled trial carried out over 64 weeks. An adherence substudy of a large multicentre treatment trial evaluating different strategies for initiating antiretroviral therapy in HIV-infected individuals.


ParticipantsAnti-retroviral therapy naive participants (n=109) in five sites in the United States of America. 51% of participants were white and 43% black. 15% of participants were women. 73% had comorbid depression. There were no significant differences in demographics between the two groups.


InterventionsParticipants were randomised to an intervention group (n=54) and a control group (n=55). Those in the intervention group received standard clinic-based patient education plus structured proactive telephone calls provided by a trained registered nurse specialist and access to a 24-hour toll free helpline. Those in the control group received standard clinic-based patient education.


OutcomesOutcomes were self-reported adherence and time to virologic failure.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskStudy states that participants 'were randomized', but no further detail is given.

Allocation concealment (selection bias)Unclear riskNo mention of allocation concealment in the study.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskBlinding of participants and personnel was unlikely to be feasible in this study, but we do not think this will have affected the results.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo mention of blinding of outcome assessment.

Incomplete outcome data (attrition bias)
All outcomes
Low riskFourteen participants in the intervention group and 20 in the control group were lost to follow up, due to dropping out of parent study or reaching the endpoint of the parent study.

Selective reporting (reporting bias)Unclear riskThe study protocol was unavailable to us.

Rotheram-Borus 2004

MethodsA randomised controlled trial


ParticipantsDrug using (must have used illicit drugs at least 5 times in the last 3 months) young people (n=175) (aged 16-29) with HIV infection. 42% were Latino, 26% black, 23% white and 8% of unspecified ethnicity. 22% were women. The in-person intervention condition had a higher proportion of protected acts across all sexual partners, among HIV-positive sexual partners, and among HIV-negative sexual partners compared with the delayed-intervention condition (P <0.05). The telephone intervention condition had a higher proportion of protected acts across all sexual partners and among HIV-negative sexual partners compared with the delayed-intervention condition (P <0.05). The telephone intervention condition also had a higher proportion of protected acts across all sexual partners and among HIV-negative sexual partners compared with in-person intervention condition (P <0.05).


InterventionsParticipants were randomised into a telephone intervention group (n=59), an in-person intervention group (n=61), or a delayed intervention control group (n=55). The intervention consisted of 18 one-on-one interventions divided up in to three modules, and delivered via telephone or in-person. Each module lasted six sessions and focused on a different target behaviour: improving physical health, reducing sexual and substance use acts, and improving mental health.


OutcomesOutcomes were: sexual risk acts; substance abuse; medication adherence; health behaviours; self-report health status; mental health outcome; and cost analysis.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNo details of sequence generation in the study.

Allocation concealment (selection bias)Unclear riskNo mention of allocation concealment in the study.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskBlinding of participants and personnel was unlikely to be feasible in this study, but we do not think this will have affected the results.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo mention of blinding to outcome assessment.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskStudy quote: 'Fewer young people completed all 18 sessions (35%): 41% of those receiving individual sessions and 29% of those receiving the telephone interventions.' No further information given.

Selective reporting (reporting bias)Unclear riskProtocol unavailable, but unlikely given the number of outcomes.

Other biasHigh riskHigh risk of bias for baseline differences between the two intervention groups and the control group. The in-person intervention condition had a higher proportion of protected acts across all sexual partners, among HIV-positive sexual partners, and among HIV-negative sexual partners compared with the delayed-intervention condition (P<0.05). The telephone intervention condition had a higher proportion of protected acts across all sexual partners and among HIV-negative sexual partners compared with the delayed-intervention condition (P< 0.05). The telephone intervention condition also had a higher proportion of protected acts across all sexual partners and among HIV-negative sexual partners compared with in-person intervention condition (P<0.05).

Stein 2007

MethodsA randomised controlled trial carried out over a six month period.


ParticipantsHIV positive persons (n=177) with depression (defined as a BDI score of 10 or more), co-enrolled with their self-identified primary informal care giver (if they were able to identify one) in the United States of America. 41% of participants were Caucasian. 44% were female. Participants had an average BDI score of 22.7.

Study quote: 'Participants randomized to treatment had significantly lower CD4 counts at baseline and reported significantly poorer mean physical function scores on SF-36. Treatment groups did not differ significantly with respect to any other variables. Baseline depression was not correlated significantly with baseline CD4 counts or months since diagnosis.'


InterventionsParticipants were randomised to a telephone-contact intervention group (n=89) and a control group (n=88). Those in the intervention group received 12 structured psychoeducational telephone calls over six months, which provided education and encouraged problem-appraisal and resolution-through-referral, with the aim of reducing problems related to mood in HIV positive patients. Those in the control group received assessment only.


OutcomesOutcome measures were: change in depressive symptoms over time (six months, measured using BDI); whether the patient goes in to remission for depression or not (measured using BDI); change in severity of depressive symptoms category; use of psychiatric medications during follow-up period; mental and physical function subscales of SF-36.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskStudy quote: 'Patients were randomly assigned to either an assessment only group or to a phone intervention group'. No further detail provided.

Allocation concealment (selection bias)Unclear riskNo mention of allocation concealment in the study.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskBlinding of participants and personnel was unlikely to be feasible in this study, but we do not think this will have affected the results.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo mention of blinding to outcome assessment in the study.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskTen participants in the intervention group, and seven participants in the control group were lost to follow up. Reasons for this were not given.

Selective reporting (reporting bias)Unclear riskProtocol of the study was unavailable.

Watakasol 2010

MethodsA pilot randomised controlled trial conducted between December 2009 and May 2010.


ParticipantsRural persons (n=42) living with HIV in the United States of America. The majority (81%) were white, 7.1% were black and 4.9% were 'other' ethnicity. 40% of participants were female.


InterventionsParticipants were randomised to a telephone-administered motivational interviewing adherence improvement intervention (n=21) and a self-monitoring comparison condition control group (n=21). Those in the intervention group received a one-session telephone-administered 60 minute structured scripted adherence improvement intervention. Those in the control group received no active intervention, but were asked to complete medication diaries over a five-week period, which was equal to the number of weeks that intervention participants completed medication diaries.


OutcomesOutcomes were: dose adherence; schedule adherence; readiness to change; intrinsic motivation; self-efficacy; depression; and reasons for non-adherence.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskStudy quote: 'Block randomization with a 1:1 allocation using a random block size of two was used to ensure equal numbers within both conditions. Specifically, every time two participants returned their pre-intervention questionnaires; a researcher drew a card from an envelope. The envelope contained two cards; one was labelled 'Intervention group' and the other one was labelled 'Control group'.

Allocation concealment (selection bias)Unclear riskIt does not mention whether the envelopes used were sealed and opaque. Assignment envelopes may have been used without appropriate safeguards.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskBlinding of participants and personnel was unlikely to be feasible in this study, but we do not think this will have affected the results.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo mention of blinding of outcome assessment in the study.

Incomplete outcome data (attrition bias)
All outcomes
Low riskStudy quote: 'All participants completed the pre-intervention questionnaire and the 2-week medication diary. However, two intervention (9.5%) and two control group participants (9.5%) did not return 3-week medication diaries. One intervention group (4.7%) and one control group participant (4.7%) did not return the post-intervention questionnaires. Contact was lost with these participants and the reasons for their termination in the study remain unknown.'

Selective reporting (reporting bias)Unclear riskProtocol was unavailable to us.

Other biasUnclear riskThere may be risk of bias due to significant baseline differences between the intervention and the control group. Study quote: 'There were significant differences between conditions at baseline on number of years on ART medications, number of different ART medications, and number of total ART pills prescribed. On average, intervention participants (15.3 years) were on ART medications for a longer period of time compared to controls (11.1 years). Compared to control group participants (2.1), intervention group participants (3.1) were prescribed a greater number of different ART medications and a greater number of total ART pills (intervention group= 5.4, control group=3.2).' However, we feel that this may have been inevitable given the small sample size.

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

Konkle-Parker 2010This RCT pilot trial assessed the efficacy of an intervention with both a face-to-face and a telephone component to improve medication adherence in PLHIV. We excluded the study because the use of the telephone was only a part of the intervention and could not be separately evaluated.

Puccio 2006This pilot study assessed the use of cell phone reminder calls for improving adherence to ART in young PLHIV. We excluded the study because of was not of an eligible study design.

Uzma 2011This RCT assessed the efficacy of an intervention combining both telephone calls and face-to-face sessions for improving ART adherence in PLHIV. We excluded the study because the use of the telephone was only a part of the intervention and could not be separately evaluated.

Wang 2010This study assessed the efficacy of a nurse delivered intervention combining home visits and telephone calls designed to improve medication adherence and quality of life. We excluded the study because the use of the telephone was only a part of the intervention and could not be separately evaluated.

 
Comparison 1. Effects of the interventions aimed at improving medication adherence

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 ART adherence3191Mean Difference (IV, Fixed, 95% CI)0.49 [-1.12, 2.11]

 
Comparison 2. Effects of the interventions aimed at improving depressive and psychiatric symptoms

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Depressive symptoms (BDI score)3447Mean Difference (IV, Fixed, 95% CI)0.02 [-0.18, 0.21]

 2 Psychiatric distress (General Severity Index of the Brief Symptom Index)117Mean Difference (IV, Fixed, 95% CI)-6.53 [-11.71, -1.35]

 3 Psychiatric distress (Outcomes questionnaire score)179Mean Difference (IV, Fixed, 95% CI)1.80 [-10.98, 14.58]

 4 Psychological distress (Total SCL-90-R score)1191Mean Difference (IV, Fixed, 95% CI)-0.10 [-0.11, -0.09]

 5 Life stressor burden (HIV-related life-stressor burden scale)1191Mean Difference (IV, Fixed, 95% CI)-0.01 [-0.02, 0.00]

 6 Support from family (PSRS)1191Mean Difference (IV, Fixed, 95% CI)0.16 [0.14, 0.18]

 7 Support from friends (PSRS)1191Mean Difference (IV, Fixed, 95% CI)0.22 [0.20, 0.24]

 8 Barriers to care (BACS)1191Mean Difference (IV, Fixed, 95% CI)-0.06 [-0.08, -0.04]

 9 Coping self-efficacy (Coping self-efficacy scale)1191Mean Difference (IV, Fixed, 95% CI)0.38 [0.34, 0.42]

 10 Social well-being (Social/family well-being subscale of the FAHI)1191Mean Difference (IV, Fixed, 95% CI)0.02 [-0.00, 0.04]

 11 Emotional well-being (Emotional well-being subscale of the FAHI)1191Mean Difference (IV, Fixed, 95% CI)0.05 [0.03, 0.07]

 
Summary of findings for the main comparison. Effects of the interventions aimed at improving medication adherence for reducing morbidity and mortality in people with HIV infection

Effects of the interventions aimed at improving medication adherence for reducing morbidity and mortality in people with HIV infection

Patient or population: People living with HIV infection
Settings: All
Intervention: Voice landline and mobile telephone delivered interventions to improve medication adherence

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Quality of the evidence
(GRADE)
Comments

Assumed riskCorresponding risk

ControlEffects of the interventions aimed at improving medication adherence

ART adherenceThe mean art adherence in the intervention groups was
0.49 higher
(1.12 lower to 2.11 higher)
191
(3 studies)
⊕⊕⊝⊝
low1,2

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: Confidence interval;

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

 1 Results of the studies are inconsistent.
2 All of the included studies had fewer than 400 participants.
 
Summary of findings 2. Effects of the interventions aimed at improving depressive symptoms for reducing morbidity and mortality in people with HIV infection

Effects of the interventions aimed at improving depressive symptoms for reducing morbidity and mortality in people with HIV infection

Patient or population: People living with HIV infection
Settings: All
Intervention: Voice landline and mobile telephone delivered interventions to improve depressive symptoms

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Quality of the evidence
(GRADE)
Comments

Assumed riskCorresponding risk

ControlEffects of the interventions aimed at improving depressive and psychiatric symptoms

Depressive symptoms
Beck Depression Inventory Score
The mean depressive symptoms in the intervention groups was
0.02 higher
(0.18 lower to 0.21 higher)
447
(3 studies)
⊕⊕⊝⊝
low1,2,3

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: Confidence interval;

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

 1 All three studies have an unclear risk of bias due to missing information.
2 Studies have wide confidence intervals.
3 All the included studies had fewer than 400 participants.