Adherence and Support

A review of nine articles and 18 abstracts until 2006 from Sub-Saharan Africa, with 12,116 patients found favorable levels of adherence, with 77% of patients achieving 95% adherence according to patient self-reports. Adherence from studies in sub-Saharan Africa showed that that more patients were adherent than patients in North America, based on 31 studies with 17,537 patients.

A systematic review and meta-analysis of 39 studies from female sex worker populations in Benin, Burkina Faso, Kenya, Rwanda, Zimbabwe, Canada, United States, India, Russia, Thailand, Vietnam, Dominican Republic, El Salvador, Brazil found that sex workers can achieve high rates of adherence.  From 9 studies from Africa, Central America, the Caribbean and North America, 76% were adherent to more than 95% of prescribed pills. The pooled estimate of adherence for studies from LMIC was 76%. Similar to women in the general population, 57% on ART were virally suppressed, with study populations in Asia, Africa and Latin America. Three studies from Canada, USA and Dominican Republic found treatment experienced female sex workers who were no longer on ART. Loss to follow-up after ART initiation based on six studies from Kenya and Burkina Faso was 6%.  Death following ART initiation based on six studies from Kenya and Burkina Faso was 6%. Female sex workers were defined as women who reported sex work within the previous six months or currently. Of the 39 studies included, 24 were prospective cohort studies, 12 were cross-sectional and 2 were retrospective case control. Median gains in CD4 count after 6 to 36 months ranged between 103 and 241 cells. However, current ART use among female sex workers was only 38%, with ever ART use greater among sex workers in high-income countries compared to LMIC.

A retrospective observational study of 79 patients (84.2% men) who accessed ART at a private clinic in Argentina and were on ART for a median of 14 years found that 87.3% had an undetectable viral load. More than 90% had a median CD4 count of 516.

A prospective cohort study of 354 Rwandan women living with HIV on ART found that all reported complete adherence in the three days prior to their clinic visit, supported by changes in lab measures. The median CD4 count within six months prior to ART was 185 compared to 264 at the post-ART visit, which occurred nine to fifteen months following ART initiation. In addition, 94% refilled their medication at the appropriate time. In Rwanda, in order for a patient to become eligible for ARVs, they must attend a two to three day treatment literacy session, accept to be visited by a health care worker, have a fixed residence in a known catchment area of a health facility and disclose HIV status to a trusted family member, along with someone who will help them with adherence. Adherence is reinforced at the community level by community health workers and/or a member of an association of people living with HIV. In addition, 96% of Rwandans have health insurance and ARVs are provided at no cost. Care was provided by an NGO with outreach efforts that included patient home visits.

A study in Botswana of 300 adult patients initiated in 2005 and found that 81.3% of patients were adherent based on four day and one-month patient recall and also based on clinic attendance for ARV medication refills during the prior three months. Adherence was defined as self-report of no missed ARV medication doses in the past four days; self-report of no missed ARV medication doses during the past one-month; and self-report of no missed ARV medication refill visits during the past three months. Adults receiving combination ART for one to six months had adherence rates of 77% and those receiving combination ART for more than twelve months had adherence rates of 79%.

A cross-sectional study with quantitative and qualitative methods in 2007 with 422 patients in Ethiopia found an adherence rate of 88.1% based on unannounced pill counts.

A meta-analysis of 19 randomized controlled trials was completed to investigate how pill burden and once-daily and twice-daily regimens of ART influenced adherence and virological outcomes.  Higher pill burden was associated with reduced adherence and reduced rates of virological suppression, for both once-daily and twice-daily ART regimens. Adherence rates were higher in the once-daily ART regimens as compared to the twice-daily regimens. There was no statistically significant relationship found between virological suppression and once- or twice-daily ART regimen.  Adherence and viral suppression decreased with longer follow-up time, but adherence decreased less with once-daily regimens than with twice-daily regimens of ART.  The studies were published between 2004 and 2011 and were located primarily in high-income countries, with some data from middle- and lower-income countries.  The trials included data from 6,312 individuals living with HIV.  Adherence was measured with either pill counts or with medication event monitoring system caps. Adherence was 2.55% higher among participants following once-daily regimens as compared to patients following twice-daily regimens.  There was no significant difference in treatment outcomes as measured by virological suppression between once- and twice-daily regimens of ART, which may be due to the minimal difference in adherence not translating to statistically significant outcomes.  Decreasing pill burden in this meta-analysis was associated with better adherence and virological suppression, and adherence was improved with once-daily regimens as compared to twice-daily regimens.

A meta-analysis of randomized controlled trials and cohort studies was done to compare the effect of fixed-dose combination ART and separate tablet regimens on adherence, virological suppression, and patient preference.  The analysis demonstrated that fixed-dose combinations were associated with better adherence and to a lesser extent, increased virological suppression as compared to those not receiving fixed-dose ART. All studies reporting on patient preference indicated that patients preferred the fixed-dose combinations.  The analysis included studies on partial fixed-dose combinations (where two pills were combined) and full fixed-dose combinations (where there was a single tablet regimen), as compared to regimens with a greater number of pills.  Twenty-two papers, including data on 27,230 people, were included in the analysis.  There were 6 randomized controlled trials, 10 prospective cohort studies, and 5 retrospective cohort studies.  The studies were published from 1999 to 2013 and most were carried out in high-income settings.  In most of the studies, less than half of the participants were women.  The meta-analysis for viral suppression which included 6 of the studies demonstrated greater virological suppression among patients on fixed-dose combinations.  The analysis for adherence which included 5 of the randomized controlled trails indicated better adherence among the group on fixed-dose combinations.  In the 4 studies that reported on quality of life, all reported better patient scores on fixed-dose combinations compared to separate pill regimens, including through quality of life score, a depression score, and self-perceived psychological fatigue.  In one study, 97% of patients on fixed-dose combination treatment perceived their regimen to be 'very easy', compared to 81% of patients not on fixed-dose combination treatment.  The authors of this meta-analysis write in favor of fixed-dose combinations that, "unlike other adherence support interventions which themselves require a degree of adherence, [fixed-dose combinations] require no further action on the part of the patient or provider as the intervention is an indivisible characteristic of the treatment."

A global review of numerous randomized controlled trials found that once daily regimens improved adherence. [See also (Ford et al., 2013b)]

A randomized controlled trial with 68 women during a six-month period in rural India found that support by village women who received training resulted in statistically significant increased adherence. These village women received a three-day training to learn the basics of HIV/AIDS progression; the importance of adherence; coping strategies; the importance of nutrition; and life skills options. Supervision was ongoing. Women in the intervention group received six classes from the trained women on how to overcome barriers to adherence; how to reduce stigma; cooking tips for good nutrition; the benefits of earning income; and received a monthly supply of 2 kilos of food to cook. The women trained to provide support visited 5 women every week and worked to mitigate any barriers to adherence, such as providing bus fare for women to get to health services and accompanying the women to health services. The usual care group just received the six classes and chickpeas to cook. Adherence was measured by pill counts.

A randomized trial of 48 women living with HIV in Nigeria who completed a six month follow up assessment found that motivational group support resulted in significantly higher levels of self-reported adherence. Among the intervention group, 93% reported never missing any medication compared to 40% of the standard of care group. Facilitators received a 24-hour training. Group support focused on topics such as discrepancies between current adherence behaviors and future goals.

A cohort study conducted from 2007 to 2011 in South Africa followed participants in adherence clubs to study the effectiveness of group-based care led by non-clinical staff to prevent loss to follow-up, death, and virologic rebound.  This data was compared to patients receiving routine care.  Participation in an adherence club reduced loss-to-care by 57% and reduced virologic rebound by 67%. Ninety-seven percent of club participants remained in care at the end of the study as compared to 85% of patients in routine care.  Participation was offered to adult patients who had been on ART for at least 18 months and had a CD4 count above 200 cells/µl.  Of the 2,829 patients in the cohort, 502 chose to participate in a club.  At the conclusion of the study, 12.8% of patients died or were lost to follow-up and 9.0% had virologic rebound.  The median age was 33, 71% of participants were women, and 88% of patients were virologically suppressed at the initiation of the study.  Adherence clubs consisted of groups of 15 to 30 patients who met for counseling and to collect their medication.  Most patients who had the option to transition to the adherence club did so as it took less time to collect medications through the club than it did at the clinic.  Medications were pre-packaged for each patient and brought to the meeting by the counselor.  Patients in normal care who were less than 25 years old and who entered the study with a CD4 count of less than 50 cells/µl had the highest rate of death and loss to follow-up.  Patients were more likely to be enrolled in a club if they had a higher CD4 count at study entry or during follow-up, were women, or had been on ART for longer than average. Adherence clubs improved adherence by decreasing loss to follow-up and virologic rebound (Luque-Fernandez et al., 2013) This model is currently being scaled up in Western Cape, South Africa.

A study of 268 patients in South Africa found that support of a treatment buddy, community health workers or support group had better ART outcomes than those patients who did not. Treatment buddies reminded the patient to take ARV drugs and 75.9% of study participants reported that their treatment buddy helped their adherence. Community health workers provided emotional support and motivation for adherence. And 89.9% of support group members reported that the support group meetings helped them by sharing knowledge and experiences. Over 85% of patients rated treatment buddies, community health workers and HIV support groups as good or excellent. Of the 268 patients, 76.4% had CD4 counts over 200 after 24 months of ART.

In Vietnam, people living with HIV served as peer assistants to help other people navigate the steps from HIV diagnosis to enrollment in HIV outpatient care clinics. National guidelines in 2010 have resulted in community teams trained to support adherence, resulting in a rate of retention of 90% in some districts. In one district, antiretroviral therapy retention improved from 66% to 85% between 2009 and 2010.

A study between 2008 and 2010 with 1,384 people living with HIV enrolled in 291 groups in Mozambique found that support groups led to increased adherence. ART distribution and monitored adherence was conducted by community groups in a model developed by Medecins sans Frontieres and provincial authorities.  Patients who were stable on ART for six months were informed about the community ART group model and invited to form groups. Group members facilitated monthly ART distribution to other group members; provided adherence and social support; monitored outcomes and ensured that each group member had a clinical consultation once every six months. Group members visited the health center on a rotational basis, so that each group member had contact with health services at least once every six months. Group members were followed for 12.9 months. Of the 1,301 patients still in community groups after 83 transferred, 97.5% remained in care, 2% died, and 0.2% were lost to follow up. In other health settings in this area, up to one in five patients are lost to follow up. To join a community support group, patients had to be clinically stable on ART for at least six months and have CD4 counts over 200. Counselors trained the newly formed groups on their roles and responsibilities.  Group members could still visit the health center for any reason at any time. At the facility, the group representative discussed each group member with a counselor or clinician, covering issues of adherence, clinical status, etc. The group monitoring form was reviewed. Upon return to the community, the group representative distributed medications and returned patient appointment cards and if needed, requested a group member to go to the health facility for follow up. All members for different community groups were invited every six months for a group session held either in the community or the health facility to discuss when to go to the clinic for unplanned consultations, patient education on TB, etc. CD4 counts are taken. The majority of community group members were female (70%). Median gain in CD4 cell count since initiation was 478.5. Of the patients, 3.5% decided to transfer back to conventional care.

A qualitative study of 18 women with four focus groups from 10 different support groups led by para-professionals in Rwanda found that the support group empowered them to become adherent: "I met with (support group) counselors and they told me how I will survive with HIV…I learned the importance of taking medicine on time…Before I attended, the doctors used to ask me if I was taking the medicine at home and I used to say, 'Yes, I take them' but I was lying to them" (p. 98).

A qualitative study of 26 women in Thailand found that joining AIDS support groups provided emotional support. One woman stated: "Joining the group activity is good...(and) makes me have emotional strength that it is not only me who has got this disease in the world" (p. 447).

A qualitative study of 15 women living with HIV in Vietnam found that women in support groups expressed that they enjoyed a lot of benefits by participating in support groups, but also found that their additional knowledge made them fear treatment failure and whether ARVs would always be available to them.

A qualitative study of women living with HIV in Colombia with six in-depth interviews and five focus groups with 47 women found that participation in support groups was associated with higher adherence.

A meta-analysis of 19 randomized controlled trials including 1,839 patients found that one-on-one counseling given by health providers, with a median of two sessions, with each session lasting 60 minutes, with patients being more than one and a half times more likely to achieve 95% adherence, compared to controls. The effect was larger in studies where pills counts were used as compared to self-reporting (Simoni et al., 2006 cited in Vergidis et al., 2009).

A review of 21 studies in Sub-Saharan Africa (including three randomized controlled trials; six mixed method studies; seven cohort studies; and five qualitative studies) on the roles and outcomes of community health workers found that community health workers effectively provided counseling and adherence support, as well as screening and referral for patients, enhancing uptake of HIV services and retention in care. Community health workers also reduced clinic wait times and reduced the workload of other health workers. Patients served by community health workers did not have increased mortality nor any differences in virologic failure compared to patients served only in facilities. Studies found that community health workers in Uganda led to timely patient referrals, reducing delays in care (Alamo et al., 2012 cited in Mwai et al., 2013); improved patient flow at clinics in Lesotho (Joseph et al., 2012 cited in Mwai et al., 2013): and increased access and improved virological suppression in Uganda and Rwanda (Kipp et al., 2012; Rich et al., 2012 cited in Mwai et al., 2013). However, community health workers lacked adequate recognition, remuneration and involvement in decision-making leading to demotivation and attrition. Community health workers were defined as those having no formal professional or paraprofessional certificate or degree in tertiary education (Lewin et al., 2010 cited in Mwai et al., 2013).

A randomized controlled trial with 297 patients living with HIV conducted from 2007 to 2009 who were ART-naïve in South Africa found that participants receiving intensive motivational counseling had similar rates of adherence to patients who received didactic adherence counseling. Participants were 43.1% male and virological suppression at nine months was achieved in 89.8% of those who received didactic counseling and 87.9% of those who received motivational counseling. Motivational counseling provided one-on-one interactive sessions and participant problem solving, including managed disclosure. Didactic counseling provided information and treatment literacy on stigma, nutrition, how ART is transmitted, the importance of adherence, side effects and resistance in three session but without engagement of individual issues faced by patients. Both groups received repeated adherence assessments, with tracking, tracing and pillboxes. Counseling was provided to any patient with problems or adherence concerns, with the additional counseling more likely among those in the didactic arm.

A randomized trial with 301 patients in Kenya from 2006 to 2008 found that early adherence counseling during the first two months of ART initiation resulted in sustained, significant impact on adherence and virological treatment failure during 18-month follow-up. Patients were randomized to 15 minute counseling or three counseling sessions of up to 45 minutes each. Patients who received three counseling sessions were 29% less likely to have monthly adherence under 80% and 59% less likely to experience viral failure compared to those who received 15 minutes of counseling. Counseling was not associated with mortality, possibly due to short follow-up time. Counseling was based on a model of successful antiretroviral adherence developed by the University of Washington and adapted for Kenya with a written standardized protocol and lasted from 30 to 45 minutes per session. The first session addressed barrier and taught patients the risks of poor adherence; the second session on a separate day reviewed the patient’s understanding and a review of the patient’s readiness to initiate ART; and the third session addressed practical and personal issues that the patient experienced on ART. Three interactive sessions, two prior to ART initiation and one session following one-month ART initiation resulted in significant outcomes 18 months later. For those who did not receive counseling, the study pharmacist explained the side effects of medication and problems associated with poor adherence in a 15-minute session prior to dispensing ART. Compared to the costs of treatment failure, employing counselors can represent significant cost-saving.

A multicenter cohort study was conducted in South Africa from 2004 to 2011 to evaluate the effect of community-based adherence-support on adherence among ART patients. The patients who received community-based support had a 35% reduction in mortality and a 37% reduction in loss to follow-up as compared to patients that did not receive community-based support.  From 57 care facilities, 66,953 patients were included; 29.4% of patients received community-based support while 70.6% did not.  Patients were initially enrolled on ART when their CD4 count fell below 200 cells/µl.  After April 2010, the criteria was expanded to include enrollment on ART when CD4 count fell below 350 cells/µl.  All adults not previously enrolled on ART with documented demographics and at least one day of follow-up were included in the study.  Community-based adherence-support included a paid patient advocate doing an initial home visit, weekly visits for the first month, and then visits at least every three months.  If the patient did not attend clinic visits regularly, the frequency of home visits by the patient advocate was increased.  Patients who were labeled “very important” included pregnant, ill, or TB co-infected patients, and were visited at least once a month.  Patient advocates provided services including counseling, adherence checks, referrals, health promotion education, and screening for opportunistic infections.  Patient advocates were trained in an initial 3-week course, subsequently participated in a 5-day refresher course every year, and were assigned 80-120 ART patients.  Patients who received community-based support and patients who did not receive community-based support had access to site-based adherence counseling at some clinics.  Patients who did not receive community-based support who missed appointments were tracked by phone or ere visited at home by a district tracing team.  Patients who received community-based support on average had more advanced HIV upon initiation of treatment, had a slightly higher baseline CD4 count, were more likely to have a TB co-infection, and were enrolled on ART during the more recent study period.  After 5 years, 13.2% patients who received community-based support were lost to follow-up, as compared to 17.7% of patients who did not receive community-based support.  The mortality rate was 9% in patients who received community-based support, while mortality was 10.6% in patients who did not.  Low base-line CD4 count was strongly associated with mortality in both groups of patients.  After 6 months of ART, viral suppression occurred in 76.6% of patients who received community-based support, while only 72% patients with no community-based support had viral suppression.  The community-based adherence-support program improved adherence with demonstrative reductions in loss to follow-up and mortality, and improved rates of viral suppression.

An ecologic study was done on HIV clinics initiating HIV-positive individuals on ART from 2004 to 2008 in Côte d'Ivoire, Ethiopia, Kenya, Lesotho, Mozambique, Nigeria, Rwanda, South Africa, Tanzania, and Zambia to evaluate the effect of different adherence support services on attrition following ART initiation. The study found that clinics with counseling services, educational materials, reminder tools, and food rations had significantly lower attrition than clinics that did not have these services. The study used data from 349 clinics that provide free ART following the national HIV treatment guidelines in each specific country.  This included data from 232,000 patients, 83,389 of which also had 6- and 12-month follow-up data that was used for cohort analysis.  The attrition measure included death, loss to follow-up, and discontinuation of ART.  A total of 300,700 person-years of data was collected.  Fifty-nine percent of patients were women.  On average, clinics reported four adherence support services (93% had at least one and 83% had more than two).  The most common adherence support service, counseling, was reported in 88% of clinics.  The least common adherence support service was food rations; 17% of clinics reported food ration services.  Overall, 72% of the patients were in care at the same clinic at the end of the follow-up period. Nine percent transferred to another clinic, 6.3% had died, 0.9% discontinued ART, and 11.9% were lost to follow-up.  The attrition rate was 14.2 per 100 person-years.  Clinics that had less than 2 adherence support services had marginally lower attrition rates than those with 2 or more services. On-site support groups, peer educator programs, pharmacy support, and active patient outreach were not significantly associated with attrition.  Clinics with less than 2 adherence support services had lower rates of loss to follow-up but not significantly lower rates of mortality as compared to their counterparts with 2 or more support services. In cohort analysis, pharmacy support services and active patient outreach were associated with lower patient attrition.

A cross-sectional study was conducted at a hospital in Jakarta, Indonesia in 2012 to explore the factors associated with ART adherence.  The study found that, "the level of social support experienced by ART-prescribed patients was positively associated with adherence," (Weaver et al., 2014: 1).  In total, 261 patients participated in the study; 74% were male, 54% were married, and 41% were unemployed.  If the patient missed one or more doses of ART medication in the last four days or three or more doses of ART medication in the last three months, they were considered non-adherent.  The difference in adherence between the male and female participants was minimal.  The patients were also asked about their social support.  A reported 52% of respondents had good social support, while 38% reported only some social support and 10% reported poor social support.  Multivariate analysis demonstrated that the level of social support reported by respondents had an effect on adherence, where, "greater levels of social support significantly improve adherence" (p. 3).

A qualitative study of 79 people (64.6% women) living with HIV in Swaziland out of 2,500 clients served by a faith-based organization that provided community home-based care found that 92% stated that with care supporters, their health had improved. From 2008 to 2011, 1,001 clients were on ART and the mortality rate of clients declined from 35.3% to 14.8% (Van Wyngaard, 2013 cited in Root and Whiteside, 2013). Caregivers’ multiplied the number of constructive and relevant conversations around HIV. Patients stated that they had commenced ART because of their caregivers’ encouragement. Caregivers encouraged clinic follow up when patients experienced medication side effects. One patient stated: “Now I’ve got the knowledge of (how to live) from the care supporter and the clinic” (Patient cited in Root and Whiteside, 2013: 5). Caregivers provided ongoing instruction regarding ART adherence. One patient noted that caregivers "helped me take the pills: eat first, wait, then take pills…I used to vomit" (patient cited in Root and Whiteside, 2013: 4). Care supporters enhanced treatment uptake and literacy, reduced stigma and assisted patients to challenge social pressures to stop ART. Caregivers, with the patients' permission, helped patients disclose to family members. With caregiver encouragement, one woman stated to men in her community that she should stop taking ART: "No, I will continue taking my medications" (patient in Root and Whiteside, 2013: 5). Caregivers also encouraged breaking down denial of HIV; other patients who were disparaged by community members for taking ARVS stated: "I am very fortunate because I know my HIV status. What about you?" (Patient in Root and Whiteside, 2013: 6). One patients stated: "Care supporters are nearer to us each and everyday…And we are open to speak to the care supporter about things that we are afraid to speak to the nurses about" (Patient in Root and Whiteside, 2013: 6). Only 11% of health facilities in the area had a health worker trained in HIV; no health facility had internet access. Care supporters were trained for one week on HIV with a religious component and received no remuneration. Care supporters traveled in pairs. Of patients, 15% did not identify as Christian but all participants wanted Christian care givers, as they perceived them to have larger "hearts," to be trustworthy and maintain confidentiality.

A model developed by the Thai Network of People with HIV/AIDS (TNP+) which involved activists living with HIV trained in treatment adherence achieved success through a comprehensive and continuous care program in Thailand. HIV-positive activists had established such centers at one-third of hospitals by 2008. The model requires that HIV-positive activists ask permission from the hospital director, be assigned a room in the hospital for their work and hospital staff agrees that the activists can join the hospital’s HIV care and treatment team. One activist manages a caseload of 15 to 20 clients, recognizing common side effects of antiretroviral drug regimens, evaluate treatment adherence, and answer questions on prevention. Activities are supported by NGOs and TNP+. Activists are now accepted as co-providers of health care rather than passive receivers.

A prospective single arm observation study of low-literacy and low-income patients over a 12 month period in 2005 in Mozambique who were provided care by a multi-disciplinary team which educates patients about HIV transmission and prevention resulted in a 95% adherence level to HAART by 69.5% of 531 patients.  An additional 16.8% reported taking between 90 and 95% of the pills prescribed, for a total of 86.4% of patients with adherence levels above 90%. Only 3.2% abandoned antiretroviral therapy. Physician, pharmacist and center coordinator each conduct counseling sessions with each patient. Peer support groups are held. Additional elements of the intervention include: health information groups and provision of illustrations on how to take HAART; nutritional support for the patient and the family, along with nutritional counseling; staff training on adherence strategies; home care; coordinator responsibility for patients' adherence; and employing trained HIV-positive community health workers. Viral loads and pill counts assessed adherence. Over 94% of the 531 patients knew that they needed to take HAART each day at precise times for the rest of their lives, despite low levels of formal education, with over 32% being illiterate and over 41% having less than five years of schooling.  Over 90% of patients said they had a good relationship with their physician. More than 94% of patients knew they could transmit HIV through sexual relationships. More than 96% took all of their pills in the three days prior to the interview and 92% of patients kept more than 95% of their appointments to collect HAART.

A prospective cohort study in Kenya was done from 2009 to 2011 to examine the relationship between pill counts and adherence. Physicians conducted 1,236 unannounced pill counts during 1,818 clinic visits among 291 patients. The number of clinician initiated pill counts was positively associated with better adherence and better treatment outcomes.  Patients who were HIV-positive, treatment-naïve adults were followed for at least one year.  The patients were categorized into three categories based on the number of pill counts they received in 6 months: no pill counts, 1 to 3 pill counts, and more than 3 pill counts.  The patients with no pill counts had a 75.9% adherence rate, the patients with 1 to 3 pill counts had an 83.9% adherence rate, and the patients with more than 3 pill counts had a 92.4% adherence rate.  Further, there was a longer time to treatment failure for those with more pill counts.  The patients with no pill counts had an average 220 days to treatment failure, the patients with 1 to 3 pill counts had an average 438 days to treatment failure, and the patients with more than 3 pill counts had an average 497 days to treatment failure.  Of the 29.2% of patients that were treatment failures, 56.5% were virologic failures, 17.6% were deaths, and 25.9% were lost to follow-up.  The physicians conducted pill counts at 68% of clinical visits, which were scheduled monthly. The number of clinician initiated pill counts in this study was positively associated with better rates of adherence and a decreased risk of treatment failure.

A survey of 299, patients in Nigeria, 72.7% female, found that use of a pillbox increased self-reported adherence.

Data obtained from an observational cohort of 245 people living with HIV from 1996 to 2000 in the United States showed that pillbox organizers were estimated to improve adherence by 4.1 to 4.5% and was associated with a decrease in viral load of .34-.37 log10 copies/mL and a 14.2% to 15.7% higher probability of achieving a viral load of lower than 400 copies/mL, with statistically significant effects. "Pillbox organizers should be a standard intervention to improve adherence to antiretroviral therapy."

A Cochrane review of Lester et al., 2010 and Pop-Eleches et al., 2011 (see studies below) found that there is high quality evidence from two randomized controlled trials that mobile phone text messaging enhances adherence to ART compared to standard care.

A randomized controlled trail was conducted on HIV-positive women in Brazil from 2009 to 2010 to assess adherence with a mobile SMS message program and their impressions and satisfaction with the program.  The trial found that participants receiving the SMS messages consistently had better adherence rates by every measurement.  Self-reported adherence resulted in an adherence rate of 84.62% for the control group and 100.00% for the intervention group.  On the other hand, the counting pills method resulted in an adherence rate of 38.46% in the control group and 50.00% in the intervention group.  Microelectronic monitoring indicated that 46.15% of the control group and 75.00% of the intervention group were adherent. In terms of acceptability, at the end of the study 81.81% of the intervention group believed that the SMS messages aided in treatment adherence, and 90.90% said that they would like to continue receiving the messages. The study group consisted of 21 HIV-positive women, and due to this small size, no statistical significance was able to be demonstrated. All study participants had to own a phone, have a viral load below 400 copies/ml for at least 3 months, and have a CD4 count above 200 cells/µl.  Thirteen were allocated and completed the study in the control group, 8 were allocated and completed the study in the intervention group.  There was no statistical difference in age or any baseline characteristic between groups except that there were a higher proportion of married women in the control group.  There were differences between groups in terms of drugs and alcohol; 53.84% of control participants reported drinking alcohol in the last 30 days while only 12.50% did in the intervention group.  No participants in the control group reported using drugs while 50.00% in the intervention group reported drug use (including cigarettes, marijuana, cocaine, and ecstasy).  All the women received routine care over a 4-month period, where they were required to attend monthly clinic visits to assess adherence as well as receive routine health checks.  Three measures of adherence were collected to determine adherence at every monthly period: self-reported adherence, pill counting, and microelectronic monitors monitoring (which records when pill containers are opened and closed).  The microelectronic monitors were only placed on one of the ARVs for each patient.  The study authors note that measuring the adherence in the study as they did was very time consuming and doubled the time of the usual consultation.  During the last visit (5th visit), after adherence checks, the women in the intervention group were interviewed about their impressions on receiving the text messages.  SMS messages were sent to the intervention group 30 minutes prior to the time for the last required medication dose in a day.  These messages were sent on alternative weekdays and on Saturdays and Sundays and included the message: ‘The UNIFESP informs: take good care of your health’.  In terms of the impressions from the SMS messages, one participant in the intervention group thought the message should be changed daily and three participants thought the message should be sent closer to the medication time.  Seven participants said it helped them remember to take their medication, three participants said that they felt that the medical center or someone else cared about them, and one participant said that the SMS messages allowed for family involvement in her treatment. Women noted that it was important to them that the text message just said to take care of their health so that their HIV serostatus was not revealed in case someone saw their phone.

A randomized controlled trial from 2007 to 2008 in Kenya with 438 patients who initiated ART found that 53% of participants receiving weekly text messaged reminders achieved adherence of at least 90% during the 48 weeks of the study, compared with 40% of participants in the control group who received standard care. Participants receiving weekly reminders were also significantly less likely to experience treatment interruptions exceeding 48 hours during the 48-week follow-up period than participants in the control group.  Patients in the intervention group were provided a mobile phone and told that they could use it as they desired. All messages were less than 160 characters and did not specify HIV or ART in order to maintain confidentiality and were not timed to coincide with dosing schedules. A sample message was: "This is your reminder. Be strong and courageous, we care about you."   Every two months, patients in the intervention group received less than US$1 of phone credit and had to show their phone to clinic staff at every visit.

A randomized clinical trial with 538 patients in Kenya found that patients who received support via text messages from mobile phones significantly improved ART adherence and rates of viral suppression. Patients in the intervention group received weekly text (SMS) messages from a clinic nurse and were required to respond within 48 hours. Suppressed viral load was reported in 156 of 273 in the SMS group and 128 of 265 in the control group. Patients had to be able to access a mobile phone on a daily basis. Patients received the message "How are you?" (Mambo) and were instructed to answer that they were doing well (Sawa) or they had a problem (Shida). The clinical then phoned the patients who said they had a problem or who failed to respond to the text within 48 hours. Only 3.8% of weekly text messages identified a requirement for follow up by responding "Shida". Participants were classified as virologically suppressed if their HIV RNA load at their 12-month visit was 400 copies per mL or less. No adverse event directly attributable to the mobile phone communication, such as breaches of confidentiality was reported. At the end of the study, 191 of 194 patients in the intervention group reported they would like the SMS program to continue. In addition, during the political turmoil and violence in Kenya after the 2007 presidential elections, mobile phones were used to request clinic staff assistance to new safe locales where drug refills could be obtained. The intervention was inexpensive, with each SMS costing about US$.05, or $20 per 100 patients per month, and follow-up voice calls which averaged $3.75 per month.

A cohort in India who had mobile phone adherence support was followed for a month in 2010.  The support utilized both pictorial text messages and phone calls and the study assessed the perceived usefulness and acceptability of the program among HIV-positive individuals.  The study found that 73% of the calls were received, 90% of the participants reported that the mobile reminders were helpful as adherence support, and despite 18% of participants reporting that someone else had inadvertently received their call (and 13% of participants reporting that someone had received their message) no participants reported discomfort or stigma.  Eighty-seven percent of the participants reported that they preferred the calls as reminders over the text messages.  A total of 744 calls were made to 150 people over the course of a month (139 people completed the follow-up visit).  Participants received both an interactive voice response call and a pictorial image through SMS messaging.  The call was weekly and the participants were asked to report '1' if they had adhered to their medication the previous day and '2' if they had not.  If the call was not received up to three calls were made in the next 24 hours in an attempt to reach the patient.  The pictorial text message was delivered once a week and three days after the call.  After a month the participants were administered a questionnaire about their experience with the intervention.  Seventy-three percent of the participants were men, 85% were educated beyond primary school, and 83% were employed.  The median CD4 count was 435 cells/µl and women and the unemployed were more likely to report sharing a cell phone with another individual (86% of participants owned their mobile phone).  Of the completed calls, 77% of the responses were '1' (that participant adhered to their regimen the previous day), 2% were '2' (the participant had not adhered to their regimen the previous day), and 21% pressed an incorrect number or did not press a number.  For the SMS messages, 59% reported viewing all of the messages and 15% of participants reported never viewing the pictorial message.  The participants were more likely to receive the call than the text message, and 87% of participants preferred the call over the text message.  The study authors note that there are benefits to both options, writing that, "SMS has the capability of deliver a message regardless of the surrounding circumstances, i.e., SMSs are silent and can be received in situations where it is not appropriate to have a conversation.  On the other hand, SMSs leave a distinct trail where [interactive voice response] may be more discreet," (Sidney et al., 2012: 616).  While 99% of participants reported knowing how to make or receive a call, only 86% knew how to receive an SMS message and only 47% knew how to send one.

A cross-sectional study in rural Uganda assessed the impact of using mobile phones to increase clinic attendance to collect refills of ARVs. Of 276 patients surveyed, 177 (64%) had access to mobile phones and all but one were willing to be contacted for missed visit reminders. Of 560 total scheduled clinic appointments for the 176 patients followed for 28 weeks, 11% of visits were missed. In 79% of episodes in which visits were missed, patients presented for treatment within a mean duration of 2.2 days after mobile phone contact. Privacy and confidentiality were not considered a problem. Total cost for using mobile phones during the 28 weeks was US$14. Adherence was also assessed every four weeks using clinic based pill counts. "Women constituted the majority of clients surveyed, while men had more access to mobile phones" (p. 1349). 

A global review of studies, including one in Uganda, found an association between decreased viral load and community-based ART programs.  The review found 21 studies (not listed); 15 of which were high quality, randomized controlled trials or quasi-experimental programs. The remaining studies were high quality program evaluations.

A systematic review found that community-based antiretroviral programs may be as effective as facility based programs, with lower costs for both patients and services. Six different programs, summarized in 18 different articles, took place in Uganda, Kenya and Mozambique between 2006 and 2013. In one study in Uganda, community health workers delivered ART at home, provided adherence support, detected side effects and referred sick patients to clinics. Each community health worker was responsible for 40 patients and had six weeks training, along with a basic salary, motorbike and cell phone. The cluster randomized controlled trial found that at 36 months on treatment, the viral load was below 500 for 84% of patients in community based ART and 83% for facility based patients. After 36 months, loss to follow up was 1% for community based patients and 2% for facility based patients; mortality was 14% for community based patients and 13% for facility based patients, with costs lower for community based patients (Jaffar et al., 2009 cited in Decroo et al., 2013). A study in Kenya found that at 12 months, of the 96 patients randomly assigned to the community based arm, 5% were lost to follow up, 1% had died and 89% had an undetectable viral load. Among the 112 patients receiving care in the health facility, 5% were lost to follow up, none died and 86% has an undetectable viral load. Those in the community based arm had 50% fewer health facility visits. Community health workers, including people living with HIV, had a seven-day training followed by a two month practicum and was responsible for 20 clinically stable patients each (Selke et al., 2010 cited in Decroo et al., 2013). Community-based antiretroviral therapy reduced costs for patients but also was more cost-effective for health services. "…Community participation requires an approach which is contradictory to the dominant provider-driven development of health services" (p. 176).

A review of the published literature on the impact of community support services was done to evaluate the power of a diverse array of community support providers on ART program delivery and outcomes.  Data was used from studies published from 2004 to 2012 and located in Uganda, Ethiopia, South Africa, Brazil, Malawi, Namibia, Lesotho, Kenya, Rwanda, Tanzania, Zambia, Nigeria, Haiti, Guyana, Botswana, Peru, and Mozambique.  The review found that community support initiatives can broaden the access of programs, as well as have a positive impact on adherence, health outcomes, patient retention, and mortality.  Challenges that the study authors note regarding success of scale-up of ART programs in many areas include, "(1) lack of integration of ART services into the general health system; (2) the growing need for comprehensive care to address the psychosocial and economic dynamics of HIV/AIDS; (3) the need to empower patients on ART towards self-management; (4) the importance of defaulter tracing to improve retention in care; and (5) the crippling shortage in human resources for health," (Wouters et al., 2012: 2).  The study authors argued that community support could help alleviate some of these challenges.  Thirty publications were analyzed, including nine descriptive studies, five randomized controlled trials, four quasi-experimental studies, five cohort studies, and two qualitative studies.  Community support ranged from community health workers who underwent short training courses and provided support to other health workers, to peer health workers who are HIV-positive and who counseled patients on adherence and health education, to field officers who supported drug delivery and monitored patients. Almost all studies reported positive impacts of community support through a diverse set of programs. Community support programs were shown to increase access to and coverage of ART programs at a facility level.  Some studies also noted that community support providers could link HIV-positive patients to other parts of the health care system, thereby further integrating HIV services with other primary care services.  Adherence levels were also demonstratively improved in many studies with the implementation of a community support program. Many studies demonstrated that with community support, a higher proportion of patients lowered their viral load and increased their CD4 counts.  Some studies also demonstrated that loss to follow-up and mortality was decreased with community support services. The articles reviewed concerning directly-observed therapy for ART did not find any association between that form of community support and improved outcomes.  Community services may also serve to empower patients and to take on the issues associated with having HIV that go beyond treatment.

A global review of studies, including studies in Uganda and South Africa, found that providing ART through community volunteers, trained and supervised by a clinical officer, resulted in adherence equal to ART provided in clinic setting.

A retrospective cohort study was done on a group of individuals living with HIV starting a community-based HIV treatment program in Rwanda from 2005 to 2006.  The community-based program included free ART with directly observed therapy at home, tuberculosis screening and treatment, nutritional support, a transportation allowance, other assistance for those in extreme socioeconomic conditions, electronic medical records, and patient support groups.  The study found that 92.3% of people were retained in care at 2 years (5% were dead and 2.7% were lost to follow-up).  The median CD4 count was to 336 cells/µl after 2 years, from 190 cells/µl at initiation.  There were originally 1,041 people enrolled in the community-based ART program.  People were initiated on ART according to national guidelines which stated that any WHO stage IV, stage III with CD4 count of less than 350 cells/µl, stage I or II with a CD4 count of less than 200 cells/µl would be started on ART.  Co-trimoxazole was prescribed to all patients with a CD4 count of less than 350 cells/µl.  Patients who had documentation of a transfer were considered retained in care (3.4% of the initial cohort).  The cohort was 66.3% women and the mean age was 38 years old.  Age of more than 50 years and enrollment in 2006 (the second year) were associated with attrition. Among women, initial clinical stage was not associated with retention in care (attrition was much more likely among men with HIV stage III or IV).  The majority of the deaths occurred in the first year of treatment.  Tuberculosis was diagnosed in 17.8% of patients, of whom 89.4% completed treatment, 7.6% died, 2.4% transferred to another clinic, and 0.6% defaulted.  There was no association found between attrition and tuberculosis diagnosis.  Only about one-fourth of participants had a viral load measurement taken, but 97.5% of those with the measurement had viral suppression at 2 years.  The study authors estimate that the extra programs would cost an additional $630 per year per person for the first year and $340 per year per person for every following year.

A retrospective cohort study was done on patients starting first-line ART between 2003 and 2008 in South Africa to develop a cohort of patients switching to second-line treatment with a viral load less than 400 copies/ml. The cohort in a workplace-based program and a community-based program were evaluated to describe outcomes and predictors of viral suppression on second-line ART.  This study found that high levels of viral suppression can be achieved in treatment programs, with much better adherence rates in the community-based program as compared to the work-based program.  After the switch to second-line ART, 48.3% of the workplace patients and 72.0% of the community patients achieved viral suppression. Non-adherence was reported in 17.9% of workplace patients, compared to 1.4% of community patients. In the workplace program, having a lower viral load at switch to second-line treatment and younger age were associated with viral suppression at 15 months.  In the community-based program, patients with a shorter duration of viral load above 400 copies/ml (viremia), patients with a higher CD4 count, and patients who transferred into the program on first-line ART were more likely to achieve viral suppression.  Of the 14,779 patients who initiated ART in the given period, 205 in the workplace program and 212 in the community program were switched to second-line treatment.  Patients in the workplace had a median age of 43 years, 91.7% were male, and initiated ART at higher CD4 counts than the patients in the community-based program. Patients in the community-based program were 38.7% male and the median age was 36 years. Patients in the workplace program also had, on average, a longer duration of viremia.  The workplace program operated in 56 clinics serving mostly mining companies.  In the workplace program, employees were eligible for ART if they had WHO stage IV, a CD4 count of less than 250 cells/µl, or WHO stage III and a CD4 count of less than 350 cells/µl.  The community program operated in 81 private general practitioner and NGO clinics serving patients with limited resources.  In the community program, patients were eligible for ART if they had WHO stage IV disease or a CD4 count of less than 200 cells/µl.  In both programs a switch to second-line ART was recommended when a patient had two viral load counts above 1000 copies/ml with good adherence.  Patients were able to collect ART at intervals of 1 to 3 months, viral load and CD4 count was measured at initiation, 6 weeks, and 6 monthly intervals thereafter, and all services were free. Adherence counseling was offered at each visit.  Treatment failure was the recorded reason for switching for 82.2% of the workplace patients and 83.8% of the community-based patients. Non-adherence was a reason for switching in 7.8% of workplace patients and only 0.5% of patients in the community program. Study authors noted that, "with no access to resistance tests and imperfect adherence assessment tools, deciding who is failing therapy and might benefit from switching is difficult," (Johnston et al., 2012: 9). Of those with measurements after the first report of viral suppression, 35.6% of patients in the workplace program and 13.2% of patients in the community program experienced at least one episode of viral rebound. At 15 months, 73.7% of workplace patients and 84.4% of community patients were alive and in care.

Use of community health workers to increase patient retention in Rwanda resulted in 92% of patients being retained in care after two years and only 3% lost to care. (Rich et al., 2012 cited in Ying et al., 2014). 

A Cochrane review of 16 studies published from 1996 to 2013 in Malawi, Ethiopia, Uganda, Kenya, Swaziland, South Africa, Lesotho, Mozambique, Rwanda, Tanzania, Thailand, and Nigeria was undertaken to assess the effects of various models of decentralized HIV treatment on initiation and maintenance of ART.  The type of care was categorized as partial decentralization (where ART is initiated at a hospital and maintained at a health center), full decentralization (where ART is initiated and maintained at a health center), and community ART provision (where ART is initiated at a health center or hospital and maintained in the community).  The outcome measured was ‘attrition’, which included loss to care and mortality.  The review of four studies with 39,090 patients in total found that in a partially decentralized system, there was less attrition and fewer patients were lost to care after one year.  The review of four studies with 56,360 patients in a fully decentralized system found that there was no difference in attrition from other models. ARVs started at a hospital and maintained at a health center had the least loss to follow-up.  Finally, a review of two studies with 1,453 patients in a community ART provision system found that there was no difference in death or loss to care as compared to patients who received care at a health center.  None of the decentralized systems were found to lead to worse health outcomes.  Two of the studies used were cluster-randomized trails and the remaining 14 were cohort studies.  This review demonstrated that different models of decentralization can lower attrition and loss to follow up and none led to worse health outcomes for the patients.

A systematic review to evaluate programs to improve linkage to care or retention in pre-ART care until initiation of ART was conducted based on studies from Cambodia, India, Kenya, Malawi, Mozambique, Rwanda, South Africa, Swaziland, Uganda, and Zambia.  The study found several interventions that improved retention in care including integration of care and peer supporters to improve linkage to care, medical incentives to improve retention, and point-of-care CD4 testing and food incentives to improve rates of completion of ART eligibility assessment and initiation.  Twenty-four studies conducted from 2004-2012 were used in the analysis.  The studies all included data on interventions provided before initiation of treatment with the goals of improving linkage or retention in pre-ART care or initiation of ART.  The systematic review included seven randomized controlled trials, seven observational studies with control groups, and ten pre/post studies.  Interventions tested included: point-of-care CD4 testing, service integration (mostly HIV services with antenatal care), health systems interventions including improved communication, referral, and treating, and incentives (food and monetary).  Most studies included data on completion of ART eligibility screening or ART initiation; fewer reported on linkage to care. Packages of health system interventions that improved referrals, communication, and teaching improved ART initiation by 15-27% in three studies.  Point-of-care CD4 testing improved rates of patients completing ART eligibility assessment and ART initiation in multiple studies.  Point-of-care CD4 testing had a less meaningful impact on retention in pre-ART care and on linkage to care. Peer support and counseling following HIV testing was found to increase linkage to care in three studies.  Food incentives in India were found to lessen time to ART initiation and increase the proportion of people retained in pre-ART care.  Medical incentives in the form of co-trimoxazole was found to increase pre-ART retention in Kenya. Overall, this systematic review demonstrated that minimizing patient facility visits, providing counseling and peer support, and providing incentives may improve rates of retention in pre-ART care and rates of ART initiation.

A retrospective cohort study of those over age 18 of 17 facilities with 4,147 patients from Tanzania, Uganda and Zambia found that community ART dispensing had significantly less attrition than other types of facilities. Patients started ART between 2003 and 2010 and 64.4% were female. Government, NGO and faith based facilities were included in the analysis. Four programs used community-based distribution of ARV drugs for stable patients. In sites with community ARV drug distribution, attrition proportions among both men and women were about 50% smaller compared with women in sites without community distribution.

Four approaches developed by Médecins Sans Frontières to simplify ART delivery for stable patients on ART through community-supported models of care were assessed: 1) appointment spacing for clinical and drug refills in Malawi; 2) peer educator-led ART refill groups in South Africa (Luque-Fernandez et al., 2013); 3) a community ART distribution point in the Democratic Republic of Congo; and 4) patient-led community ART groups in Mozambique.  All programs reduced the burden for patients via reduced travel time and reduced lost income and the burden on the health system via reduced patient volume.  The retention in care was high for all four programs: 1) 94% of patients retained at 36 months in appointment spacing system in Malawi; 2) 97% at 40 months in a peer education program in South Africa; 3) 89% at 12 months in a community ART distribution point programs in the Democratic Republic of the Congo; and 4) 92% at 48 months in community ART groups in Mozambique.  Separating ART delivery from clinical assessment was found to benefit both patients and programs in these diverse settings.  In Malawi, 8,528 patients between 2008 and 2013 were enrolled in an appointment system where stable patients received ART in the community and only attended clinical visits once every 6 months.  In South Africa from 2007 to 2011, 502 patients in adherence clubs where groups of around 30 patients met every 2 months for general health assessment and to pick up medication were evaluated against 2,327 patients in standard care.  In the Democratic Republic of the Congo from 2010 to 2013, 2,162 patients were followed who had access to community ART distribution points where ART was distributed to stable patients at locations managed by people living with HIV.  These distribution points were closer to the patient’s homes (saving the patient’s time and money) and also provided adherence support and basic health checks.  In Mozambique, 5,729 community ART group members were followed from 2008 to 2012, where members were organized into groups of 6 and alternated which of the six patients went to the clinic to pick up medication for all six patients while having their own clinical check-up.  These diverse programs and settings all separated ART delivery from the clinical assessment of the patient, in order to make the ART more accessible.  In every case, the rates of retention were higher in the community-supported models than the control group (when available) or other published local data.

A study was done on two rural hospitals with twelve nurse-led health centers in Lesotho to compare outcomes of patients starting ART between facilities.  The study found that the health centers had fewer early losses and similar long-term retention to care as the hospitals.  When stratified by sex, the rates for women were similar between the hospitals and health centers; however, men had a higher retention to care at health centers as compared to hospitals.  There were 3,747 patients above 16 years old that initiated ART at the hospitals or health centers from 2008 to 2011.  Of these patients, 54.5% started ART at one of the health centers and the remaining patients started treatment at the hospitals.  The hospitals were staffed with physicians and provided basic lab services. The health centers were staffed with nurse-clinicians, nurse-assistants, and lay or community counselors. The counselors provided HIV counseling and monitoring and traced patients that were lost to follow-up.  The nurses initiated ART and managed the clinical monitoring of the patient.  Retention in care, no follow-up (not returning after initiation of ART), and lost to follow-up (not returning to facility for 6 months after at least one follow-up visit) were all measured as part of the analysis.  All patients initiating care in Lesotho are given a unique national identifier number, which is used in the instance of transfer from one center to another.  At health centers retention to care was 68.8% at three years and at hospitals retention to care was 64.1% at three years. Three-year retention for women was 69.4% in hospitals and 68.7% in health centers - a difference that is not statistically significant.  For men, three-year retention was 54.7% at hospitals and 68.8% at health centers.  Men who were treated at health centers were significantly more likely to be retained in care at three years. Among women, 2.9% of hospital patients and 1.8% of health center patients had no follow-up visit. Among men, 2.9% of hospital patients and 1.7% of health center patients had no follow-up visit. The proportion of patients who died or were lost to follow-up were slightly lower in health centers as well.  The proportion of patients who died was 10.1% at health centers and 13.5% at hospitals.  The proportion of patients who were lost to follow-up were 12.7% at health centers and 13.9% at hospitals.  The health center enrollee population did have less advanced disease and better immunological status.  The only statistically significant difference was the rates of loss to follow-up for men between the health centers and hospitals.  The study authors hypothesize that this difference, "may be employment related," as many men, "work as subsistence farmers or as labourers...the easier accessibility of [health centers] may be of particular benefit to men," (p. 7).

Longitudinal analysis of HIV patients from 2001 and 2008 was conducted on 4,322 patients who received centralized and 11,090 patients who received decentralized care in rural Malawi and found that program attrition after two years was lower among patients receiving decentralized care (9.9 per 100 person years, compared to 20.8 per 100 person years), with no statistically significant differences in one-year levels of viral suppression observed between the two health care levels.  Routine CD4 counts were conducted annually, with viral load testing only if a provider suspected treatment failure. Median CD4 cell counts were higher in decentralized facilities, meaning that patients were accessing care earlier in decentralized facilities, possibly due to reduction of patient travel time and costs. A cross-sectional survey of 312 patients in centralized and 390 patients in decentralized care was also conducted in 2009. Between 2001 and 20002, 78.5% of ART initiation took place at the centralized level; by 2008, 86.5% occurred in health centers. In each time period, higher proportions of men were initiated on ART in centralized rather than in decentralized facilities.

Retrospective cohort data from vertical and integrated Médecins sans Frontières programs in the Central African Republic, Côte d'Ivoire, the Democratic Republic of the Congo, Ethiopia, Nigeria, the Republic of the Congo, Uganda, Zambia, and Zimbabwe was used to compare the survival probabilities between programs from 2003 to 2010.  The review showed that after adjusting for baseline population characteristics, there were similar outcomes for mortality and lower risk of loss to follow-up in the integrated programs compared to the vertical programs.  Vertical programs were programs that specifically treated HIV and integrated programs provided comprehensive health care including HIV services.  Greater program experience was shown to be protective against death, although risk of loss to follow-up increased in programs with greater experience.  Initiating ART at WHO stage IV disease was the characteristic most associated with death.  A total of 15,403 HIV-positive people were followed in this analysis; with 14,124 patients in 7 vertical programs and 1,279 in 10 integrated programs.  The process for HIV testing and treatment was uniform across the different programs.  The study authors noted many reasons that integrated HIV programs may have potential advantages over vertical programs, including, "increasing the access of patients to HIV care in areas where vertical HIV programs may not be feasible; supporting retention in care by bringing services closer to patients... normalizing HIV as one illness among many, with potential stigma reduction; [and] allowing the treatment of patients for multiple conditions in the same facility by the same staff," (p. e92).  Patients older than age 15 years and who had more than 6 months of follow-up time were included in the analysis.  Vertical programs had 64% female patients compared to 71% in integrated programs. Vertical programs also had a lower proportion of patients classified as WHO clinical stage IV at ART initiation, 18% compared to 36% in integrated programs.  Average CD4 count at ART initiation was higher at the integrated program, but CD4 measurement was not uniformly measured in the integrated programs. Without adjusting for these variables, the proportion of patients who died was lower in vertical programs.  After multivariate analysis, the adjusted hazard of death was similar in both programs.  The integrated programs also had a 29% lower risk of loss to follow-up when compared to the vertical programs.  The study authors noted also that, "a standard package of protocols and training materials helped staff, as did the engagement of an experienced HIV clinician to help start the program and train staff," (p. e96).  

A meta-analysis was done to examine the efficacy of viral load monitoring as a tool to identity patients who need adherence interventions, and their subsequent outcomes with implementation of these interventions.  Eight studies published from 2004 to 2013 in South Africa, the United States, Thailand, Mali, Burkina Faso, Swaziland, India, and France were included in analysis.  Five of the studies, which reported percentages of patients with viremic resuppression, had a combined resuppression rate of 70.5% after the implementation of an adherence intervention following viral load testing.  The remaining 3 studies all reported declines in mean viral load.  Viral load monitoring identified patients who experienced an increase in viral load after previous suppression and who benefited from an adherence intervention.  Patients in need of adherence interventions and those in need of second-line therapy could be identified through virological monitoring.  Six of these studies were retrospective cohort studies and 2 were prospective cohort studies.  Follow-up periods ranged from 3 months to 4 years.  The threshold of viral suppression ranged from less than 50 copies per milliliter to less than 1,000 copies per milliliter.  There was also variability in the adherence interventions, which included peer support, adherence counselors, peer counselors, educational sessions, home visits, adherence support tools, and short-term directly observed therapy.  Although there was a good amount of variability between each cohort, the rates in the five studies reporting on proportion of patients resuppressing after an adherence intervention ranged from 54.2% to 89.1%.  The remaining 3 studies reported significant declines in mean viral load. All studies showed that a majority of patients experienced resuppression with adherence interventions. When viral load testing began later, more than 60% of patients had already developed resistance mutations, suggesting that early viral load monitoring is vital to identify patients in need of additional adherence interventions.  Studies showed that 78% of patients suspected of failing second-line treatment experienced viral resuppression with adherence interventions, which was particularly important where third-line treatment options continue to be expensive and not widely available.

A meta-analysis and systematic review of studies published between 2001 and 2009 on resistance patterns of the most commonly used ART regimens showed a higher resistance at 2 years in patients whose viral loads were less frequently monitored. A total of 8,376 patients Cameroon, Haiti, Malawi, South-Africa, Uganda, Zimbabwe, Europe and Canada were included in the studies. Group comparisons were made based on viral-load monitoring frequency: infrequent or none versus frequent. Resistance at virological failure to NNRTI in patients monitored at intervals of more than 3 months or none was 88.3% compared with 61% in patients monitored more frequently. Other mutations (thymidine analogue and lamivudine) were also higher in those who received no monitoring or the less frequently monitored group of patients. Mutations in all of the three classes of commonly used drugs in resource limited countries were substantially higher in those with virological failure in patients on NNRTI-based first-line ART for which viral load monitoring was absent or infrequent compared to patients who were monitored intensively.

A study conducted in Asia (the TREAT Asia Observational Database) found that viral load testing at intervals of less than annual frequency increased disease progression with poorer prognosis at low-income sites. Data of 2,333 patients from more than 11 Asian countries initiating on HAART since 2000 was analyzed.  Sites were categorized according to the World Bank country income criteria as high-income or low-income and a yearly frequency of VL testing (3 or more, 1 to 2 or less than 1) or CD4 measurements (3 or more and less than 3). Low-income sites contributed to 61% of the eligible patients. Women constitute 35% of the low-income and 20.1% of high-income site patients. Patients at low-income sites and with sites reporting less frequent viral load testing had a poorer prognosis. Less than 1 viral load testing per year was significantly associated with higher rates of progression to CDC class C event or death. Patients at sites with less than annual reported viral load testing had lower odds of achieving virological suppression at 12 months than high-income sites which reported viral load testing frequency of 3 or more per year. In conclusion, low measures of site resourcing were associated with 35% increase in progression to AIDS and hence poor outcomes in patients from sites with viral load testing with intervals of less than once per year.

A study of a high risk express care program for 635 patients who were just initiating ART or who had a CD4 count of less than 100 in Kenya between 2007 and 2009 found that those enrolled in the high risk express care program were much more likely to be alive and in care following ART initiation for up to two years later. Following ART initiation by a clinical officer or physician, a nurse was responsible for interim weekly visits either physically or by phone for a period of three months, with a monthly visit by the patient to a clinical officer or physician. Patients in the high-risk express care program did not need to wait in any lines but could go directly to one-stop care. Nurses followed up any patient who missed an appointment.

A treatment program in Malawi that initiated active follow up when ARV patients were three weeks overdue for their appointments reduced the loss to follow up by 59% among 1,158 loss to follow up patients who had not died, transferred out or declined an interview.