Recent evidence on challenges in patient retention
Although the scale-up of treatment programmes in resource-limited settings has dramatically improved the health, well-being, and survival of most of the people with HIV now receiving antiretroviral therapy (ART), some of the rapidly expanding programmes are having difficulty keeping their patients on treatment and coming back for follow-up clinic visits, according to reports at the Fourteenth Conference on Retroviruses and Opportunistic Infections last month in Los Angeles.
What’s even more troubling, according to operational research from the antiretroviral programme in Botswana, is that most of the patients who are lost-to-follow-up are dying.
Despite great success scaling-up, researchers from some of the vanguard ART programmes in resource limited settings described key challenges facing their programmes during a poster discussion session at the conference. Barriers to care remain in many settings, such as paying for treatment, and the need to take time off work to travel what are sometimes great distances to clinics only to wait whole day for service. And with increasing numbers of people joining programmes, it becomes increasingly difficult to give them the sort of individualised attention each patient needs for appropriate symptom and opportunistic infection diagnosis and management. There is a serious risk that many will ‘fall through the cracks’ or not receive the adequate support to keep them in care — partly because human resource constraints limit a clinic’s ability to actively follow up their patients.
However, given the human resource challenges confronting ART programmes, and health programmes in general in the developing world, more aggressive management to keep people in care may require hiring and training new cadres of lay-professional health care workers to provide more active follow-up.
ART programmes make remarkable progress
Several years into the global roll out of ART, the key lesson learned is that it is indeed possible to get life-saving care and treatment to hundreds of thousands of people living in very diverse resource-limited settings. There were several reports at the conference of projects that have demonstrated this.
For instance, since July 2004, the International Center for AIDS Care and Treatment Programs at Columbia University (ICAP) has built upon a network of projects, which initially offered services to prevent mother-to-child transmission (PMTCT) of HIV, to establish an enormous comprehensive HIV care programme with 155 sites in Ethiopia, Kenya, Mozambique, Nigeria, Rwanda, South Africa and Tanzania. According to Dr. Wafaa El-Sadr, by December 2006, these sites had enrolled a total of 171,259 HIV-infected patients, 71,482 of whom have been initiated on ART.
Most of the programme’s patients on ART remain on their first-line regimen and have had good immunological responses. Across the countries the programme serves, 4 to 6% of the ART recipients have died and 1-17% have been lost to follow-up (more on this below). Of the remaining patients on ART, nearly all the adults (98%) and children (93%) are still on their first-line regimens. The average increase in median CD4 cell counts after six and twelve months of ART was 116 and 149 cells/mm3, respectively.
Likewise, a poster by Richard Marlink and colleagues described the phenomenally rapid scale-up of Project HEART (Help-Expand-ART), which was initiated in March 2004 by the Elizabeth Glaser Paediatric AIDS Foundation with funding from PEPFAR/CDC and private donors. In just 31 months, this programme enrolled over 167,000 adults and 87,000 children into care at sites in Cote d’Ivoire, South Africa, Tanzania, Zambia and recently Mozambique. Again, treatment outcomes in the programme are comparable to those in developed settings. The poster’s authors concluded “care and treatment was possible in a variety of settings, including areas of conflict, remote settings and very poor public sites.”
Many of these programmes might not be possible were it not for the example of one of the earliest and most catalytic ART programmes in the world, the Khayelitsha township project in the Western Cape of South Africa. Dr. Gilles Van Cutsem of Médecins Sans Frontières presented an analysis of the outcomes and emerging challenges at the project’s three clinics. Since it was one of the very first to offer ART in a developing country (beginning in 2001), the Khayelitsha project is now one of the first to be able to report outcomes for patients after four and five years of treatment.
Dr. Cutsem reported on 3373 patients enrolled between 2001 and 2005 with follow-up until September 2006. Although only a relatively small number of people entered the programme in 2001, of those to reach five years of follow-up, 70% were still alive and in care (about 16.9%, (95% confidence interval (CI) 13.6-21.0) are known to have died). Most (79%) of the patients on treatment at five years are still on their first line regimen.
Responses on treatment for the programme as a whole (and thus including those on first-line and second line regimens) have been exceptional. The mean increase in CD4 cell count from baseline after four and five years on ART was 429 (95% CI 388-469) and 440 (95% CI 341-541) cells/mm3 respectively. The programme is also unusual in resource limited settings in that it has been able to document viral load outcomes in its participants, with the proportion of patients [on treatment] with viral loads <400 copies per ml at years 1, 2, 3, 4 and 5 at 88.4%, 85.8%, 88.5, 79.1% and 82.1% respectively.
There were also oral reports from the national ART programme in Cameroon, Botswana’s treatment programme (the first national public health programme to offer free ART), and Uganda (where a variety of government clinics and other projects provide treatment). However, now that the rollout has been demonstrated as quite possible, the focus has shifted to the challenges, some of which have only emerged as programmes begin to reach the limits of their infrastructural capacity, to providing ongoing high quality care.
Challenges keeping patients in care despite reducing cost of ART in Cameroon
Several studies in resource limited settings have previously reported that requiring patients to pay for ART is a major barrier to adherence (see http://www.aidsmap.com/en/news/3E8B197B-3939-46E6-BF00-0CDE13F722D5.asp) and access to care that is associated with a high rate of losses to follow-up (see http://www.aidsmap.com/en/news/AFB128B5-AEF7-4FAF-9A0B-E0322C3BA99F.asp) and poorer clinical outcomes (see http://www.aidsmap.com/en/news/4B6E4B14-E87D-4630-AEB3-55ADB431D908.asp). And in one Ugandan study presented at the Retrovirus Conference, 13.7% of almost 700 patients receiving ART at two clinics in Kampala discontinued treatment for at least one month — and the commonest reason for discontinuation, reported by 43%, was drug cost (see http://www.aidsmap.com/en/news/4A668DC7-AC9F-41CC-861F-CA0F3271E2BE.asp).
Data presented Dr Jembia Mosoko of the CDC in Cameroon suggest that even when fees are dramatically reduced, it can still be difficult to keep patients in care and on treatment.
Cameroon’s ART programme began in 2002 at the Limbe Provincial Hospital in the Southwest Province. Dr. Mosoko and colleagues performed a retrospective analysis of the clinical and pharmacy payment records for 2920 patients who enrolled in the programme between February 2002 and December of 2005.
The median CD4 count at entry was 107 cells/mm3 and as seen in so many other programmes, those who stayed in care had good increases in their CD4 counts (to a median of 294 at 6 months, and 340 at 12 months).
However, loss of patients to this programme was rather substantial — 22% of patients dropped out of care after their very first clinic visit and even more patients skipped subsequent clinic visits. Adherence, defined by this analysis as attendance at scheduled clinic visits, was also poor. In fact, 77.4% of the participants attended less than four out of five of their scheduled visits. The probability of remaining alive and in care was 0.66 (95% CI 0.64-0.68) at 6 months, 0.58 (95% CI 0.56-0.60) at 1 year, 0.47 (95% CI 0.45 to 0.49) at 2 years and 0.35 (95% CI 0.32 to 0.38) at 3 years.
One of the key reasons for the poor retention and adherence was that 96% of the patients paid for ART out of pocket — either by themselves or with help from their families. Only a small proportion of patients had other help, such as an employers paying for the program. In October 2004, there was a major price reduction for ART from $30-50 per month down to $6 per month so one of the issues investigated was whether the price reduction enhanced enrolment or treatment outcomes.
After the price reduction was announced, the number of patients enrolling each month almost doubled. But surprisingly, the reduction in cost did not improve retention in care [hazard ratio (HR) 1.1, (95 CI 0.92-1.2).
In a multivariate analysis, patients were more likely to remain alive and in care if they didn’t have to pay for the treatment (adjusted HR 3.34, 1.76 to 6.24; p <0.001).
According to one researcher at the conference the take home message is that “free is good and paying is bad.”
The study’s findings may have already changed public policy in Cameroon. According to Dr Mosoko, “We had the opportunity to listen to the Minister of Health who publicly promised the Cameroonian population that within the next few weeks to come, treatment will be free.”
Does gender affect patient retention?
A couple of other factors were also associated with better outcomes in the multivariate analysis of the Cameroonian cohort. For instance, female gender was also associated with remaining in care (HR 1.2, 1.1 to 1.3; p = 0.003).
Others studies, though not all, have made similar observations concerning gender’s role in treatment outcomes. For instance, male sex was independently associated with death within the first 12 months after initiating ART (adjusted RH 1.8, 95% CI 1.0 to 2.6) in a report from the Botswana ART programme, but in the Khayelitsha project there was no association between gender and mortality on treatment.
According to Dr. Gregory Bisson, of the University of Pennsylvania, who presented the data from Botswana, there were significant differences in the baseline CD4 counts between the men and women. “The median CD4 cell count for men was about 65 and for women it was about 95, he said. “But even after adjusting for that, the effect was still there.”
Nevertheless, these data suggest that in some settings there could be culturally based differences in health-seeking behaviour between men and women that persist even after going on ART. These could be addressed by targeted interventions.
Travelling distances to the clinic: one reason to decentralise care
The final factor in the multivariate analysis associated with remaining in care in the Cameroonian cohort was having good access to the clinic (HR 1.5, 1.3 to 1.8; p <0.001). The definition of good access to the clinic was rather liberal — living within less than 40 kilometres of the clinic (or as far away as 80 km in the nearby city of Douala if on a main road). In fact, some of the patients in the programme actually lived more than 150 kilometres away (54.3% of these participants dropped out of the programme or died).
The great distances - often hundreds of kilometres - that some people in resource-limited settings must travel to reach pilot ART clinics have long been known to serve as a barrier to care. And as in most countries, Dr Mosoko and colleagues recommended decentralisation to establish clinics in other districts. “This treatment centre was just one of 23 initial ART centres that we created around the country. Later on the scale up process continued to create smaller treatment units to cover some of the other district populations.”
As ART programmes have scaled up and decentralised by opening new clinics to service other regions, access has of course improved. However many people with HIV in developing countries are still too far away to have convenient access to their nearest ART clinic — frequently in areas without good roads or affordable transportation, which can significantly add to the cost of getting refills and staying on treatment.
According to a recent article in IRIN Plus News loss-to-follow-up is an emerging problem in Namibia, where “concern has been growing over possible treatment interruptions for rural dwellers who start taking antiretrovirals and are then unable to make follow-up visits. ‘It is not unheard of for patients to walk up to 20 kilometres to Eenhana Hospital because they are short of cash to pay for a ride. Sometimes they come to the clinic and sometimes they don't,’” said one staff member at the clinic. (see the entire article at http://www.plusnews.org/AIDSreport.asp?ReportID=6733&SelectRegion=Southern_Africa&SelectCountry=NAMIBIA).
Without reliable transport — or the ability to pay for it — it can even be difficult to make it to a clinic in one’s own town, particularly for someone who feels ill.
“I notice that there’s a great reluctance by programmes today [to address] the transportation cost for the patients,” noted Dr Wafaa El-Sadr during the discussion of this study. “And often it’s a big barrier, even within cities themselves because of the required number of visits and so on. Somehow it’s not traditionally the kind of thing you do, to give the patients vouchers or something to come to the clinic. [Whereas] in the United States, for example, you can provide poor patients that I see in my clinic in Harlem with transportation vouchers. It’s just a global problem.”
High death rates among those lost to follow-up
And what happens to most of these patients who don’t make it back to the clinic?
“We tried to do some analysis looking at factors associated with this early dropout,” said Dr Mosoko. “And one of the factors was actually death.”
Another study designed to explore what was happening to patients lost-to-follow-up in Botswana’s national ART programme reached precisely the same conclusion — most were dead.
The study was designed to determine whether outcomes as measured by passive follow-up could accurately reflect the success of ART programmes. Dr Gregory Bisson of the University of Pennsylvania and colleagues in Botswana performed a retrospective study among 410 patients who went onto ART during the first year of the Botswanan programme at the Infectious Disease Care Clinic (IDCC) in Gaborone, comparing outcomes collected by passive follow-up to those subsequently collected by active follow-up. With passive follow-up, patients were classified as dead if their death was recorded anywhere in the clinical record, and as lost to follow-up if their last contact with the clinic was more 30 days past their last visit. By active follow-up, any lost patients were actively traced by phone calls or home visits.
The median duration of follow-up for the 410 individuals in the analysis was 44 weeks. Again, as in the other studies: “The virologic and immunologic response rate of patients who survive long enough to be monitored [at six months] is incredibly good.” 94% of those with six months follow-up data had undetectable viral loads at that time and 80% had CD4 counts above 200 despite a median baseline CD4 count of only 80.
However, 29 (7%) individuals were classified as dead, by passive follow-up, and 68 (17%) were lost to follow-up. Then, after tracking down the lost patients, 69 (17% of the entire cohort) were found to be dead and 22 (5%) remained lost. Thus passive follow-up missed 40 patients (59%) who had died (p <0.001).
While the observation that outcomes are poor among those lost-to-follow-up is nothing new, the proportion who died fairly soon at starting ART in this model programme is about as high as has been reported anywhere, and suggests there could be even higher death rates in programmes with higher rates of loss to follow-up.
Unfortunately the study provided little new insight into the causes of death.
“Clearly if you are no longer alive, it is highly unlikely that you will be coming back to the clinic,” said Dr James Hakim of the University of Zimbabwe, who co-chaired the poster discussion session. But he suggested that it would be more useful to get to ”the root of the causes of death — it’s probably germane to know what the causes are early and late, so that those can be addressed in the programme —and what proportion of patients died because they defaulted as opposed to those who defaulted because they died.”
The study identified a couple of risk factors for death within the first year of starting ART, including male gender (as mentioned earlier) and a CD4 cell count below100 at baseline (adjusted RH 4.8, 95% CI 2.4 to 9.3) — and, the median survival time of those who died was about 44 days. “But it’s going to take careful prospective studies to really determine what’s going on, whether it is immune reconstitution or overwhelming infection from opportunistic infections or adverse events,” said Dr Bisson. “We don’t know the causal pathway. So, we don’t know if patients are defaulting when they’re dying or if they’re dying on HAART. We don’t have data to really answer that question.”
However, Dr Bisson believes that implementing active follow-up in ART could save lives.
“If you could follow up intensively on all patients in a setting of rapid scale up - make calls, home visits — if you had the resources to do that, it would actually make a difference in outcomes and I think that gets at the causal mechanism of why patients are dying,” he said. At the IDCC site, “it took about two calls per patient - several patients needed more than two calls. About 50% needed about three calls, and about 20% of the patients needed home visits.”
However, simply following up on patients may not be enough to keep them in care. “There’s other data from Zambia where they followed up on a significant proportion of the patients that were lost to follow up there and they found that they could make many calls and they were unable to get patients back into care,” he noted.
Loss to follow-up increasing in some ART sites reaching capacity
Dr Bisson also conceded that this analysis was performed early in the course of the Botswana ART programme. Like the Limbe Provincial Hospital in Cameroon, the IDCC in Gaborone was one of the first ART clinics to open in the national programme and thus some of the findings of this study may be specific to the early stages of the programme. Programmes probably need to perform such operational assessments on a regular basis. As a programme matures and more ART sites open, one would hope that factors such as the travel distance to the clinic would have less of an impact on patient retention.
Unfortunately, data presented from several of these programmes suggest just the opposite may be occurring — largely because of the difficulty in managing large numbers of patients
For example, even though loss to follow-up in the ICAP programme was low overall (ranging between 1-17% between each national project), it was higher at some of the sites in South Africa, and Tanzania (over 20%), and in some sites in Kenya, it was over 35%. According to Dr El-Sadr and colleagues. “Sites with higher death and lost-to-follow-up rates tend to be those that have been offering services for longer periods of time.”
Marlink et al also noted that Project HEART needed to improve “overall patient follow-up, management of toxicities/opportunistic infections and providing basic HIV care to the large number of patients both on treatment or not yet eligible for treatment. Many treatment sites have reached capacity, emphasising the need to add more primary care facilities and to train [more] physicians in HIV/AIDS care and treatment.”
Likewise, although there has been a reduction in early reported deaths in the Khayelitsha programme (partly explained by an increase in CD4 cell counts when patients are going on ART), losses to follow-up have also been increasing — occurring earlier and in a greater proportion of patients with each successive year. In 2001 and 2002, there was virtually no loss to follow-up within the first twelve months. However, loss to follow-up within the first year on treatment occurred in 2% in 2003, 3% in 2004, and 6% in 2005.
According to Dr Cutsem in his poster presentation, “Losses to follow-up have increased dramatically since the clinics reached a critical patient load. These losses include patients who had been on therapy for a few years. [This could be] a reflection of the challenges facing the clinics in Khayelitsha currently where the rapid increase in patient numbers has outstripped the service’s ability to maintain consistent community-based follow-up of defaulting patients. It is also likely that other components of the care and counselling pathways have been compromised by the rapid scaling up of care.”
When the project in Khayelitsha began, and access to ART was extremely limited, adherence support was prioritised and patients were even selected for participation in the programme based upon their motivation and likelihood of being able to adhere to treatment. This usually involved designating a friend or family member partner to help them take their medication regularly. However, in order to keep putting more patients on care, it may become more difficult to continue giving the same intensive adherence support programme.
“When we started to increase roll out massively in 2004 and 2005, where you had a thousand people, then 1500 people yearly enrolled on ARVs, the clinics became saturated because we didn’t have the authorisation from government to open more clinics,” said Dr Cutsem.
“So we were stuck with three clinics that became monster clinics with now almost 2000 patients on ARVs per clinic with a relatively small number of staff and that’s when the loss to follow-up started to increase even among patients who have been adherent for two to four years in the beginning of the programme. And that’s what most programmes are going to face now with massive scale-up.”
Dr Cutsem recommends that ART-scale up be accompanied “by decentralisation of treatment to the most peripheral level and to a greater number of service points in order to avoid the creation of “monster” treatment sites where adequate follow-up cannot be maintained.”
In order to improve follow-up, Dr Cutsem and colleagues also recommended that: “appropriate systems need to be created to accommodate the ever-increasing patient load. This includes capacitating all levels of health care workers (in the South African settings especially nurses) to deliver and monitor ART and ensuring additional adherence support by non-medical staff.”
ICAP is conducting cross-programme surveys “trying to learn from one site to another” to assess just what sort of staffing might be needed (how many staff members of every different category, whether they are doctors, nurses, “lay-professional staff or volunteers) to run a site well, in terms of enrolment, and the proportion of various staff members needed per patient to keep them in care or on ART.
“Obviously it has an impact,” said Dr El-Sadr, “especially in growth, in terms of enrolment at these very large sites that are reaching the capacity at this point in time.”
El-Sadr W et al. Variability in populations enrolled and their outcomes in HIV care and treatment programs across countries in sub-Saharan Africa. Fourteenth Conference on Retroviruses and Opportunistic Infections, Los Angeles, abstract 534, 2007.
Van Cutsem G et al. Clinical outcomes and emerging challenges after 5 years of ART in a South African township. Fourteenth Conference on Retroviruses and Opportunistic Infections, Los Angeles, abstract 535, 2007.
Mosoko J et al. Survival and adherence to ART in an era of decreasing drug cost in Limbe, Cameroon. Fourteenth Conference on Retroviruses and Opportunistic Infections, Los Angeles, abstract 536, 2007.
Bisson G et al. High rates of death among patients lost to follow-up in Botswana’s national ART program: implications for monitoring and evaluation. Fourteenth Conference on Retroviruses and Opportunistic Infections, Los Angeles, abstract 537, 2007.