A Successful Model of Expedited Antiretroviral Therapy for Clinically Stable Patients Living with HIV in Haiti

Abstract

Background

Recommendations for universal antiretroviral therapy (ART) have greatly increased the number of HIV-infected patients who qualify for treatment, particularly with early clinical disease. Less intensive models of care are needed for clinically stable patients.

Setting

A Rapid Pathway (RP) model of expedited outpatient care for clinically stable patients was implemented at the Haitian Group for the Study of Kaposi’s Sarcoma and Opportunistic Infections (GHESKIO) Center, Port-au-Prince, Haiti. Expedited visits included nurse-led assessments and point-of-service ART dispensing.

Methods

We conducted a retrospective analysis including patients who initiated RP care between June 1, 2014 and September 30, 2015, comparing outcomes of patients with timely visit attendance (never >3 days late) with patients with ≥1 non-timely visit within 6 months prior to RP enrollment. We calculated retention in care and adherence at 12 months, and assessed predictors of both outcomes.

Results

Of the 2361 patients who initiated RP care during the study period, 1429 (61%) had timely visit attendance and 932 (39%) had ≥1 non-timely visit prior to RP enrollment. Among RP-enrolled patients, 94% were retained at 12 months and 75% had ≥90% adherence, with higher proportions in those with timely pre-RP visits (95% vs. 92%; 87% vs. 55%). In multivariable analysis, pre-RP visit timeliness was associated with both retention (adjusted OR [aOR]: 1.67;95% confidence interval [CI]: 1.08-2.59) and adherence (aOR: 4.53; 95% CI: 3.58-5.72).

Conclusion

Rapid Pathway care was associated with high levels of retention and adherence for clinically stable patients. Timeliness of pre-RP visits was predictive of outcomes after RP initiation.

INTRODUCTION

The Joint United Nations Program on HIV/AIDS (UNAIDS) has set global targets of diagnosing 90% of people living with HIV, initiating antiretroviral therapy (ART) for 90% of those found to be infected, and achieving viral suppression in 90% of those on ART by the year 2020 to curb the AIDS epidemic.¹ In addition, in 2015 the World Health Organization (WHO) updated their guidelines to recommend ART for all persons living with HIV, based on evidence that earlier treatment improves outcomes and decreases transmission.² As countries work to achieve these goals, the number of HIV-infected patients on ART will substantially increase, particularly those with early clinical disease. However, funding is already under threat, meaning that few if any additional resources will be available to care for larger numbers of patients.

Furthermore, retention and adherence rates were already sub-optimal throughout the continuum of HIV care, even prior to this influx of new patients.^3–9 Over one-third of those who test newly positive for HIV in resource-poor settings are lost to follow-up prior to ART initiation.^6,7 Retention after ART initiation is also low, with about 80% of patients in care at 12 months, and adherence is sub-optimal even among patients remaining in care.^5,9–12 Predictors of attrition include male sex, younger age, and low socio-economic status.^5,13 Obstacles to retention and/or adherence include transportation costs and distance to clinic, competing demands for time, long clinic waiting times and inconvenient hours, poor trust in services, food insecurity, forgetfulness, and stigma.^9,14–16

New models of care with greater efficiencies in service delivery are needed. The WHO, the US President’s Emergency Plan for AIDS Relief (PEPFAR), and the Global Fund to Fight AIDS, Tuberculosis and Malaria (Global Fund) are all recommending a differentiated care framework.^2,17–19 Differentiated care is adapted to the needs of the patients, and simplifies services for those with fewer clinical needs through strategies such as fewer clinic visits, task-shifting from physicians to other types of providers, multi-month prescriptions, out-of-clinic drug refills, community adherence groups, and adherence clubs.^{2,17,18,20–23} Treatment outcomes with these less intensive models of care have been outstanding.^19,24–38 We report on the rate and predictors of retention in care and adherence with a facility-based model of expedited care for clinically stable patients in Haiti.

METHODS

Study Setting and Participants

This study was conducted at the Haitian Group for the Study of Kaposi’s Sarcoma and Opportunistic Infections (GHESKIO) Center in downtown Port-au-Prince, Haiti. Haiti is the poorest country in the Western Hemisphere, and is ranked 163 out of 188 on the 2016 Human Development Index.³⁹ Haiti has an adult HIV prevalence of 1.7%.⁴⁰ GHESKIO is a Haitian non-governmental organization, and the largest provider of HIV and tuberculosis (TB) care in the Caribbean. GHESKIO diagnoses approximately 3000 HIV-infected patients per year and treats up to 700 adult patients per day for HIV and/or TB. Most patients who seek care at GHESKIO live on less than $1 US per day, and all services are provided free of charge.

Prior to 2014, standard care for adult patients on ART at GHESKIO included monthly physician visits, with assessment of symptoms, clinical stability, and medication adherence, laboratory testing as needed, and pharmacy visits for dispensing of ART. Due to the high volume of patients and limited staff, patients spent several hours in clinic at each visit. In 2014, GHESKIO implemented a new model of Rapid Pathway (RP) care that included nurse-led, expedited services for clinically stable, adult patients on ART. Eligibility criteria for RP included HIV-infected patients who: (1) were ≥18 years of age, (2) had no evidence of active WHO Stage 3 or 4 disease, and (3) had been on ART for at least 6 months prior to RP enrollment.

“Rapid Pathway” Model of Care

The RP model of care differed from standard non-RP care which consisted of monthly physician visits and a single month’s medication supply. RP care included a first enrollment visit, where potentially eligible RP patients met with a physician to confirm RP eligibility. Follow-up nurse visits were then scheduled every two months for all patients in RP care, with a two-month supply of medications.

To remain in RP care, patients needed to be timely for visits (no more than three days late) and remain free of symptoms. On a weekly basis, a community health worker (CHW) reviewed an electronically generated list of all RP patients with scheduled appointments during that week. The CHW narrowed the list to include only RP patients who were eligible for RP care that week based on previous timeliness of visits and CD4 count evolution. Patients who were late for their previous visit (i.e., came to clinic more than 3 days after scheduled appointment) and/or who had decreasing CD4 counts since their last laboratory assessment were removed from the list and referred for standard care during their upcoming visit. A physician then reviewed the list generated by the CHW, and evaluated the clinical and immunologic status of each patient to confirm RP visit eligibility.

Patients who were eligible for RP care were phoned by a CHW one day in advance of each appointment to remind them of their appointment and to query them about symptoms. Patients were confirmed for RP care if they were successfully contacted and reported no symptoms. Additionally, if a patient gave prior notice to the CHW that they needed to re-schedule their RP visit, the CHW verified the patient had enough medication to last until the rescheduled date, and made the proper arrangements to receive them. Before the end of the clinic day, the CHW finalized the next day’s RP visit list and the RP nurse packaged patients’ medications for the subsequent visit date.

If the patient attended their RP visit on the scheduled date, or within the subsequent three days, they received RP care. Upon arrival, they were checked in to clinic by a CHW, and had vital signs checked by a nurse assistant. They were then referred to the RP nurse who conducted a brief visit with a symptom review, dispensed ART and other standard medications at point-of-service, and documented the prescription in the pharmacy electronic medical record. First-line ART included a single tablet formulation of efavirenz, tenofovir disoproxil fumarate and lamivudine, and second-line ART included a protease inhibitor with two nucleoside reverse transcriptase inhibitors. Other medications included isoniazid, pyridoxine, co-trimoxazole and iron sulfate.

Patients who were more than three days late for a visit or reported symptoms to the social worker or nurse were referred for physician evaluation. If patients had been timely for visits but reported poor adherence to the RP nurse, they were also referred to the physician. Patients who were late or symptomatic re-qualified for RP care at the next visit, as long as they were on time (within three days of scheduled visit date) and symptom-free. If a patient developed any active WHO Stage 3 or 4 condition, they received standard (non-RP) care until completion of treatment. Patients who were referred for physician evaluation or back to standard non-RP care were escorted to the appropriate clinic by a CHW.

Statistical Analysis

Demographic, clinical, and laboratory data, including age, education, income, marital status, residence zone, CD4 counts, ART initiation date, RP visit dates, and ART prescription dates were de-identified and exported from the electronic medical record into an Excel spreadsheet and analyzed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA.). Data analysis was limited to adult (≥18 years) ART patients enrolled in RP care from June 1, 2014 to September 30, 2015.

We evaluated retention and adherence at 12-months after the date of RP enrollment. Retention was defined as at least one clinic visit from nine to 15 months after the date of the first RP visit. Twelve-month adherence was measured using medication possession ratio, which is measured by adding up the number of ART pills dispensed during the 12-month period, and dividing this by the number of pills that should have been dispensed with perfect adherence. Baseline characteristics were summarized using simple frequencies and proportions and medians with interquartile ranges (IQRs). We conducted multivariable logistic regression to determine predictors or retention in care and adherence including all covariates listed in Table 1.

Table 1.

Baseline Characteristics of Patients in Rapid Pathway Care

	≥1 Late Visit Prior to RP (n=932)*	Timely Visits Prior to RP (n=1429)*	Total (n=2361)	p-value
Female gender – no. (%)	529 (57)	783 (55)	1312 (56)	0.347
Age – median (IQR)	44 (37, 51)	47 (40, 53)	46 (39, 52)	<0.001
Education primary school or less – no. (%)	454 (49)	662 (46)	1116 (47)	0.256
≤$US 1 per day	752 (81)	1170 (82)	1922 (81)	0.468
Marital status – (%)
Single	199 (21)	257 (18)	456 (19)	0.043
Currently married or cohabitating	499 (54)	786 (55)	1285 (54)	0.486
Formerly married	228 (25)	378 (27)	606 (26)	0.280
Residence zone – no. (%)
Lives in PAP residence zone that includes a slum**	531 (57)	807 (56)	1338 (57)	0.810
Lives in PAP residence zone that does not include a slum	291 (31)	466 (32)	757 (32)	0.480
Lives outside of PAP metropolitan area	87 (9)	132 (9)	219 (9)	0.936
Time on ART (years) – median (IQR)	4.6 (2.0, 7.1)	4.9 (2.6, 7.6)	4.8 (2.3, 7.5)	<0.001
CD4 count at Rapid Pathway enrollment	482 (328, 660)	508 (351, 699)	497 (344, 684)	0.007

^*Definition of late visit is >3 days after scheduled date during the six months prior to Rapid Pathway enrollment; p-values compare the late and timely patients.

^**PAP = Port-au-Prince

A time and motion study was conducted using Samsung Galaxy Nexus 4G Android mobile phones equipped with a “time and motion” application for a two-week period (September 2016). The CHWs and nurses used the Android devices, which were equipped with a customized time and motion application, during clinic hours to measure the total amount of time patients spent in clinic once enrolled in RP care. Waiting time was measured at check-in, vital sign testing, and dispensing of ART at point-of-service by the RP nurse. At check-in, the CHW created a QR bar code which represented the patient’s GHESKIO ID and scanned the patient into the application. At each subsequent station, the staff scanned the patient’s QR code to measure waiting time and stop time collection after the patient had completed the station-specific activities. At the end of each RP visit, patients were scanned out to document the total time spent in clinic. Time and motion data were stored on the devices offline and then at the end of each workday, uploaded to the GHESKIO server and exported into Excel files for data analysis.

RESULTS

From June 1, 2014 to September 30, 2015, a total of 2361 patients were determined by a physician to be eligible for RP care, and had at least one RP visit during the study period. Of these, 1429 (61%) were never more than three days late for a scheduled visit in the six months prior to RP enrollment, and 932 (39%) were at least three days late for at least one visit during that period. Table 1 describes the baseline characteristics of the study population, stratified by timely visit attendance prior to RP enrollment. Patients who were timely for all visits in the six months prior to RP enrollment were more likely to be older (median age 47 vs. 44 years; p-value <0.001), with longer time on ART (median of 4.9 vs. 4.6 years; p<0.001), and higher CD4 counts at RP enrollment (median of 508 vs. 482; p=0.007); they were less likely to have single marital status (18% vs. 21%; p=0.043).

A total of 2214 (94%) of the 2361 patients were retained at 12 months after RP enrollment. Among the 1429 patients who had been timely for all visits for the six months prior to RP enrollment, 1359 (95%) were retained in care at 12 months; among the 932 patients who had been late at least once during this period, 855 (92%) were retained in care. In multivariable analysis, predictors of 12-month retention in care included timeliness for visits for the six months prior to RP enrollment (aOR: 1.67; 95% CI: 1.08 – 2.59) time on ART (aOR 1.16 per year on ART; 95% CI: 1.07 – 1.26) and CD4 count at RP enrollment (aOR 1.08 per increment of 50 CD4 cells; 95% CI: 1.03 – 1.14) (Table 2).

Table 2.

Predictors for Twelve-Month Retention in Care

Variable	Reference Group	Unadjusted OR (95% CI)	p-value	Adjusted OR (95% CI)	p-value
Timely for visits prior to Rapid Pathway	Late for ≥1 visit	1.61 (1.13, 2.31)	0.009	1.67 (1.08, 2.59)	0.021
Female gender	Male gender	1.27 (0.89, 1.81)	0.189	1.07 (0.68, 1.69)	0.771
Age (per decade)	---	1.08 (0.91,1.29)	0.360	1.00 (0.81, 1.24)	0.968
At least some secondary school	No school or primary only	1.20 (0.84, 1.72)	0.311	1.45 (0.91, 2.29)	0.115
Income >$US 1 per day	≤$US 1 per day	0.64 (0.42, 0.97)	0.037	0.78 (0.45, 1.34)	0.366
Married or cohabiting	Single	1.12 (0.78, 1.59)	0.549	1.22 (0.78, 1.89)	0.384
Lives in PAP residence zone that does not include a slum	Lives in PAP residence zone that includes a slum*	1.02 (0.70, 1.49)	0.916	1.31 (0.80, 2.16)	0.285
Lives outside of PAP*		0.84 (0.47, 1.49)	0.552	0.75 (0.38, 1.48)	0.405
Time on ART (per year)	---	1.20 (1.13, 1.28)	<0.001	1.16 (1.07, 1.26)	<0.001
CD4 count at RP enrollment (per increment of 50 CD4 cells)	---	1.10 (1.05, 1.15)	<0.001	1.08 (1.03, 1.14)	0.002

^*PAP = Port-au-Prince

Among the 2214 patients retained in care, adherence was ≥90% in 1655 (75%), 80 to 89% in 275 (12%), and 60 to 79% in 200 (9%), as measured by the medication possession ratio (Figure 2A). Among the 1359 patients retained in care who had been timely for all visits for the six months prior to RP enrollment, 1184 (87%) had ≥90% adherence; among the 855 patients who were late for at least one visit during this period, 471 (55%) had ≥90% adherence (Figures 2B and 2C). Predictors of adherence included timeliness for visits in the six months prior to RP enrollment (aOR 4.61; 95% CI: 3.65-5.82), age (aOR 1.14 per decade; 95% CI: 1.01-1.28), education with at least some secondary school (aOR 1.30, 95% CI: 1.02-1.66), time on ART (aOR 1.07 per year on ART; 95% CI: 1.03-1.11), and CD4 count at RP enrollment (aOR 1.05 per increment of 50 CD4 cells; 95% CI: 1.02-1.08).

Figure 2.

A. Twelve-Month Adherence for Total Cohort

B. Twelve-Month Adherence for Cohort with Timely Visits Prior to Rapid Pathway Initiation

C. Twelve-Month Adherence for Cohort with Late Visits Prior to Rapid Pathway Initiation

Time in clinic was measured for RP patients for two weeks during the study period. From September 5 to 16, 2016, 395 patients had RP visits. Patients spent a median time of 4 minutes (IQR: 1, 7) at vital sign testing and 9 minutes (IQR: 4, 12) at RP nurse visit, including dispensing of medications. The median duration of a RP visit, from check-in to departure from clinic, was 31 minutes; median waiting time was 18 min (IQR: 10, 42).

DISCUSSION

The results of this study demonstrate high retention rates with fast-track follow-up care in Haiti. These results are important as countries seek new models of care to increase the efficiency of ART services. In addition to making care logistically easier for patients, expedited follow-up strategies are likely to save human resources. We also found that timeliness of visits prior to Rapid Pathway enrollment was predictive of retention in care and adherence with RP care. Among the nearly two-thirds of patients with timely visits prior to RP enrollment, 95% were retained, and 87% had adherence levels of at least 90%. However, among those with at least one late visit prior to RP enrollment, 92% were retained, and 55% had adherence of at least 90%.

Our findings add to the evidence base in support of the high quality of care provided by nurses, and of the potential efficacy of expedited facility-based care for clinically stable patients.^{27–29,31,32,35,38} In Uganda and Kenya, the Sustainable East Africa Research on Community Health (SEARCH) test-and-treat trial found high retention with a streamlined model of care that included quarterly follow-up visits for stable patients, with reduced waiting time.³³ In Myanmar, after the implementation of a differentiated care model, patients who had been on ART for at least 12 months with good clinical and immunologic response had high retention with 6-month nurse visits and 3-month medication pick-ups.³⁷ In Malawi, retention was found to be higher among clinically stable patients receiving 6-monthly nurse visits and 3-monthly pharmacy-only ART refill visits, compared to standard care.²⁵ Community-based models of differentiated care have also been found to be very successful. ART adherence clubs (managed by a health care worker), community adherence groups, and community drug distribution points have all been implemented with high rates of adherence and viral suppression.^19,24,25,36

Rapid Pathway visits were very short in duration, generally taking about 30 minutes, including dispensing of ART. This is much shorter than standard visits at GHESKIO and other HIV clinics, which generally take several hours. The short duration of these visits was possible due to four interventions. First, patients were called one day in advance of their visit, and queried about symptoms. Those with symptoms were referred to a physician and did not receive RP care. Second, ART and prophylactic medications were pre-packaged in advance; this was feasible because patients generally took similar medications. Third, the visits were quick, with a vital sign check and then a rapid assessment of symptoms and adherence by a nurse, and point-of-service dispensing of ART. Point-of-servicing dispensing of ART contributed to the greatest decrease in clinic time as patients did not have to travel to a different location for medication pick up nor wait in other pharmacy lines to pick up medications. Fourth, the RP clinic was adequately staffed, so waiting times were very short. We believe that the short visit provided incentives for clinic attendance and contributed to the high rate of retention in care.

For this study, RP visits were scheduled at two-month intervals. GHESKIO is in the process of extending visits to four-month and six-month intervals, with monthly phone calls. Based on the results of this study, GHESKIO has also added a requirement for timely pre-RP visits, in order for patients to qualify for RP care. In addition, HIV-1 RNA testing has recently become routinely available in Haiti, so an undetectable viral load is now required for RP eligibility. We will evaluate the impact of the decrease in visit frequency, and added eligibility requirements for RP care, on retention, adherence, and clinic visit duration in further studies. Currently, 4075 patients are receiving Rapid Pathway care at GHESKIO. Due to the success of this program, the Centers for Disease Control and Prevention and the Ministry of Health have requested the scale-up of RP care in Haiti nationwide.

The strongest predictor of retention and adherence with Rapid Pathway care was the timeliness of visits (never >3 days late) in the six months prior to RP enrollment. This suggests that timeliness of visits may be an important requirement for eligibility for fast-track care, in addition to time on ART, stable clinical status, and viral suppression (where viral load is widely available). We also found that a longer duration of time on ART and higher CD4 count at RP enrollment were predictive of retention and adherence with RP cares. Using these criteria for predicting good adherence on RP care is very useful and can predict those less likely to be adherent and requiring more support. Though male gender has been associated with worse HIV treatment outcomes, gender was not associated with either retention or adherence with RP care.^41–43

This study was conducted in a large urban clinic, which may limit the generalizability of our findings. In addition, we note that patients who were timely prior to RP enrollment were older, with longer time on ART, and higher CD4 counts, so they may have been more likely to conform to timely clinic visits after initiating RP care.

In conclusion, we found that Rapid Pathway care can be provided with very short visits, medical care provided entirely by nurses, and point-of-service dispensing of ART. Timeliness of pre-RP visits was predictive of retention and adherence after RP initiation. The cost of RP care using CHW and nurses is likely to be much less expensive than the standard of care, which includes physician visits.

Figure 1.

Rapid Pathway Model of Care

Table 3.

Predictors of Adherence of at Least Ninety Percent^*

Variable	Reference Group	Unadjusted OR (95% CI)	p-value	Adjusted OR (95% CI)	p-value
Timely for visits prior to Rapid Pathway	Late for ≥1 visit	5.01 (4.11, 6.10)	<0.001	4.61 (3.65, 5.82)	<0.001
Female gender	Male gender	0.90 (0.75, 1.08)	0.254	0.95 (0.75, 1.22)	0.706
Age (per decade)	---	1.23 (1.13, 1.35)	<0.001	1.14 (1.01, 1.28)	0.031
At least some secondary school	No school or primary only	1.19 (0.99, 1.43)	0.061	1.30 (1.02, 1.66)	0.033
Income >$US 1 per day	≤$US 1 per day	0.98 (0.77, 1.25)	0.886	1.14 (0.83, 1.57)	0.423
Married or cohabiting	Single	1.03 (0.86, 1.24)	0.761	0.97 (0.77, 1.23)	0.802
Lives in PAP residence zone that does not include a slum**	Lives in PAP residence zone that includes a slum**	1.05 (0.86, 1.27)	0.660	1.00 (0.77, 1.28)	0.972
Lives outside of PAP**		0.78 (0.57, 1.06)	0.110	0.72 (0.48, 1.07)	0.101
Time on ART (per year)	---	1.09 (1.06, 1.13)	<0.001	1.07 (1.03, 1.11)	0.001
CD4 count at RP enrollment (per increment of 50 CD4 cells)	---	1.06 (1.04, 1.08)	<0.001	1.05 (1.02, 1.08)	<0.001

^*Denominator includes patients who were retained in care;

^**PAP = Port-au-Prince