Implementing the package of CDC and WHO recommended linkage services: Methods, outcomes, and costs of the Bukoba Tanzania Combination Prevention Evaluation peer-delivered, linkage case management program, 2014-2017

Although several studies have evaluated one or more linkage services to improve early enrollment in HIV care in Tanzania, none have evaluated the package of linkage services recommended by the Centers for Disease Control and Prevention (CDC) and the World Health Organization (WHO). We describe the uptake of each component of the CDC/WHO recommended package of linkage services, and early enrollment in HIV care and antiretroviral therapy (ART) initiation among persons with HIV who participated in a peer-delivered, linkage case management (LCM) program implemented in Bukoba, Tanzania, October 2014 –May 2017. Of 4206 participants (88% newly HIV diagnosed), most received recommended services including counseling on the importance of early enrollment in care and ART (100%); escort by foot or car to an HIV care and treatment clinic (CTC) (83%); treatment navigation at a CTC (94%); telephone support and appointment reminders (77% among clients with cellphones); and counseling on HIV-status disclosure and partner/family testing (77%), and on barriers to care (69%). During three periods with different ART-eligibility thresholds [CD4<350 (Oct 2014 –Dec 2015, n = 2233), CD4≤500 (Jan 2016 –Sept 2016, n = 1221), and Test & Start (Oct 2016 –May 2017, n = 752)], 90%, 96%, and 97% of clients enrolled in HIV care, and 47%, 67%, and 86% of clients initiated ART, respectively, within three months of diagnosis. Of 463 LCM clients who participated in the last three months of the rollout of Test & Start, 91% initiated ART. Estimated per-client cost was $44 United States dollars (USD) for delivering LCM services in communities and facilities overall, and $18 USD for a facility-only model with task shifting. Well accepted by persons with HIV, peer-delivered LCM services recommended by CDC and WHO can achieve near universal early ART initiation in the Test & Start era at modest cost and should be considered for implementation in facilities and communities experiencing <90% early enrollment in ART care.


Introduction
The considerable benefits of ART to reduce HIV-related morbidity and mortality, and HIV transmission risk to uninfected sexual partners and offspring are now well established [1][2][3]. To achieve HIV epidemic control, the United Nations AIDS Program established goals by the end of 2020 for countries to diagnose �90% of HIV-infected persons, initiate and retain �90% of diagnosed persons on ART, and achieve viral load suppression on �90% of persons on ART (90-90-90) [4]. Achieving optimal impact of the second and third "90" on reducing HIV transmission risk and HIV morbidity and mortality, however, hinges on early ART initiation following diagnosis [1,2]. In sub-Saharan Africa, although HIV testing and care and treatment programs have increased access to ART remarkably, many HIV-infected persons still initiate ART late in the course of their HIV disease [3,5,6]. Delayed enrollment in HIV care following diagnosis is a particularly important factor contributing to late ART initiation, particularly among young adults, men, and persons diagnosed in community settings [1,3,[6][7][8][9].
To improve early enrollment in care and ART initiation following HIV diagnosis, CDC and WHO released recommendations on linkage services in 2014 and 2016, respectively [1,10]. No studies in Tanzania, however, have evaluated whether providing the CDC/WHO package of recommended services can routinely achieve �90% early enrollment in HIV care overall, and for groups that are more likely to delay enrollment in care. Evaluating the effects and costs associated with implementing the package of recommended linkage services is fundamental for informing scale up of these services. This evaluation is particularly important in light of recent WHO recommendations to initiate ART for all persons with HIV within 7 days of diagnosis (rapid ART) [2].
As part of the Bukoba Tanzania Combination Prevention Evaluation (BCPE), we developed a peer-delivered, linkage case management (LCM) program to provide the package of linkage services recommended by CDC and WHO. The goal of LCM was to enroll in HIV care within three months of diagnosis �90% of HIV-infected, out-of-care persons diagnosed in facility and community settings. At the time our intervention was developed in 2014, many reports from sub-Saharan Africa including Tanzania suggested that fewer than 50% of persons enroll early in HIV care following diagnosis, particularly persons diagnosed in community settings [7,8,11,12].
We report on the uptake, per-client program cost, and impact of LCM on early (�90 days) enrollment in care and ART-initiation among persons diagnosed in facility and community settings during the intervention phase of BCPE. LCM services and outcomes are reported by client demographic characteristics and participation during three different ART-eligibility periods.

Study population
BCPE is an evaluation of the population-level impact of a community-wide, integrated scale up of HIV testing and counseling (HTC), LCM, and defaulter-tracing programs in Bukoba Municipal Council, a Lake Victoria community in northwest Tanzania of approximately by the U.S. Centers for Disease Control and Prevention to provide data-management services only. Author Johnita Byrd is employed by ICF International. ICF International provided support in the form of salary for author JB, but did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. The specific role of this author is articulated in the 'author contributions' section.
150,000 persons [13]. The LCM study population included consenting eligible clients who  tested HIV-positive from October 2014 through March 2017 at outpatient departments and  other clinics at 11 health facilities (facility clients), and at homes, venues, or community events in all municipal wards (community clients). Excluding military and police facilities, all municipal government health facilities participated in BCPE including one regional referral hospital, two health centers, and five dispensaries. Three private, ministry-supported health centers also participated in BCPE. Clients who tested HIV-positive in accordance with the national rapidtesting algorithm were routinely screened for eligibility. Clients were eligible for LCM if they were newly or previously diagnosed and had not received HIV care in the past 90 days (HIVpositive, out-of-care). HIV-positive, out-of-care clients were ineligible for LCM if they requested referral to non-participating military or police facilities within and all facilities outside of Bukoba Municipal. national guidelines expanded the ART-initiation threshold twice during BCPE, resulting in the three ART-eligibility periods (CD4 count): October 2014 -December 2015 (<350); January 2016 -September 2016 (�500); and October 2016 -May 2017 (any count, Test & Start). For clients who were ready and did not present contraindications, a 14-day course of antiretroviral drugs was routinely provided by licensed providers at the first CTC visit beginning in December 2015, and at the point of diagnosis in community settings beginning in November 2016.
Retraining and changes in EC compensation and LCM practices. During the CD4�500 period, ECs were retrained on LCM service expectations, and LRCs were retrained to provide closer supervision and to help ECs manage difficult cases (e.g., calling and counseling clients who did not enroll early in HIV care). During CD4�500 and Test & Start periods, ECs were provided a monthly monetary incentive of 10,000 Tanzania Shillings (approximately $5.00 USD) when telephone support and other monthly service indicators were met. The base monthly compensation for part-time ECs, 125,000 Tanzania Shillings (approximately $60 USD), did not change during the intervention. During Test & Start, ART-initiated community clients were routinely transported to the CTC to register for HIV care on the same day (if possible) to provide blood specimens for baseline tests, and transition case-management services from community-to facility-based ECs.

Data collection and management
LRCs recorded delivered linkage services from program forms and clinical outcomes from CTC charts onto linkage registers maintained at each participating facility. BCPE monitoring and evaluation (M&E) staff visited each facility monthly to validate and count (compile) service and outcome indicators of cases that were closed that month. Depending on the service or clinical indicator, data were compiled by sex, age group, and HIV-diagnostic status (new or prior) and setting (facility or community). Senior investigators periodically validated monthly counts against register records, and register records against source forms including medical charts. Compiled data from monthly summary forms were analyzed using SAS 9.4 (SAS Institute Inc., Cary, NC, USA).

Outcome measures, analyses, and costing
Enrollment in care was defined as registering for and receiving HIV medical services at least once at a CTC. LCM service and clinical outcomes were analyzed for facility and community clients separately and combined, by demographic characteristics and ART-eligibility period. Outcomes are reported as percentages for nominal and ordinal variables, and as medians and interquartile ranges (Q1-Q3) for ratio-scaled variables. Statistical tests for group differences were not used because LCM clients represent the population of interest.
Total and per-client LCM program costs were estimated from data collected retrospectively from financial records and staff interviews. Costs were estimated for personnel, travel, commodities, training, and equipment. Vehicle costs were annuitized at 3% a year over an expected useful life of five years. All costs were collected in Tanzanian Shillings (TZS), inflated to 2017 price levels using the annual Tanzania consumer price index (CPI) ratio of 1.05, 1.11 and 1.17 for 2014, 2015 and 2016, respectively, and converted to 2017 U.S. Dollars (USD) using the 2017 average market exchange rate (USD 1 = TZS 2,269.27) [14,15]. We also estimated total and per-client costs for implementing a facility-based model that excluded transportation services and that substituted senior EC supervisors for nurse LRCs. We estimated costs for these hypothetical models because LCM (1) can be implemented in health facilities alone for clients who do not require transportation services, and (2) has been successfully implemented without employing LRC nurse supervisors [16].

Human subjects
Provision of peer-delivered LCM services, and collection of linkage-service and clinical-outcome data were reviewed and approved by institutional review boards at the National Medical Research Institute, Dar es Salaam, Tanzania, and at Columbia University Medical Center, New York City, United States. All adult clients and parents or legal guardians of children age <15 years who participated in LCM provided written informed consent on the content and duration of LCM services, and the collection of contact, service, and clinical data. As part of their training, counselors were educated on the rights of human subjects and routinely informed clients as part of the consent process that they could refuse or discontinue LCM at any time without reducing their ability to receive the same health-care services.

Service uptake
Most clients received recommended services including counseling on the importance of early enrollment in care and ART initiation (100%), CTC escort (83%), treatment navigation (94%), telephone support (77% among clients with cellphones), and counseling on disclosure and partner/family testing (77%) and on real and perceived barriers to HIV care (69%). Receipt of LCM services was similar by client gender (Table 2). Proportionally more facility than community clients were escorted to CTCs on foot (79% vs. 3%), and proportionally fewer were escorted by car (9% vs. 54%). Of clients who received treatment navigation services, 83% were escorted to the CTC on foot or by car, and 11% (8% facility, 25% community) were met at the CTC by appointment. In all ART-eligibility periods, proportionally more facility than community clients received follow-up counseling on disclosure and partner/family testing, and on real and perceived barriers to HIV care (Table 2).
Compared with the CD4<350 period, proportionally more clients in CD4�500 and Test & Start periods received CTC escort, treatment navigation, and telephone-support services ( Table 2). During Test & Start, nearly all facility and community clients received CTC escort (95% and 93%, respectively) and treatment navigation (98% and 96%, respectively) services. Of facility and community clients with cellphones (85% and 80%, respectively), a large majority were contacted by phone at least once in the CD4�500 and Test & Start periods, and the monthly median (IQR) telephone contacts per client increased from 1.3 (0.8-2.0) during the CD4<350 period to 3.7 (2.8-4.5) during Test & Start.
Partner and family member testing and barriers to care. Of 3256 clients who completed the session on disclosure, 1617 (50%) reported having a main partner, and of 229 partners tested for HIV as part of LCM services, 120 (52%) were newly HIV diagnosed. An additional 462 family members of clients also tested for HIV as part of LCM services, of whom 56 (12%) were newly HIV diagnosed. Of 2882 facility and community clients who completed the session on barriers to HIV care, proportionally more community than facility clients (43% vs. 37%) reported at least one barrier to care. Similar proportions of female (39%) and male (36%) clients reported at least one barrier to care.

Enrollment in HIV care
Within 90 days of HIV diagnosis, 93% of all clients, and 95% of 3538 facility and 82% of 668 community clients enrolled in HIV care. From CD4<350 to Test & Start periods, enrollment in care increased from 93% to 98% and from 78% to 96% among facility and community clients, respectively (Table 3). Of 752 clients during Test & Start, 98% and 97% of males (n = 310) and females (n = 442), and 96% and 98% of clients aged 15-24 (n = 150) and 24-49 (n = 535) years, respectively, enrolled in care.  ART eligibility and initiation among clients enrolled in HIV care. Before Test & Start, of 3186 clients who enrolled in HIV care, 3032 (95%) overall, and 95% and 94% of facility and community clients, respectively, had a baseline CD4 test result recorded in the linkage register. During CD4<350 and CD4�500 periods, of clients who enrolled in HIV care on whom CD4 test results were available, 48% and 71%, respectively, were ART eligible by CD4 count alone. During CD4<350, CD4�500, and Test & Start periods, of all 3918 clients who enrolled in care, 52%, 70%, and 89%, respectively, were initiated on ART within 90 days of diagnosis. Proportionally more enrolled facility than community clients were ART eligible by CD4 count and initiated on ART overall and during both CD4<350 and CD4�500 periods (Table 4).

Discussion
Of over four thousand persons HIV diagnosed in facility and community settings in Bukoba Municipality who participated in a peer-delivered, linkage-case-management program, nearly  (Table 3). e Of all LCM clients (Table 3) all received linkage services recommended by CDC and WHO, and within three months of diagnosis, most (93%) enrolled in HIV care and many (60%) were initiated on ART during a period (October 2014 -May 2017) when ART-eligibility thresholds were expanded from CD4<350 to Test & Start. Although proportionally fewer community than facility clients received linkage services and enrolled early in HIV care, delivery and uptake of escort, treatment navigation, and telephone-support services improved over time. During Test & Start, early enrollment in HIV care and ART initiation exceeded 95% and 85%, respectively, for both facility and community clients.
Our finding of near universal early enrollment in care among LCM clients are similar to those reported from a peer-delivered LCM program in Eswatini that also provided the package of CDC/WHO recommended services [16], and stands in contrast to findings from studies conducted in sub-Saharan Africa on persons who are provided a referral as the only linkage service following HIV diagnosis. In fifteen studies conducted before Test & Start, for example, only 18-51% and 39-71% of persons diagnosed in community and facility settings, respectively, enrolled early in HIV care following referral [7]. Even with expansion of ART eligibility during Test & Start, when referral to care is the only linkage service, reported early enrollment often falls well below 90%, particularly among community clients [17][18][19][20]. Notably, in two large combination prevention trials in South Africa and Zambia, of the nearly 7,000 intervention-arm participants who were informed that they would be initiated on ART, only 37% and 41%, respectively, enrolled in care within three months of diagnosis following referral alone, and <50% in both studies had initiated ART by six months of diagnosis [19,20].
Consistent with a systematic review and more recent studies in Tanzania, proportionally fewer community than facility clients before Test & Start were ART-eligible by CD4 count, and proportionally fewer were initiated on ART [7,17,18]. Because community clients at the Peer-delivered linkage case management, Bukoba Tanzania, 2014-2017 time of participation were not seeking health care and were less likely to be ART eligible, they may have been less likely than facility clients to perceive the need to enroll early in HIV care [17,18,[21][22][23]. Including point-of-care CD4 testing as part of linkage services might have helped some community clients overcome this important barrier to care before Test & Start [24][25].
Not surprisingly, when point-of-diagnosis ART-initiation combined with same-day transport and treatment navigation services were routinely provided during Test & Start, similarly high proportions of community and facility clients initiated ART and enrolled in HIV care. Encouragingly, in the last three months of the rollout of Test & Start, >90% of over 450 facility and community clients were initiated on ART. Our findings suggest that escort, treatment navigation, and point-of-diagnosis ART initiation can achieve near universal rapid ART initiation and enrollment in facility-based care among persons diagnosed in community settings. Rapid ART initiation within seven days diagnosis is now recommended for all persons without contraindications who are ready to begin ART [2,26].
As recommended by CDC and WHO, peer-delivered services were initiated for all consenting clients at the point of diagnosis [1,10]. Linkage programs that require either referral forms or documented missed appointments to initiate follow-up may lose important opportunities to provide timely and effective linkage services [8,20]. In KwaZulu-Natal, for example, of over 1,000 persons diagnosed at home who were supposed to receive linkage services after failing to present for care within three months, enrollment in care 3 and 6 months after diagnosis increased from only 37% to 47% [20]. As a proactive program, LCM expert client counselors initiate services at diagnosis to build rapport, assess and understand individual circumstances, and use their personal experiences in living with HIV to help clients cope with their disease and overcome barriers to HIV care [27][28][29]. These proactive, peer-delivered services may be particularly helpful to younger persons and men who are more likely to delay their enrollment in care because of real or perceived barriers to HIV care [1][2][3][7][8][9].
The importance of routinely assessing and mitigating barriers to care is underscored by the considerable magnitude and multidimensionality of reported barriers, and in a recent study, that persons who report one or more barriers have significantly increased mortality one year after diagnosis [1,10,[21][22][23][27][28][29][30][31][32][33]. During Test & Start, nearly a third of enrolled clients who completed the third counseling session reported at least one barrier to care.
ECs were expected to identify and help resolve barriers to care during all counseling sessions. However, because we routinely assessed barriers in the third session, usually conducted several weeks after enrollment in care, many clients did not receive the assessment and our findings on the proportion of clients with at least one barrier is likely an underestimate. Reasons for the decrease in the delivery or uptake of the barriers counseling session are unclear, but might be attributed, in part, to clients who did not keep their CTC appointment. To identify and mitigate barriers throughout the entire LCM period, barriers should be routinely identified and addressed during all linkage sessions as part of an individual-level M&E system. While our standard form assessed 12 specific barriers to help target counseling, we could not monitor and evaluate the frequency and distribution of specific barriers with our registerbased system.
Non-disclosure of HIV status is a particularly important barrier to both enrollment and retention in HIV care [17,18,21,23,[27][28][29][30][31][32]. Integrated as part of LCM, counseling on the importance of HIV-status disclosure and partner and family-member testing was provided to most clients. Although our evaluation was unable to assess the efficacy of counseling on disclosure practices, ECs reported helping many clients disclose to and test partners and family members. Consistent with many studies, of the nearly 700 partners and family members who tested through LCM services, many were newly HIV diagnosed, including approximately half of over 200 partners tested [34-37]. Our finding on high yield of new diagnoses but relatively low uptake by partners underscores the importance of fully implementing LCM partner-notification services. Greater uptake of these services might be achieved by adopting best practices from other studies and ensuring that all recommended partner elicitation and notification strategies are routinely employed [34][35][36][37][38].
Studies conducted in sub-Saharan Africa that have reported costs for linkage services including point of diagnosis CD4 testing, transportation assistance, and follow-up telephone or face-to-face counseling also included HTC-related costs for identifying HIV-positive clients (per-client costs in these studies ranged from $111 to $2102 USD) [39][40][41][42]. Because HTCrelated costs were included in these studies, comparative linkage-program costs are not available. Based on our experiences in Bukoba and elsewhere, reducing LCM service costs while maintaining >90% early enrollment in care is possible. A similar LCM program in Eswatini did not employ nurses to supervise ECs and enrolled in HIV care 98% of 651 clients [16]. In health facilities, implementing an EC-managed model without travel and transportation-service costs would enable routine delivery of peer-delivered, LCM services at relatively low cost (~$18 USD per client with similar personnel-compensation, training, commodity, and equipment costs). Not providing transportation services for facility clients should have minimal impact on enrollment in care as few clients requested transportation services to another ART clinic.
Although a facility-based, escort-only model might achieve similarly high enrollment in care rates at even lower per-client costs, peer-delivered LCM might have additional durable effects such as reducing barriers associated with defaulting from ART care [27][28][29]. Given CDC and WHO recommendations to routinely address barriers to care, particularly non-disclosure, further research is needed to identify optimal methods for integrating barrier assessment and mitigation as part of LCM services, and evaluating corresponding costs and effects on both enrollment and retention in ART care [1,10].
Findings in this report are subject to several additional limitations not previously mentioned. First, some linkage-service and clinical data may not have been recorded completely or accurately onto linkage registers. Senior-investigator audits throughout the intervention period, however, found that register data consistently matched source forms, including patient medical records. Second, although manual compilation of outcomes is bound to include some errors, quality-assurance procedures routinely prompted re-compilations that identified few and relatively minor discrepancies. Nonetheless, outcomes of approximately 2% of consented clients either were not compiled or compilation of consented records were in error. Third, although consent rates were high, not all eligible persons chose to participate in LCM, particularly clients who were younger and diagnosed in the community. Enrollment in HIV care among LCM non-participants is unknown. Fourth, ART initiation was measured during the 90-day LCM period only. After Test & Start was approved, however, a comprehensive callback and defaulter-tracing program was conducted as a good-faith effort to inform and initiate on ART all clients who were not initiated on ART in the 90-day LCM period. Finally, because our community-wide program evaluation did not include a control group, we could not estimate LCM intervention effects and improvements in services on enrollment in care overall and by time period. Although our enrollment in care findings far exceed most historical cohorts in Tanzania and elsewhere in sub-Saharan Africa, observed differences may also be attributed to increased awareness of the benefits and availability of ART and improvement in HIV-related societal norms. LCM enrollment-in-care outcomes, however, were substantially higher than historical cohorts in the CD4<350 era in which most cohorts were evaluated, and remain higher compared with several cohorts in the Test & Start era [7,8,11,12,[17][18][19][20].
We found that implementing the package of CDC and WHO recommended linkage services was feasible and well accepted, and during Test & Start, achieved near universal early enrollment in HIV care and ART initiation at modest cost. In 2017, peer-delivered LCM services was approved by the Tanzania Ministry of Health, Community Development, Gender, Elderly, and Children as a new service delivery model, and in 2018, was implemented as facility-and community-based linkage interventions in ten regions in Tanzania by four HIV prevention organizations [43]. As a recommended solution by the United States President's Emergency Plan for AIDS Relief, peer-delivered LCM services should be considered for implementation in other countries and programs with <90% early enrollment into ART care among persons diagnosed in facility and community settings [44].