Secondary Logo

Journal Logo

Retention Among ART Patients in the Highlands of Papua New Guinea: Evaluating the PAPUA Model

Das, Sarthak DrPH*; Carmone, Andy CNM, MPH*; Franke, Molly F. ScD; Frank, Dale MD, MPH*; Kiromat, Hannelly BSc*; Kaima, Petronia MD; Kiromat, Mobumo MD*

JAIDS Journal of Acquired Immune Deficiency Syndromes: February 1st, 2014 - Volume 65 - Issue 2 - p e67–e73
doi: 10.1097/QAI.0b013e3182a14f7e
Implementation and Operational Research: Epidemiology and Prevention
IOR articles

Background: Despite more than 10,000 patients on antiretroviral therapy (ART), there remains a dearth of operational research in Papua New Guinea related to HIV service delivery. This study examined the effectiveness of a locally developed model of HIV service delivery called PAPUA (Patient and Provider Unified Approach) in the Highlands of Papua New Guinea. The model emphasizes coordinated patient and provider support along with decentralized services to rural districts in the Highlands.

Methods: We conducted a chart review among HIV-infected adults on ART at clinics in Eastern Highlands Province, where the PAPUA model was implemented in addition to the standard of care, and in Western Highlands Province, where the standard of care was implemented. We calculated yearly retention rates and used multivariable Cox proportional hazards regression analyses to compare retention rates across the provinces.

Results: Data for 2457 patients from the 2 provinces were analyzed. Among patients receiving ART under the PAPUA model in Eastern Highlands, the 12-, 24-, 36-, and 48-month retention proportions were 0.79, 0.73, 0.68, and 0.63, respectively. When we compared retention probabilities across the 2 provinces, patients receiving care under the PAPUA model had a 15% lower rate of attrition from care during the first 4 years of ART (hazard ratio, 0.85; 95% confidence interval: 0.74 to 0.99; P = 0.03), after adjusting for age, gender, and year of enrollment.

Conclusions: The PAPUA model seems to be a promising intervention although it is inextricably linked to the limitations posed by a resource-constrained health system.

Supplemental Digital Content is Available in the Text.

*Clinton Health Access Initiative, Goroka, Papua New Guinea;

Department of Global Health and Social Medicine, Harvard Medical School, Boston, MA; and

National Department of Health, Highlands Region, Papua New Guinea.

Correspondence to: Sarthak Das, DrPH, Clinton Health Access Initiative, 599 Goroka, Eastern Highland Province, Papua New Guinea 441 (e-mail:

Funding for CHAI PNG program is through the Australian Agency for International Development.

The authors have no conflicts of interest to disclose.

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Web site (

Received March 05, 2013

Accepted June 12, 2013

Back to Top | Article Outline


Although the Highlanders of Papua New Guinea (PNG) lived unknown to the rest of the world until the early 20th century, HIV has reached some of the most remote communities on earth. From 1998 to 2007 PNG had the fastest growing epidemic in the Asia-Pacific region; 19 of every 20 HIV cases reported in the Pacific between 1984 and 2007 were Papua New Guineans.1 Although the epidemic seems to have slowed with a nationwide reported prevalence of 0.8% in 2009,2 a continued strong national response to the epidemic remains an urgent need, particularly among rural communities in the Highlands, which account for 40% of PNG's population of 7 million and which reported more than 60% of new HIV cases in 2009.1

In 2006, the National Department of Health (NDOH) requested the Clinton Health Access Initiative to undertake a pilot program to develop a rural model for HIV service delivery, which could be replicated in other high-prevalence provinces. Called PAPUA (Patient And Provider Unified Approach), the model seeks to decentralize HIV services in high-prevalence rural districts of the Highlands of PNG while emphasizing patient case management, clinical mentorship, quality assurance, and laboratory support. There has been minimal documentation of antiretroviral therapy (ART) patient outcomes in PNG despite more than 10,000 people being currently on ART.3

This article reports on patient retention among HIV-infected persons enrolled for ART under the PAPUA intervention in the Eastern Highlands Province of PNG. We also compared these outcomes to those from the nearby Western Highlands Province, where the standard NDOH model is implemented.

Back to Top | Article Outline


Study Setting

PNG reported its first case of HIV in 1987; since then there has been an average annual increase in reported cases per annum of more than 30%. A high tolerance of informal and formal transactional sex, social practices of polygamy or multiple partners during courtship, and the absence of knowledge of HIV transmission, particularly among rural communities, underpin the spread of HIV in PNG.4,5 The high prevalence of untreated sexually transmitted infections in the general population may also drive the epidemic.6,7

Until late 2006, the primary efforts in the national response have been focused on major urban centers.8,9 Rural public health infrastructure in PNG is severely challenged with the closing of nearly half of government-run primary health facilities.10 Where they remain open, poor road access, human resources, and weak logistics for delivery of diagnostics and drugs challenge delivery of basic health services.11 Tribal conflict and acute issues of interpersonal and gender-based violence, and severe HIV-related stigma further complicate service delivery.12,13 Subsequently, capacity to deliver HIV services, let alone decentralize HIV testing, care and treatment to rural communities, is tenuous.

Back to Top | Article Outline

Patient Data

The patient data was collected from 2 provinces: Eastern Highlands Province (population: 432,972) which covers an area of 11,200 km2 and is divided into 8 districts and Western Highlands Province (population: 440,025) that covers an area of 8500 km2 and is divided into 7 districts. Both provinces have 1 main provincial referral hospital, each with its own ART clinic: in Eastern Highlands, this is Michael Alpers Clinic, Eastern Highlands Provincial Hospital and in Western Highlands this is Tininga Clinic, Mt Hagen Hospital. These facilities are roughly 200 km apart or 4–5 hours apart by road. The PAPUA program was undertaken in Eastern Highlands Province. Maps showing ART sites in the Eastern and Western Highlands Provinces are shown in the Supplemental Digital Content (

Back to Top | Article Outline

PAPUA Model Overview

The goal of the PAPUA pilot in Eastern Highlands was to complement the standard model of care provided by NDOH by decentralizing high-quality HIV testing, care, and treatment to the districts while revitalizing basic primary health services. When the pilot program began in 2006, ART was available only in Eastern Highlands Provincial Hospital. By the end of 2011, services were decentralized to all 8 districts of the province, each of which had at least 1 health center. The PAPUA model was locally designed for the PNG context with 3 primary objectives: (1) to address major barriers to patient retention through case management, (2) to decentralize services to district-level health centers, and (3) to strengthen health system capacity to deliver HIV care through a focus on laboratory support, quality assurance, and clinical supervision. All aspects of PAPUA are delivered through existing government-operated health facilities. Table 1 summarizes the PAPUA model and the additional components designed to complement the existing NDOH model.



With its emphasis on both the coordination of referral pathways and retention of patients through a focus on reducing the material obstacles to care, case management is a critical aspect of strong HIV treatment programs, and studies have shown fewer unmet needs and better uptake of and adherence to medical regimens among HIV-positive clients with case managers.14–16 By maximizing retention, daily adherence to antiretroviral therapy, and viral load suppression, the case management model may also contribute to a reduced risk of HIV transmission to discordant partners.17–20 There has not, however, been sufficient investigation into the efficacy of the case management approach within a developing country context.14

As part of PNG's first effort to decentralize HIV services to rural populations at the district level through government facilities,21 a case management model was proposed and implemented within the PAPUA model that uses additional health workers as part of a case management team that seeks to:

  1. Coordinate a package of patient support services to reduce material obstacles to care and adherence, including transportation and nutritional support.22–24
  2. Augment existing health care providers in the provision of ART and the monitoring of current ART patients.
  3. Support clinic staff in the collection and maintenance of patient data.
  4. Conduct clinical mentoring and supervision for new ART providers both at the main referral hospital and, on a rotational basis, at the district ART clinic level.

At the Provincial Hospital level, case management teams consist of 2 health extension officers (ie, mid-level providers, the equivalent of nurse practitioners in many Western countries), 2 nurses, 2 community health workers, and a driver. District-level facilities receive support from case management teams at the provincial level but do not currently have extra human resources for case management activity. Clinical mentoring, supervisory visits, assistance with ensuring adequate supplies of essential drugs and test kits, and support in the maintenance of patient data are all functions of the case management team.

Back to Top | Article Outline

Data Collection

Study staff retrospectively reviewed charts for the following data: age, gender, history of tuberculosis (TB), CD4 count, ART initiation date, transfer out status, and visit dates. Sociodemographic data and other clinical variables were not consistently available in clinical charts.

Back to Top | Article Outline

Outcome Definitions

The primary outcome of interest was retention in care at 12-, 24-, 36-, and 48-months. Because we included individuals who initiated ART as far back as 2006 and as recently as 2010, potential follow-up time varied from person to person. All individuals included in this report had the potential to be followed for 1 year and therefore contributed to the 12-month retention proportion. For the 24-, 36-, and 48-month retention proportions, we considered only those individuals who initiated ART early enough to be followed for 24, 36, and 48 months, respectively. We defined an individual as retained if she/he had a visit date within 90 days of the chart review and not retained (ie, deceased or lost-to-follow-up) otherwise. Individuals who transferred to an outside clinic did not contribute to the numerator or the denominator of retention proportions. Follow-up time ended 90 days after the last visit date for people who were lost-to-follow-up and on the date of the last visit for individuals who transferred out. Among those who died, follow-up ended on the date of death. When this date was unavailable, as was often the case, follow-up ended on the last visit date.

Back to Top | Article Outline

Data Analyses Methods

We calculated overall retention proportions and corresponding exact confidence intervals among individuals who received the PAPUA model and separately for each district site. Because early attrition (ie, death or loss-to-follow-up) in an ART program may be a result of advanced HIV disease, we conducted secondary retention calculations among those who were retained after the first 90 days of treatment.25,26

To compare retention rates among individuals who received the PAPUA model in Eastern Highlands and those who received the standard NDOH model in the Western Highlands, we conducted Cox proportional hazards regression analyses and plotted Kaplan–Meier curves. In both analyses, we adjusted results for age, gender, and year of enrollment and censored individuals who transferred out and those who had a visit within 90 days of the chart review.

Back to Top | Article Outline

Ethics Approval

Approval was obtained from University of North Carolina Chapel Hill, the Research Advisory Council of the National AIDS Council Secretariat, and the Medical Research Advisory Council of PNG. Research approval was also obtained from Mt Hagen Hospital, and Eastern Highlands Provincial Hospital.

Back to Top | Article Outline


A total of 993 patients initiated ART under the PAPUA model in the Eastern Highlands during the study period, during which time the number of persons initiating ART increased steadily from 59 in 2006 to 284 in 2010. The majority of these patients were women (n = 625; 62.9%) and younger than 40 (n = 811; 81.8%), and 10.4% had a history of TB. Baseline CD4 cell count up to 180 days before or 60 days after ART initiation was available for only 247 (24.7%) individuals, and the median value was 158 cells per microliter (interquartile range, 86–248). Forty-two percent of ART patients were seen at district-level facilities.

Back to Top | Article Outline

Patient Retention

The 12-, 24-, 36-, and 48-month retention proportions for patients receiving ART under the PAPUA model in the Eastern Highlands were 0.79, 0.73, 0.68, and 0.63, respectively (Table 2). The majority of attrition (70.4%) during the first 12 months occurred within 6 months of ART initiation, and 57 (5.7%) individuals did not return to the clinic after their first visit. As expected, when we calculated retention among those patients who were retained after the first 90 days, the proportions retained were notably higher: 0.91 at 12 months, 0.84 at 24 months; 0.78 at 36 months, and 0.75 at 48 months (Table 2). Retention across districts was generally good (Table 3), although some heterogeneity across sites was observed.





Back to Top | Article Outline

Multivariable Analyses and Kaplan–Meier Estimates

In univariable analyses, receiving care in the Eastern Highlands, where the PAPUA model was implemented in addition to the NDOH standard of care, was associated with a 9% lower rate of attrition relative to those receiving the standard NDOH model in Western Highlands, although this was not statistically significant (hazard ratio: 0.91:95% confidence Internal [CI]: 0.78 to 1.05; P = 0.18). After adjusting for age, gender, and year of enrollment, receipt of the PAPUA model was associated with a statistically significant 15% lower rate of attrition from care during the first 4 years of ART (hazard ratio: 0.85; 95% CI: 0.74 to 0.99: P = 0.03; Table 4). Kaplan–Meier retention curves adjusted for age, gender, and year of enrollment are shown in Figure 1.

Table 4

Table 4



Back to Top | Article Outline


In a relatively large sample, we found that the PAPUA model yielded moderate retention levels in a setting where the delivery of care is extremely challenging. Most attrition from care occurred early on in treatment and nearly 6% of patients never returned for a second visit. When we excluded attrition occurring in the first 90 days of ART, retention proportions were high at 0.91 and 0.84 at 12- and 24-months, respectively. Further investigation is needed to confirm whether early attrition is because of death from advanced HIV disease27 or whether other factors such as fear or stigma or domestic violence may contribute to early attrition.28

A comparison of outcomes from the PAPUA model to those from a nearby province in which ART was delivered according to the standard NDOH model found that the addition of the PAPUA model was associated with higher retention rates. Although we cannot discount the possibility that an important unmeasured difference between the 2 groups (ie, disease severity, quality of care, and sociodemographic characteristics) accounted for the observed differences in retention, the 2 provinces are comparable in a number of ways. The main ART centers and provincial hospitals in Mt Hagen and Goroka are government-operated and therefore have similar human resource, infrastructure, and health system constraints. Although there are differences in cultural practices within Eastern and Western Highlands Provinces, they remain more akin to one another than to other non-Highlands–based communities, such as those in coastal areas or on the islands where cultural practices are vastly different. On the other hand, Western Highlands tend to have better health and sociodemographic indicators and better access to national roadways than Eastern Highlands.11 If these factors were related to retention among patients on ART, adjustment for these variables could yield an even stronger association between the PAPUA model and retention. Although a more rigorous comparison that includes adjustment for additional potential confounders is warranted, our findings provide preliminary evidence that the PAPUA model may lead to improved retention on ART.

Decentralization of ART to district-level services is a key element to the PAPUA model; however, district sites deliver HIV care and treatment services in a setting that is often more resource constrained than the provincial referral hospital, and therefore, providing care at these sites introduced additional obstacles. For example, the transport of samples to the provincial laboratories from district sites can be difficult given the poor road conditions; this also affects the flow of essential drugs, diagnostics, test kits, and results to health centers from the provincial center. HIV testing, care, and treatment with its need for consistent supplies and referral pathways to manage complications and/or coinfections (eg, TB) presents an additional layer of complexity for the health system to manage.

Although it may be true that bringing services closer to where patients live may improve retention, the fact that these HIV care and treatment services are part of a severely resource constrained health infrastructure should not be underestimated. An additional challenge of district-level service provision relates to training of health care providers. Neither Integrated Management of Adult Illness nor Provider Initiated Counseling and Testing are included in preservice medical training in PNG, leading to a health workforce with constrained capacity for HIV service delivery, particularly at the district level, where more severe clinical presentations are referred to tertiary care centers for treatment.

Although district sites tended to perform well, we did observe some differences across sites. Variations may be at least partially explained by community-specific characteristics, such as the level of tribal conflict and community-based stigma, which affect both patients and providers. In the case of tribal conflict, patients in some districts have been forced to seek treatment at alternate locations or miss appointments altogether because of ongoing ethnic clashes. With respect to providers, in 1 district an ART provider was forced out of the facility because of cultural disputes with the local community. Despite the challenges, these are among the first data from PNG to indicate the feasibility of decentralization of ART services to the district level.

As is convention in retention analyses, we excluded individuals who transferred out from the numerator and denominator of our retention proportions. Transfer was common and may not have been consistently documented in patient charts. Therefore, some people who transferred may have been considered lost-to-follow-up in these analyses. A second limitation relates to our outcome definition, which considered retention at the time of data collection and could not consistently identify individuals who may have been lost and returned to care. Another potential limitation is that, although the PAPUA model was implemented fully and consistently to the best of our knowledge, we lack formal assessments of the true fidelity to the PAPUA model throughout the 5-year study period. Training curricula for case management, standard operating procedures, job descriptions, referral pathways, intake forms, and integration with clinic staff, all evolved over time as the model matured.

Back to Top | Article Outline


To the research team's knowledge, this is the first study to examine the feasibility and effectiveness of decentralizing care while providing a package of patient/provider support to the district level through rural government-operated health facilities. Although there is room to improve the retention proportions observed under the PAPUA model in Eastern Highlands, particularly early on in treatment, it seems a strong foundation has been established for HIV service delivery; the PAPUA model seems a promising intervention in a complex resource poor setting with tremendous challenges to successful care delivery.

Future research should include qualitative investigations to understand patient and provider perceptions of the benefits and drawbacks of the PAPUA model. The authors believe that PAPUA is replicable by NDOH, by reallocation of current resources in the national HIV response. A formal costing analysis is currently being undertaken. Given the urgent need for a rural model for HIV service delivery, PAPUA should strongly be considered by key stakeholders as a viable scalable approach to improving access to ART in rural PNG and other settings facing similar health care delivery challenges.

Back to Top | Article Outline


The authors thank Dr Joseph Apa and Mr Ben Haili of the Eastern Highlands Provincial Health Authority, Dr James Kintwa, of the Western Highlands Provincial Health Authority, and the Hospital and District level ART Providers and Staff in Michael Alpers and Tininga Clinics. They also sincerely thank the National Department of Health, National AIDS Council Secretariat, and Medical Research Advisory Council. Finally, the authors thank AusAID for their continuing partnership and support.

Back to Top | Article Outline


1. Coghlan B, Millan J, Malau C, et al.. The HIV epidemic in Papua New Guinea. J Acquir Immune Defic Syndr. 2011;58:e48–e50.
2. AusAID. Impacts of HIV/AIDS 2005–2025 in Papua New Guinea, Indonesia and East Timor. Final Report of HIV Epidemiological Modelling and Impact Study. Canberra, Australia: Australian Agency for International Development, Commonwealth of Australia; 2006.
3. Kelly A, Worth H, Man N, et al.. Barriers and facilitators for adherence to antiretroviral therapy in Papua New Guinea. Curr HIV Res. 2010;8:630–637.
4. Wardlow H. Men's extramarital sexuality in rural Papua New Guinea. Am J Public Health. 2007;97:1006–1014.
5. Schwarz N. Blessed Are the Virtuous: Evangelical and Pentecostals in Papua New Guinea Speak About HIV/AIDS. Goroka, Papua New Guinea: Melanesian Institute; 2011.
6. UNAIDS/WHO Oceania. AIDS Epidemic Update. Regional Summary. Geneva, Switzerland: Joint United Nations Programme on HIV/AIDS; 2008.
7. Wardlow H. Passenger-women: changing gender relations in the Tari Basin. P N G Med J. 2002;45:142–146.
8. National AIDS Council. National HIV/AIDS Support Project, Milestone 103 Review of ART Pilot Report. Port Moresby, Papua New Guinea: National AIDS Council; 2006. Available at: Accessed August 25, 2013.
9. Raynes P, Maibani G. The challenges to scaling up antiretroviral therapy in Papua New Guinea. P N G Med J. 2006;49:32–42.
10. Izard J, Dugue M. Moving Toward a Sector-Wide Approach: Papua New Guinea, the Health Sector Development Program Experience. Manila, Philippines: Asian Development Bank; 2003.
11. Papua New Guinea District and Provincial Profiles. New Guinea, Indonesia: The National Research Institute of Papua; 2010.
12. Mgone C, Passey M, Anang J, et al.. Human immunodeficiency virus and other sexually transmitted infections among female sex workers in two major cities in Papua New Guinea. Sex Transm Dis. 2002;29:265–270.
13. Zimmer-Tamakoshi L. Nationalism and Sexuality in Papua New Guinea. Pac Stud. 1993;16:61–97.
14. Katz MH, Cunningham WE, Fleishman JA, et al.. Effect of case management on unmet needs and utilization of medical care and medications among HIV-infected persons. Ann Intern Med. 2001;135:557–565.
15. Sherer R, Stieglitz K, Narra J, et al.. HIV multidisciplinary teams work: support services improve access to and retention in HIV primary care. AIDS Care. 2002;14(suppl 1):S31–S44.
16. Thurman TR, Haas LJ, Dushimimana A, et al.. Evaluation of a case management program for HIV clients in Rwanda. AIDS Care. 2010;22:759–765.
17. Quinn TC, Wawer MJ, Sewankambo N, et al.. Viral load and heterosexual transmission of human immunodeficiency virus type 1. N Engl J Med. 2000;342:921–929.
18. Donnell D, Baeten JM, Kiarie J, et al.. Heterosexual HIV-1 transmission after initiation of antiretroviral therapy: a prospective cohort analysis. Lancet. 2010;375:2092–2098.
19. Cohen MS, Chen YQ, McCauley M, et al.. Prevention of HIV-1 infection with early antiretroviral therapy. N Engl J Med. 2011;365:493–505.
20. Anglemyer A, Rutherford GW, Baggaley RC, et al.. Antiretroviral therapy for prevention of HIV transmission in HIV-discordant couples. Cochrane Database Syst Rev. 2013;4:CD009153.
21. Heywood A. Evaluation Report AidWorks ING918. St. Lucia, Australia: UniQuest Pty Limited; 2009. Available at: Accessed August 25, 2013.
22. Tuller DM, Bangsberg DR, Senkungu J, et al.. Transportation costs impede sustained adherence and access to HAART in a clinic population in southwestern Uganda: a qualitative study. AIDS Behav. 2010;14:778–784.
23. Hardon AP, Akurut D, Comoro C, et al.. Hunger, waiting time and transport costs: time to confront challenges to ART adherence in Africa. AIDS Care. 2007;19:658–665.
24. Franke MF, Murray MB, Muñoz M, et al.. Food insufficiency is a risk factor for suboptimal antiretroviral therapy adherence among HIV-infected adults in urban Peru. AIDS Behav. 2011;15:1483–1489.
25. Zachariah R, Teck R, Buhendwa L, et al.. Community support is associated with better antiretroviral treatment outcomes in a resource-limited rural district in Malawi. Trans R Soc Trop Med Hyg. 2007;101:79–84.
26. Jerene D, Næss A, Lindtjørn B. Antiretroviral therapy at a district hospital in Ethiopia prevents death and tuberculosis in a cohort of HIV patients. AIDS Res Ther. 2006;3:10.
27. Tassie JM, Baijal P, Vitoria MA, et al.. Trends in retention on antiretroviral therapy in national programs in low-income and middle-income countries. J Acquir Immune Defic Syndr. 2010;54:437–441.
28. Magnus M, Herwehe J, Murtaza-Rossini M, et al.. Linking and retaining HIV patients in care: the importance of provider attitudes and behaviors. AIDS Patient Care STDS. 2013;27:297–303.

HIV/AIDS; case management; rural; Highlands; decentralized; delivery

Supplemental Digital Content

Back to Top | Article Outline
© 2014 by Lippincott Williams & Wilkins