Introduction

Measurement of absolute cluster of differentiation type 4 (CD4+) cells/μL blood levels is important for initiating and monitoring anti-retroviral therapy (ART) in Human Immunodeficiency Virus (HIV) infected people. WHO guidelines recommend CD4 testing at ART initiation and at six month intervals in settings where viral load measurement is not available [1]. Zimbabwe is a resource-limited country in Sub-Saharan Africa (SSA) with an HIV prevalence of 15% among the adult population [2]. In line with WHO recommendations, national HIV guidelines between 2010 and 2013 recommended ART initiation for all HIV positive adults with CD4 levels below 350 cells/μL, or with WHO clinical stages three or four [3]. In mid-2013, this threshold was raised to 500 CD4 cells//μL, and viral load testing is now encouraged by WHO, if possible [4]. Wide scale implementation of these new recommendations in countries like Zimbabwe, however, is limited by financial constraints. In Zimbabwe, national guidelines thus recommend prioritizing patients with CD4 levels below 350 cells/μL and continue to recommend CD4 testing as the main diagnostic method for treatment initiation and monitoring [4]. For the initiation of preventive treatment of HIV coinfections such as toxoplasmosis or cryptoccocal meningitis, CD4 testing remains the method of choice to define treatment eligibility according to WHO recommendations [5,6].

In Zimbabwe, as in other SSA countries, CD4 testing capacity has not kept up with the continuous ART decentralization process from urban to peripheral health facilities [7]. This is due to high costs of diagnostic equipment, poor facility infrastructure and lack of skilled staff [8,9]. As a result, CD4 diagnostics remain limited to central laboratories in urban settings, which requires sample transportation between rural and central facilities [7,10,11]. As in most SSA countries, Ethylenediaminetetraacetic Acid (EDTA) tubes are widely used in Zimbabwe for CD4 blood collection. Samples collected in EDTA tubes must be analysed within 24 hours, requiring same-day transportation and processing from rural to central facilities. This, however, is often not possible, particularly in remote areas or during the rainy season.

The effects of these constraints on access to CD4 testing for patients from rural areas in SSA remain poorly understood. The challenges of providing HIV diagnostics and care in rural SSA are frequently reported in the literature, but most often without explicitly comparing rural versus urban areas [12]. Studies that do assess HIV diagnostic issues lack a quantification of the impact at an individual patient level [11,13], and tend to focus on patient outcomes [9,14,15], ART coverage [16–19], or voluntary testing and counselling [9,20–22]. CD4 testing in rural patients for treatment initiation and monitoring is often overlooked, or where it is assessed, explicit quantification of the difference between rural and urban patients is lacking [8,10,14,23]. To our knowledge, there are no studies that compare and quantify differences in WHO recommended CD4 testing practices between HIV patients from rural and urban areas during the first year on treatment.

Evidence regarding the existence and magnitude of differences in access to CD4 testing in disadvantaged populations is crucial for public health planning and to support changes in existing service delivery. For instance, alternative diagnostic procedures that could circumvent sample transport constraints such as use of dried blood spots, blood fixatives, point-of-care CD4 enumeration, or CD4 Stabilization Tubes are often associated with higher costs and hence not affordable for many national AIDS programmes. Their introduction would require sound justification. In addition, if routine viral load testing is to become a reality as a treatment monitoring tool in Zimbabwe and other SSA countries soon, it will likely require overcoming similar challenges in specimen transport [10]. Hence the need for more detailed evidence on the differences in access.

In a Médecins sans Frontières / Doctors Without Borders (MSF) HIV project in Beitbridge district, Zimbabwe, we compared access to CD4 testing among HIV patients from rural and urban areas at ART initiation, at six and at 12 months of treatment in a retrospective cohort analysis.

Material and Methods

Results

There were 2,222 records that matched the inclusion criteria. Of these, 77 (3%) were excluded (34 duplicate records; 43 records with missing data), leaving records of 2,145 patients for the analysis. Patient baseline characteristics are described in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of rural vs urban HIV patients in Beitbridge, Zimbabwe, at ART initiation between January 2011 and December 2012.

https://doi.org/10.1371/journal.pone.0129166.t001

Adjusting for socio-demographic and clinical differences between patient groups, the risks for missing CD4 testing were 0.8 (95% CI 0.7–0.9) times lower, 9.2 (95% CI 5.5–15.3) times higher and 7.6 (95% 3.7–17.1) times higher for rural patients compared to urban patients at ART initiation, six months and twelve months, respectively. The inclusion of covariates had no substantial effect on the association between area of residence and CD4 testing (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Association between rural vs urban residence and CD4 testing among HIV patients in Beitbridge, Zimbabwe, initiated on ART between January 2011 and December 2012.

https://doi.org/10.1371/journal.pone.0129166.t002

While the proportion of patients accessing medical consultations decreased at six and 12 months of treatment progressively, the rate of decline in visits was similar in rural and urban patients (Fig 1). Of those patients coming for consultations, the proportion with CD4 results recorded at six and at 12 months (+/-30 days) dropped precipitously in both groups, particularly in rural patients at six and 12 months. At these two visits, almost no patients from rural areas had a CD4 result recorded (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Proportions of rural vs urban patients accessing consultations, and proportions receiving CD4 testing during consultations, during first year of anti-retroviral therapy in Beitbridge, Zimbabwe, between January 2011 and December 2012.

https://doi.org/10.1371/journal.pone.0129166.g001

Among all patients initiated on ART, the proportion of patients that received all CD4 tests as recommended during the course of their first year on treatment decreased from 43% and 53% at initiation to 15% and 2% at six months, and to 1% and 0% at 12 months, for urban and rural patients, respectively.

Discussion

This is the first study to compare and quantify differences in CD4 testing between HIV patients from rural and urban areas in a SSA setting during the first year of ART. While our findings show that rural patients among the studied population are much less likely to have CD4 testing done when they should, according to WHO recommendations, at six and 12 months compared to urban patients, we found startlingly low CD4 testing in both patient groups.

The main strengths of this study are that i) a large number of patients, enrolled over two years, were included and followed up during their first year of treatment; ii) data were sourced from a standardized and internationally used database; iii) given the low number of excluded records, we are confident that our study sample is representative of the MSF HIV cohort initiated in Beitbridge in 2011 and 2012.

A number of limitations are recognized. First, our study only included patients from the six MSF-supported RHCs out of 13 RHCs districtwide, plus BDH. Treatment coverage among ART eligible HIV patients in Beitbridge district was about 35% (MSF project monitoring tool (logframe) 2013). While MSF chose to support RHCs based on remoteness and the intention to have fair geographical coverage, we cannot presuppose representativeness of our cohort for the the district or Zimbabwe as a whole. In fact, we may even underestimate the disparity in CD4 testing between urban and rural areas since MSF offered reliable and comprehensive care at RHCs during outreach days, which might have stimulated uptake of health services in these areas.

Second, data were collected for routine reporting purposes and we were unable to cross-check our study data with patient records for this analysis; nonetheless, data are believed to be robust due to built-in data entry checks in FUCHIA.

Third, although we controlled for basic demographic and clinical differences between patient groups, we cannot rule out residual confounding by other factors such as socio-economic status, health seeking behavior, number of patient clinic visits, or differences in service provision between urban and rural health facilities. In addition, mobility might have been different, being potentially more cyclical in rural areas depending on farming seasons and more constant in urban areas for commercial activities.

Our findings raise a number of important points for discussion:

Until now, the disadvantaged situation in rural areas in SSA with regards to CD4 testing had only been explored from a health system or laboratory infrastructure perspective. It had not yet been quantified at a patient level [11,13]. Our findings from Zimbabwe reveal that rural patients have substantially lower access to CD4 testing at time points recommended by WHO under the current system with a surprisingly large effect size. This provides much-needed evidence to lobby for changes in health care delivery at national and international level even if it requires financial investment in new diagnostic technologies such as use of dried blood spots or point-of-care devices [13]

By differentiating between consultations accessed and test results received, we were able to suggest that the reasons for poorer access to CD4 testing in rural compared to urban areas are likely to be within the health care system rather than with the patients. The current EDTA tube based testing method relies on timely specimen transportation and as such, given the transport problems during days when no MSF/MoHCW outreach visits were done, probably explains the bulk of the rural-urban discrepancy. Other explanations could be human resources constraints or lack of phlebotomy equipment. However, since MSF was supplying materials to RHCs during stock-out periods, we think that the latter factors are unlikely to explain much of the observed rural-urban discrepancy. The observed higher proportion of patients receiving CD4 testing at baseline in rural compared to urban areas is striking. A possible explanation for this finding may be related to differences in levels of training between rural and urban health care workers. Rural health care workers often have a lower level of training compared to urban staff, and as such might place greater reliance on the certainty of CD4 testing, instead of clinical assessment, to initiate ART compared to health staff in urban areas. This would lead to a relative increase the proportion of patients initiated on ART who receive CD4 testing at baseline in rural areas.

Overall, proportions of being CD4 tested were very low in urban as well as in rural patients, suggesting severe underlying health system constraints in the provision of timely CD4 testing that go beyond transport issues. Parts of the gap might be due to CD4 tests being done outside the precision window of +/-30 days. Patients tested outside this time period who returned within the time window shortly thereafter might not have been re-tested during that subsequent consultation. This implies rational clinical practice, but might have led to an overestimation of the problem in our study. However, given the magnitude of the observed gaps, it is unlikely that this fully explains the extent of our findings. Data capture of CD4 testing into the FUCHIA database was done from returned laboratory result slips in patient files and as such we could not determine where between patient contact and filing of result slip the processing chain broke down. Possible reasons for patients in both areas not receiving effective CD4 testing when they should have (according to WHO recommendations) may have included: patients not being bled during consultations; specimen handling issues in the laboratory; errors in returning of result slips into patient files; stock ruptures of cytometer consumables in the national supply system; or intermittent cytometer breakdowns.

The setting of this study—one of the districts in Zimbabwe with the highest HIV prevalence [2]—has a rural-urban pattern comparable to other districts in the country, and the public health care system suffers from financial, human resources and logistics constraints that are similar to other resource-limited high HIV prevalence countries. However, due to Zimbabwe´s unique political and economic situation, our findings should only be extrapolated to other settings in SSA with utmost caution.

To improve CD4 testing for rural patients, either sample transportation needs to be drastically improved or alternatives to the current testing methods need to be considered [27]. Point-of-care enumeration is a promising alternative to currently-used EDTA tubes [28,29] and is increasingly being implemented in rural SSA settings [30–33]. However, there have been mixed results from its performance in operational settings in SSA [34,35]. The use of CD4 Stabilization Tubes (Becton Dickinson, Franklin Lakes, New Jersey, USA) circumvent the need for same-day sample analysis, but this approach requires more evaluation in operational settings [36,37]. Viral load enumeration is now being encouraged as the cornerstone of treatment monitoring [38]. If it becomes affordable for widespread use in SSA soon, it should be done based on dried blood spot (DBS) techniques in rural areas, as any system that relies on time-restricted specimen transportation is unlikely to be successful in such settings [28,39]. Dried blood spot based viral load enumeration has been successfully tested in SSA settings but still relies on a well-functioning health system during routine use [38,40,41]. Ideally, a viral load Point-of-care test providing immediate results would be needed.

Conclusions

In conclusion, this study found low access to CD4 testing at time points recommended by WHO (+/- 30 days) in a high HIV burden district in Zimbabwe. Rural patients were particularly disadvantaged, most likely due to difficulties in specimen transportation, and might benefit from the use of new context-adapted diagnostic methods. However, any innovation will have limited overall impact on the majority of patients without a well-managed health system to support it. Hasty roll out of universal viral load monitoring may face a similar fate as current CD4 testing methods.

Supporting Information

Acknowledgments

We would like to thank staff from MoHCW who collaborated with MSF in Beitbridge during the project time, in particular all staff who participated in joint outreach visits. We would also like to thank MSF programme managers at national and international level who supported this research.