The South African National Health Laboratory Service provides laboratory services for public sector health facilities, utilising a tiered laboratory model to refer samples for CD4 testing from 255 source laboratories into 43 testing laboratories.
The aim of this study was to determine the impact of distance on inter-laboratory referral time for public sector testing in South Africa in 2018.
A retrospective cross-sectional study design analysed CD4 testing inter-laboratory turn-around time (TAT) data for 2018, that is laboratory-to-laboratory TAT from registration at the source to referral receipt at the testing laboratory. Google Maps was used to calculate inter-laboratory distances and travel times. Distances were categorised into four buckets, with the median and 75th percentile reported. Wilcoxon scores were used to assess significant differences in laboratory-to-laboratory TAT across the four distance categories.
CD4 referrals from off-site source laboratories comprised 49% (
Variability in inter-laboratory TAT was noted for all inter-laboratory distances, especially those exceeding 300 km. The correlation between distance and laboratory-to-laboratory TAT suggests that interventions are required for distant laboratories.
Laboratory services in South Africa are provided within the public sector by the National Health Laboratory Service (NHLS).
A CD4 count is used to assess the patient’s immune status (level of immune suppression) at presentation and during therapy.
Previous work revealed that the placement of CD4 testing equipment and testing capacity, within an area identified with longer TAT attributable to pre-analytical causes, led to a marked shortening of the associated pre-analytical TAT and noticeably shorter overall TAT.
To achieve prompt sample referral within the CD4 network to address pre-analytic TAT, the NHLS employs a ‘hub and spoke’ approach whereby each testing CD4 laboratory receives referred samples from multiple source laboratories where the samples are first accepted into the laboratory network (but where there are no facilities to perform the testing).
It is anticipated that the distance and travel time for inter-laboratory referral could potentially impact lab-to-lab TAT performance. Laboratories located more than 300 km from a testing laboratory are expected to have longer lab-to-lab TAT. Conversely, laboratories within a 100 km radius are expected to have shorter travel times and lab-to-lab TAT. With a decentralised laboratory service model, it would be assumed that an inter-laboratory referral distance would not exceed 250 km. The 250 km radius is based on the study by Cassim et al. that used the integrated tiered service delivery model (ITSDM) coverage precinct approach to address ART-related testing service coverage gaps.
The objective of this study was to determine the impact of distance on inter-laboratory referral time in 2018 for public sector testing in South Africa.
Ethical clearance was obtained from the University of the Witwatersrand (M1706108). This study was a retrospective analysis of NHLS laboratory data, did not use any patient identifiers and did not involve any direct patient contact.
This study used a cross-sectional study design. Microsoft Access and Excel (Microsoft Corp., Redmond, California, United States) were used to prepare data
Skewness can be quantified as a representation of the extent to which a given distribution varies from a normal distribution.
For each CD4 sample, the requesting health facility is captured to facilitate the delivery of patient reports. In the LIS, each health facility has a designated location code. This code is captured in the LIS based on health facility details provided on the laboratory request form. Unfortunately, very few health facilities had latitude and longitude data making it difficult to calculate the distance to the local source laboratory.
CD4 samples that were not referred (i.e. that were received directly at testing laboratories) were excluded for the purpose of this study. Data were reported as descriptive statistics, and visualised using histograms. A scatter plot was used to demonstrate the relationship between referral times (hours) and distances (km), with descriptive statistics reported. The inter-laboratory distances between source and testing sites were categorised into four buckets; (1) ≤ 100 km, (2) 101–200 km (3) 201–300 km and (4) > 300 km. The aim of the analysis was to determine whether lab-to-lab TAT was influenced by distance. The analysis could also reveal whether decentralisation within a 200 km radius would decrease lab-to-lab TAT. The number of health facilities associated with local source laboratories in each of the four inter-laboratory distance buckets was reported. The lab-to-lab TAT component and descriptive statistics (median and 75th percentile) were reported, using the non-parametric Wilcoxon scores (rank sum) test to determine differences in TAT reported among the distance categories. In addition, inter-laboratory referral routes longer than 300 km that referred 1000 or more samples in 2018 are reported as a separate table reporting: (1) source laboratory, (2) testing laboratory, (3) distance, (4) travel time (hours), (5) percentage within 12-h cut-off and (6) 75th percentile, using source and testing laboratory aliases. The projected daily referral volumes were calculated based on the assumption of 21.73 working days per month.
There were 2 844 242 CD4 samples performed in 2018, of which 1 390 510 (49%) were referred for centralised testing within the NHLS laboratory network.
A linear relationship exists between referral times and distances. lab-to-lab data reported a positively skewed distribution (skewness = 2.83). A lab-to-lab median of 11 h with an interquartile range of 7–17 h was reported. The mode was 8 h, with a range of 96 h (
Distribution of inter-laboratory CD4 testing referral laboratory-to-laboratory turn-around time in the 2018 calendar year for National Health Laboratory Service laboratories in South Africa. Median, interquartile range, mode and range are indicated (inset).
The inter-laboratory CD4 testing referral distance data also reported a positively skewed distribution (skewness = 1.66) (
Relationship between inter-laboratory CD4 testing referral distance and travel time between source and CD4 testing laboratories in the 2018 calendar year across the National Health Laboratory Services in South Africa.
Overall, 67% of samples were referred to testing laboratories within a 100 km inter-laboratory referral distance, increasing to approximately 93% of samples within a 200 km radial precinct. Approximately 7% of samples were transported 201 km or more from their source laboratory to their sister centralised testing laboratory. Overall, for referral distances under 300 km, all inter-laboratory precincts (both the mode and median lab-to-lab TAT) met the organisational inter-laboratory TAT cut-off of 12 h.
Longer lab-to-lab TATs were noted from referral or source laboratories further than 300 km from testing laboratories. Analysis of the 75th percentile lab-to-lab TAT, unlike the mode and median parameters, revealed that only inter-laboratory referral distances of under 100 km met the 12-h cut-off (
Inter-laboratory CD4 testing referral laboratory-to-laboratory performance by distance categories in the 2018 calendar year for National Health Laboratory Services laboratories in South Africa.
Distance category |
Number of health facilities |
Number of referred samples (%) |
Number of source laboratories (%) |
Cut-off | Mode | Median | 75th percentile | |||
---|---|---|---|---|---|---|---|---|---|---|
% | % | % | ||||||||
≤ 100 km | 2169 | 45.7 | 936 667 | 67.4 | 124 | 49 | 12 | 7 | 11 | 8 |
101–200 km | 1644 | 34.6 | 356 162 | 25.6 | 86 | 34 | 12 | 8 | 11 | 17 |
201–300 km | 576 | 12.1 | 73 596 | 5.3 | 29 | 11 | 12 | 9 | 10 | 14 |
> 300 km | 357 | 7.5 | 23 815 | 1.7 | 16 | 6 | 12 | 12 | 15 | 27 |
Note: Wilcoxon rank sum test –
CD4, cluster of differentiation 4.
, Indicates the distance between the CD4 source laboratory that receives samples from clinics via a courier network and the testing laboratory. The further the distance, the higher the potential for an extended turnaround time affecting service delivery.
, Number of laboratory information system location codes.
, Percentage of samples referred.
Laboratories serviced the majority of health facilities (80%) with an inter-laboratory referral distance within 200 km. There were 12% and 7.5% of health facilities that were serviced by laboratories with an inter-laboratory referral distance of 201 km – 300 km and more than 300 km.
For five source laboratories, distance to the nearest testing facility exceeded 300 km (
Laboratotry-to-laboratory performance of inter-laboratory referral routes over 300 km and involving at least 1000 referred samples in the 2018 calendar year for National Health Laboratory Service laboratories in South Africa.
Source laboratory |
Testing laboratory |
Number of samples referred | Projected daily referrals |
Distance (km) |
% within lab-to-lab cut-off | 75th percentile lab-to-lab |
---|---|---|---|---|---|---|
S1 | T1 | 6328 | 24 | 402 | 36 | 17 |
S2 | T1 | 6312 | 24 | 350 | 37 | 15 |
S3 | T2 | 4903 | 19 | 319 | 6 | 36 |
S4 | T3 | 2378 | 9 | 304 | 83 | 10 |
S5 | T4 | 1622 | 6 | 560 | 8 | 29 |
CD4, cluster of differentiation 4.
, Laboratory that receives CD4 samples from health facilities that is delivered by a courier.
, Laboratory that performs the CD4 test.
, Number of samples referred from a source to a testing laboratory.
, Distance in kilometres between a source and a testing laboratory (inter-laboratory referral).
In this study, the impact of distance and travel time on lab-to-lab TAT as a pre-analytical component of total TAT was assessed to categorise laboratory service efficiency and, in national service, understand where bottlenecks of logistics may occur. We observed moderate correlation between distance travelled and the overall recorded inter-laboratory lab-to-lab TAT. However, across all service precincts, especially those exceeding a radius of 300 km, variable inter-laboratory TAT was recorded. Generally, median lab-to-lab TAT of referrals within a 200 km radius of a testing facility (
Despite previous decentralisation drives and a national median inter-laboratory TAT of 11 h for at least 93% of referred samples, there are still outliers noted. For example, where distances travelled to a centralised testing facility was between 100 km and 200 km, a 75th percentile inter-laboratory TAT of 17 h was noted. Therefore, although most of the referring times reported here were within the inter-laboratory transfer cut-off of 12 h, service gaps are evident (where referral laboratories failed to deliver their samples to the testing laboratory within the stipulated time frame).
Glencross et al. developed the ITSDM that consists of five tiers to facilitate sustainable ‘full-service coverage’ across South Africa.
This study demonstrated a median lab-to-lab TAT of 15 h, with a right skewed distribution and a 75th percentile of 27 h for referrals outside a 300 km service precinct to the nearest testing laboratory. This finding, again, reiterates the need to apply the ITSDM approach described above to assist with identifying sites for capacity development and alleviation of the longer inter-laboratory TATs identified in this work. At least four sites with low test volumes and exceeding 300 km to the nearest testing facility warrant the introduction of a local service that could be accomplished with lower throughput platforms (daily capacity of 40 samples).
With only about 50% of tested CD4 samples referred between laboratories, the NHLS strives to have adequate testing facilities for maximum coverage especially in high-burden and remote areas,
A limitation of this analysis is that it focuses on the inter-laboratory referral after samples are received within the laboratory network. The laboratory pre-analytical phase of TAT is measured from the first registration in an NHLS laboratory to receipt in the testing facility; it does not include the time from venepuncture to first registration on the LIS system. However, as the latter time periods are poorly recorded, it is difficult to determine if a longer lab-to-lab TAT is due to a delay in time from sample collection to registration at the source laboratory. Challenges in the pre-analytical and analytical arms would have to be addressed separately due to differences in the nature of challenges and the involvement of different personnel. Therefore, separate corrective action interventions need to be set in motion. This may include redesigning the specimen referral network which could include additional routes or potentially identifying some sites for point-to-point routing instead of a multi-stop route, or vice versa.
To adequately address all delays of inter-laboratory referral TAT, all aspects of this phase should be addressed. Not all of the processes undertaken during the transfer of samples are documented and traceable, that is, not all steps generate a time and date stamp on the LIS or are date-stamped at the point of venesection. The introduction of electronic monitoring of all total TAT would facilitate a more complete view of possible points of delay across the entire sample journey and identify aspects for corrective action. Vertical audits can be used to identify sources of delay on different days of the week to calculate average (delay) times for each process.
The study used predominantly laboratory data to assess the performance of CD4 test referrals across the NHLS. This study did not report data for other tests as the inter-laboratory referral network is integrated. All the concepts reported in this study for CD4 testing apply to other referred tests as well. No information was available to review time from sample collection to registration onto the laboratory network, as well as time after sample registration at the source laboratory to collection by the courier.
This study demonstrates that most referrals for CD4 testing reach their testing facilities within the expected 12-h window, with some outliers identified. Here, differences in inter-laboratory referral distances and lab-to-lab TAT performance suggest that there are inconsistent systems and practices in use to transfer samples between centres. Further investigation to understand the root causes would assist in aligning efficient delivery of all samples between facilities. There is a need for additional data collection for the inter-laboratory referral process to better understand where service bottlenecks exist. This study identified the need for electronic data recording at multiple stages of sample inter-laboratory referral such that bottlenecks can easily be identified and resolved to optimise timely referrals. For distances exceeding 300 km, the establishment of additional community CD4 laboratories is recommended.
The authors would like to thank the corporate data warehouse, and staff at the National Priority Programme unit for their assistance.
The authors have declared that no competing interests exist.
D.K.G. supervised the study by providing leadership and oversight and was also the project leader. N.C. and L.M.C. designed the study, developed the methodology, conducted the research, conducted the data analysis and prepared the maps. D.K.G. provided editorial comments and technical input. All authors contributed to the manuscript development.
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
Data sharing is not applicable to this article as no new data were created or analysed in this study.
The views and opinions expressed in this article are those of the authors and do not necessarily reflect the official policy or position of any affiliated agency of the authors.