Newly implemented community CD4 service in Tshwaragano, Northern Cape province, South Africa, positively impacts result turn-around time

Background The Northern Cape is South Africa’s largest province with an HIV prevalence of 7.1% versus a 13.5% national prevalence. CD4 testing is provided at three of five National Health Laboratory Service district laboratories, each covering a 250 km precinct radius. Districts without a local service report prolonged CD4 turn-around times (TAT). Objective This study documented the impact of a new CD4 laboratory in Tshwaragano in the remote John Taolo Gaetsewe district of the Northern Cape, South Africa. Methods CD4 test volumes and TAT (total, pre-analytical, analytical, and post-analytical) data for the Northern Cape province were extracted for June 2018 to October 2019. The percentage of CD4 results within the stipulated 40-h TAT cut-off and the median and 75th percentiles of all TAT parameters were calculated. Pre-implementation, samples collected at Tshwaragano were referred to Kimberley or Upington, Northern Cape, South Africa. Results Pre-implementation, 95.4% of samples at Tshwaragano were referred to Kimberley for CD4 testing (36.3% of Kimberley’s test volumes). Only 7.5% of Tshwaragano’s total samples were referred post-implementation. The Tshwaragano laboratory’s CD4 median total TAT decreased from 24.7 h pre-implementation to 12 h post-implementation (p = 0.003), with > 95.0% of results reported within 40 h. The Kimberley laboratory workload decreased by 29.0%, and testing time significantly decreased from 10 h to 4.3 h. Conclusion The new Tshwaragano CD4 service significantly decreased local TAT. Upgrading existing community laboratories to include CD4 testing can alleviate provincial service load and improve local access, TAT and efficiency in the centralised reference laboratory.


Introduction
The Northern Cape province occupies 30.0% of the land area of South Africa (372 889 km 2 ) yet is the most sparsely populated due to the arid climate, with a population of 1.26 million people In South Africa, a universal 'test-and-treat' approach was adopted for the diagnosis and treatment of HIV/AIDS patients. 4,5,6,7,8,9 Late HIV presentation and a high disease burden 8,9,10,11 remain challenges in the South African HIV programme. CD4 testing, although no longer required for eligibility for antiretroviral therapy initiation, still plays a valuable role in the identification of HIV-positive patients with advanced disease (CD4 counts < 200 cells/µL) or very advanced disease (CD4 counts < 100 cells/µL). 4,5,6,7,8,10,12,13,14,15,1 6,17,18,19,20,21 The local standard of care in South Africa recommends patients returning for care within seven days of HIV diagnosis. 12 CD4 counts from this visit are used to fasttrack patients onto antiretroviral therapy and for the treatment of other HIV-related opportunistic diseases such as tuberculosis and cryptococcal disease. 15,18,22 In South Africa, CD4 services are offered through the National Health Laboratory Service (NHLS) at 46 testing laboratories placed strategically to service public healthcare facilities consisting of hospitals and primary healthcare clinics in both urban and rural settings. In 2019, 2.8 million CD4 tests were performed across South Africa, with 2.2% done in the Northern Cape province, compared to 34.0% in KwaZulu-Natal and 21.0% in the Gauteng province. The 2020 mid-year population statistics and annual NHLS CD4 testing data (2019) reported that an average of 4.4% of all South Africans received a CD4 test, with 4.7% in the Northern Cape compared to 8.0% per capita in the KwaZulu-Natal province of South Africa. 23 As only 46 of the 268 NHLS laboratories provide a CD4 testing service, samples are not always tested locally. All samples get registered onto the NHLS laboratory information system (LIS) upon arrival at one of these laboratories, even if CD4 testing is not provided. For this paper, two types of sample referrals are described: one where samples from primary health care facilities and hospitals within respective service precincts are received and registered at smaller, nontesting source laboratories and referred to a CD4 testing laboratory (inter-laboratory referral), and another where samples from hospitals and primary healthcare facilities within the immediate service precinct or catchment area are received, registered and tested at the same laboratory.  Figure 1) and provides general pathology services to 38 public health facilities within a 200 km service precinct radius.
Laboratory service efficiency and overall performance are typically measured by turn-around time (TAT). 26,27,28,29,30 In the NHLS, total TAT is reported and includes pre-analytical or laboratory-to-laboratory TAT, that is from first registration onto LIS to sample receipt at the testing laboratory; the analytical registered-to-test TAT component, reporting the

Ethical considerations
Ethics clearance was obtained from the University of the Witwatersrand (M1706108, approved for five years from 13 July 2017). This was a retrospective study using CD4 sample data from the corporate data warehouse, with no patient consent required and no patient identifiers captured.

Study design
A cross-sectional study design was used to analyse TAT data from CD4 testing laboratories in Kimberley, Upington, De Aar and Tshwaragano, Northern Cape, South Africa. Only data with a reviewed CD4 result between June 2018 to October 2019 were analysed. CD4 testing at all of the Northern Cape laboratories was done using the Beckman Coulter Aquios CL (continuous loader) cytometer system (Beckman Coulter, Miami, Florida, United States). 32 This system is ideal for low to medium-high test volumes (10-120 samples per day) and is user-independent, with sample preparation and analysis performed in a closed system requiring minimal operator hands-on time, an especially useful feature for sites with limited dedicated/skilled staff. The Aquios CL platform performance correlated well with previously used systems such as the COULTER EPICS XL-MCL™ 33 and FC500 multi-plate loader (MPL) 34 as per validations done before the national roll-out 35,36 and instrument verifications done per testing site as part of the implementation process. 37,38,39

Data preparation
Data were extracted from the Central Data Warehouse of the NHLS as password-protected files. The extracted data included the following variables: episode number, referral laboratory, testing laboratory, facility name, province, health district, sub-district, date first registered on the LIS, date tested, date reviewed (result released), total TAT, laboratoryto-laboratory TAT, registered-to-test TAT, and test-to-review TAT. Data were categorised as 'Not referred' where the referral and testing laboratory was the same and as 'Referred' where the testing and referral laboratories differed. Data with invalid dates and a total TAT > 120 h or < 1 h were excluded. A total TAT of 1 h is set as the minimum analytical acquisition and reporting time on the Aquios platform, while a total TAT > 120 represented the upper limit of the CD4 testing window (five days), beyond which a specimen is no longer suitable for analysis. 40,41 Data were categorised to distinguish between the pre-implementation (June 2018 to October 2018) and post-implementation (November 2018 to October 2019) periods; the month of implementation (October 2018) was categorised as pre-implementation as samples were still being referred to Kimberley.

Implementation procedures at the Tshwaragano laboratory
All of the procedures undertaken to ensure that the Tshwaragano laboratory was ready to offer quality CD4 testing were documented. Typically, a site inspection was done by the service provider and an on-site operational checklist was completed by a CD4 trainer to assess requirements for implementation. Several readiness checks were confirmed before the commencement of instrument implementation, training, verification, and reportable patient testing ( Figure 2).

Satisfaction questionnaire
A Likert scale satisfaction survey was circulated to the 15 health facilities with the highest test volumes referred to Tshwaragano laboratory for CD4 testing. On a scale from '1' (poor) to '5' (very good), the facility manager/senior nurse at these clinics rated the overall service rating, average TAT and CD4 test result delivery of the CD4 service pre-and postimplementation of local CD4 testing at the Tshwaragano laboratory.

Data analysis and statistics
Microsoft Excel (Redmond, Washington, United States) and SAS ® statistical software 9.4 (Cary, North Carolina, United States) were used to analyse the data. Graphs were created with GraphPad Prism software, version 7.0 for Windows (GraphPad Software, San Diego, California, United States).
Pre-implementation of the CD4 testing laboratory at Tshwaragano, the overall and monthly number of referred samples tested by the three then active CD4 laboratories in the region (Kimberley, De Aar and Upington) were assessed. TAT parameters to determine laboratory performance/ efficiency included the total TAT, the interquartile range http://www.ajlmonline.org Open Access (IQR) of the total TAT, the percentage of all tested samples within the NHLS-stipulated 40-h TAT cut-off, and the median TAT for the three TAT components. Due to the non-Gaussian distribution of TAT data, 28 the median and 75th percentile of total TAT was calculated and reported for this study; this was to allow the analysis of the majority of samples without the influence of outliers or, in Gaussian terms, the long right tail, that is CD4 samples with a total TAT exceeding 120 h (typically outside the 99th percentile). The monthly median and 75th percentile total TAT for the Tshwaragano and Kimberley laboratories were reported.
Pre-and post-implementation test volumes and TAT parameters (total TAT and TAT components) were compared at the provincial level (all Northern Cape results) and individual laboratory level using either the unpaired, nonparametric Mann-Whitney analysis for two groups or the unpaired, non-parametric Kruskal-Wallis test for three or more groups. P-values of < 0.01 were considered statistically significant and p < 0.001 highly significant.

Results
Exclusion criteria identified 10 131 samples (1.5%) with a total TAT exceeding 120 h and two samples with a total TAT of less than 1 h. After applying the exclusion criteria, CD4 data for 83 232 samples collected during the study period were analysed. Of all CD4 samples tested in the Northern Cape province, the median CD4 count was 437 cells/µL (IQR: 237-658 cells/µL), the median total TAT was 19.8 h (IQR: 11.9-31.9 h), and 96% of all tested samples were within the NHLS-stipulated 40-h TAT cut-off.

Pre-and post-implementation CD4 test volumes
During the pre-implementation and implementation months

Referral patterns during the pre-and postimplementation periods
During the pre-implementation phase, 32

Turn-around time: Pre-versus postimplementation
The median total TAT across the four Northern Cape province testing laboratories during the whole testing period was 19. For the Tshwaragano laboratory, there was a significant decrease in median total TAT from 24.7 to 12 h pre-versus post-implementation (Mann-Whitney p = 0.0009), mainly  LAB-to-LAB, laboratory-to-laboratory; REG-to-TST, registered-to-test; TST-to-RVW, test-to-review.

Pre-implementation Post-implementation
The test-to-review TAT component did not impact the total TAT as all laboratories reported TAT times < 1 h (median: 10 min).

Questionnaire results
The pre-implementation assessment score was 3 (performance rated as average), with delayed result delivery being the most common reason for the rating provided. The postimplementation score was 5 (very good), with faster result delivery (most within 24 h) being the most cited improvement. Respondents observed that patients returned earlier than the stipulated seven days to collect their results (within 2-4 days). Nonetheless, feedback received highlighted ongoing challenges of poor connectivity that affected access to results on the Internet-based LIS and delivery of SMS printer results in remote outlying facilities. The test-to-review TAT component at all testing laboratories was < 1 h and did not impact total TATs, that is, there were no delays associated with waiting for qualified staff members to review results. Further streamlining of this process was made possible by the introduction of autoreview in March 2020. 42 There was a slight increase (~4%) in the monthly average CD4 tests conducted in the Northern Cape province during the pre-and post-implementation periods. The latter could be attributed to the availability of local testing at Tshwaragano and the addition of six additional referring clinics (personal communication with laboratory management) since the commencement of CD4 services.

Discussion
The outcomes of the satisfaction rating questionnaire were a testament to the success of this implementation, with an overall post-implementation rating of 5 (vs 3 pre-implementation). Improved local CD4 report delivery (due to shortened local TAT) could see patients return for their results earlier than the stipulated seven days. This is especially relevant in patients with severe HIV disease (CD4 < 100 cells/µL) to enable the fasttracking of serum cryptococcal antigen testing and rapid initiation of treatment in cryptococcal antigen-positive cases.
As monitored through the NHLS weekly/monthly TAT dashboard reports, the Tshwaragano laboratory maintained a median total TAT and 75th percentile total TAT of < 20 h for the last financial year (April 2020 to March 2021), with > 98% of samples tested within the 40-h cut-off.
This study supports the findings of previous studies 24,28,31,43 and demonstrates how decentralising CD4 testing services to district or community facilities according to a tiered service delivery model 25 can improve the TAT at the decentralised site. Smaller community laboratories with existing laboratory infrastructure offering more generalised pathology services can reliably implement CD4 services, effectively eliminating delays due to the transport of samples historically sent to centralised testing facilities. NHLS community laboratories are usually located in the seat of the respective districts and generally serve large service precincts, typically supporting up to 40 or more health facilities. With minimal training effort and cost 25,43 (benchtop installation, installation of LIS interface with instruments, staff training, and testing platforms on national service level agreement), laboratories can be upgraded to include CD4 testing in their repertoire of services. Decentralised services also improve coordinated technical efforts across the parent province (the Northern Cape in this instance), where CD4 laboratories can, in line with the integrated service delivery model service plan, 25 operate as a consolidated network with two-way support between the main reference laboratory (in Kimberley) and sister community laboratories, thereby ensuring uninterrupted self-sustaining service.

Limitations
The reported TAT only reflects laboratory TAT and not the full sample journey from a patient to a result-in-hand, which may impact patient management (i.e., leading to patients returning for results, loss to follow-up, and/or delays in the initiation of therapy). However, this paper aimed to show the impact on TAT at the laboratory level and did not assess the impact on patient management.

Conclusion
Using existing infrastructure, upgrading an existing district NHLS laboratory to include CD4 testing is possible with minimal intervention. District CD4 testing laboratories can service their local health clinics and hospitals without delays due to sample referral, reducing the total TAT and allowing earlier patient support/intervention. The implementation of a fourth district CD4 testing facility in the Northern Cape province resulted in the local testing of > 90% of samples from their catchment areas while maintaining excellent performance.