Reduction in cycle time for a rapid polymerase chain reaction diagnostic test at the point of care

Abstract

Background

Rapid testing facilitates safe and effective diagnosis, but the true speed of the process is the time from collection of a sample to delivery of an accurate and reliable test result – ‘end-to-end’ time. Transport, unpacking and relaying of information can extend this time considerably beyond the minimum laboratory turnaround times as stipulated by PCR testing protocols.

Aim/Objective

This study aimed to minimise time needed to ascertain SARS-CoV-2 status prior to treatment in a UK Dental Hospital using a novel, mobile, direct to polymerase chain reaction (PCR) workflow.

Methods

Process flow analysis and PDSA (Plan, Do, Study, Act) cycles for rapid continuous improvement were employed in a service improvement programme. Primerdesign™ q16 rapid PCR instruments and PROmate® COVID-19 direct assays were used for molecular testing.

Findings/Results

We showed a reduction in real-world end-to-end time for a diagnostic test from 240 min to 85 min (65% reduction) over a 4-week period.

Discussion

New rapid technologies have become available that reduce analytical time to under 90 min, but the real-world clinical implementation of the test requires a fully integrated workflow from clinic to reporting.

Background

The time from clinical sample collection to availability of a test result (‘end-to-end time’) is critical for successful rapid-testing implementation, and optimal utilisation of available clinical resources. During the COVID-19 pandemic, healthcare bodies have highlighted the need for improved infection control and use of COVID-19 testing (International Federation for Emergency Medicine, 2020; Irish Association for Emergency Medicine, 2020; Royal College of Emergency Medicine, 2020). However, prolonged end-to-end testing times reduce acceptability and adoption of testing (Kozel and Burnham-Marusich, 2017; Ravi et al., 2020). This may result in increased patient and staff waiting times, impeded patient flow, and crowding that jeopardises infection prevention measures (Boyle and Henderson, 2020).

Rapid COVID-19 testing with reduced end-to-end times allows service providers to optimise treatment pathways, reducing risk of disease transmission to staff and other patients, minimising workforce illness and isolation time, and reduce clean-down time of clinical areas, freeing up staff time for patient-facing work.

Faster access to diagnostic test results means less time spent by patients in assessment areas prior to transfer to the appropriate clinical area (COVID-19-positive or -negative), improving clinical outcomes and infection control (Brendish et al., 2020). Implementation of routine testing could enhance patient and staff safety and improve cross infection control, as well as facilitating resumption of elective procedures.

This service evaluation aimed to implement a rapid-testing workflow that would be feasible, scalable, and consistently fall within a turnaround time of 90 min from sample collection to test result. While others have focused on the benefits of testing in advance of healthcare procedures, this study demonstrates how continuous improvement methodology can be used to implement a rapid-testing protocol and successfully achieve the turnaround time goal.

For any process, we can define starting conditions, and subprocess steps required to reach the end state. We can then identify steps on the critical path, defining the cycle time for the overall process, by asking:

• • What is the required outcome?

• • Which steps must occur in what sequence to achieve that outcome?

• • What defines the time taken for each step?

Efforts toward cycle time reduction focus on decreasing the number and duration of steps on the critical path. We can eliminate certain steps from the process entirely, take steps off the critical path by doing them ‘outside the cycle’, perform steps in parallel, or reduce time for individual steps.

Substantial reductions in cycle time can be achieved through continuous improvement, whereby the process is repeatedly trialled and analysed, and inferred changes subsequently implemented. Here, the PDSA (Plan, Do, See, Act) framework was used to realise this improvement.

Methods

Clinical setting

The clinical setting is a Dental teaching hospital with multiple operatories in large open-plan clinical areas over 4 floors; thus an exemplar of how implementation of rapid testing might also operate in large clinics or small hospitals (and office buildings).

An untested asymptomatic carrier of SARS-CoV-2 undergoing an aerosol generating procedure (AGP) may transmit the infection to other patients and staff in both treatment rooms and open-plan areas (Han et al., 2020; Scottish Dental Clinical Effectiveness Programme, 2021; Sugano et al., 2020). Recent public health guidelines have compelled the Dental Hospital to reduce occupancy of clinical areas and implement enhanced personal protective equipment (PPE) protocols, significantly reducing overall patient capacity.

Safely increasing patient capacity requires a SARS-CoV-2 virus-free environment, and testing of all staff and patients ensures only SARS-CoV-2-negative individuals are involved in AGPs. Rapid point-of-care PCR testing is currently the most reliable technology to offer both the required accuracy (clinical sensitivity) and rapid turnaround time, thus helping restore clinical activity to pre-pandemic levels (Dinnes et al., 2021).

PCR testing

Polymerase chain reaction (PCR) testing detects specific nucleic acid (DNA or RNA) sequences, and is considered the gold standard for diagnosis of an active SARS-CoV-2 infection (Al-Muharraqi, 2020). The method exponentially amplifies nucleic acids during repeated thermal cycles, allowing detection of tiny quantities of the target virus in the sample.

Asymptomatic carriers may have high viral loads but fail to show signs or symptoms of COVID-19; or those with low viral loads may be in an early stage of infection and not yet developed clinical symptoms. Given the risk of releasing infectious viral particles during AGPs, we considered it necessary to identify individuals in either group. We required a test with high analytical sensitivity to detect samples containing very low viral loads (Saah and Hoover, 1997). Diagnostic sensitivity was also critical, so we could restrict entry to clinical areas to virus-free individuals only (Stites and Wilen, 2020). A high rate of false negatives would undermine this goal. Diagnostic specificity was of secondary importance, because any positive result would cause rescheduling of the individual’s clinical session, subject to a repeat test using standard NHS protocols.

Given these requirements, rapid PCR testing was identified as the most appropriate and accessible way to meet the need for both analytical and clinical diagnostic sensitivity (Dinnes et al., 2021). A novel mobile rapid PCR testing platform and reagents (Primerdesign™ q16 rapid PCR instrument and PROmate® COVID-19 direct assay) were used. This system offered high clinical diagnostic sensitivity and processing up to 14 samples per hour on each PCR instrument, allowing throughput to be scaled up by using multiple instruments in parallel (Reginald et al., 2021).

Process

The process in this service evaluation is outlined in Figure 1. A similar process was followed for staff testing, and carried out within the 72 h prior to the clinic.

Figure 1.

Evaluation of patient service. The goal of this service evaluation was to minimise the time from patient arrival at the Dental Hospital to start of treatment, provided that a negative PCR test result for SARS-CoV-2 was received.

The starting condition is the patient’s arrival, with subprocess steps being:

• • Obtain patient consent for undergoing a swab test

• • Take anterior nasal swab sample

• • Set up and run PCR test in laboratory

• • Analyse and approve result of PCR test

• • Communicate result to treating clinician

Once the clinician has a patient’s negative test result, the patient can be summoned for treatment. Those who test positive would be rescheduled, if treatment was not urgent, and offered a separate confirmatory nasopharyngeal swab test and advised to self-isolate in line with government guidance.

Our goal, therefore, was to reduce turnaround time for the overall process so that treatment could begin as early as possible after patient arrival. This was to be achieved using the PDSA cycle framework.

PDSA cycles (Plan, Do, Study, Act)

The PDSA cycle is a framework widely used in industry and increasingly applied to healthcare settings to drive continuous improvement of systems and processes (Taylor et al., 2014), with the following methodology:

Plan

A cross-functional team (hospital managers, administrative staff, clinicians, nurses, process engineers and laboratory staff) formulated a plan for how the process should work, including physical maps (e.g. floor plans), decision flow schemes and training guidelines (e.g. swabbing instructions).

Do

Iterative trials of the process were run (in whole and in part) to model the real conditions, capturing data on cycle time for each step. Relevant facts about aspects that worked well or caused problems were also recorded.

Study

Trial run data was reviewed daily. Process steps were adapted based on available facts and data, focusing on improvements with the greatest impact on cycle time and process repeatability.

Act

Useful learnings from daily reviews were captured and systematised. Problematic aspects and opportunities to further reduce cycle time were prioritised and new approaches planned (triggering the next Plan phase).

Results

Total turnaround time has two key phases (Figure 2(a)): the patient-facing sample collection subprocess, from patient arrival to sample receipt in the laboratory; and the sample processing subprocess, when the laboratory test is run, concluding when results are communicated. Figure 2(b) shows the relative contribution of each stage to total time for 31 runs over a 4-week period. In the following discussion, we consider individual subprocess steps and discuss key changes which drove the overall cycle time reduction.

Figure 2.

Processes, subprocesses and overall improvement of patient service. (a) Gantt chart for the process showing relationships between individual steps, and the ‘sample collection’ and ‘sample processing’ subprocesses. (b) Total time from patient entering the Dental Hospital to results received by hospital staff. In runs 11–18, no representative clinical swabbing pathway was trialled. In order to compare turnaround times, a nominal swabbing time of 21 min (the minimum demonstrated time before run 11) has been indicated on the graph.

‘Sample collection’ subprocess

This phase has two sub-steps (Figure 3(a)). The swabbing step involves obtaining consent and collecting swab samples, and timing depends on the number of samples per PCR test batch, with full batches of 14 maximising machine capacity. The second step was transport of the sample batch to the testing area.

Figure 3.

Analysis and improvement of sample collection subprocess. (a) Total time from patient arrival to samples received at the testing area. (b) Time for the swabbing step varies depending on the number of individuals swabbed per batch (indicated by column labels). As multiple swabbing streams ran in parallel, the graph is not representative of actual contact time between each swabber and patient. Dates highlighted in grey were part of a training period during which different members of staff were learning the process in each run; cycle time improvements were not necessarily expected in this period. (c) Changes implemented using the Plan, Do, See, Act (PDSA) framework to reduce cycle time of the sample collection subprocess, with the checkpoint pathway. (d) Changes implemented via the PDSA framework to reduce cycle time of the sample collection subprocess, with the point of care pathway.

For the swabbing phase, two pathways were trialled (Figure 3(b), showing swabbing time per person). The ‘checkpoint’ pathway trialled in runs 1–10 used a team of trained swabbers taking samples in a dedicated clinical area. The ‘point of care’ (PoC) pathway saw dentists and nurses take swab samples in their own clinics (runs 19–31).

The process changes identified and implemented through the PDSA cycles (Figure 3(c)) demonstrate the team collecting a full batch of 14 swab samples in 21 min, with confidence that further daily practice and embedding the system would make swabbing possible within 15 min.

While the PoC pathway included a longer swab time per patient (3.3 min vs. 2.7 min for checkpoint), this was offset by significant patient flow benefits: dedicated swabbing and patient waiting areas would not be needed, meaning spaces could be fully utilised, nor would a dedicated swab team be required. However, the pathway came with a higher initial investment of staff time including the training of the swabbing process to all dentists and nurses.

We therefore ran a ‘pressure test’ trial in Run 19, showing it was indeed possible to swab a batch of 14 samples within a 15-minute period. Subsequent runs were carried out with a different group of clinical staff each day, so a reduction in cycle time was not expected in this period. Instead, the runs were used to embed the system and refine the approach to training so that clinic staff could develop confidence in executing the process.

During this period, variations in staff arrival time led to notably extended swabbing times. Furthermore, each individual was being trained separately rather than within a group. Also, if clinic attendance was much above or below a PCR batch size of 14 samples, we either waited for enough to run a full batch, causing delays in total swabbing time, or ran a partial batch, causing reduced machine utilisation. These challenges were reviewed through the PDSA cycle to find swift resolutions (Figure 3(d)).

The sample testing location was a nearby external laboratory, requiring a sample transport time of approximately 7 minutes. A further improvement to end-to-end turnaround time would eliminate this transport time by implementing in-house testing within the dental clinics.

‘Sample processing’ subprocess

This consists of three key steps (Figure 4(a)). Received samples are processed and set up in the PCR reaction. The thermal cycler then runs the test, and an operator reads out the results. These are then communicated to a contact at the Dental Hospital, who notifies clinicians as to who can begin treatment. While the first two steps were consistently completed within 90 min with 1 week’s training, result communication was surprisingly the area for most improvement.

Figure 4.

Analysis and improvement of sample processing subprocesses. (a) Total time from samples received at the testing area to results reported to Dental Hospital staff, at which point clinicians can be notified as to which patients’ treatments can begin. (b) PCR setup time varies depending on number of samples per batch (indicated by column labels). (c) Changes implemented using the PDSA framework to reduce cycle time of the sample processing subprocess.

Figure 4(b) elucidates the incremental improvements in cycle time for PCR set-up accepting that the total step time is dependent on the number of samples included in the batch.

The key changes outlined in Figure 4(c) were implemented in the sample processing phase to reduce total time from over 200 min in early runs to 57 min per batch by run 22 (Figure 4(b)), a reduction of over 70%.

‘In-run’ calling can determine a result of negative/positive/repeat needed. Only in the case of a confirmed negative test result can treatment begin. In early trials, repeats to confirm results were originally run before reporting. As confidence in the system and operator skill level grew, it was possible to report findings immediately and only run repeats where necessary, based on end-of-run results (entry 3, Figure 4(c)).

Towards this end, a key step was agreeing that any patients with positive or inconclusive results needed to be rescheduled, hence ending the process flow. Any repeats needed for confidence in the diagnosis could then happen ‘outside the cycle’ and not on the critical path. Only patients who were confirmed negative at that time-critical decision point would proceed to treatment.

After implementing these three key changes, we demonstrate a cycle time of 57 min for sample processing in run 22 (Figure 4(a)). An ongoing goal for this service improvement programme is to transfer both technology and skills to maintain this rapid turnaround time when testing capability is brought in-house within the Dental Hospital.

Discussion

For successful implementation of a rapid diagnostic test in a clinical setting, it is necessary to plan the end-to-end process from patient arrival to result reporting. Here we demonstrate a significantly reduced end-to-end cycle time using rigorous attention to process analysis and a PDSA cycle improvement framework. The targeted continuous improvement methodology drove both a 65% reduction in turnaround time and built confidence in a scalable system which could be rolled out throughout the Hospital. A particular benefit of this approach was the speed of implementation of the system, achieving the turnaround time of 90 min within 1 week and end-to-end target cycle time within 3 weeks from a ‘cold start’. We believe it is possible to reach the target cycle time even faster in future efforts by using learnings documented through the PDSA improvement cycle. Given the time-critical nature of the COVID-19 pandemic, speed and ease of system setup is crucial to allow continuation of clinical operations. Furthermore, as the rapid-testing system only requires 14 samples in a single PCR batch, the waiting time per sample to start a testing run is significantly lower than high-throughput large batch size technologies (96 or 384 samples per batch).

Rapid testing facilitates effective continuation of critical clinical operations during the COVID-19 pandemic. This service improvement programme demonstrated the implementation of a fully integrated rapid-testing workflow within a very short timeframe. This paves the way for a significant increase in utilisation of clinical resources to return to pre-pandemic levels, and offers a model for analysing and improving usage of rapid testing in divergent clinic use cases.

ORCID iDs

Talia R Pettigrew https://orcid.org/0000-0003-2841-9914

Paul Ryan https://orcid.org/0000-0002-7091-9974

Mark J Coldwell https://orcid.org/0000-0002-6243-3886