The Logistics of Covid-19 Testing

Lehan Edirisinghe and Elle Trump July 28, 2020

Watch Facts Elle's

Key takeaways

- By July 2020, COVID-19 testing had expanded to millions of tests daily, yet some results took two weeks or longer to arrive. The delays undermined infection control because people with mild or no symptoms might continue spreading the virus while awaiting results.

“tests now take up to two weeks, or sometimes more, to come back.”

- Quest Diagnostics processed approximately 17% of US COVID-19 tests using 28 laboratories and a fleet of roughly 20 aircraft. With capacity for 130,000 tests daily, the company flew samples overnight from distant collection sites to centralized facilities, particularly serving the American South during its outbreak surge.

- Competition among thousands of private laboratories made critical testing supplies more expensive and difficult to secure. Each test required approximately 20 consumable items, including swabs, tubes, and reagents, meaning a shortage of any single component could halt processing; centralized government purchasing was presented as a potential cost-saving alternative.

“a bottleneck in any one of those twenty supply chains stops testing.”

- Quest received the first US FDA emergency authorization for pooled COVID-19 testing, allowing four samples to be screened together. At an assumed 8% positivity rate, the episode estimated that 100 people could be screened using just 57 processing runs, with individual retesting required only for positive pools.

“this pooled testing technique used 57 processing slots to generate 100 results”

- Pooled testing becomes less economical as infection rates increase, making uniform pool sizes unsuitable for every region. Low-positivity populations can support larger batches of eight, 12, or 16 samples, while high-positivity areas require smaller pools or individual testing to avoid excessive retesting.

- Wastewater surveillance could identify COVID-19 outbreaks approximately one week before conventional clinical testing. Measuring viral genetic material at sewage treatment plants offered community-wide information with fewer clinical sampling biases, potentially allowing governments to respond before outbreaks became clearly visible in reported cases.

“sewage testing indicates spikes about a week before clinical testing”

- Berkeley researchers proposed using machine learning to determine testing pool sizes according to geographic and occupational infection risks. Their model would group low-risk samples together while testing higher-risk samples individually or in smaller batches, with a hypothetical testing cost of just $3–$5 per person.

“Hypothetically, a pooled testing system designed by machine learning could lower the cost of a test to just $3 to $5”

- The approaching 2020 influenza season threatened to overwhelm testing systems even without an increase in COVID-19 infections. Similar symptoms would send more people for testing between October and December, while laboratories were consuming available supplies immediately rather than building reserves for winter.

“All the supplies for testing produced now are being used now, there is no stash being built up”

All points

2047 words

- COVID-19's ability to spread without obvious symptoms made widespread testing essential. The virus could resemble a common cold, while some infected people experienced no symptoms at all. Even a relatively low mortality rate could produce hundreds of thousands of deaths when infections became widespread.

- Testing was presented as one of the most important tools for controlling the pandemic. Until effective treatments, vaccines, or herd immunity became available, identifying infected people quickly was considered critical to reducing transmission and deaths.

- COVID-19 testing expanded from hundreds to millions of daily tests within six months. The first genetic sequence of the virus was published on January 5, 2020, and researchers released testing protocols within days. Global testing increased from hundreds of people daily in late January to thousands in February, hundreds of thousands in March, and millions by July.

- The central problem was no longer simply how many tests could be performed, but how quickly useful results could be delivered. By July 2020, testing volumes had grown enormously, yet delays and supply shortages prevented the system from controlling infections as effectively as it could.

- Standard COVID-19 laboratory testing depended on polymerase chain reaction technology. PCR amplifies genetic material so that it can be detected and studied. For an RNA virus such as SARS-CoV-2, reverse transcription is used as part of the testing process.

- PCR testing required specialized equipment capable of rapidly changing temperatures. The process involves denaturation, annealing, and elongation, each requiring specific temperatures controlled by a machine called a thermal cycler.

- Laboratory equipment costs made decentralized PCR testing difficult. Basic thermal cyclers cost thousands of dollars, while high-throughput machines commonly used for COVID-19 testing cost many tens of thousands. Many also required trained technicians, making installation in every doctor's office impractical.

- Rapid point-of-care tests offered convenience but involved trade-offs. The alternatives available at the time were generally described as less accurate, more expensive, and capable of processing fewer samples simultaneously. Collecting samples was therefore much easier than processing large numbers of them.

- Quest Diagnostics processed approximately 17% of all COVID-19 tests conducted in the United States. At the time of the episode in July 2020, the company had a reported testing capacity of 130,000 samples per day across 28 laboratories nationwide.

- Quest Diagnostics depended on an extensive transportation network to connect collection sites with centralized laboratories. Large parts of the United States had no nearby Quest laboratory, making a combination of road transportation and aircraft necessary to move samples quickly.

- Quest operated a fleet of approximately 20 aircraft dedicated to overnight sample transportation. Most were based in Reading, Pennsylvania, and Lawrenceville, Georgia. Aircraft generally began operating around 7 p.m., after samples had been collected during the day.

- One Quest aircraft visited seven collection destinations during a single overnight journey on July 16, 2020. The Pilatus PC-12 registered N338QD departed Reading, Pennsylvania, at 8:45 p.m., stopping at Elmira, Syracuse, Rochester, Niagara Falls, Allegheny County, Manassas, and Baltimore before returning to Reading by 3:30 a.m.

- Samples arriving in Baltimore were transferred by road to a nearby laboratory. The laboratory was located a few miles from Baltimore/Washington International Airport, illustrating how air and ground logistics worked together to deliver samples to processing facilities.

- Another Quest aircraft completed a complicated overnight circuit through the southeastern United States. Its route began in Lawrenceville, Georgia, and included Jacksonville, Gainesville, Tampa, Orlando, Valdosta, a return to Lawrenceville, then Roanoke, Manassas, Greensboro, and finally Lawrenceville again.

- Quest concentrated much of its aircraft activity in the American South during the observed period. The region was experiencing a surge in COVID-19 cases while having relatively few Quest laboratories. Aircraft allowed the company to serve geographically dispersed collection sites without operating laboratories in every city.

- Centralized laboratories created economies of scale but introduced logistical dependence. Large processing facilities could handle samples more efficiently than numerous small laboratories, but that model required reliable transportation, equipment, supplies, and coordination across distant locations.

- Approximately 20 different consumable supplies were needed to perform a COVID-19 test. These included nasal swabs, sample transport tubes, reagents, and other materials. A shortage of any one essential item could stop the entire testing process, even when every other component was available.

- Competition for scarce testing supplies became a major economic obstacle. In the United States, thousands of private testing providers independently competed to purchase equipment and consumables, giving suppliers numerous alternative customers.

- Testing suppliers gained pricing power because demand greatly exceeded available supply. A manufacturer could reject one laboratory's offer and sell to another provider willing to pay more. Laboratories urgently needing supplies had incentives to accept higher prices to maintain operations.

- The episode compared testing procurement with the concentrated British grocery market. It cited Tesco, Sainsbury's, Asda, and Morrisons as controlling approximately 75% of UK grocery sales. Their collective purchasing power gave suppliers such as dairy farmers fewer alternative buyers, helping explain how concentrated buyers can force prices downward.

- Centralized government purchasing was proposed as a way to lower testing supply costs. If the US federal government became the sole or overwhelmingly dominant buyer, manufacturers would have fewer alternative customers and less bargaining power. The narrator acknowledged that centralized allocation might be less efficient in other respects.

- Even major laboratories struggled to secure supplies against international competition. Quest Diagnostics and LabCorp had greater purchasing power than smaller providers and processed more tests daily than many countries, yet still competed with buyers worldwide for limited materials.

- Testing providers accumulated more samples than they could process. By July 2020, major laboratories were collecting samples faster than their processing capacity could handle, creating a growing backlog.

- Some COVID-19 test results took two weeks or longer to arrive. Such delays sharply reduced the usefulness of testing because asymptomatic or mildly symptomatic individuals might not isolate throughout the waiting period. A result delivered after an infectious period offered much less opportunity to prevent transmission.

- Quest Diagnostics received the first US Food and Drug Administration emergency use authorization for pooled COVID-19 testing. The technique was intended to increase laboratory capacity without requiring equivalent increases in equipment or supplies.

- Pooled testing combined as many as four individual samples into a single laboratory test. If the combined sample tested negative, all four individuals were considered negative. If it tested positive, each sample was tested separately to identify which individual or individuals were infected.

- The episode calculated that pooling could nearly double processing capacity at an 8% positivity rate. Its illustration grouped 100 samples into 25 pools of four, estimated that 17 pools would be negative and eight would require individual retesting, and calculated 57 processing runs instead of 100. This was the video's simplified estimate, not an exact probability calculation.

- Pooled testing became less efficient as the proportion of positive samples increased. The episode's example for a 20% positivity rate estimated that 20 of 25 pools would require retesting, resulting in 105 processing runs for 100 people. That calculation was also a simplified illustration rather than a statistically exact expectation.

- High-infection hotspots were less suitable for conventional pooled testing. When many samples were positive, combined batches frequently required individual retesting. The additional processing could eliminate the efficiency advantage.

- Low-positivity areas could potentially use larger sample pools. Rather than combining only four specimens, laboratories could consider groups of eight, 12, 16, or more when positive samples were sufficiently uncommon.

- Expanding testing could improve the efficiency of pooling itself. Testing a wider population, including more low-risk individuals, could reduce the percentage of positive samples. Lower positivity would permit larger pools, potentially creating a reinforcing cycle of greater capacity and more widespread testing.

- PCR case reporting was inherently delayed relative to actual transmission. After exposure, individuals often needed several days to develop symptoms, additional time to seek testing, and still more time to receive results. Consequently, significant changes in transmission might take weeks to become clearly visible in clinical testing data.

- Wastewater testing emerged as a potential early-warning system for COVID-19 outbreaks. Researchers could test sewage for the virus's genetic signature to detect transmission across an entire community rather than relying exclusively on individual clinical tests.

- Sewage infrastructure naturally aggregated biological material from large populations. Samples collected at relatively few wastewater treatment facilities could provide information about the areas they served, reducing dependence on large numbers of individual clinical samples.

- Wastewater surveillance could help overcome sampling biases in clinical testing. Clinical data depended on who sought testing, who could access it, and how quickly results were processed. Sewage analysis offered a broader population-level signal of whether transmission was occurring and its approximate scale.

- Research indicated that wastewater signals could reveal infection surges approximately one week before clinical testing data. This advance warning could help governments identify worsening outbreaks before conventional case reports fully reflected the increase.

- Earlier detection could give governments more time to introduce restrictions. Wastewater surveillance was presented as a way to identify emerging transmission increases while preventive measures could still have a greater effect.

- Researchers at Berkeley proposed combining wastewater surveillance with pooled clinical testing. Their approach would use information about infection prevalence and individual risk to determine how samples should be grouped, rather than applying the same pool size everywhere.

- Geographic infection rates could determine which samples should be tested individually. The episode gave a hypothetical 25% positivity probability for a sample from Florida, making individual testing preferable, compared with a 1% probability for a sample from North Dakota, making larger pools more attractive.

- Occupation was identified as another useful indicator of infection risk. Samples from healthcare and retail workers could be assigned higher expected positivity rates than samples from businesspeople working at home. This information could influence how laboratories grouped specimens.

- The Berkeley proposal used machine learning to adjust pool sizes according to predicted risk. Low-risk samples would enter larger groups, while high-risk samples would enter smaller groups or be processed individually. The objective was to maximize capacity while avoiding the inefficiencies of pooling samples with high probabilities of testing positive.

- Risk-based pooling was intended to work across regions with very different infection rates. Instead of abandoning pooling whenever positivity increased, laboratories could selectively reduce pool sizes for higher-risk samples while retaining the efficiency gains available from lower-risk specimens.

- The episode argued that frequent testing could become dramatically more resource-efficient through advanced pooling. It presented a theoretical claim that testing a population daily might require only twice the resources of testing it monthly under a sufficiently optimized system. This was a proposed possibility, not a demonstrated operating result.

- Machine-learning-designed pooling was projected to reduce testing costs to approximately $3–$5 per person. The figure was hypothetical and depended on achieving substantially greater efficiency. Lower costs could make repeated population-wide testing more practical.

- Autumn 2020 was identified as a predictable period of exceptional pressure on testing systems. Seasonal influenza infections typically began increasing substantially between October and December, creating a new challenge for COVID-19 diagnosis.

- Influenza and COVID-19 shared many symptoms, making laboratory testing necessary to distinguish them. People experiencing ordinary seasonal flu symptoms would have reasons to request COVID-19 tests, even when they were not infected with the coronavirus.

- Demand for COVID-19 tests could therefore increase even if coronavirus infections declined. A substantial influenza season could generate large numbers of additional people seeking to rule out COVID-19, increasing laboratory workloads independently of actual coronavirus prevalence.

- Testing systems were not building meaningful reserves of supplies ahead of winter. According to the episode, testing materials being manufactured in July 2020 were being consumed immediately rather than stockpiled for the expected autumn demand surge.

- Laboratories and healthcare systems faced simultaneous shortages of materials, processing capacity, and time. Increasing the number of sample collection sites alone could not resolve the problem. Supplies, laboratory throughput, transport networks, and turnaround times all needed to improve together.

- The episode's central operational argument was that smarter use of existing resources could matter as much as increasing laboratory capacity. Combining centralized laboratory logistics, risk-adjusted sample pooling, and wastewater surveillance was presented as a way to process more tests, identify outbreaks earlier, and improve the public-health value of testing before the anticipated winter surge.