Pooled Adaptive PCR Testing | Hacker Noon

Author profile picture

This article proposes adaptive pool testing which involves testing a combined sample from multiple people as a methodology to increase testing capacity for Sars-COv2 using PCR. Pool tests involve that if a pool is detected positive, each sample from the pool will be tested separately. Also, this article proposes an optimization algorithm for the pool size-determine the optimum number of samples that can be combined into a single test in order to obtain the best results with a minimum number of tests.

  • Romania and other countries use PCR testing for diagnosing COVID -19; this method offers the best detection rate compared with other testing methods but is expensive and is limited by availability of kitting tests and PCR testing machines; current capacity in Romania is around 4000 test per day;
  • Currently testing in Romania and other countries with limited testing capacity are conducted only for peoples with specific symptoms and for contacts of infected persons; A PCR test examines the presence of a unique genetic sequence of viruses in a sample taken from the patient and takes several hours.

Researchers at Technion – Israel Institute of Technology and Rambam Health Care Campus have successfully tested pooling methodology for COVID 19 that enables simultaneous testing of dozens of samples. “According to the new pooling approach that we have currently tested, molecular testing can be performed on a “combined sample,” taken from 32 or 64 patients. This way we can significantly accelerate the testing rate. Only in those rare cases, where the joint sample is found to be positive, will we conduct an individual test for each of the specific samples.”

According to Prof. Roy Kishony, head of the research group in the Faculty of Biology at Technion, “This is not a scientific breakthrough, but a demonstration of the effectivity of using the existing method and even the existing equipment to significantly increase the volume of samples tested per day. This is done by pooling multiple samples in a single test tube. Even when we conducted a joint examination of 64 samples in which only one was a positive carrier, the system identified that there was a positive sample. (https://www.technion.ac.il/en/2020/03/pooling-method-for-accelerated-testing-of-covid-19/)

Similar research was performed in Germany by Goethe-Universität Frankfurt am Main with combined sample size from 5 patients and results were 10% accurate.

(more details here: https://idw-online.de/en/news743899)

In mathematics, group testing was first studied by Robert Dorfman in 1943; is a procedure that breaks up the task of identifying certain objects into tests on groups of items, rather than on individual ones. This methodology was applied in epidemiology for other diseases like malaria and HIV for PCR testing; a related article for malaria:

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4301903/

Pooled adaptive PCR testing strategy based on prevalence estimation

Facts:

  • It’s possible to test a pool of up to 64 people for SARS-CoV-2 at once and the test will say whether at least one person in the pool is positive;
  • Based on previous testing results we can calculate the prevalence rate; for example in Romania (as of April 6, 2020): we have 3.864 detected cases and we performed 38.629 tests to detect these cases (only peoples with specific symptoms and contacts of infected persons were tested); in this case prevalence rate is around 10% (10% of total test are positive);

Proposed solution:

Develop a pooling testing protocol (algorithm) based on prevalence estimation; this algorithm will be adaptive and pool size will be changed frequently based on estimated prevalence rate and changes in prevalence rate in previous days.

There are two potential strategies:

  • One pool strategy in cases where prevalence rate is relatively high (between 1% and 30%); if the prevalence of positive samples is greater than 30% it is never worth pooling and the recommended strategy is individual tests; for the cases when prevalence rate is between 1% and 30% the objective is to determine the optimal pool size in order to obtain the minimum number of tests; strategy in this case is to test each pool of tests and if the results of a pool is positive to test individual each test from the pool;
  • Two pools strategy: second strategy proposed it is recommended for a prevalence rate below 1%; a huge decrease in number of tests required to test a large population can be obtained using two pools; in this case first round of testing will use large pools (between 32 and 64 samples) and a second round of testing will be performed for positive large pools with smaller pools; only for positive small pools will be performed individual testing.

Tests and pool calculator

Based on the above data I developed a calculator used to estimate the number of tests required based on prevalence rate and pool size; also, this calculator allow to determine the optimum pool size in order to use a minimum number of tests;

This calculator uses as input prevalence rate and pool size and the output is the estimated number of tests required using pool methodology; calculator was developed in Excel and is available for download. Also, this calculator determines the optimum pool size for a selected prevalence rate

Figure 1: Calculator

Proposed methodology for targeted tests

  • Calculate prevalence rate in testing region based on previous data:
  • Prevalence rate = number of positive tests divided by total number of tests performed; prevalence can be calculated at country level or region/state based
  • Prevalence rate can be calculated using data from the beginning of testing; maximum benefits can be obtained by calculating prevalence rate based on last three days test results; also better results can be obtained if we calculate based on region prevalence
  • Use Excel calculator to determine optimal pool size: enter number of cases and number of positive tests; prevalence is automatically calculated and also optimal pool size; you can use also a fixed pool size; calculator will show the number of test required for both situations (fixed and optimum pool size);
  • Perform pool testing and individual tests for samples from positive pools

Proposed methodology for large population tests

  • Calculate prevalence rate in testing region based on previous data:
  • Prevalence rate = number of positive tests divided by total population; prevalence can be calculated at country level or region/state based; adjust number by adding a coefficient for undetected case (asymptomatic cases)
  • Use Excel calculator to determine optimal pool size
  • Perform pool testing and individual tests for samples from positive pools

Advantages

Pooled adaptive PCR testing strategy based on prevalence estimation can offer a major increase in testing capacity with a big impact also in testing costs:

  • Reduce the number of tests required for targeted testing with up to 50%
  • Test large population with low prevalence rates using large pools (up to 64 tests); for example, you can test 100.000 persons with around 12.000 tests using a pool size of 30 at a prevalence rate of 0.25%.
  • This methodology can be used also for prevention (i.e. test all passengers from a plane with only few pools and several individual tests)
  • A “Two pools strategy” can be also developed; this strategy it is recommended for a prevalence rate below 1%; a huge decrease in number of tests required to test a large population can be obtained using two pools; in this case first round of testing will use large pools (between 32 and 64 samples) and a second round of testing will be performed for positive large pools with smaller pools (4-8 tests); only for positive small pools will be performed individual testing.

Below table is based on Worldometer data (https://www.worldometers.info/coronavirus/) as of April 8, 2020 show the optimum pool size for several countries and US states; also the table show the estimated number of tests required using adaptive pool methodology for the same number of patients and the gain in tests:

Figure 2: Individual testing versus adaptive pool methodology

  • The number of tests required can vary based on positive tests distribution (number can be higher or lower);

It is obvious that major gains can be obtained for lower prevalence rates:

Figure 3: Reduction rate using pool testing

Very good results can be obtained for prevalence rate less than 20%.

From a testing resources perspective with less than 2,000 tests we can test 10,000 persons (2% prevalence rate) which is a huge gain for testing capacity.

Tags

Join Hacker Noon

Create your free account to unlock your custom reading experience.

read original article here