Testing and test data, what it can and cannot tell us

Testing is an essential component of a comprehensive response to the COVID-19 pandemic, and one of the four components of the Resolve to Save Lives-recommended Box It In strategy. There are several factors that should be considered in order to implement and monitor a successful testing strategy. This includes the types of tests available, when they should be used, metrics that can be useful to monitor testing, and additional contextual considerations related to testing and test results. In this insight, we review some of these key questions around testing for COVID-19.

1. What can COVID-19 PCR testing tell individuals?

   The two most commonly discussed types of COVID-19 tests are RT-PCR tests and antibody tests. From the perspective of an individual taking the test, the RT-PCR test can tell you whether you have COVID-19 now; antibody tests may tell you whether you have had a COVID-19 infection in the past. Although this seems simple, interpreting COVID-19 test results can be more complicated. The RT-PCR test looks for viral RNA (or viral genetic material), generally in the back of the nasal cavity. Here is one simple table on how to interpret COVID-19 test results.

   Positive RT-PCR test results: The chance of a false positive result is low and people with a positive test result generally have an active infection. However, it is possible to detect viral particles despite no longer being infectious or sick with COVID-19. South Korea recently found that people who retested as positive after testing negative were not experiencing reactivations or reinfections of COVID-19. Instead, the test was likely picking up viral RNA from dead viruses.

   Negative RT-PCR test results: According to the CDC, if someone has a negative test, they probably are not infected with COVID-19 at the time they are tested. However, false negatives are not uncommon. Early reports from China indicated a sensitivity (probability of a positive test result given someone has the disease) of around 70%, or a 30% chance of a false negative result. Data on the sensitivity of commonly available tests is very limited, but there is some evidence that not all perform equally. Testing results can also depend on sample collection and handling and timing since infection. Researchers from Johns Hopkins University found that the likelihood of an infected person testing negative went from 100% four days before symptom onset, to 38% the day symptoms started, to 20% three days later and then began rising again. Negative test results should be evaluated in the context of the likelihood of infection, including symptoms, prevalence of COVID-19 in the area and whether or not an individual has a high risk of exposure.

 

2. What can COVID-19 PCR testing tell communities and larger populations?

   At the population level, metrics around testing are critical to understand both the epidemic and our response. Testing is a cornerstone of the Box it In approach, which also includes isolating all infected people, tracing their contacts and quarantining the contacts for 14 days after last exposure. Without widespread testing, it will be difficult to identify those that are infected in time to prevent spread of the disease. Information on testing is also critical to understand the extent of the epidemic. As noted in a previous data insight, trends in daily case counts or new cases per population are critical indicators for assessing the decline or potential resurgence of COVID-19. However, these data depend on the number of tests performed: as testing goes up, case counts may also go up, reflecting the fact that more cases are being found and not necessarily an increase in the rate of infections overall.

   Although widespread testing is critical, it needs to be targeted towards the right people and result in the right action. A number of groups, including RTSL, have released lists of priority groups for testing. CDC has identified the highest priority groups as people hospitalized with symptoms; healthcare workers, congregate facility workers and first responders with symptoms; and congregate facility residents with symptoms and those without symptoms whenever there a case of COVID-19 in the facility. Other high priority groups, identified by CDC and others, include persons with symptoms who are not high risk, contacts of cases, and residents and workers in congregate facilities. However, as employers and others start to screen low-risk populations for COVID-19, the number of tests may go up without necessarily improving the detection of COVID-19 in the population. Ideally, each jurisdiction would know and track the proportion of priority people being tested (e.g., all patients admitted to the hospital, all symptomatic people, all asymptomatic people living or working in congregate facilities with cases of COVID-19, all contacts). In practice, this is difficult.

   Broadly, two different indicators have been used to measure whether the amount of PCR testing in a community is appropriate: tests per population and test positivity (that is, the proportion of tests with a positive result). Both indicators measure the quantity of tests without necessarily indicating whether tests are targeted appropriately and if appropriate action is taken with the results. When calculating these indicators, results from antibody testing and PCR testing should not be combined, as they measure separate things (current infections vs. past infections). A different type of indicator measures if testing is happening quickly enough to prevent transmission as part of a Box it In strategy.

   • Test positivity
     • Why: Lower test positivity can be a sign that you are testing widely and catching most people with the virus. Countries that have done a good job controlling their epidemics such as Taiwan, Republic of Korea, Australia and New Zealand have test positivity less than 1.5%. Places with severe epidemics, such as New York City, had a very high test positivity at the peak of the epidemic (Figure 1).
     • Who uses it: The CDC includes declining rates of test positivity in its gating criteria for reopening.
     • Recommended ranges: CDC recommends it should be <10%.  Other groups have recommended <3% and the best-performing programs have positivity rates of 1.5% or less.
     • Challenges:
       • Should be measured in terms of people, not tests as some people may be tested multiple times.
       • No guarantee that you are testing the right people.
     • Tests per population:
       • Why: This indicator is most useful when comparing number of actual tests conducted against a target, potentially one that is based on an estimate of the number of people in priority groups. It could also be a proxy for testing capacity.
       • Who uses it: The recent US Health and Human Services strategic testing plan provides targets for number of tests per 1,000 population per month.
       • Recommended ranges: RTSL estimates that between 350 to 700 thousand tests per day would be needed in the United States to reach the highest priority groups. That is 1.5 per 1,000 per day. An action plan from the Rockefeller Foundation recommends 12 per 1,000 per day.
       • Challenges:
         • The difficulty of this metric is to define what the target should be. Testing targets are often based on the size of the epidemic and are difficult to standardize.
         • It’s important to ensure that the way the indicator is defined matches the target (people tested vs. those targeted).
         • No guarantee you are reaching the people you need to reach.
   • Time from symptom onset to positive test result
     • Why: It is critical that people with symptoms of COVID-19 are tested soon after symptom onset and receive their results in a timely manner. This ensures that they know to stay isolated and that their contacts can be elicited and traced before there is an opportunity for onward spread. This can be a single indicator or two sub-indicators can be measured instead:
       • Time from symptom onset to test
       • Time from test order to result

   Related indicators that measure how well the Box it In system works after the test result comes back include time from test result to isolation and time from test result to contact notification.

   • Who uses it: Timeliness indicators are used by both the CDC and the state of Utah among others to assess the robustness of testing programs. Both assess the indicator of time from test order to result.
   • Recommended ranges: For states to get to phase three as defined by the CDC, the time from test to result should be less than two days on average.  Utah aims for <1.5 days.  Overall, from symptom onset to result should be <2.5 to three days given that the time from symptom onset in a case to symptom onset in an infected contact is six days.
   • Challenges:
     • Ensuring that time of symptom onset is correctly identified and recorded to be incorporated into a reporting system will be challenging.
     • Test processing times may vary widely by site (clinic, hospital, etc.) and may need to be stratified in order to take corrective action.

   Figure 1: Percent of people with positive results by date, New York City

   

   Source: NYC

 

3. What does SARS-CoV-2 antibody testing tell individuals?

   As mentioned earlier, antibody tests check for infection in the past. They do this by looking for evidence of the body’s immune system response to an infection. For SARS-CoV-2, evidence of the body’s immune system response can appear as early as one week after infection, but generally takes closer to two weeks. To the best of our current knowledge, once positive, these test results can persist for at least months, although it is too early to assess the maximum duration.

   For several reasons, there are limitations on the accuracy of current antibody testing. The most valid and reliable antibody test, a serologic test, is performed on a blood draw and tested in a laboratory. Other faster “on-the-spot” tests have been developed for use outside of this setting, however their accuracy is still being reviewed. Antibody tests should not be done to diagnose disease nor to make decisions about “immunity,” but they may serve some role in clinical and public health scenarios.

   Positive test results: A positive test for antibodies for the SARS-CoV-2 virus generally means that a person has antibodies that have formed to the virus as a result of prior infection. There is a chance that the test could falsely measure antibodies to other closely related coronaviruses such as those that cause the common cold. There is not currently enough information to say reliably that the presence of the body’s immune system response is equivalent to being protected from reinfection in the future, or how long this protection could last. Therefore, testing positive does not mean that a person should stop taking precautions to protect themselves and others from infection. There is some evidence to suggest that if serologic testing is going to be used in a setting where it may be relevant, two separate independent tests should be performed in sequence to be able to best interpret a positive result.

   Negative test results: A negative test for SARS-CoV-2 antibodies means that you mostly likely have not had COVID-19 in the past, however it does not rule out that you may have a recent infection as there may not have been adequate time for antibodies to develop.

   Special considerations: Some instances have emerged in which antibody testing can be useful as a diagnostic tool. In cases where patients present late, after one to two weeks of illness, PCR tests may be less useful in detecting virus, whereas antibody tests may be increasingly useful as antibody levels rise in the second and third weeks. A second consideration is to assist in the evaluation of patients with late complications of COVID-19 without a prior known diagnosis. Antibody testing has been important in the identification of multisystem inflammatory syndrome in children and its association with COVID-19.

 

4. What does SARS-CoV-2 antibody testing tell communities?

   Focusing more on communities rather than individual patients, antibody testing can have a role in a number of ways. It may be a key component of surveillance efforts to provide a more comprehensive picture of the local epidemic, and a critical component identifying antibodies in recovered plasma donors (convalescent plasma) for research into using antibodies from recovered persons to use in the treatment of the severely ill.

   Serologic surveillance: Relying on PCR testing alone to understand the extent to which SARS-CoV-2 affected a population would not be adequate. We now know that likely a third, but up to half, of all infected persons show no symptoms during their active infection, and in most cases would not have qualified for PCR testing. Others with mild manageable symptoms may have elected to, or been counseled to avoid testing so as to not stress a stretched health care system. Public health experts, therefore, will need to look back with antibody testing to determine how the first wave of the pandemic affected communities, and how to plan moving forward. In the US, local efforts have already been underway in New York State, Chelsea, Massachusetts and elsewhere. On the national level in the US, the CDC currently has a COVID-19 Serology Surveillance Strategy to better understand how many infections have occurred. This could provide a more complete picture of the impact of COVID-19 and can assist in planning control measures. Similar national efforts to expand serologic surveillance are underway in the UK, Germany, and elsewhere. Special populations such as health care personnel are also being considered, such as in Austria, where seroprevalence in the occupational setting is under study.
    

5. What are the main challenges and limitations of antibody testing?

   False-positive results: When a disease has a low prevalence and has not affected a large number of people in a community, the chance that any single positive test result is accurate is lower. This is further affected by inherent inaccuracies in testing. In a scenario where 5% of a community of 100 are affected by a disease (5 true cases), and a test has 95% specificity for what it is designed to detect, the test result would be positive for 10 people—five true positives and five false positives. That means that half of the persons testing positive have not actually had the disease. A clinical or public health expert who is well versed in this type of limitations should be involved in any interpretation of a positive test antibody result. Cross-reactivity, or having antibodies to another similar common-cold coronavirus, further complicates antibody testing and interpretation of positive results.

   Protective Immunity: Additional research is necessary to state confidently that the presence of antibodies to SARS-CoV-2 indicate that a person is immune or protected from reinfection. Currently, there is not enough information available to make such a statement, and studies in this area are ongoing. Although the presence of antibodies confers protective immunity for many viruses, it does not do so for all. If and when there is data to support protective immunity against COVID-19 after initial infection, additional research is necessary to determine how long this would last.

   Immunity certificates: We examined immunity certificates and the many challenges around them in a previous issue, which can be accessed here.
    

6. What are some testing considerations around the world?

   It is recognized that widespread testing is foundational to COVID-19 pandemic control efforts all over the world. In sub-Saharan Africa (SSA), the Africa CDC has announced plans to significantly increase testing coverage as part of a ‘Trace, Test, & Track’ initiative. Within SSA, approaches to testing vary by country, from a testing and tracing approach (Ghana) to a focus on areas with the biggest outbreaks (Nigeria) to a more widespread testing approach (South Africa). But no matter the strategy, successful testing strategies in SSA will need to both accomplish the herculean task of building the capacity to sufficiently track and contain SARS-CoV-2 and address potentially significant social barriers to testing.

   Infectious disease control efforts have been challenged by a deep-seated stigma against people or populations who are perceived to pose an infectious risk. There are examples of this from many outbreaks, including Ebola, HIV, influenza, polio, and SARS. The situation in Kenya can serve as a case study in the potentially serious economic, social, and mental health consequences of getting tested for or diagnosed with SARS-CoV-2 infection. In Nairobi, voluntary public testing for SARS-CoV-2 has been made available, and there are reports of employers rejecting potential employees who don’t present their test results. On the other hand, getting tested for SARS-CoV-2 may have its own consequences. Kenya Ministry of Health guidelines mandate facility-based isolation for those diagnosed with COVID-19, as well as potential quarantine for case contacts. Such measures may have significant economic costs for those who enter isolation or quarantine, both in terms of lost wages and facility fees. Patients who have recovered from COVID-19 and those released from isolation/quarantine may not be accepted by their communities. Given this, those with mild or no symptoms may be particularly deterred from being tested. Relatedly, there have been instances of discrimination, in-person and online harassment, and/or physical violence against healthcare workers in SSA who may have contact with COVID-19 patients. In Ivory Coast, a SARS-CoV-2 testing center was destroyed due to fear that the virus could spread from the center to people’s homes.

   Stigma against those with, or perceived to have, COVID-19 is not unique to SSA. There have been numerous reports from around the world of discrimination against groups perceived to carry SARS-CoV-2, of recovered COVID-19 patients being ostracized, and of healthcare workers being accused of spreading SARS-CoV-2. The consequences of fear and stigma can be magnified in settings where there is widespread misinformation about SARS-CoV-2, when high population density and poverty preclude physical distancing and basic hygiene measures, where there are limited public health and medical resources, and if livelihoods are threatened in a largely cash economy in which many people work in the informal sector. The World Health Organization (WHO) has outlined ways to prevent and address stigma related to COVID-19, and much can be learned from past experiences with other epidemics. Keys to combating stigma include assessing and addressing what people know and don’t know about COVID-19, partnering with the community and local leaders, and implementing practical interventions and skill-based trainings to reduce stigma. Without addressing social barriers to testing, even the most well-resourced and best-designed testing strategies will fall short.