AGA recently released a new guidance document: AGA Institute Rapid Review and Recommendations on the Role of Pre-Procedure SARS-CoV2 Testing and Endoscopy. Our guidance is focused on three key factors: local prevalence, diagnostic test performance data and available personal protective equipment (PPE).
To aid your decision making, we have developed a tool that combines local prevalence information with diagnostic test performance data to calculate the proportion of true versus false positives and negatives to help endoscopy centers understand the downstream consequences of implementing a pre-procedure testing strategy.
Follow the step-by-step instructions and videos below to use the online tool. Refer to AGA’s complete recommendations document for help interpreting your results.
Step 1: Calculate the prevalence of asymptomatic infection
There are several approaches that can be used, and they are listed in order of preference, acknowledging the limitations of each approach.
Approach 1. Use locally available data from health systems (near your endoscopy area) that have been conducting nucleic acid tests in asymptomatic individuals in outpatient settings. For example, if your health system has employed a pre-testing strategy for asymptomatic individuals undergoing elective surgeries or for outpatient women prior to childbirth, this information on the rate of positive and negative tests may be generalizable to your endoscopy setting. This is likely the most accurate estimation of the prevalence of asymptomatic individuals in your area.
Approach 2. Use locally available data from state public health departments that can provide information on the prevalence of positive tests among asymptomatic individuals. This is likely not going to be reported.
Approach 3. Use publicly available data about the state or county through the CDC website, The COVID Tracking Project or the COVID-19 Dashboard by the Center for Systems Science and Engineering at Johns Hopkins University. These websites provide the proportion of positive tests in an area. This data can be used to calculate an estimate of the prevalence of infection among asymptomatic individuals.
Watch video tutorial on calculating state prevalence through The COVID Tracking Project.
Watch video tutorial on calculating county prevalence through Johns Hopkins University data.
Unpublished data and the available models suggest that for each diagnosed symptomatic patient there may be 10 asymptomatic or undiagnosed individuals in the area. Using the number of diagnosed cases over the past two weeks and the publicly available data about population count, or census data, in your area, you can calculate the prevalence of asymptomatic infection in your area. The use of the proportion of positive tests alone is likely an overestimate of the prevalence of infection in asymptomatic individuals and should not be used to assess the prevalence.
Watch video tutorial on calculating population in your area through the U.S. Census Bureau.
Case example of how to calculate state prevalence of asymptomatic population in the state of Missouri using the COVID Tracking Project:
1. Visit the COVID Tracking Project website at https://covidtracking.com.
2. Look up your state. In the example illustrated below, Missouri was selected.
3. Look up the cumulative number of cases as of Aug. 7.
4. Look up the cumulative number of cases as of July 25 (14 days ago).
5. Subtract “4” – “3” (i.e. cases from Aug. 7 – July 25) = number of new cases in the past 14 days.
6. Adjust for asymptomatic population by multiplying the number of new cases in the past 14 days x 10.
7. Go to U.S. Census Bureau website at https://www.census.gov/quickfacts/ and look up estimated state population.
8. Divide numbers above: “6” / “7” (asymptomatic population / state population) = estimated prevalence in asymptomatic population.
Step 2: Determine the test characteristics of your locally available test
If you are using a commercially available test or an institutional laboratory-derived test, find out the diagnostic test accuracy of the specific test used in your individual setting.
In the tool, the diagnostic test characteristics are defaulted to the pooled sensitivity and specificity derived from the meta-analysis which encompasses data from commercially available U.S. tests (see guideline document). Of note, we use the lower end of the 95% confidence interval as a better estimate of the performance of the test as testing accuracy is likely to be lower for asymptomatic individuals (as compared to symptomatic individuals).
Step 3: Use the interactive tool to determine the false negative rate and false positive rate in your local setting
Based on assumptions of prevalence and test accuracy, the online tool will provide information on the rate of true positives, false positives, true negatives and false negatives in your local setting. The tool displays the results of testing a hypothetical sample of 1,000 individuals and the expected number of individuals who test positive (true positives and false positives) and the expected number of individuals who test negative (true negatives and false negatives). See guideline for how to use this information to inform decision-making regarding pre-testing before endoscopy. Finally, as local prevalence may change over time and require reassessment of the implemented strategy, periodic re-evaluation of prevalence to inform local endoscopy center practices will be necessary.
