Communicating the Imperfect Diagnostic Accuracy of COVID-19 Rapid Antigen Self-Tests: An Online Randomized Experiment

Abstract

Objective To investigate the potential impacts of optimizing coronavirus disease 2019 (COVID-19) rapid antigen test (RAT) self-testing diagnostic accuracy information. Design Online randomized experiment using hypothetical scenarios: in scenarios 1 to 3 (RAT result positive), the posttest probability was considered to be very high (likely true positives), and in scenarios 4 and 5 (RAT result negative), the posttest probability was considered to be moderately high (likely false negatives). Setting December 12 to 22, 2022, during the mixed-variant Omicron wave in Australia. Participants Australian adults. Intervention: diagnostic accuracy of a COVID-19 self-RAT presented in a health literacy-sensitive way; usual care: diagnostic accuracy information provided by the manufacturer; control: no diagnostic accuracy information. Main Outcome Measure Intention to self-isolate. Results A total of 226 participants were randomized (control n = 75, usual care n = 76, intervention n = 75). More participants in the intervention group correctly interpreted the meaning of the diagnostic accuracy information ( P = 0.08 for understanding sensitivity, P < 0.001 for understanding specificity). The proportion who would self-isolate was similar across scenarios 1 to 3 (likely true positives). The proportion was higher in the intervention group than in the control for scenarios 4 and 5 (likely false negatives). These differences were not statistically significant. The largest potential effect was seen in scenario 5 (dinner party with confirmed cases, the person has symptoms, negative self-RAT result), with 63% of the intervention group and 49% of the control group indicating they would self-isolate (absolute difference 13.3%, 95% confidence interval: −2% to 30%, P = 0.10). Conclusion Health literacy sensitive formatting supported participant understanding and recall of diagnostic accuracy information. This may increase community intentions to self-isolate when there is a likely false-negative self-RAT result. Trial registration: Australia New Zealand Clinical Trial Registry (ACTRN12622001517763) Highlights Community-based diagnostic accuracy studies of COVID-19 self-RATs indicate substantially lower sensitivity (and higher risk of false-negative results) than the manufacturer-supplied information on most government public Web sites. This online randomized study found that a health literacy–sensitive presentation of the imperfect diagnostic accuracy COVID-19 self-RATs supported participant understanding and recall of diagnostic accuracy information. Health literacy–sensitive presentation may increase community intentions to self-isolate after a negative test result where the posttest probability is still moderately high (i.e., likely false-negative result). To prevent the onward spread of infection, efforts to improve communication about the high risk of false-negative results from COVID-19 self-RATs are urgently needed.