Table 4. Pooled estimates of the proportion of asymptomatic infection at initial testing for COVID-19 and proportion asymptomatic at the end of follow-up.
| Author | Proportion of Asymptomatic Infection at Initial Testing | Asymptomatic Infection Throughout Follow-up | ||||
|---|---|---|---|---|---|---|
| Asymptomatic COVID-19 Positive / Total COVID-19 Positive | Proportion at Testing (95% CI) |
I2 | Remained Asymptomatic Through Follow-up / Total COVID-19 Positive | Proportion Asymptomatic at End of Follow-up (95% CI) |
I2 | |
| Obstetric Patients Presenting to Hospitals | ||||||
| Ochiai, D. | 2/2 | 100.0% (15.8% to 100.0%) | — | 2/2 | 100.0% (15.8% to 100.0%) | — |
| Goldfarb, I. | 9/20 | 45.0% (23.1% to 68.5%) | — | 9/20 | 45.0% (23.1% to 68.5%) | — |
| Andrikopolou, M. | 63/75 | 84.0% (73.7% to 91.4%) | — | 46/75 | 61.3% (49.4% to 72.4%) | — |
| London, V. | 10/10 | 100.0% (69.2% to 100.0%) | — | — | — | — |
| Bianco, A. | 24/24 | 100.0% (85.7% to 100.0%) | — | — | — | — |
| Pooled estimate | 108/131 | 95.1% (45.1% to 99.8) | 92% | 74/97 | 58.8% (48.8% to 68.1%) | 0% |
| Nursing Home Residents | ||||||
| Patel, M. | 14/27 | 51.9% (31.9% to 71.3%) | — | 13/27 | 48.1% (28.7% to 68.1%) | — |
| Dora, A. | 14/19 | 73.6% (48.8% to 90.8%) | — | 6/19 | 31.6% (12.61% to 56.6%) | — |
| Graham, N. | 54/126 | 42.9% (34.1 to 51.9%) | — | 49/126 | 38.9% (30.3% to 47.9%) | — |
| Aarons, M. | 27/47 | 57.4% (42.1% to 71.7%) | — | 3/47 | 6.4% (1.3% to 17.5%) | — |
| Roxby, A. | 2/3 | 66.7% (9.4% to 99.2%) | — | — | — | — |
| Pooled estimate | 111/222 | 53.6% (42.0% to 64.7%) | 40% | 71/220 | 27.9% (13.0% to 49.8%) | 84% |
| Nursing Home Staff | ||||||
| Patel, M. | 8/19 | 42.1% (20.3% to 66.5%) | — | — | — | — |
| Dora, A. | 4/8 | 50.0% (15.7% to 84.3%) | — | — | — | — |
| Graham, N. | 3/3 | 100.0% (29.2% to 100.0%) | — | — | — | — |
| Roxby, A. | 0/2 | 0.0% (0.0% to 84.2%) | — | — | — | — |
| Pooled estimate | 15/32 | 46.9% (30.6% to 63.0%) | 0% | — | — | — |
Notes: 95% CI, 95% confidence interval; proportions are calculated using a logit transformation of a random effects meta-analysis.