Table 10 Effects of baseline incidence, proportions diseased, risk ratios, and baseline symptom correlations on the specificities obtained from the maximal area under the receiver operating characteristic curve using at most 40 symptoms for disease diagnosis.
From: Diagnostic accuracy of symptoms for an underlying disease: a simulation study
Coefficients | (95% CIs) | p | |
|---|---|---|---|
Incidence not reaching 1 among those diseased | |||
(Intercept) | 0.72 | (0.706 to 0.734) | < 0.0001 |
Correlation between diseases | 0 | (− 0.014 to 0.014) | 0.968 |
Proportions diseased | 0.01 | (− 0.005 to 0.025) | 0.201 |
Baseline incidence | − 0.899 | (− 0.921 to − 0.877) | < 0.0001 |
Risk ratio | − 0.013 | (− 0.014 to − 0.012) | < 0.0001 |
At-risk incidence | 0.616 | (0.6 to 0.632) | < 0.0001 |
Baseline symptom correlation | − 0.116 | (− 0.129 to − 0.103) | < 0.0001 |
Dependent variable mean = 0.69; SD = 0.29; median = 0.77; min = 0; max = 1 | |||
Adjusted R-square = 0.69 | |||
Incidence reaching 1 among those diseased | |||
(Intercept) | 1.249 | (1.23 to 1.268) | < 0.0001 |
Correlation between diseases | 0.008 | (− 0.008 to 0.024) | 0.33 |
Proportions diseased | − 0.015 | (− 0.032 to 0.002) | 0.084 |
Baseline incidence | − 0.511 | (− 0.535 to − 0.487) | < 0.0001 |
Risk ratio | − 0.005 | (− 0.006 to − 0.004) | < 0.0001 |
Baseline symptom correlation | − 0.354 | (− 0.368 to − 0.34) | < 0.0001 |
Dependent variable mean = 0.88; SD = 0.2; median = 0.98; min = 0.33; max = 1 | |||
Adjusted R-square = 0.71 | |||
