SS_diagnostics.R

# Title: Choosing Sample Size for Evaluating a Diagnostic Test
# Step 1 specifications
sn <- 0.90  # Substitute your value for sensitivity here
sp <- 0.85  # Substitute your value for specificity here
p <- 0.20   # Substitute your value for prevalence here
w <- 0.10   # Substitute your value for width here
Disease <- "Disease Name"
CI <- "95%" #### or 90%
# Step 2: Calculate TP+FN
a_c <- (1.96^2) * sn * (1 - sn) / (w^2) #### 1.645 for 90% or 1.96 for 95% CI
# Step 3: Calculate N1
n1 <- a_c / p
# Round up to the next whole integer
n1_int <- floor(n1)
if (n1 != n1_int) {
  n1 <- n1_int + 1
}
# Step 4: Calculate FP+TN
b_d <- (1.96^2) * sp * (1 - sp) / (w^2) #### 1.645 for 90% or 1.96 for 95% CI
# Step 5: Calculate N2
n2 <- b_d / (1 - p)
# Round up to the next whole integer
n2_int <- floor(n2)
if (n2 != n2_int) {
  n2 <- n2_int + 1
}
# Step 6: Get final sample size
if (n1 > n2) {
  n <- n1
} else if (n2 > n1) {
  n <- n2
} else {
  n <- n1
}
# Print the sample size
mySS = (data.frame(Disease = Disease,Precision = w, Sensitivity = sn,
  Specificity = sp, Prevalence = p, N1 = n1, N2 = n2, Total_Subjects = n,
  CI = CI))
print(mySS[,c(2:8)], row.names = F)