Table 1:

Effects of screening for chlamydia on pelvic inflammatory disease among general-risk individuals

Outcome	Approach	No. of studies and design	Follow-up period, mo	Rate in unscreened individuals	Rate in screened individuals (95% CI)	Absolute difference (95% CI)	Certainty of evidence
Pelvic inflammatory disease	Offer of screening,* all eligible	2 RCTs5,6 n = 141 362	12–36	27.0 per 1000†	27.3 per 1000 (19.4 to 38.0)	0.30 more in 1000 (7.60 fewer to 11.0 more)	⊕⊖⊖⊖ VERY LOW‡,§
	Offer of screening,* selected individuals	1 RCT31 n = 2607	12	27.0 per 1000†	11.6 per 1000 (5.70 to 24.0)	15.4 fewer per 1000 (3.00 to 21.30 fewer)	⊕⊕⊖⊖ LOW§,¶
	Acceptors of screening	2 RCTs, 1 CCT8,28,33 n = 30 652	12–18	27.0 per 1000†	21.3 per 1000 (16.2 to 28.1)	5.70 fewer per 1000 (10.8 fewer to 1.10 more)	⊕⊕⊖⊖ LOW‡,§

Note: CCT = controlled clinical trial, CI = confidence interval, CT = Chlamydia trachomatis, PID = pelvic inflammatory disease, RCT = randomized controlled trial.

^*These analyses represent results of studies that examined the effect of offering chlamydia or gonorrhea screening to all eligible individuals, regardless of level of uptake. One study used an offer of screening approach in a preselected population of individuals interested in screening (offer of screening, selected individuals).

^†The effects without screening assumed that about 6% of the female population would have chlamydia (general-risk prevalence). For the outcome of PID, it was assumed that about 13% of females with chlamydia would develop PID (0.78% of the total population), and that about 25%–30% of all-cause PID is attributed to chlamydia (all-cause PID = 3.5 times PID from CT); 0.78% × 3.5 = 2.7% prevalence of PID owing to chlamydia in the unscreened group.

^‡Serious concerns about indirectness.

^§Serious concerns about imprecision.

^¶Serious concerns about risk of bias.