Table 5.

Overall strength of evidence by main outcome.

Outcome and reference standard (and additional comparator)		Relevant studies	Evidence statement	Strength of evidence
Diagnostic accuracy
	GPa (F2Fb consultation)	Berry et al [16]c	Despite great variations among OSCsd, overall diagnostic accuracy was deemed to be low and always lower than that of the reference standard.	Moderate
	Attributed to the vignette	Semigran et al [14]c Ceney et al [21]c	Despite great variations among OSCs, overall diagnostic accuracy was deemed to be low and always lower than that of the reference standard.	Strong
	Attributed to the vignette and confirmed by GPs	Hill et al [22]c Gilbert et al [23]c	Despite great variations among OSCs, overall diagnostic accuracy was deemed to be low and always lower than that of the reference standard.	Strong
	Attributed to the vignette GPs as additional comparator	Gilbert et al [17]c Semigran et al [15]c	Despite great variations among OSCs, overall diagnostic accuracy was deemed to be low and always lower than that of the reference standard and of the comparator (GPs).	Strong
Triage accuracy
	Triage nurses	Verzantvoort et al [26]c Trivedi et al [24]c Dickson et al [20]c Yu et al [27]c	Despite some variations among OSCs, including relatively high levels of triage accuracy in the study by Verzantvoort et al [26], triage accuracy was always lower than that of the reference standard. However, the study by Verzantvoort et al [26] and the study by Yu et al [27] both had 1 area each with a high risk of bias.	Moderate
	GPs	Poote et al [19]c Delshad et al [18]e	Findings were inconsistent, with great variations between studies that evaluated 2 different OSCs. However, the reference standard chosen in the study by Poote et al [19]c has a high risk of bias.	Inconsistent
	GPs Patients’ self-triage as additional comparator	Chan et al [25]e	Overall triage accuracy was lower than that of the reference standard but considered high and higher than that of the additional comparator (patients).	Moderate
	Attributed to the vignette	Semigran et al [14]c Gilbert et al [23]c	There was great variation among OSCs. Overall triage accuracy was always lower than that of the reference standard and considered to be low.	Strong
	Attributed to the vignette & confirmed by GPs	Hill et al [22]c	Despite some variations among OSCs, triage accuracy was deemed to be low and always lower than that of the reference standard. However, the index test chosen has a high risk of bias.	Weak
	Attributed to the vignette GPs as additional comparator	Gilbert et al [17]f	There was great variation among OSCs. Overall triage accuracy was always lower than the reference standard, but some OSCs performed almost as well as the additional comparator (GPs).	Moderate
	Patients’ self-triage	Berry et al [16]c	Overall triage accuracy was deemed to be low and always lower than reference standard. However, the reference standard chosen has a high risk of bias.	Weak
	NICEg guidance	Ceney et al [21]c	Overall triage accuracy was deemed to be low, with a few exceptions and great variations among OSCs but always lower than the reference standard.	Moderate

^aGP: general practitioner.

^bF2F: face-to-face.

^cWorst outcome with online symptom checkers.

^dOSC: online symptom checker.

^eBetter outcome with online symptom checkers.

^fVarying results within study.

^gNICE: National Institute for Health and Care Excellence.