Systematic review of predictive performance of injury severity scoring tools

Table 2 Results of quality assessment

Internal validity
Q1	Were selection criteria clearly described?
	Yes	61	95.3%
	No	3	4.7%
Q2	Were any quality assurance measures for managing and/or collecting data described?
	Yes	24	37.5%
	No	40	62.5%
Q3	Were missing data adequately managed?
	Yes	38	59.4%
	No	28	43.8%
	Two studies were double-counted because a part of variable were excluded and the rest of variables were estimated.
Q4	Was the length of follow-up described?
	Yes	35	54.7%
	No	29	45.3%
Q5	Was the version of the reference code systems used described?
	Yes	24	37.5%
	No	40	62.5%
Q6	Was the derivation of coefficients of TRISS or weights of ICISS described?
	Yes	41	34.5%
	No	11	9.2%
	NA	14	11.8%
	Two studies described the derivation of only a part of scores studied.
Q7	Were the new coefficients or weights validated?
	Yes	25	89.3%
	No	3	10.7%
External validity
Q8	Was the description of the study population reported?
	Yes	62	96.9%
	No	2	3.1%
Q9	Was the study conducted using multi-institutional population?
	Yes	28	51.9%
	No	36	48.1%
Q10	Was the precision of AUROC, such as standard error, reported?
	Yes	31	48.4%
	No	33	51.6%

NA, not applicable; AUROC, area under the Receiver Operating Characteristic curve; TRISS, Trauma and Injury Severity Score; ICISS, International Classification of Diseases-based.

ISSN: 1757-7241