Title-plus-abstract versus title-only first-level screening approach: a case study using a systematic review of dietary patterns and sarcopenia risk to compare screening performance

Table 2 Study screening approach contingency tables and performance results

A. Title-only contingency table and performance results
	Qualified	Unqualified	Potential
Included	22	132	154
Excluded	16	8356	8372
Total	38	8488	8526
Sensitivity (included ‘trues’)			57.89%
Excluded ‘trues’ rate			42.11%
Specificity			98.44%
Accuracy			98.26%
Predictive power			14.3%
Chi-squared statistic^a			2.9208
p value^a			0.874
B. Title + abstract contingency table and performance results
	Qualified	Unqualified	Potential
Included	36	105	141
Excluded	2	8383	8,385
Total	38	8488	8,526
Sensitivity (included ‘trues’)			94.74%
Excluded ‘trues’ rate			5.26%
Specificity			98.76%
Accuracy			98.75%
Predictive power			25.5%
Chi-squared statistic^a			5.208
p value^a			0.0225^c
C. Average performance results contingency table^b
	Qualified	Unqualified	Potential
Included	29	119	148
Excluded	9	8370	8,379
Total	38	8488	8,526
Average sensitivity (included ‘trues’)			76.32%
Excluded ‘trues’ rate			23.68%
Average specificity			98.60%
Average accuracy			98.50%
Average predictive power			19.7%

^aChi-squared statistics and associated p values were calculated to compare each first-level screening approach against the average performance of either screening approach
^bThe average performance is defined as the average screening performance across both screening approaches taken together. For example, it is expected that 29 studies (or (22 + 36)/2) = 29) will be correctly identified and included in the next stage of evaluation
^cIndicates that the difference in screening performance versus the average performance results is statistically different with + 95% confidence

ISSN: 2046-4053