Sensitivity and specificity of various digit span scores in the detection of suspect effort

Digit Span Age-Corrected Scaled Score (ACSS) and Reliable Digit Span (RDS) have been suggested as effective in assessing credibility. The purpose of this study was to confirm the efficacy of suggested cutoffs for ACSS and RDS and to explore the utility of other Digit Span variables in a large sample (N = 66) of "real-world >= suspect effort patients versus clinic patients with no motive to feign (N = 56) and controls (N = 32). With specificity at >= 90%, sensitivity of ACSS increased from 32% to 42% when a <= 5 cutoff was used instead of the recommended <= 4. The RDS recommended cutoff of <= 7 resulted in a sensitivity of 62% but with an unacceptably high false positive rate (23%); dropping the cutoff to <= 6 raised the specificity to 93% but sensitivity fell to 45%. Cutoffs for other Digit Span scores did not exceed 45% sensitivity with the exception of 50% sensitivity (11% false positive rate) for average time per digit for all attempted items > 1.0 second. A criterion of ACSS <= 5 or RDS <= 6 was associated with 51% sensitivity (91% specificity) while RDS <= 6 or longest string with at least one item correct <= 4 was associated with 54% sensitivity (88% specificity). While only moderately sensitive, Digit Span scores, including new time variables, may have a unique and effective role in the detection of suspect effort.