Comparison of diagnostic test accuracy estimates using the Multiple Thresholds Model based on complete and published cut-off data with a bimodal reporting pattern
2Institute of Medical Biometry and Statistics, Faculty of Medicine and Medical Center – University of Freiburg, Freiburg, Germany, Germany
The multiple thresholds (MT) model allows for diagnostic test accuracy (DTA) meta-analysis of multiple cutoffs simultaneously. Previous analyses demonstrated the ability of the MT model to approximate DTA estimates of a complete cutoff dataset, based on reported cutoffs. This has been done for the Patient Health Questionnaire-9 dataset (Benedetti 2020), which had a symmetrical cutoff reporting, and the most frequently reported cutoff were mostly the study-specific optimal cutoff. The optimal cutoffs defined by the authors are only a subset of the reported cutoffs, so the most frequently reported cutoff is not necessarily the most frequent optimal cutoff. It has been speculated that the cutoff reporting pattern might influence the extent to which this approximation is possible.
Objectives: To evaluate how well MT approximates DTA of a complete cutoff dataset based on cutoffs with bimodal reporting in which the most frequently reported cutoff is not the most frequent optimal cutoff defined by authors using Youden’s index.
Methods: We used data from an ongoing review on DTA of the Hospital Anxiety and Depression Scale questionnaire, for which information about published and all cutoffs was available. We present cutoff reporting pattern and DTA estimates obtained through the MT model for both published and all cutoffs.
Results: We analyzed 251 cutoffs from 13 studies. Eighty-one cutoffs were published (Figure 1), whereas 170 were provided by authors. Reporting on cutoffs was bimodal, 8 and 11 were the most frequent cutoffs, reported in 10 and 8 studies, respectively. Only one optimal cutoff each was at cutoff 8 and 11. Using reported versus full dataset, absolute difference of MT DTA estimates were as follows (Figure 2): sensitivities were underestimated by use of reported cutoffs up to 1.6% with cutoffs ≤10 and overestimated by a maximum of 1.4% above cutoff 10. Specificities were overestimated by a maximum of 3% with cutoffs ≤14 and underestimated by a maximum of 0.03% above cutoff 14.
Conclusions: Our preliminary analysis suggests that MT model approximates accurately the complete cut-off dataset despite the use of cutoffs with a bimodal reporting pattern. Patient, public, and/or healthcare consumer involvement: none. Additional resources: Figure 1, Figure 2.