Carolien Zeetsen

129 CHAPTER Summary and general discussion 7 correlated significantly to results on the corresponding NPA domain when administered subsequently (correlation coefficients ranging from small [.229] to medium [.542]), providing evidence for low to moderate construct validity of the MoCA for use in addiction care. Another important form of validity is criterion validity, which is the extent to which the result of a test is related to an outcome (criteria). In other words, how well do results on the MoCA correspond to the classification of cognitive impairments based on an NPA? Criterion validity can be determined concurrently, by administering the MoCA and an NPA on the same day, and predictively, by administering the MoCA several days/weeks previous to an NPA. On both instances, the validity is determined by calculating at several possible MoCA cut–off points: the sensitivity (the percentage of patients with NCD who scored below the cut–off score on the MoCA) and specificity (the percentage of patients without NCD who scored above the cut–off score on the MoCA), by which the adequacy of the screening test is being determined. But also by calculating at the same cut–off points: the positive predictive value (PPV; the percentage of patients scoring below the cut–off on the MoCA who actually have NCD), and the negative predictive value (NPV; the percentage of patients scoring above the cut–off on the MoCA who actually do not have NCD), by which the patient is being assessed. In Chapter 3, this was done in a heterogeneous sample of 82 patients with SUD. Predictively, a MoCA–TS cut–off score of 24 yielded the most optimal sensitivity (55.6%), specificity (62.2%), PPV (64.1%) and NPV (53.5%). Concurrently, a MoCA–TS cut–off score of 25 yielded the most optimal sensitivity (66.7%), specificity (73.0%), PPV (75.0%) and NPV (64.3%). These findings are partly in line with other MoCA studies, which have some important differences to the current study. Firstly, only homogeneous groups of patients with AUD were included in those studies (as opposed to a heterogeneous group of patients with SUD), limiting their external validity (Wester et al., 2013; Oudman et al., 2014; Alarcon et al., 2015). Secondly, patients in these earlier studies were abstinent for only one week (Alarcon et al., 2015) or for more than six months (Oudman et al., 2014), while in clinical practice patients are often not abstinent at intake. One study that also related MoCA performance directly to an NPA (Ewert et al., 2018) found a higher cut– off score in a sample of hospitalized patients with AUD. The first–known study on the MoCA, performed in a heterogeneous group of patients with SUD, found slightly better psychometric properties than were currently found, but were most in line with the present findings (Copersino et al., 2009).