The General Practitioner Assessment of Cognition (GPCOG)

Reference

% dementia

Sensitivity

Specificity

PPV

NPV

AUC

Two-stage English [2, 13]^{a, b}

246

0.85

0.86

0.71

0.93

14.2 %

0.89

Chinese [5]^{a, b}

456

0.97

0.89

0.72

0.99

13.4

0.97

French [14]^c

280

0.96

0.62

0.83

0.90

Italian [6]^{a, b}

200

0.82

0.92

0.95

0.70

17.4 %

0.96

Korean [15]^b

131

0.88

0.75

0.85

Portuguese/Brazilian [16]^a

0.91

0.78

Sub-sets of the original Australian sample [2, 13]^a

Aged <75 [13]

0.82

0.94

0.90

0.88

11.1 %

Aged 75 ≤ 80 [13]

128

0.81

0.95

0.77

0.96

7.9 %

Aged >80 [13]

123

0.88

0.72

0.67

0.90

21.9 %

Edu ≤8 year [13]

0.82

0.89

0.78

0.91

13.5 %

Edu >8 year [13]

0.86

0.85

0.68

0.94

14.8 %

Other Australian cohorts

Basic et al. [17]^b

151

38 %

0.98

0.77

0.97

Pond et al. [18]^{a, $}

1717

0.79

0.92

0.44

0.98

8.9 %

0.92

N sample size, % dementia prevalence, PPV positive predictive value, NPV negative predictive value, MC misclassification rate, AUC Area under the curve, Edu education; ^$ unpublished data

Recruitment/setting: a GP/primary care, b memory clinic/specialist, c psychogeriatric inpatients

The GPCOG’s ability to differentiate between various dementia subtypes or dementia and mild cognitive impairment has not been established yet. However, the GPCOG total score as well as its patient and informant sub-scores were found to differentiate between varying stages of dementia severity as defined by the Clinical Dementia Rating Scale (CDR; [19]) scores of 0, 0.5 and ≥1 [6]. This was still true when the authors controlled for confounding variables such as age and education [6].

10.7 Demographic and Other Biases

Cognitive screening tools are often affected by patients’ age, gender, education or cultural background [20, 21]. While being associated with patient age in some [2, 6] but not all studies [17], the GPCOG was independent of patient gender [6, 17], cultural and linguistic background [17] and education [13, 17] in populations with average educational attainment. However, threshold effects may exist whereby illiterate patients and those with less than 4 years of formal schooling perform systematically worse compared to more educated individuals [16].

The GPCOG informant interview, on the other hand, was found to be entirely free of any demographic (patient and informant) bias [13]. Likewise, cognitive performance on the GPCOG seems largely unrelated to patients’ physical and mental health [2, 6], even though results are mixed [17].

10.8 Patient and GP Acceptability of the GPCOG

The vast majority of surveyed GPs rate the GPCOG as practical (87.8 %), economically viable (87.8 %), and most importantly acceptable to their patients (98 %) [2]. Most GPs were also either satisfied or very satisfied with the GPCOG (83.7 %) and indicated they would use it again (89.8 %) [2].

Only gold members can continue reading. Log In or Register to continue