The prevalence of dementia is rapidly increasing, currently reaching 45 million cases worldwide. This number is expected to triple by 2050 (Batsch and Mittelman, Reference Batsch and Mittelman2015). Dementia poses a great burden on individuals, their families, and on society. In 2015, the cost of dementia worldwide is estimated to be $818 billion, with informal care and direct costs contributing equally to the total cost (Wimo et al., Reference Wimo2017). Recognizing the impact of dementia has led the World Health Organization to prioritize dementia diagnosis and treatment and to recommend every country develop a national Alzheimer’s and dementia plan (World Health Organization and Alzheimer’s Disease International, 2012).

Despite lack of cure or disease-modifying treatments, early diagnosis of dementia, even at the symptomatic stages, is advantageous (Rodriguez-Gomez et al., Reference Rodriguez-Gomez2019); it enables patients and families to gain access to available treatment and support systems, while avoiding unnecessary pharmacological treatments or medical procedures that may hasten cognitive decline or worsen other dementia-related symptoms (Dubois et al., Reference Dubois, Padovani, Scheltens, Rossi and Dell’Agnello2016). Additionally, timely diagnosis allows for planning and decision-making on health, familial, legal, and financial matters and may delay institutionalization (de Vugt and Verhey, Reference de Vugt and Verhey2013). It may also provide informal caregivers relief from the impact of uncertainty and the opportunity to adapt to their changing role (de Vugt and Verhey, Reference de Vugt and Verhey2013). Early diagnosis of dementia, especially when combined with early treatment and caregiver support (Kawakita et al., Reference Kawakita2020), has been associated with reduced hospitalization (Bernardes et al., Reference Bernardes, Massano and Freitas2018), significant reduction of treatment cost (Dubois et al., Reference Dubois, Padovani, Scheltens, Rossi and Dell’Agnello2016), amelioration of caregiver burden, and increase in insurance service utilization in patients at all severity levels of dementia (Kawakita et al., Reference Kawakita2020). Despite these advantages, underdiagnosis of dementia is still common; previous studies have demonstrated that less than 50% of cases are recognized (Black et al., Reference Black2019). A recent study based on 600 Australian patients treated in primary care or in specialists’ clinics demonstrated that the mean time from the appearance of the first dementia-related symptoms to the first consultation was approximately 6 months, and that 20% of patients already had moderate-to-severe cognitive impairment at the time of diagnosis (Black et al., Reference Black2019). Moreover, the majority of patients (76%) progressed beyond prodromal stages by the time they arrived to consultation with a healthcare practitioner (Black et al., Reference Black2019). The factors leading to underdiagnosis are complex. Lack of knowledge on dementia, its course, and consequent legal matters (Pettigrew et al., Reference Pettigrew2019), as well as fear from stigma and loss of autonomy, may prevent help-seeking in dementia diagnosis from the perspective of patients and their caregivers (Parker et al., Reference Parker, Barlow, Hoe and Aitken2020). Delay in dementia diagnosis may reflect the medical community’s nihilistic attitudes toward the efficacy of existing treatments for dementia, or result from lack of resources (e.g. overburdened healthcare systems or limited access to specialized clinics). Moreover, the primary care setting, in which most patients with cognitive complaints are first examined, lacks the expertise to differentiate between early dementia and its subtypes, Mild Cognitive Impairment (MCI), subjective cognitive complaints, and potentially modifiable situations that can mimic dementia (e.g. depression). Thus, there is an urgent need for reliable and valid brief screening instruments, which can be implemented in the primary care setting, for the detection of AD, other dementia types, and their preceding clinical presentations.

Help-seeking in dementia diagnosis may be facilitated by recognition of symptoms as problematic by the patients themselves or their proxies, a gradual worsening of initial symptoms, and realization that the symptoms are no longer temporary (Parker et al., Reference Parker, Barlow, Hoe and Aitken2020). Nevertheless, reliance on patients’ awareness for self-referral to dementia diagnosis may be problematic, since the loss of awareness is associated with the core pathology of AD, and precedes the clinical diagnosis of dementia even among cognitively normal individuals (Hanseeuw et al., Reference Hanseeuw2020). Thus, in addition to its benefits from the primary care providers’ point of view, routine cognitive screening may facilitate help-seeking in dementia diagnosis by objectively measuring patients’ performance in comparison to the normal range, thus potentially compensating for the loss of awareness.

The role of imaging and CSF-based biomarkers in the early diagnosis of AD is increasingly recognized (McKhann et al., Reference McKhann2011). However, their universal use as screening tools for AD is limited due to their invasiveness and cost. Thus, cognitive assessment tools remain the first-line screening method for the detection of cognitive impairment and dementia (McKhann et al., Reference McKhann2011).

Numerous cognitive screening tools – both paper and pencil and computerized tests–have been developed (De Roeck et al., Reference De Roeck, De Deyn, Dierckx and Engelborghs2019). Computerized tests have several advantages over paper and pencil tests; these include recording, standardization, possibility to adapt to participants’ abilities, potential availability of multiple versions – thus decreasing the risk of practice effect (Silverberg et al., Reference Silverberg2011) and automatic score calculation. Additionally, computerized tests can be easily administered bedside or in other, nonclinical settings, by people with no neuropsychological background (de Oliveira and Brucki, Reference de Oliveira and Brucki2014). Nevertheless, the superiority of computerized over paper and pencil tests in all relevant populations is yet to be demonstrated, specifically in older age groups, whose performance may be affected by visual and motor impairments, lack of basic computer skills, and language difficulties due to immigration status.

One of the most commonly used screening tests in clinical practice is the Montreal Cognitive Assessment (MoCA) (Nasreddine et al., Reference Nasreddine2005), with a high ability to discriminate between normal aging, MCI, and mild dementia (Nasreddine et al., Reference Nasreddine2005). The MoCA is more comprehensive and overcomes some of the limitations of the well-known Mini-Mental State Examination (MMSE) – by better assessing executive functions, language, and visuospatial abilities (Folstein et al., Reference Folstein, Folstein and McHugh1975). Nevertheless, it has several limitations; the average time of administration is 10–15 min, it is sensitive to language, education level, and cultural background (Rossetti et al., Reference Rossetti, Lacritz, Cullum and Weiner2011), and in many languages, paper and pencil administration is still required, thus reducing the practicality of its use as a screening tool in primary care settings.

In this issue of International Psychogeriatrics, Hicks et al. (Reference Hicks2021) present their study aimed at examining the association between global cognition, measured using the MoCA test, and performance on the computerized Paired Associates Learning (PAL) task, and to assess the usefulness of the PAL as a stand-alone tool to distinguish between healthy controls (HC) and patients with AD. The PAL test is an object location memory task, developed to examine visual episodic memory, a cognitive domain that heavily relies on the integrity of mid-temporal brain structures, affected at the earliest stages of AD. The version used in the study is part of the Cambridge Neuropsychological Test Automated Battery (CANTAB). Previous studies have demonstrated poor performance in the PAL test in individuals with prodromal or mild AD (Junkkila et al., Reference Junkkila, Oja, Laine and Karrasch2012), and an association between PAL scores and hippocampal structure and function in cognitively normal individuals and in people with MCI (de Rover et al., Reference de Rover2011; Nathan et al., Reference Nathan2017). In cognitively normal middle-aged individuals, CSF p-TAU levels were predictive of PAL scores 10 years later (Soldan et al., Reference Soldan, Pettigrew, Moghekar, Albert and Team2016). Participants in the study of Hicks et al. were patients with AD (n = 25) and normal controls (n = 22). All participants were relatively highly educated, and importantly, both groups included people whose first language was not English. The control group was younger and more highly educated than the AD group. Several PAL outcome measures are available, but the authors chose the PAL Total Errors (adjusted) outcome, which is more widely applicable as it does not depend on the number of stages completed. In Hicks’ study, PAL Total Errors (adjusted) was highly correlated with the MoCA total score (r = −0.9, p < 0.001) and its subdomains, especially the delayed recall, visuospatial executive function, attention, and orientation. These cognitive domains are primarily affected in AD (Kaur et al., Reference Kaur, Edland and Peavy2018; McKhann et al., Reference McKhann2011), but may also be sensitive to other types of dementing processes, e.g. Lewy body disease (van de Beek et al., Reference van de Beek2020) or vascular cognitive impairment (Sudo et al., Reference Sudo, Amado, Alves, Laks and Engelhardt2017). A linear regression analysis, controlling for age, sex, and education, demonstrated that PAL task Total Errors (adjusted) is a good predictor of MoCA Total Score. Moreover, a ROC analysis demonstrated that the PAL Total Errors Score (adjusted) has an excellent ability to differentiate between the AD and cognitively normal groups with high sensitivity (92%) and specificity (86%).

The results of this study are in line with previous reports demonstrating the potential of the PAL task to differentiate between patients with AD and normal controls, especially when used in conjunction with brain imaging (de Rover et al., Reference de Rover2011) or with other cognitive tests (Junkkila et al., Reference Junkkila, Oja, Laine and Karrasch2012). Patients with MCI performed poorly on this test, but not on other cognitive tests, compared to normal controls (Egerhazi et al., Reference Egerhazi, Berecz, Bartok and Degrell2007), suggesting its potential role as a screening tool for MCI. The Hicks et al. study is innovative by demonstrating the PAL as a stand-alone tool for AD screening.

The PAL test takes several minutes to administer, it does not require specific training (Barnett et al., Reference Barnett, Blackwell, Sahakian and Robbins2016) and can be modified for online administration. Thus, it can potentially serve as a screening tool for AD in overburdened primary care settings. Its successful implementation in this role will require further validation in larger samples, with a wider range of demographic characteristics (including very old and less educated people), dementia, and pre-dementia subtypes, while taking into consideration the significant decline in PAL performance with aging, in the absence of clinically overt dementia (Abbott et al., Reference Abbott2019), as well as its correlation with the level of education (Meng and D’Arcy, Reference Meng and D’Arcy2012). Optimal screening for dementia in the primary care setting will require the development of policies that determine the clinical settings in which further diagnostic workup and treatment will be carried out in people with suspected dementia following the screening procedure, as well as the allocation of resources. Furthermore, more massive detection of dementia cases, at the early stages of the disease, will require educational programs on advanced care planning (Pettigrew et al., Reference Pettigrew2019) and will have to be complemented by condensation of informal networks and the establishment of dementia-friendly communities (Parker et al., Reference Parker, Barlow, Hoe and Aitken2020)- services that may attenuate persons’ fear of stigma and loss of autonomy.