Brief cognitive tests validated in Peru for detection of cognitive impairment A systematic mapping of the scientific literature

Abstract. Brief cognitive tests (BCTs) are necessary for early detection of cognitive impairment, particularly in primary care settings. Objective: This report describes a systematic review of BCTs evaluated in Peruvian populations. Methods: We used systematic mapping techniques to identify articles on screening tests for cognitive impairment involving Peruvian subjects. We included studies published in English and Spanish up to 2018. We reviewed 6 reference databases within the Virtual Health Library network, as well as the Web of Science, Scopus (MEDLINE), and EMBASE databases. Results: Ten out of 447 articles met the inclusion criteria. Studies included both outpatient (9) and community-based (2) samples. Eligibility criteria of the studies were similar. Although different protocols were applied, the diagnostic criteria were standardized. For discrimination between dementia and controls, IFS (AUC: 0.99) and ACE (AUC: 0.95 to 1.00) showed superior performance, as did the M@T (AUC: 1.00) and CDT-Mv (AUC: 0.94 to 1.00) for discriminating between Alzheimer’s disease (AD) and controls. Conclusion: The available evidence is limited. However, our analysis of national data suggests that the ACE may be a good choice whenever it can be applied to Peruvian patients. Alternatively, the M@T and IFS can be used for screening patients with suspected AD or FTD, respectively.

D ementia is defined as a marked decline in cognitive function that is severe enough to affect social and occupational functioning. 1 The most common form of dementia is Alzheimer's disease, although many other conditions can cause dementia, including vascular dementia, Frontotemporal dementia (FTD), dementia with Lewy bodies, Parkinson's with dementia, and mixed dementia. Other less common disorders can also cause symptoms of dementias. 2 At the end of 2010, there were 35.6 million people with dementia globally; 58% of whom were in middleincome countries. Dementia is projected to rapidly increase in the near future, by nearly 63% (relative to 2010 levels; i.e. to 65.7 million cases) in 2030 and by 71% (relative to 2010 levels; i.e. to 115.4 million cases) in 2050. 3 Dementia is therefore a growing global health concern, especially in low-and middle-income countries (LMIC) where there are major gaps in the availability of services. In Latin America, population ageing further adds to poor socioeconomic conditions in producing a critical scenario. 4 Early identification of cognitive impairment can allow for timely intervention, thereby slowing down the progression of the disease. This can be especially beneficial in cases where the treatment does not modify the natural course of the disease. 5 Thus, the need for further research on early diagnosis of Alzheimer's disease, as well as on identifying the root cause of neurodegeneration and cognitive impairment, has been stressed in various studies as a means of prevention and control of disease progression globally. 6,7 A study reported that generalists detect <40% of patients with cognitive impairment, with failure rates proving highest in preclinical and prodromal stages than in patients with established dementia (56% versus 19%). 8 In this context, access to brief cognitive tests (BCTs) is essential to increase detection in outpatient consultations, whether as a screening test or to confirm a diagnosis in persons suspected of having cognitive impairment. [9][10][11] One of the main drivers behind this study was to inform the medical community of existing validated BCTs for cognitive impairment. Understanding which BCTs have been validated would also help to identify existing research gaps and hence the research needs which might be prioritized in the future.
The study objective was to create a systematic map to identify and describe the properties of BCTs in Peru.

Study design
We conducted a systematic mapping review of scientific articles on BCTs published in indexed journals based on a protocol that addressed the overarching objective of the study with the following research questions: RQ1. What  The properties of BCTs are: 1) ease of use, simplicity and short administration time (no more than 5-10 minutes for primary care doctors and specialists, respectively) and training; 2) objective evaluation, unequivocal (free of complex instructions) and straightforward assessment (without the need for additional instruments); 3) user-friendly and ecologically friendly (does not generate rejection in the individual being evaluated; 4) validated within the context in which it will be applied (independent of personal , sociodemographic, ethnic, or cultural characteristics); and 5) flexible to be adapted for use in other contexts. 12 For the purpose of this study, the following BCTs were considered: 1) Addenbrooke's Cognitive Examination (ACE); 13

Data collection
The following databases were queried: Web of Science (primarily including MEDLINE and SciELO Citation Index), Scopus, MEDLINE, Excerpta Medica data-BASE (EMBASE), and Virtual Health Library (including LILACS, IBECS, CUMECS, BINACIS, PERNAL and MINSA). We used a broad set of search terms as follows: (root "Peru") AND (complete and abbreviated search terms of each of the BCTs considered in this study) OR ("cognitive impairment" OR "cognitive decline" OR "dementia"). The latter terms were used in order to avoid excluding studies that evaluated different BCTs to those identified a priori. No temporal limitations were applied.

Search process
The following search string was used on each of the reference databases. A total of 883 articles were obtained. Search results were put into an electronic database and duplicate articles eliminated. The remaining articles were reviewed by two independent researchers -a neurologist specialized in cognitive disorders (NC) and an epidemiologist specialized in systematic reviews (EHP), through the screening of title, abstract, and keywords.
At this point, the same two researchers filtered the 17 references, independently screening them for the full content of each paper. A final set of 15 references were produced and included in the study.
The following data were collected: a) study subjects (population, participant eligibility criteria, sample and comparison group); b) methodology (study design type); c) diagnostic procedures (ruling out other causes of cognitive impairment, tests, procedures and diagnostic criteria utilized; and d) BCT performance according to comparison groups (cut-off points, area under the ROC curve, sensitivity, specificity, % correctly classified, probability coefficient, predictive values, Cronbach's alpha, and Pearson's coefficient).
The Pfeffer Functional Activities Questionnaire (PFAQ) was also included for evaluating functioning and for the screening of dementia. 28,29 Statistical analysis All studies selected and relevant information extracted was put into a database and further analyzed using STATA version 15. With the goal of comparing BCT performance, we utilized the area under the receiving operating characteristic (ROC) curve (AUC).

Ethical aspects
The protocol was approved by the ethics committee of the Instituto de Medicina Tropical "Daniel Alcides Carrión" of the Universidad Nacional Mayor de San Marcos under approval number CIEI-2018-020.

Study selection
The search resulted in the retrieval of 883 articles, of which 436 duplicates were removed. A further 432 were excluded based on screening of the title and abstract. A total of 15 eligible articles were reviewed using the full-text version of the papers. Of these, five were subsequently excluded (two for insufficient information and three that were not focused on evaluating psychometric test performance). A total of 10 studies were thus selected for inclusion in the systematic mapping [ Figure  1]. These studies assessed only seven BCTs, namely, the MMSE, ACE, CDT-Mv, FAB, IFS, M@T, and the Peruvian monetary test (PMT).

Study sample characteristics
Of the total studies included, two were secondary analyses from databases produced for other studies with different objectives. The studies included involved clinical samples from outpatient settings, where all but two were conducted using community-based residential samples 39 or in elderly day care centers, all of which were located in the capital city of Lima. Differential diagnosis included dementia, dementia of the Alzheimer type (DAT), frontotemporal dementia (FTD), Parkinson's disease dementia (PPD), mild cognitive impairment (MCI), and depression (Table 1).
We characterized studies in terms of the type of controls: 1) healthy participants; 2) participants who did not have the study disease; and 3) participants that had the study disease ruled-out (negative controls). We evaluated whether the comparison groups within each  Figure 1. Flowchart of study selection process study were "comparable" in relation to basic criteria of sociodemographic variables ("intergroup homogeneity"). Results revealed that some study groups were not "comparable" in terms of age (four studies), sex (one study), and educational level (one study). However, for three studies, the comparison groups were comparable for all of these sociodemographic variables (Table 1). Eligibility criteria common to the included studies were: age (particularly >60 years of age), years of education received, the absence of disorders that could affect performance in executing the psychometric tests applied in each of the studies, sensory disorders (vision and hearing) or non-sensory disorders (motor or physical). Only one study included as exclusion criteria use of medications that affect cognitive performance (analgesic opioids, decongestants, anti-spasmodics, anticholinergics, antidepressants, antiarrhythmics, antipsychotics, antiemetics, anxiolytics, and valproate) ( Table 2).
With regard to the diagnostic protocols applied in each of the studies included, the presence of ruling out potential causes for secondary cognitive impairment was reviewed, revealing the Beck's Depression Inventory as the main tool used to rule out depression. Similarly, study protocols for carrying out laboratory tests, neuroimaging, and specialized evaluations were reviewed. One of the studies failed to report information on these data tests. Serological testing for syphilis and ELISA-HIV was also included.  (Table 3).
Regarding the diagnostic criteria applied in each of the studies, the DSM-IV and the National Institute of Neurological and Communicative Disorders and Stroke-Alzheimer's Disease and Related Disorders Association (NINCDS-ADRDA) were the standards for diagnosing dementia and Alzheimer's disease. Similarly, we report the criteria utilized in a consistent manner for diagnosing other types of dementia. In this respect, the criteria utilized to classify the severity of the patients were the Clinical Dementia Rating (CDR), Alzheimer's Disease Assessment Scale-cognitive subscale (ADAS-cog), and the Global Deterioration Scale (GDS) ( Table 4).    to 1.00). For discriminating between FTD and controls, the IFS (AUC: 0.98) showed superior performance to that of the FAB (AUC: 0.73) ( Table 5).

DISCUSSION Implications
This study assessed seven BCTs (the MMSE, ACE, CDT-Mv, FAB, IFS, M@T, and the PMT). According to our results, the IFS, ACE, M@T and CDT-Mv showed acceptable performance levels for the discrimination of various types of cognitive impairment. Likewise, a data comparison reported in a recently published systematic review 30 observed that the performance of the ACE in Peruvian samples was similar for the discrimination between DAT and controls (AUC: 0.98).
The majority of the BCTs included evaluate global cognitive efficiency (MMSE, ACE, CDT-Mv and PMT), which have a distinct value in LMICs such as Peru given that the prevalence of cognitive impairment of vascular origin and mixed dementia tends to be greater in these countries. 31 In turn, the remaining tests focus on evaluating executive functions (FAB and IFS), that better discriminate cases of FTD and vascular dementia or on memory function (M@T) that best discriminate typical cases of DAT. 32 The level of scientific evidence suggests that in Andean countries such as Peru, there is a lack of community-based studies. Of the total number of studies included, only one was conducted in a community setting, including persons with low levels of education. This represents a valuable opportunity to improve future studies in that diagnostic accuracy for BCTs varies in community and clinical settings. 33 We evaluated for homogeneity with regard to key sociodemographic variables. Six out of 10 studies were heterogeneous for one of these variables, primarily agewhich is considered a direct reflection of age and dementia prevalence. 34 We considered that the heterogeneity observed for other variables does not compromise validity in the studies given that sex is not a determinant of cognitive performance and that education effects can be adjusted for (in the case of CDT-Mv). [35][36]   It has been recommended that after a cognitive complaint suggesting the presence of dementia, a generalist should conduct a physical exam, request bloodwork and urine testing to rule out reversible causes for cognitive impairment, all in addition to applying a BCT. In countries like Canada 37 and Chile, 38 the primary care physician has a major role in diagnosing dementia. In the case of clinical specialists, their initial evaluation includes a neurologic evaluation and applying a BCT. 39 In this regard, it is fundamentally important to report the evidence available to help support decision-making in selecting a BCT.
The NICE guide, an international reference, considers a specific BCT set for the initial diagnostic evaluation of a patient suspected of having dementia that includes the 10-point Cognitive Screener (10-CS), 6-item Cognitive Impairment Test (6CIT), 6-item screener, Memory Impairment Screen (MIS), Mini-Cog, and the Test Your Memory (TYM) assessment. 39 None of these tests have been evaluated in communities of Andean countries, probably due to the fact that they were not designed to be used in settings with zero to low levels of education in different cultural settings, such as is the case with Peru.
Even though there is limited evidence, data suggests that given the broad spectrum of global cognitive tests and their performance observed, the ACE can be a first line option in screening for cognitive impairment, particularly when the optimal time to administer the test is available (around 20 minutes). 40,41 Moreover, the M@T can be a good option to confirm suspected cases of AD, whereas the IFS is suited for detecting FTD or vascular disease. 42 The clinical utility of the BCTs needs to be analyzed in a rigorous manner. It is indeed a valuable tool for opportune identification of cognitive impairment in primary care settings. It is also important to rule out that obtaining normal results after conducting a BCT should not be considered the only input to rule out a diagnosis of dementia. 39 This is particularly important in individuals with high levels of education and a good cognitive reserve, or even in cases of recurring depression where a complete neuropsychological evaluation and longitudinal follow-up tend to be clues for establishing a diagnostic.

Limitations
Akin to all systematic investigations, we faced the challenge of a lack of data in the original studies available. 30 Consequently, we cannot rely on the validity of the results for rural populations or populations whose native language is not Spanish, since original studies did not include these populations. Additionally, the studies selected did not report data about the reliability (interobserver and test-retest) and practicality (training demand, administration time, implementation costs) of the tests assessed. Finally, another means of improvement relates to the uniformity of the data reported, since none of the studies covered all the information required for this systematic mapping (see tables).

Strengths
A recently published systematic review addressed the BCT topic for cognitive impairment, 30,33 but does not comprise publications conducted in the Spanish language or indexed in databases such as LILACS. Another distinguishing point in our study is the performance mapping of a BCT set in different subtypes of dementia. Another strength is in reference to the inclusion criteria and detailed evaluation of the rates of diagnostic accuracy, an observation made in other publications. 30,43 Conclusions Our investigation further supports evidence for the IFS, ACE, M@T and CDT-Mv as viable options for discriminating various types of cognitive decline and identified the ACE as potentially the best alternative screening test evaluating global cognitive impairment. However, more reliable studies are needed that incorporate the diversity of the local setting in Andean countries to provide sound recommendations.

Recommendations
We stress that there is no assessed BCT providing optimal diagnostic performance to discriminate between the different forms of cognitive impairment. Therefore, primary care physicians should focus on the strategies of selection and prioritization of the BCTs based on their clinical suspicions. Considering the available data, in the Peruvian setting we recommend starting with the ACE whenever possible, subsequently opting for the M@T and IFS in cases of suspected AD or FTD, respectively. This first recommendation is particularly valid for urban contexts in Peru. For environments outside the Peruvian context, its validity depends on the similarity in sociodemographic conditions. Thus, according to previous reports, we anticipate better validity for Andean countries and Mexico. 44,45 Additionally, further investigation on this topic is necessary, particularly for adapting, validating and assessing the several BCTs in each region. 46 Taking into account the Latin-American context, future investigations should collect data on subjects with different levels of education (including low-educated and illiterate individuals), geographical location (including rural zones), 47 and mother tongue (including indigenous languages), 43 as well as report all data about the psychometric properties of BCTs (performance, reliability and practicality). Our study could be used as the basis for generating a checklist for research protocols.
In settings with limited resources for the administration and scoring of BCTs, it is recommended to have digital assessment tools in place that can be self-administered. 43 Unfortunately, none of the studies included had applied BCTs with these characteristics. Given their greater usability potential for population-based screening, it is necessary to evaluate the performance of computerized BCTs, such as the Computer Assessment of Mild Cognitive Impairment or the UCSF Brain Health Assessment, as well as those that are not yet computer-ized may be administered with the use of information technology such as the IFS, ACE, or the M@T.