Decision Tree for Early Detection of Cognitive Impairment by Community Pharmacists

Abstract

Purpose: The early detection of Mild Cognitive Impairment (MCI) is essential in aging societies where dementia is becoming a common manifestation among the elderly. Thus our aim is to develop a decision tree to discriminate individuals at risk of MCI among non-institutionalized elderly users of community pharmacy. A more clinically and patient-oriented role of the community pharmacist in primary care makes the dispensation of medication an adequate situation for an effective, rapid, easy, and reproducible screening of MCI.Methods: A cross-sectional study was conducted with 728 non-institutionalized participants older than 65. A total of 167 variables were collected such as age, gender, educational attainment, daily sleep duration, reading frequency, subjective memory complaint, and medication. Two screening tests were used to detect possible MCI: Short Portable Mental State Questionnaire (SPMSQ) and the Mini-Mental State Examination (MMSE). Participants classified as positive were referred to clinical diagnosis. A decision tree and predictive models are presented as a result of applying techniques of machine learning for a more efficient enrollment.Results: One hundred and twenty-eight participants (17.4%) scored positive on MCI tests. A recursive partitioning algorithm with the most significant variables determined that the most relevant for the decision tree are: female sex, sleeping more than 9 h daily, age higher than 79 years as risk factors, and reading frequency. Moreover, psychoanaleptics, nootropics, and antidepressants, and anti-inflammatory drugs achieve a high score of importance according to the predictive algorithms. Furthermore, results obtained from these algorithms agree with the current research on MCI.Conclusion: Lifestyle-related factors such as sleep duration and the lack of reading habits are associated with the presence of positive in MCI test. Moreover, we have depicted how machine learning provides a sound methodology to produce tools for early detection of MCI in community pharmacy.Impact of findings on practice: The community of pharmacists provided with adequate tools could develop a crucial task in the early detection of MCI to redirect them immediately to the specialists in neurology or psychiatry. Pharmacists are one of the most accessible and regularly visited health care professionals and they can play a vital role in early detection of MCI.

Importance of Increasing Modifiable Risk Factors Knowledge on Alzheimer’s Disease Among Community Pharmacists and General Practitioners in Spain

“…Pharmacists are those who must inform and advise on drugs, detect breaches, problems related to medicines and prevent diseases, offering pharmaceutical care that improves the quality of life of the patient. Such professionals can be trained to detect signs or characteristic symptoms of any pathology, performing tests and screenings to determine the possible referral to the practitioners (Climent et al, 2018). This turns pharmacies into excellent spaces to inform and disseminate best practices.…”

Episodic-Memory Performance in Machine Learning Modeling for Predicting Cognitive Health Status Classification

“…Machine learning has an inherent capacity to reveal meaningful patterns and insights from a large, complex inter-dependent array of clinical determinants and continue to "learn" from ongoing utility of practical predictive models. Thus, we are confident that our models will improve with more and more diverse clinically validated health status data (e.g., a broad multifactorial scope including genomics, promising biomarkers, and other functional, behavioral, and lifestyle indicators) to train the models [2,11,23]. A robust, multi-faceted, and externally validated model can uniquely complement and measurably enhance the sensitivity and specificity of MemTrax as a valid cognitive health screen tool and thus greatly assist in clinical decision support and patient management.…”

“…Dabek and Caban [10] also utilized SVM in predictive modeling of military service members developing post-traumatic stress disorder after traumatic brain injury. And Climent et al [11] conducted a cross-sectional study including an extensive array of clinically relevant variables and two screening tests while using decision tree machine learning modeling and complementary ensemble techniques to detect early mild cognitive impairment and associated risk factors in older adults. This new approach in utilizing machine learning to address the complexity of various human health challenges is only recent; but the demonstrated advantages in more aptly considering myriad interrelated factors that reflect the multiple domains of real-world systems biology are increasingly being realized.…”

Utility of MemTrax and Machine Learning Modeling in Classification of Mild Cognitive Impairment

“…The advent and evolving practical application of artificial intelligence and machine learning in screening/detection have already demonstrated distinct practical advantages, with predictive modeling effectively guiding clinicians in the challenging assessment of cognitive/brain health and patient management [7,[9][10][11]. In our study, we chose a similar approach in MCI classification modeling and cognitive impairment severity discrimination as confirmed by clinical diagnosis from three datasets representing selected volunteer inpatients and outpatients from two hospitals in China.…”