Background SymptomGuide Dementia (DGI Clinical Inc) is a publicly available online symptom tracking tool to support caregivers of persons living with dementia. The value of such data are enhanced when the specific dementia stage is identified. Objective We aimed to develop a supervised machine learning algorithm to classify dementia stages based on tracked symptoms. Methods We employed clinical data from 717 people from 3 sources: (1) a memory clinic; (2) long-term care; and (3) an open-label trial of donepezil in vascular and mixed dementia (VASPECT). Symptoms were captured with SymptomGuide Dementia. A clinician classified participants into 4 groups using either the Functional Assessment Staging Test or the Global Deterioration Scale as mild cognitive impairment, mild dementia, moderate dementia, or severe dementia. Individualized symptom profiles from the pooled data were used to train machine learning models to predict dementia severity. Models trained with 6 different machine learning algorithms were compared using nested cross-validation to identify the best performing model. Model performance was assessed using measures of balanced accuracy, precision, recall, Cohen κ, area under the receiver operating characteristic curve (AUROC), and area under the precision-recall curve (AUPRC). The best performing algorithm was used to train a model optimized for balanced accuracy. Results The study population was mostly female (424/717, 59.1%), older adults (mean 77.3 years, SD 10.6, range 40-100) with mild to moderate dementia (332/717, 46.3%). Age, duration of symptoms, 37 unique dementia symptoms, and 10 symptom-derived variables were used to distinguish dementia stages. A model trained with a support vector machine learning algorithm using a one-versus-rest approach showed the best performance. The correct dementia stage was identified with 83% balanced accuracy (Cohen κ=0.81, AUPRC 0.91, AUROC 0.96). The best performance was seen when classifying severe dementia (AUROC 0.99). Conclusions A supervised machine learning algorithm exhibited excellent performance in identifying dementia stages based on dementia symptoms reported in an online environment. This novel dementia staging algorithm can be used to describe dementia stage based on user-reported symptoms. This type of symptom recording offers real-world data that reflect important symptoms in people with dementia.

Background Individuals with dementia and mild cognitive impairment (MCI) experience a wide variety of symptoms and challenges that trouble them. To address this heterogeneity, numerous standardized tests are used for diagnosis and prognosis. myGoalNav Dementia is a web-based tool that allows individuals with impairments and their caregivers to identify and track outcomes of greatest importance to them, which may be a less arbitrary and more sensitive way of capturing meaningful change. Objective We aim to explore the most frequent and important symptoms and challenges reported by caregivers and people with dementia and MCI and how this varies according to disease severity. Methods This cross-sectional study involved 3909 web-based myGoalNav users (mostly caregivers of people with dementia or MCI) who completed symptom profiles between 2006 and 2019. To make a symptom profile, users selected their most personally meaningful or troublesome dementia-related symptoms to track over time. Users were also asked to rank their chosen symptoms from least to most important, which we called the symptom potency. As the stage of disease for these web-based users was unknown, we applied a supervised staging algorithm, previously trained on clinician-derived data, to classify each profile into 1 of 4 stages: MCI and mild, moderate, and severe dementia. Across these stages, we compared symptom tracking frequency, symptom potency, and the relationship between frequency and potency. Results Applying the staging algorithm to the 3909 user profiles resulted in 917 (23.46%) MCI, 1596 (40.83%) mild dementia, 514 (13.15%) moderate dementia, and 882 (22.56%) severe dementia profiles. We found that the most frequent symptoms in MCI and mild dementia profiles were similar and comprised early hallmarks of dementia (eg, recent memory and language difficulty). As the stage increased to moderate and severe, the most frequent symptoms were characteristic of loss of independent function (eg, incontinence) and behavioral problems (eg, aggression). The most potent symptoms were similar between stages and generally reflected disruptions in everyday life (eg, problems with hobbies or games, travel, and looking after grandchildren). Symptom frequency was negatively correlated with potency at all stages, and the strength of this relationship increased with increasing disease severity. Conclusions Our results emphasize the importance of patient-centricity in MCI and dementia studies and illustrate the valuable real-world evidence that can be collected with digital tools. Here, the most frequent symptoms across the stages reflected our understanding of the typical disease progression. However, the symptoms that were ranked as most personally important by users were generally among the least frequently selected. Through individualization, patient-centered instruments such as myGoalNav can complement standardized measures by capturing these infrequent but potent outcomes.

No abstract

BACKGROUND Individuals who live with dementia or mild cognitive impairment (MCI) experience a variety of symptoms and challenges that trouble them and/or their carers. The usual remedy for this heterogeneity is to employ several standardized tests to cover the variety of problems in cognition, behaviour and function. These tests are used for diagnosis, prognosis, and to track effects of treatment. A complementary approach is to employ individualized measures. MyGoalNav™ Dementia is one such: an online tool that allows impaired individuals and their caregivers to identify and track outcomes of greatest importance to them. Such individualized outcome measurement can be a less arbitrary and more sensitive way of capturing meaningful change. OBJECTIVE To explore the most frequent and important symptoms and challenges reported by caregivers and people with dementia and MCI, and how this varied by disease severity. METHODS This cross-sectional observational study involved 3909 online myGoalNav™ users (mostly caregivers of people with dementia or MCI), who completed symptom profiles between 2007-2019. Users chose from a library of common dementia-related symptoms and challenges their most personally important or troublesome to track over time. Users were also asked to rank their chosen symptoms from least to most important, which we called the symptom potency. As the stage of disease for these online users is unknown, we applied a supervised staging algorithm, previously trained on clinician-derived data, to classify each profile as MCI, into these four stages: MCI, Mild, Moderate and Severe dementia. Across these stages, we compared symptom tracking frequency, symptom potency, and the relationship between frequency and potency. RESULTS The staging algorithm classified 917 MCI, 1596 Mild, 514 Moderate, and 882 Severe dementia profiles. The most frequent symptoms in MCI and Mild profiles were similar and consisted of early hallmarks of dementia (e.g. recent memory, language difficulty). As the dementia stage increased to Moderate and Severe, the most frequent symptoms were characteristic of loss of independent function (e.g. incontinence) and behavioural problems (e.g. aggression). The most potent symptoms were similar between stages, and generally reflected disruptions in everyday life (e.g. problems with hobbies/games, travel, looking after grandchildren). Symptom frequency was negatively correlated with potency at all stages, and the strength of this relationship increased with increasing disease severity. CONCLUSIONS Our results underscore the feasibility and interpretability of patient-centricity in MCI and dementia studies. They illustrate the valuable real-world evidence that can be collected with digital tools. Here, the most frequent symptoms across the stages reflected our understanding of the typical disease progression. The symptoms ranked as most personally important by users, however, were generally among the least frequently selected. Through individualization, patient-centered instruments like myGoalNav™ can complement standardized measures by capturing these infrequent but potent outcomes.

BACKGROUND SymptomGuide® Dementia (SG-D) is a publicly available online symptom tracking tool. The value of these data are enhanced when the specific dementia stage is identified. OBJECTIVE We aimed to develop a supervised machine learning algorithm to classify dementia stages based on the symptoms tracked in SymptomGuide® Dementia (SG-D). METHODS We employed clinical data from 717 people from three sources: 1) a memory clinic; 2) a long-term care study30; and 3) the VASPECT31 clinical trial. Symptoms were captured with SymptomGuide® Dementia (SG-D), a web-based symptom tracking tool aimed to support caregivers of persons living with dementia. A clinician-rated dementia stage classified four groups using either the Functional Assessment Staging Test or the Global Deterioration Scale: Mild Cognitive Impairment, or mild, moderate, or severe dementia. Individualized symptom profiles from the pooled data were used to train machine learning models to predict dementia severity. To ensure unbiased evaluation of model performance, models trained with 6 different machine learning algorithms were compared using nested cross-validation to identify the best performing model. The best performing algorithm was used to train a model optimized for balanced accuracy. Model performance was assessed using measures of balanced accuracy, precision (Positive Predictive Value), sensitivity (recall), Cohen’s Kappa, Area Under the Receiver Operating Characteristic Curve (AUC-ROC), Area Under the Precision-Recall Curve (AUPRC). RESULTS The study population was mostly female (59%), older adults (77.3 ± 10.6 years, range 40-100 years) with mild-moderate dementia (46%). Age, duration of symptoms, 37 unique dementia symptoms and 10 symptom-derived variables were used to distinguish dementia stages. A model trained with a Support Vector Machine Learning algorithm using a one-versus-rest approach showed the best performance. The correct dementia stage was identified with 83% balanced accuracy (Cohen’s Kappa=0.81, AUPRC=0.91, AUC-ROC=0.96). The best performance was seen when classifying severe dementia (AUC-ROC= 0.99). CONCLUSIONS A supervised machine learning algorithm exhibited excellent performance in identifying dementia stages based on dementia symptoms reported in an online environment. This novel dementia staging algorithm can be used to describe dementia stage based on user-reported symptoms. This type of symptom recording offers real-world data that reflect important symptoms in people affected by dementia.