The discriminant validity of single-question assessments of subjective cognitive complaints in an Asian older adult population

Ting Pang, Xuhao Zhao, Xindi He, Cheuk Ni Kan, Narayanaswamy Venketasubramanian, Ching‐Yu Cheng, Changzheng Yuan, Christopher Chen, Xin XuFrontiers in Aging Neuroscience 2022 · DOI

TL;DR

This study investigates the effectiveness of single-question assessments for identifying subjective cognitive complaints (SCC) in older adults, a potential early indicator of dementia. The research found that while single questions alone have limitations, combining them with a standard cognitive screening tool like the MoCA significantly improves the accuracy of dementia detection in a community setting.

What Problem Does This Paper Address?

Dementia is a growing global health challenge, and early detection is crucial for timely intervention and management. Subjective cognitive complaints (SCC) – a person's own or their caregiver's report of declining memory or cognitive abilities – are recognized as an early sign of preclinical Alzheimer's disease and an increased risk for dementia. However, relying solely on subjective reports can be unreliable due to factors like depression or an individual's inability to accurately self-assess. This paper addresses the need to validate simple, single-question SCC assessments for their ability to discriminate between individuals with and without dementia, and to explore how these simple tools can be best combined with objective cognitive assessments for efficient community-based screening.

Methods

The study employed a two-phase approach. In Phase I, a large cohort of older adults (≥60 years) in Singapore underwent initial screening using the Abbreviated Mental Test (AMT) and the Progressive Forgetfulness Question (PFQ). Individuals who showed signs of cognitive impairment in Phase I were invited to Phase II. In Phase II, these participants underwent a more comprehensive assessment, including various single-question SCC assessments (GDS-10, AD8-patient, AD8-informant), the Montreal Cognitive Assessment (MoCA), and a formal neuropsychological battery. This allowed for a research diagnosis of dementia based on DSM-IV criteria. The researchers then calculated discriminant validity indices (sensitivity, specificity, positive predictive value, negative predictive value, and overall accuracy) for each SCC tool and the MoCA, both individually and in combination. They explored two combination strategies: compensatory (either tool positive) and conjunctive (both tools positive).

Datasets

The study utilized data from the Singapore Epidemiology of Eye Diseases (SEED) study, a community-based cohort of multi-ethnic subjects aged 60 years or older in Singapore. The dataset included 3,780 participants assessed in Phase I, with 957 proceeding to Phase II for comprehensive cognitive and clinical investigations. Of these, 911 were dementia-free and 46 had dementia.

Key Insights and Ideas

  • Informant Reports Shine: The AD8-informant (AD8-info) question, where a caregiver reports on the participant's memory problems, demonstrated superior specificity and overall accuracy compared to self-reported SCC assessments. This highlights the value of caregiver perspectives in identifying cognitive decline, especially in individuals who may not accurately self-assess.
  • Single Questions Aren't Enough (Alone): While SCC assessments are valuable indicators, using them as standalone tools for dementia screening has limitations in discriminant validity. This is likely due to factors like depressive symptoms influencing self-reports or the inherent difficulty older adults have in accurately assessing their own cognitive changes.
  • The Power of Combination: The most significant finding is that combining a single-question SCC assessment (specifically AD8-info) with an objective cognitive tool like the MoCA dramatically enhances diagnostic accuracy. This synergy is key to efficient and reliable community screening.
  • Context Matters for Combinations: The study suggests that the choice of combination strategy (compensatory vs. conjunctive) depends on the screening context. Compensatory combinations are better for maximizing sensitivity (capturing more potential cases), while conjunctive combinations improve specificity (reducing false positives).

Key Results

  • The MoCA alone showed good sensitivity (95.6%) and specificity (81.5%) for dementia detection.
  • Among the single-question SCC tools, AD8-informant had the highest specificity (83.2%) and overall accuracy (82.5%).
  • Combining MoCA with AD8-informant using a compensatory approach achieved optimal sensitivity (100%) and positive predictive value (100%), ideal for broad initial screening.
  • Using a conjunctive approach with MoCA and AD8-informant resulted in excellent specificity (96.3%) and overall accuracy (94.8%), which is highly effective for narrowing down the pool of individuals needing further investigation.

Limitations and Future Work

The authors acknowledge several limitations:

  • Gold Standard Diagnosis Subset: The gold standard diagnosis of dementia was only administered to a subset of individuals screened positive in Phase I, potentially leading to an underestimation of dementia prevalence.
  • Small Dementia Sample Size: The relatively small number of dementia cases limited further analysis of dementia subtypes.
  • Potential Selection Bias: A significant proportion of participants (39.9%) refused to participate in the comprehensive cognitive assessment, potentially introducing selection bias.
  • Estimation of PPV/NPV: The generally lower prevalence of dementia in the community setting may have affected the estimation of positive and negative predictive values.

Future work could focus on:

  • Predicting pathology-related cognitive changes among older adults at risk of dementia using single-question SCC measures.
  • Further validation of SCC tools in preclinical stages of dementia and across different dementia subtypes.
  • Adjusting predictive values of SCC tools based on census data for more accurate estimations.

Summary updated

Abstract

Combining a reliable single-question SCC assessment with an objective tool can efficiently discriminate dementia patients from healthy older adults in the community.
BibTeX
@inproceedings{pang2022discriminant,
  author = {Ting Pang and Xuhao Zhao and Xindi He and Cheuk Ni Kan and Narayanaswamy Venketasubramanian and Ching‐Yu Cheng and Changzheng Yuan and Christopher Chen and Xin Xu},
  title = {{The discriminant validity of single-question assessments of subjective cognitive complaints in an Asian older adult population}},
  booktitle = {Frontiers in Aging Neuroscience},
  year = {2022},
  doi = {10.3389/fnagi.2022.901592},
}

Loading related content...

Figure