New Risk Score Predicts Dementia Probability

Megan Brooks

November 22, 2022

Researchers have developed a point-based score that can help clinicians predict an individual's risk of developing dementia over 13 years, allowing for early action to delay or prevent cognitive problems. However, some experts are skeptical.

The risk score does not require neuroimaging and "may achieve nearly 100% prediction accuracy," study investigator Xi-jian Dai, PhD, with The Second Affiliated Hospital of Nanchang University in China, told Medscape Medical News.

The study was published online November 17 in JAMA Network Open.

High Accuracy

The researchers developed the risk model for dementia using UK Biobank data for nearly 445,000 men and women (mean age, 56 years) who were cognitively normal at baseline. During 13 years of follow-up, 0.7% of men and 0.5% of women developed dementia.

Not surprisingly, increasing age was strongly associated with higher dementia risk. Other dementia risk factors included socioeconomic adversity, sleep disorders, and several comorbidities including respiratory disease, cerebrovascular disease, diabetes, hypertension, and cardiovascular disease.

Men and women shared some modifiable dementia risks and protective factors, but they also had some independent risk factors that accounted for about 32% of men and 53% of women developing dementia, the investigators report.

The total point score of the risk model ranged from –18 to 30 for men and –17 to 30 for women.

For example, a 70-year-old (10 points) man with underweight (3 points), low educational level (1 point), and a history of diabetes (1 point) and cerebrovascular disease (5 points) would have a total risk score of 20 points, and the corresponding risk of dementia is 9% at 5 years, 31% at 9 years, and 54% at 13 years.

The discriminative Cox proportional hazard regression model to predict 5-, 9- and 13-year dementia risk had a C statistic of 0.86 for men and 0.85 for women in the training set, and 0.85 for men and 0.87 for women in the testing set.

The risk score had a predictive accuracy for 9-year dementia risk of 97.59% in men and 99.59% in women, and close to 100% predictive accuracy for 13-year dementia risk both in men and women, the researchers report.

They caution that the results have not been externally validated with other independent cohorts.

Also, individuals in the UK Biobank have a restricted age range and some measurements were based on a single-item subjective assessment, which could lead to misclassification.

Other limitations include that all participants were hospitalized, which could lead to selection bias, and some potentially important predictors, such as diet, were not considered.

Despite these limitations, they say this "proposed" risk prediction tool "may help individuals to identify their potential risk profile and provide guidance on precise and timely actions to prevent or delay dementia."

Dementia Experts Skeptical

In a statement, the UK-based nonprofit and independent Science Media Centre urged caution in drawing firm conclusions on the predictive power of the tool.

"The claim 'the risk score model yielded 100% prediction accuracy of 13-year dementia risk' is extremely misleading," said David Curtis, MBBS, MD, PhD, with University College London's Genetics Institute.

"The score does not accurately predict whether or not one will develop dementia in 13 years, rather it provides the probability that somebody will develop dementia," Curtis said.

"This would be like me claiming that I can predict the risk of getting heads when I toss a coin with 100% accuracy — the risk of getting a head is 0.5. The score is a poor predictor of whether somebody will get dementia or not, it only predicts their chances of getting dementia," Curtis added.

Curtis noted that the results of this research largely "recapitulate already well-established risk factors such as age, low education, and vascular disease along with a number of features which may indicate that the subject already has early dementia but that it is yet to be formally diagnosed."

Echoing Curtis, Tom Russ, PhD, director of the Alzheimer Scotland Dementia Research Centre in Edinburgh, noted these risk factors have been previously highlighted in "great detail" by the 2020 Lancet Commission article, as previously reported by Medscape Medical News.

"Importantly," said Russ, the Lancet Commission report "highlighted the different points in the lifespan when particular risk factors might be most important."

Ivan Koychev, PhD, senior clinical researcher at the University of Oxford in Oxford, England, had a slightly different take on this research.

He noted that the risk model for dementia with a 5-to-13-year horizon has "sound methodology and has the benefit of working with one of the largest datasets currently available."

"The implications are that models such as the one described are in good position to be rolled out in clinical practice to complement screening methods for dementia such as blood tests. The cost-benefit of such screening programs for dementia is yet to be established," Koychev told the Science Media Centre.

This study was supported by grants from the Guangdong Basic and Applied Basic Research Foundation, Natural Science Foundation of Hunan Province, and University of Macau. The authors, Curtis, and Russ have disclosed no relevant financial relationships. Koychev is a medical adviser to Five Lives, a digital technology company developing a solution to offer dementia risk prediction and lifestyle modification to aging adults.

JAMA Netw Open. 2022;5(11):e2242596. Full text

For more Medscape Psychiatry news, join us on Twitter and Facebook.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.