TL;DR
Turn quiet afternoons into listening time
- Thousands of audiobooks, podcasts and originals
- Listen on your phone, tablet or Echo — also offline
- Cancel anytime
Researchers at Rensselaer Polytechnic Institute developed Quest2Dx, an AI framework that analyzes health questionnaire responses for patterns associated with Parkinson’s disease. In a study published in Nature Aging, it performed better than comparison models and nearly as well on a targeted 30-question survey as on longer questionnaires. The researchers say it could support earlier referrals for testing, but it is not a diagnostic tool.
Researchers at Rensselaer Polytechnic Institute have developed an AI framework called Quest2Dx that analyzes health questionnaires for patterns associated with Parkinson’s disease, with the aim of helping clinicians refer people for testing earlier. In a study published in Nature Aging, the team reported that the system outperformed comparison models; they emphasize that it is designed to screen for possible cases, not diagnose the disease.
The researchers tested Quest2Dx using responses from the Parkinson’s Progression Markers Initiative database and used the Fox Insight Study as a second dataset to validate the model. The questionnaires are lengthy: the PPMI survey studied by the team contains 187 questions, while the Fox Insight questionnaire has more than 300. Respondents may skip items, and different surveys can phrase or order questions differently.
According to the study, Quest2Dx was designed to handle missing responses without relying on conventional imputation, which fills in unanswered items and can introduce errors. It also uses a large language model to interpret question meaning, allowing it to work across questionnaires with different wording, and identifies which question-and-answer combinations are most informative for an individual. The team reported stronger results than earlier models, including an approach based on the Llama 2 language model.
The researchers also assembled a shorter survey from the 30 questions the system ranked as most relevant. They reported that Quest2Dx’s predictions using this reduced set were nearly as accurate as its predictions using the full questionnaires. The source report does not provide specific performance figures, such as sensitivity or false-positive rates, so those cannot be stated here.
Shorter Surveys Could Support Earlier Referrals
Parkinson’s disease can be present before the movement-related symptoms that commonly prompt evaluation become apparent. The RPI team’s proposed use for Quest2Dx is to flag questionnaire patterns that may warrant follow-up, potentially helping primary care providers refer some patients to specialists sooner. The value would be in supporting clinical attention, not replacing a clinician’s assessment.
A shorter, more focused questionnaire could also make screening easier to use in research or primary care than surveys containing hundreds of items. Still, the study reports model performance on existing datasets; it does not establish that using Quest2Dx in clinics will lead to earlier diagnosis, better care, or improved patient outcomes.
Products Worth Considering
Fitbit Versa 2 Health and Fitness Smartwatch with Heart Rate, Alexa, Music, Sleep and Swim Tracking
LEVOIT Vital Pet Smart Air Purifier
Philips Sonicare 5950 Rechargeable Electric Toothbrush with 2 Brush Heads, Pink
As an affiliate, we earn on qualifying purchases.
Why Questionnaire Patterns Are Difficult
Clinical assessment of Parkinson’s relies in part on observing and measuring motor symptoms, while some diagnostic procedures can be invasive. The disease may precede visible physical symptoms by years, and the researchers argue that early signs can be overlooked or attributed to other causes. They cite changes such as loss of smell, sleep problems, handwriting difficulties, and mood changes as potential clues, not as signs that establish a diagnosis.
Long questionnaires pose a separate technical challenge. Respondents may leave items unanswered, while standard machine-learning systems can be sensitive to question order or wording. Quest2Dx was built to interpret the meaning of questions and answers and to consider combinations of responses. The study’s examples of informative responses included difficulty with everyday activities, exposure to industrial solvents or degreasers, sleep behavior, anxiety, and stress. The authors caution that these factors overlap with other conditions and should be considered together rather than treated as diagnostic individually.
“We built this AI model to screen for potential cases earlier.”
— Pingkun Yan, RPI biomedical engineering department head and study team leader
Clinical Performance Still Needs Testing
The reported results do not show how Quest2Dx would perform prospectively in a routine primary-care population, where the mix of patients and conditions may differ from the datasets used in the study. The source material provides no numerical accuracy measures, thresholds for referral, or details on how the system might balance missed cases against unnecessary follow-up.
It is also unclear whether the shorter questionnaire will be tested directly with patients in clinical settings, how clinicians would receive or act on a screening result, and whether performance is consistent across different groups. Because symptoms and risk factors can overlap with other illnesses, a flagged response pattern cannot confirm Parkinson’s disease.
Validation Before Primary-Care Use
The researchers say the findings could help guide the design of simplified questionnaires for Parkinson’s screening and future studies in primary care. The next practical step is further evaluation of the shortened survey and Quest2Dx in settings beyond the datasets used in the paper, including assessment of how results affect referrals and follow-up testing.
No deployment timeline or planned clinical trial is specified in the source material. Until additional validation establishes how the tool performs in practice, Quest2Dx should be understood as a research screening framework, not a stand-alone diagnostic system.
Key Questions
Can Quest2Dx diagnose Parkinson’s disease?
No. The researchers describe it as a screening tool intended to identify questionnaire patterns that may merit testing or referral. A screening result does not establish a diagnosis.
What information does Quest2Dx analyze?
It analyzes answers to health questionnaires, including surveys with hundreds of items. The system is designed to interpret question meaning, handle unanswered items, and identify combinations of responses that may be informative.
What did the study report about the 30-question survey?
The team reported that Quest2Dx was nearly as accurate using a targeted set of 30 questions as it was using the full questionnaires. The source report does not give numerical performance measures.
Does a response about sleep, anxiety, or solvent exposure indicate Parkinson’s?
No single response is diagnostic. The researchers note that these factors can overlap with other conditions and say potential indicators need to be considered alongside other signs.
When could the system be used in clinics?
The source material does not give a deployment date. It describes future studies and possible primary-care use as goals, with further testing needed before routine clinical use can be assessed.
Source: news-medical.net
