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 to flag people who may need testing for Parkinson’s disease. In a study published in Nature Aging, it performed better than comparison models and came close to its full-questionnaire performance using a targeted set of 30 questions. The system is a screening tool, not a diagnostic test, and its usefulness in routine primary care has not yet been established.
Researchers at Rensselaer Polytechnic Institute have developed an artificial intelligence framework that analyzes health questionnaires to flag people who may need further testing for Parkinson’s disease. In a study published in Nature Aging, the team reported that Quest2Dx outperformed earlier models on the questionnaires it analyzed, but stressed that the system is intended to support screening, not diagnose the disease.
The RPI-led team trained Quest2Dx using data from the Parkinson’s Precision Medicine Initiative database and evaluated it using a separate dataset from the Fox Insight Study. The questionnaires were lengthy: one contained 187 questions, while the other had more than 300. Participants often left some items unanswered, a challenge for conventional machine-learning systems that may fill in missing responses and introduce errors.
According to the paper, Quest2Dx was designed to handle missing answers, interpret questionnaires with different wording or question order, and identify the question-and-answer combinations most relevant to each person. The researchers said it performed better than earlier approaches, including a model based on the large language model Llama 2. They also tested a shortened survey of 30 questions selected by the framework; its predictive performance was nearly as strong as with the complete questionnaires.
Questions the system ranked as informative included difficulties with everyday tasks such as getting out of bed or eating, as well as reported exposure to industrial solvents and degreasers. The authors also identified factors such as sleep problems, anxiety and stress-related feelings. These are possible indicators in combination, not proof of Parkinson’s: the same experiences can occur for many reasons.
A Shorter Route to Specialist Assessment
The work addresses a practical challenge in identifying people who may benefit from further assessment: some signs associated with early Parkinson’s can appear before prominent movement symptoms, but they may be subtle or attributed to other causes. The researchers say patients and caregivers may overlook issues such as changes in smell, sleep, handwriting or mood, delaying referral to a neurologist.
If validated for use in clinics, a system that can find patterns in routine questionnaires could help primary-care teams decide who should receive more detailed evaluation. The 30-question result also points to a possible way to reduce the time and effort required of patients. However, the study does not establish that using Quest2Dx in primary care will lead to earlier diagnosis, improved outcomes or fewer unnecessary referrals. Those effects would need to be tested in real-world clinical settings.
Products Worth Considering
Fitbit Versa 2 Health and Fitness Smartwatch with Heart Rate, Alexa, Music, Sleep and Swim Tracking
Fitbit Inspire 3 Health and Fitness Tracker Bundle
Levoit Core 400S-P Smart Air Purifier for Large Rooms (Up to 1,718 sq ft)
As an affiliate, we earn on qualifying purchases.
Questionnaires Before Motor Symptoms
Parkinson’s assessment relies in part on clinicians observing and measuring motor symptoms, alongside other examinations and, in some cases, invasive procedures such as spinal taps. The disease may be present for years before physical symptoms become apparent, according to the study’s background. The RPI researchers say their goal is to help identify people who might need further testing earlier, rather than replace clinical evaluation.
Long health surveys can contain relevant details but are difficult to interpret when they are lengthy, incomplete or phrased differently across studies. Quest2Dx uses a large language model to interpret the meaning and context of questions and answers, the authors said, instead of treating each item only as a fixed position in a form. The research team included RPI department head Pingkun Yan, Juergen Hahn, Diego Machado Reyes and University of Pennsylvania professor Li Shen.
“We built this AI model to screen for potential cases earlier.”
— Pingkun Yan, head of RPI’s Department of Biomedical Engineering
Clinical Use Still Needs Testing
The reported results come from analyses of existing questionnaire datasets. The source material does not provide the model’s exact accuracy figures, confidence intervals, or performance across particular patient groups, so the scale of its advantage cannot be quantified here. It is also unclear how it would perform among people with other conditions that can cause similar symptoms, or in clinics whose patient populations differ from the study datasets.
Quest2Dx has not been shown in the material provided to diagnose Parkinson’s, replace neurologists, or improve patient outcomes. A positive screening result would need to be followed by appropriate clinical assessment. The authors’ findings about particular questions describe patterns in their analysis; they do not mean that any one symptom, exposure or answer confirms the disease.
Validation Before Routine Screening
The researchers say their findings could help shape shorter questionnaires for future Parkinson’s screening studies and potentially for primary-care use. The next step is further evaluation of the tool in settings where it would be used with patients, including whether it consistently identifies people who need specialist assessment without generating an unmanageable number of false alarms.
The source material does not give a timeline for such studies or indicate that Quest2Dx is currently available in clinics. Until clinical validation and implementation are reported, it remains a research framework rather than an established screening service.
Key Questions
Does Quest2Dx diagnose Parkinson’s disease?
No. The researchers describe it as a screening framework that may flag people for further testing. It is not a diagnosis and does not replace evaluation by a qualified clinician.
What information does Quest2Dx analyze?
It analyzes answers to health questionnaires, including responses about symptoms, everyday activities and possible risk factors. The study tested it on data from the Parkinson’s Precision Medicine Initiative and Fox Insight Study.
How did the shorter questionnaire perform?
The team created a survey containing the 30 questions it considered most pertinent. The model’s predictive performance with that shorter survey was reported as nearly as strong as with the full questionnaires.
Can a symptom or solvent exposure confirm Parkinson’s?
No. The authors say factors such as sleep problems, anxiety or solvent exposure must be considered alongside other information. None of these questionnaire responses is diagnostic on its own.
Is Quest2Dx being used in primary-care clinics now?
The available report describes potential future use in primary care, but does not say the system is currently deployed there. Its performance and effects in routine clinical settings remain to be tested.
Source: news-medical.net
