Evidence-based guidance and an interactive catalogue of 44+ question types for building effective clinical research surveys.
Great data starts with great survey design. Learn evidence-based best practices, then explore 44+ question types purpose-built for clinical research.
Evidence-based principles from 40+ years of survey methodology research
Define exactly what data you need before writing a single question. Every question should directly support a research objective — if it doesn't, remove it.
Questions should be written at a reading level accessible to your full participant cohort. Avoid jargon, double-barrelled questions, and leading phrasing.
Order questions from general to specific. Group related topics together and use information blocks to transition between sections.
Completion time is the biggest predictor of dropout. Research shows surveys over 10 minutes lose 5–20% of respondents per additional minute.
Leading questions, double negatives, and loaded terms introduce bias. Ensure each question measures exactly one thing.
Match your question type to the data you need. Closed questions give structured data; open questions give richness. Use scales for intensity and frequency.
Run your survey with 5–10 participants from your target population before launch. Measure completion time, ask about confusing questions, and check for ceiling/floor effects.
In clinical research, informed consent, data minimisation, and participant welfare must guide every design decision. Ensure triage questions can escalate when needed.
Each category includes design guidance to help you choose the right question type
Use for capturing names, comments, and narrative responses. Keep open-text questions focused — provide placeholder text to guide expected length and format.
Short free-text answers with optional validation (email, phone, NHS number).
Longer narrative responses with configurable row count.
Touch-friendly signature capture, saved as an image.
Freeform drawing for symptom mapping or visual responses.
Ideal for categorical data. Randomise option order to reduce primacy bias. Keep lists under 7 items where possible — use ranking for longer preference lists.
Radio-button style with customisable options and colour coding.
Checkbox selection allowing multiple answers.
Binary choice with large, touch-friendly buttons.
Boolean question variant for factual assertions.
Live-result polling with percentage bars.
Drag-and-drop ordering of options by preference.
Scales produce the richest quantitative data. Use 5-point scales for simple opinions, 7-point for nuanced attitudes, and 0–10 NRS for clinical measures like pain. Always label anchor points clearly.
Configurable 0–10 scale with dot, button, or slider display modes.
Continuous slider with custom min/max, units, and labels.
Visual radial gauge with gradient fill and unit display.
Multi-row agreement scale with customisable columns.
Emoji-based pain/severity rating (😊 to 😫).
Visual battery-style energy level indicator.
Experiential sampling methods work best when used repeatedly over time. Pair these with scheduled reminders for ecological momentary assessment (EMA) studies.
20 emoji options capturing emotional snapshots in real-time.
Circular emotion selector with colour-coded mood segments.
Place demographics at the end of your survey to avoid stereotype threat. Use validated category sets (e.g. ONS ethnicity categories) and always offer 'Prefer not to say'. ProNova52's on-server ML automatically recognises demographic answers and persists them to the participant's profile — so returning participants never have to re-enter information they've already provided.
Title, first name(s), and last name. Auto-persisted to participant profile via on-server ML.
Full UK-format address capture. Recognised and stored for future surveys automatically.
Searchable dropdown with 60+ options. Answer persisted across studies.
ONS-standard categories. Pre-filled on repeat surveys from stored profile data.
Multi-select from 25 options. Remembered for future participation.
Dropdown with common nationalities. Automatically carried forward.
Media questions dramatically increase engagement but also increase survey length. Use sparingly and always provide clear instructions. Test recording on both iOS and Android.
In-browser audio capture with playback preview.
Camera-based photo with instant upload and preview.
Video capture with duration hints and progress tracking.
General file upload with type and size restrictions.
Embed YouTube, Vimeo, or uploaded videos with pre/post questions.
Clinical question types require careful governance review. Ensure triage thresholds are validated against your protocol and that consent text has been approved by your ethics committee.
Full consent text with checkbox confirmation — required before proceeding.
Data privacy notice with acknowledgement checkbox.
Red flag assessment with configurable severity actions.
Body map for selecting affected areas.
Region-based intensity mapping with touch-friendly controls.
Date-based frequency tracking for symptom occurrence.
Numeric input for clinical values like heart rate, BP, temperature.
Genetic testing questions should always be optional with clear 'Find out more' links. Ensure your participant information sheet covers the implications of genetic data.
Optional Illumina GSA v3 genotyping for lifestyle trait insights. Kit ordering built in.
Clinical-grade 30x WGS with ClinVar-validated variant reporting.
Information blocks are essential for setting context between sections. Use A/B comparisons for preference research and multi-entry lists for medication or symptom logging.
Display instructions, context, or guidance between questions.
Participants add multiple structured entries (e.g. medications).
Side-by-side comparison for preference testing.
Multi-step wizard for complex data collection.
Multiple toggle switches for feature/symptom checklists.
ProNova52 uses private, on-server ML models to automatically identify demographic answers as participants complete surveys. Recognised fields — name, address, ethnicity, language, nationality, and country of birth — are persisted to each participant's secure profile and pre-filled on any future survey, eliminating repetitive data entry and improving completion rates.
No public LLMs
All ML processing runs entirely on our infrastructure. Participant data is never sent to third-party AI services like OpenAI, Google, or Anthropic.
Answer once, use forever
When a participant returns for a follow-up or new study, demographic fields arrive pre-populated. They can review and update, but never have to start from scratch.
Automatic field recognition
The system identifies which answers map to profile fields without manual configuration. Investigators don't need to tag questions — it just works.
A participant completes demographic questions (name, address, ethnicity, etc.) in a survey.
Private ML models running on ProNova52 infrastructure identify which answers are demographic profile fields.
Recognised fields are securely written to the participant's encrypted profile — no data leaves your tenant.
When the participant returns for a follow-up or new study, demographics are pre-populated. They review and confirm.
Everything you need to know before getting started.