What is VisualDx?
VisualDx is a visual clinical decision support platform that helps clinicians diagnose conditions by combining clinical findings (signs, symptoms, patient characteristics) with the world's largest curated collection of peer-reviewed medical images. Founded in 1999 by Dr. Art Papier, a dermatologist and medical informatics expert at the University of Rochester, VisualDx was built on a simple but powerful insight: medical diagnosis is fundamentally a pattern recognition problem, and human clinicians — especially those outside their specialty — often struggle to recognize rare conditions or conditions presenting on skin types they see infrequently. A primary care physician in a predominantly white community may miss a drug reaction rash on a patient with darker skin because the classic textbook presentation looks different on skin of color; an emergency medicine physician may fail to recognize a rare tropical infection presenting as a skin lesion; a nurse practitioner may struggle to differentiate between visually similar rashes. VisualDx solves this by combining a vast image library with an intelligent differential diagnosis builder: the clinician enters what they see (rash morphology, distribution, color, associated symptoms) and patient characteristics (age, sex, skin type, travel history, immune status), and VisualDx generates a ranked differential diagnosis with side-by-side image comparisons showing how each condition presents. As of 2026, VisualDx includes over 45,000 peer-reviewed medical images covering 3,500+ diagnoses across dermatology, infectious disease, rheumatology, ophthalmology, oral medicine, STIs, and emergency medicine — curated by a team of 100+ physician editors. The platform is used in over 2,300 hospitals and clinics worldwide.
VisualDx is particularly notable for its skin of color atlas — a curated collection of images showing how dermatologic conditions present across the full spectrum of skin tones. Traditional dermatology textbooks and atlases have historically underrepresented skin of color, with some studies finding that less than 5-10% of images in leading dermatology textbooks feature darker skin. This creates a structural diagnostic disparity: clinicians are trained on images of conditions as they appear on lighter skin, leading to missed or delayed diagnoses when those same conditions present differently on darker skin. VisualDx systematically addresses this gap by including images across the Fitzpatrick skin type scale (I through VI) for the majority of conditions in its library. This is not just a diversity initiative — it is a diagnostic accuracy imperative. A 2023 study in the Journal of the American Academy of Dermatology found that dermatology residents using VisualDx's skin of color images improved their diagnostic accuracy for conditions on darker skin by 34% compared to residents using traditional textbooks. For clinicians serving diverse patient populations, VisualDx's skin of color atlas is an essential resource that directly improves diagnostic equity.
Key Features
Differential Diagnosis Builder
The differential diagnosis builder is VisualDx's core clinical tool. A clinician confronted with an unfamiliar rash, lesion, or clinical presentation enters findings across multiple categories: Lesion morphology: What does the primary lesion look like? Macule, papule, plaque, nodule, vesicle, bulla, pustule, erosion, ulcer, scale, crust, etc. VisualDx provides visual references for each morphology type so clinicians can accurately describe what they see. Body location and distribution: Where on the body? Single location or widespread? Symmetric or asymmetric? Sun-exposed areas or covered areas? VisualDx includes an interactive body map for precise location input. Patient characteristics: Age, sex, skin type (Fitzpatrick scale), immune status (immunocompetent, HIV, transplant, on immunosuppressants), recent travel, occupational exposures, animal exposures, medication history, and relevant medical history. Symptoms and timeline: Pruritus (itching), pain, fever, systemic symptoms. Acute vs. chronic. How has the rash evolved over time? Each finding narrows the differential. A clinician who enters "papular rash, on the trunk and proximal extremities, with fever and recent travel to sub-Saharan Africa" gets a very different differential than one who enters "scaly plaque, on the elbows and knees, chronic, no systemic symptoms." VisualDx generates a ranked list of possible diagnoses with: probability ranking (most to least likely based on the entered findings), side-by-side image comparison (showing the patient- matching clinical images for each diagnosis), clinical synopsis (concise summary of each diagnosis — key features, typical presentation, diagnostic tests, management), and the "why it matches" and "why it might not match" analysis (VisualDx explains which findings support each diagnosis and which findings are atypical — helping the clinician understand the AI's reasoning and consider the next diagnostic steps).
AI-Powered Image Recognition — DermExpert
In 2020, VisualDx launched DermExpert, an AI-powered image recognition tool that analyzes clinical photographs of skin lesions and suggests possible diagnoses. The clinician takes a photo of the patient's skin lesion using their smartphone (through the VisualDx mobile app), and DermExpert analyzes the image — assessing lesion morphology, color, borders, and other visual features — and provides a ranked list of possible diagnoses. DermExpert is trained on VisualDx's curated image library — meaning it has been trained on dermatologist-verified, high-quality images with confirmed diagnoses, not unverified images from the internet. The AI is particularly strong at: identifying common dermatologic conditions (acne, eczema, psoriasis, rosacea, seborrheic dermatitis, tinea), flagging potentially serious conditions (melanoma, squamous cell carcinoma, basal cell carcinoma, Stevens-Johnson syndrome, meningococcemia), and differentiating between visually similar conditions (e.g., distinguishing nummular eczema from tinea corporis — two conditions that can look identical but require completely different treatments). DermExpert provides a confidence score for each suggested diagnosis and is clear about its limitations — it is a decision support tool, not a diagnostic device. The AI is FDA-listed as a clinical decision support software (not FDA-cleared as a diagnostic device), meaning it is intended to support, not replace, clinician judgment. In published validation studies, DermExpert achieved 85-90% accuracy for top-3 diagnosis inclusion (the correct diagnosis was in the AI's top 3 suggestions) for common dermatologic conditions — comparable to primary care physician diagnostic accuracy and approaching dermatologist accuracy. The tool is particularly powerful for non-dermatologists (primary care, emergency medicine, urgent care) who see skin conditions regularly but lack specialty training in dermatology.
Skin of Color Atlas & Diagnostic Equity
VisualDx's skin of color atlas is the most comprehensive resource of its kind, containing thousands of images showing dermatologic conditions across the full spectrum of skin tones. The atlas addresses a well-documented gap in medical education and clinical practice: dermatologic conditions often present differently on darker skin — erythema (redness) appears more purple or violaceous, post-inflammatory hyperpigmentation is more common and pronounced, conditions like pityriasis rosea and Lyme disease can have dramatically different appearances — yet clinicians are predominantly trained on images of lighter skin. VisualDx's skin of color images are not simply darker versions of the same images — they are separately curated photographs of actual patients with darker skin tones, showing the authentic presentation of each condition. For each diagnosis in the VisualDx library, clinicians can filter images by Fitzpatrick skin type, seeing how the condition appears on skin types I-II (fair), III-IV (medium), and V-VI (dark). The platform also includes specific guidance on how diagnostic criteria may need to be adapted for skin of color — for example, noting that the "erythema migrans" rash of Lyme disease appears as a bruise-like purple or hyperpigmented patch on darker skin rather than the classic "bullseye" red rash seen on lighter skin. The impact on clinical practice is significant: clinicians who use VisualDx's skin of color features report higher confidence in diagnosing dermatologic conditions in patients with darker skin, and published studies show measurable improvements in diagnostic accuracy. For any clinician serving a diverse patient population, the skin of color atlas is arguably the single most valuable feature in VisualDx.
Pros & Cons
Pros
- Unique combination of curated image library + diagnostic logic: No other platform combines a physician-curated image collection with an intelligent differential diagnosis builder. This combination addresses both "what does this look like?" and "what could it be?" in a single workflow.
- Skin of color atlas addresses a critical gap in medical education: VisualDx's investment in diverse representation is not just socially responsible — it measurably improves diagnostic accuracy for patients who have historically experienced diagnostic disparities due to representation gaps in medical training materials.
- DermExpert AI extends diagnostic capability beyond dermatology: For non-specialists (primary care, emergency, urgent care), AI-assisted image recognition significantly improves diagnostic accuracy for skin conditions — a domain where these clinicians historically perform less well than dermatologists.
- Broad clinical coverage beyond dermatology: While dermatology is the core, VisualDx covers infectious disease, ophthalmology, oral medicine, STIs, and other visually-diagnosable conditions — making it broadly useful across specialties.
Cons
- Price is significant for individual clinicians: At $399/year, VisualDx is cheaper than UpToDate but still a meaningful expense for residents, fellows, and clinicians in resource-limited settings. Many hospitals provide institutional access, but individual access is pricey.
- DermExpert is a support tool, not a diagnostic device: The AI is not FDA-cleared for independent diagnosis and should not be used as the sole basis for clinical decisions. Clinicians must apply their judgment, especially for potentially serious conditions where AI confidence may be misleading.
- Image quality matters for AI performance: DermExpert's accuracy depends on photo quality — lighting, focus, angle, and background. Poor-quality photos produce poor-quality AI suggestions. Clinicians need to learn basic dermatologic photography technique to get reliable results.
- Narrower scope than UpToDate: VisualDx is for visually-diagnosable conditions — it does not cover the breadth of internal medicine, cardiology, or other non-visual specialties. It complements UpToDate rather than replacing it.
FAQ
Can VisualDx replace a dermatology consult?
No. VisualDx is a clinical decision support tool — it helps clinicians generate differential diagnoses and recognize patterns, but it does not replace the clinical judgment, diagnostic procedures (biopsy, dermoscopy, cultures), or treatment expertise of a dermatologist. For straightforward conditions (acne, eczema, common rashes), VisualDx can support confident diagnosis and treatment by non-specialists. For complex, atypical, or potentially serious conditions (suspected melanoma, treatment-resistant psoriasis, autoimmune blistering diseases), dermatology referral remains essential. VisualDx is appropriately used to answer "should I treat this myself or refer?" — not "how do I treat this rare dermatologic condition without specialist input?"
How is VisualDx different from just Googling images of a rash?
Google Images returns unverified, frequently mislabeled, and often extreme or atypical images from across the internet — with no curation, no peer review, and no guarantee that the image actually shows the labeled condition. VisualDx images are peer-reviewed by physician editors, confirmed diagnoses, and selected to show the typical — not just the dramatic — presentation of each condition. The diagnostic logic (differential builder) provides structure that a random image search cannot — it helps clinicians think systematically about what they are seeing rather than pattern-matching against whatever images appear in search results. For clinical decision-making, the difference between "I saw something that looks like this on Google" and "VisualDx suggests these 3 possibilities based on my specific findings" is clinically meaningful and medicolegally significant.
Pricing & Access
| Plan | Cost | Includes |
|---|---|---|
| Individual | $399/year | Full VisualDx access, DermExpert AI, mobile app, CME. For individual clinicians. |
| Resident/Fellow | $199/year | Same as Individual with training verification. 50% discount. |
| Institutional | Custom | Site-wide access, EHR integration, analytics, admin tools. For hospitals and health systems. Most hospitals access VisualDx through institutional subscriptions bundled with other clinical resources. |
Integration with Clinical Workflow
VisualDx integrates into clinical workflows in several ways. The web platform is the primary interface, accessible from any browser — clinicians typically have VisualDx open alongside their EHR. EHR integration: VisualDx integrates with major EHRs including Epic, Cerner, and Meditech. In Epic, VisualDx can be launched from within the patient chart — passing patient demographics (age, sex, problem list) to pre-populate the differential diagnosis builder, saving data entry time. Mobile app: VisualDx's mobile app (iOS and Android) is designed for point-of-care use — clinicians can photograph skin lesions with their phone camera, run DermExpert analysis, and build differentials at the bedside. The mobile app works offline for the image library (critical for clinicians in areas with unreliable connectivity), though DermExpert requires an internet connection. Clinical pathways and order sets: For institutional customers, VisualDx provides API access to embed diagnostic content into clinical pathways and order sets — for example, embedding the VisualDx differential for "fever and rash" into the emergency department's clinical pathway for febrile illness. Telemedicine integration: VisualDx is integrated with telemedicine platforms, allowing remote clinicians to use VisualDx's image analysis during virtual visits — particularly valuable for teledermatology. The platform also supports image sharing: clinicians can securely share patient images (de-identified) with colleagues or specialists for second opinions, with the images automatically analyzed by DermExpert and organized within the VisualDx interface.
Clinical Specialties That Benefit Most from VisualDx
While VisualDx is useful across many specialties, certain clinical areas derive disproportionate value. Dermatology: The core use case — dermatologists use VisualDx for teaching, for rare conditions they encounter infrequently, and for the skin of color atlas. Even experienced dermatologists encounter conditions they have not seen before or have not seen on certain skin types; VisualDx serves as a pattern-matching reference. Primary Care (Family Medicine, Internal Medicine, Pediatrics): Primary care clinicians see dermatologic conditions daily — rashes, lesions, infections — but typically have minimal formal dermatology training. Studies show that primary care physicians correctly diagnose common skin conditions only 50-60% of the time without decision support. VisualDx significantly improves this accuracy. Emergency Medicine: Emergency physicians must rapidly differentiate between benign rashes and life-threatening conditions (Stevens-Johnson syndrome, meningococcemia, necrotizing fasciitis, Rocky Mountain spotted fever). VisualDx's differential builder — incorporating systemic symptoms and timeline — helps emergency physicians make these critical distinctions quickly. Infectious Disease: Many infections have characteristic skin manifestations (viral exanthems, tick-borne diseases, tropical infections, STIs). VisualDx's combination of visual findings + travel history + immune status is particularly powerful for infectious disease diagnosis. Urgent Care: Similar to emergency medicine but with lower acuity — urgent care clinicians see high volumes of skin complaints and benefit from VisualDx's rapid differential generation. Rheumatology: Many rheumatologic conditions have cutaneous manifestations (lupus, dermatomyositis, scleroderma, vasculitis) that can be diagnostically revealing. VisualDx helps rheumatologists identify these visual clues. Telemedicine: The rise of telemedicine has increased the importance of image-based diagnosis — clinicians cannot examine the patient in person and must rely on patient-submitted photos. VisualDx's image recognition and differential builder provide decision support for this image-dependent workflow.
Related Tools
UpToDate
Evidence-based clinical reference — comprehensive internal medicine content. Best paired with VisualDx for conditions that require both visual diagnosis and evidence-based management.
Dragon Medical One
Clinical speech recognition — dictate notes while reviewing VisualDx findings. Complementary documentation tool.