Frequently Asked Questions (FAQ)

RadWise is a comprehensive online platform designed to help radiology trainees prepare for the FRCR Part 2A examination. We offer a large bank of practice questions, detailed explanations, a helpful AI assistant and performance analytics to target your revision effectively.

Our questions are designed to reflect the style, difficulty, and syllabus of the actual FRCR 2A exam. The curriculum has been mapped and appropriate topics have been carefully selected. Each question goes through a rigorous process of checking the information at 3 different setpoints to ensure accuracy and relevance of the questions, distractors and study notes so you can focus on revision knowing it's high-yield content, highly relevant for the exam.

Different Large Language Models have been used to draft and refine the questions, explanations and answers to ensure clarity and format into easily digestible information. Wherever LLMs were used, the content has been verified afterwards by a human expert on at least 3 separate occassions. The knowledge is consistent with other radiological resources and the authors believe the accuracy of the content is on par or better than conventional question books.

Careful selection of AI applications with sophisticated model choice and prompting has been helpful in ensuring initial quality. Each question data is then reviewed by a human expert at 3 separate checkpoints against their own knowledge and the available radiology resources to ensure accuracy. The authors are radiology trainees who are or have been recently involved in the exam revision therefore possess accurate and contemporary understanding of the exam and the tested knowledge.

No, unfortunately much less so. Due to the fast paced nature of the question bank a faster AI model has been used to answer the follow-up questions - who would like to wait 40 seconds for answer? Also, the content is not reviewed by a human expert at all, not to say 3 times! In our practice however, the AI bot follow-up responses have been very accurate. We estimate the accuracy of the AI follow-up bot at >99% and believe it is extremely useful as it allows rapid fact checking and speeds up the revision process significantly. However, we accept no responsibility for the AI generated content. Always double check the generated message.

Each question has been tagged with appropriate data to indicate what knowledge and from which radiological subdomain is being tested. The topics involve the commonly tested or important associations and the amount of repeated knowledge tested has been minimised, so we advise to make every question count and consider all the other distractors, plus review the provided study note carefully.

This is intended. The exam has undergone some changes recently and not all of the resources have been updated in line with this. For example, we have removed questions that test the eponynomous names as the main piece of knowledge for the exam. They're not relevant, and won't be tested!

Those questions have been carefully crafted to simulate the hardest questions encountered in the FRCR 2a exam. They can be very vague with little information to go on and more than one plausible answer. They can be very frustrating to answer but please don't be discouraged as they are a fantastic learning material and prepare the candidates well for the actual sitting.

We believe radiology trainees are very busy and we believe the worst preparation one can do for the exam is waste time on poor quality content, given the demands of the training programme. That's why our questions are very rigorously scrutinised on their learning value before making their way to the active question pool. We believe more in quality than quantity of questions as after all we understand we are not the only resource available for exam preparation. However, we strive to ensure each candidate comes out with better knowledge after finishing our question bank. Currently there are on average 3-4 questions that have been taken off the database for 1 question you are seeing during your revision.

We have taken a lot of care to make Radwise a revision platform that encourages early exam preparation with 'little and often' revision over a longer period of time. We would like to strongly encourage our users to start the revision early and we would like to discourage last minute cramming of the questions in our question bank. Last minute cramming is not beneficial to the candidates, does not utilise the amazing spaced repetition features we have, and introduces congestion on the platform in the immediate period before the exam. That's why the pricing structure is flat with the emphasis on early revision start.

Your Dashboard and Analysis pages provide detailed insights into your performance. You can see your overall accuracy, performance by category relative to your peers, performance by subdomains relative to your peers, difficulty level, and track your progress over time. The statistics compare you to all peers so you can see where you stand within the cohort preparing for the exam.

This is because there isn't enough meaningul statistical data collected yet. We suggest you engage more with the question bank and once you answer enough questions more advanced statistics will start to show!

We strive for accuracy and each question content has been reviewed at least 3 times. However, if you spot an issue, please use the "Provide Feedback" button available under each question. Your input is valuable in helping us maintain the quality of our question bank.

You can reach out to us via our Contact Us page.