Welcome to OpenForum. Join our global community: Sign up, Login or Learn more

Osmosis: Knowledge Diffusion

Osmosis - an evidence-based learning science platform to improve medical education, reduce medical errors, and improve patient outcomes.

Photo of Vishal Punwani
1 6

Written by

Name of applicant

Shiv Gaglani

Job title of applicant

CEO and Co-Founder of Osmosis

Organization name

Osmosis: Knowledge Diffusion

Financial structure of organization (choose one)

  • For profit

City, State, Country of organizational headquarters

Baltimore MD USA

Your social media links


Names, titles and affiliations of teammates

1. Shiv Gaglani, CEO and Co-Founder. MD(JHU)/MBA(HBS) Student.
2. Ryan Haynes PhD, CTO and Co-Founder. MD(JHU) Student.
3. Rishi Desai MD/MPH, Chief Medical Officer, ID Pediatrician(Stanford).
4. Vishal Punwani, Lead Medical Editor, MD(Melbourne) Student.
5. Tanner Marshall MSc, Lead Video Developer.
6. Caleb Furnas, User Experience and Growth.
7. Sam Miller, UI/UX, Design.
8. Thasin Jaigirdar, Project Manager.
9. Kyle Slinn BSN, Video Content Lead.

Who is the primary purchaser of your innovation? (choose one)

  • Providers
  • Employers

In which care setting will your innovation primarily be used? (choose one)

  • Multiple

If you answered "other" or "multiple" to the previous question, please describe.

Osmosis is used in medical schools, and will be used in hospitals as we expand into CME.

How would you characterize your innovation? (choose one)

  • Product or service innovation (i.e., this type of product or service does not currently exist)

Osmosis is a web- and mobile- learning platform that uses evidence-based learning science principles to drive health and medical students' knowledge retention. We currently specialize in medical education - students can create and gain access to tens of thousands of questions, flashcards, and teaching videos, but Osmosis is much more than an online library of resources. Similar to evidence-based medicine, we believe in evidence-based education. Our team includes PhDs in neuroscience & education, practicing clinicians, and medical students who have built research-proven cognitive techniques into the Osmosis platform to maximize learning and retention. These include test-enhanced learning, spaced repetition, memory anchors, collaborative learning, the Fogg Behavior Model, and gamification, which all complement our robust content recommendation engine. In addition to being featured in the bestselling book, Make it Stick: The Science of Successful Learning (p 245), Osmosis has published peer-reviewed papers (www.osmosis.org/institutions) and is actively engaged in research with a view to improving education delivery, including a cross-institutional study with UIC and UCF on its intelligent formative assessment system - this is being funded by a Stemmler Grant from the National Board of Medical Examiners.

In our paper in the high-impact Annals of Internal Medicine, “What can medical education learn from Facebook and Netflix?” we described a vision for the future of education. After lamenting that the tools that our future clinicians use to socialize (Facebook) and watch TV (Netflix) are managed by more sophisticated algorithms than the tools they use to learn medicine, we decided to build Osmosis so that we could bring data-driven recommendations and experiences to our future clinicians. Osmosis analyzes what the students are learning in class and generates custom quizzes and recommendations that help them simultaneously prepare for school, board exams, and most importantly clinical practice. By solving the problem for medical education, we believe Osmosis can also be applied to many other fields of education.

Currently, there are two major problems in the delivery and learning of medical education, The first problem is that there is an overwhelming amount of critically important content to be delivered in medical school. In fact, learning in medical school has rightfully been likened to trying to drink water from a firehose. The second problem is that medical students utilize incredibly ineffective learning techniques such as cramming. To make things worse, there are essentially 3 different curricula that students need to learn in med school - and they don't always overlap much if at all. Medical students require clinical knowledge to work with our patients, class knowledge to do well on shelf exams, and an often entirely separate set of knowledge to pass our board exams. Osmosis' goal is to allow our future clinicians to be able to learn more efficiently and enjoyable: we consolidate these 3 curricula, and use evidence-based learning principles to drive long-term retention of what students need to know.

We know that strong performance in medical school is correlated with strong performance as a clinician. Osmosis demonstrably improves knowledge retention, thus in the long run we are confident that it will help to prevent medical errors and therefore improve patient outcomes.

Impact: Describe the value proposition of your innovation. How will this approach improve health care delivery? (200 words)

Medical errors made by physicians are the 3rd leading cause of death in the US. In 2008, medical errors cost the US $19.8bn, and a recent Harvard Medical Practice Study showed that diagnostic errors account for 17% of preventable errors, many of which could be traced back to lapses in physician knowledge.

Osmosis is a web- and mobile- platform that uses evidence-based learning science principles to drive knowledge retention. Our team is comprised of PhDs in neuroscience and education, practicing clinicians, and medical students. To maximize learning and retention, we’ve utilized test-enhanced learning, spaced repetition, memory anchors, collaborative learning, and the Fogg Behavior Model. The system allows future clinicians to learn more efficiently, by focusing on a mastery-based model and offering personalized video and question recommendations that are synchronized with classroom lessons. Osmosis also challenges students to assess their confidence in what they know, a metacognitive skill that is deliberately aimed at reducing medical errors. Osmosis demonstrably improves knowledge retention, thereby preventing medical errors and improving patient outcomes.

Evidence: What data do you have to illustrate that your solution works? The more specific, the better. Please share any non-proprietary results including cost savings, improved outcomes, patient satisfaction, etc. (300 words)

Osmosis is used by 40,000+ medical students from 300+ schools around the world. To date, over 8.5 million USMLE-style questions and flashcards have been answered on the Osmosis platform. Much of the learning has been social, with 480,000+ crowdsourced questions created by students themselves. Students have also uploaded 62,000 documents totalling 2.2m pages of medical school lectures for our machine-learning algorithm to analyze. Finally, it’s helpful to remember that a lot of the learning has occurred on-the-go, by users engaged on mobile devices, using the Osmosis app (4.5 star app rating).

The learning science and the high user engagement have shown real-world results in the form of short-term and long-term knowledge gains. Students who use our platform have seen improvements in exam grades by as much as 40%, and have seen their time spent studying decrease by as much as 50%. Among students who use Osmosis for more than 3 months, the median USMLE score was 247 (85th percentile), well above the national median of 229.

To date, Osmosis have been validated by 4 peer-reviewed journal articles about Osmosis, as well as organizations such as Harvard-Macy Foundation, the Robert Wood Johnson Foundation, and the National Board of Medical Examiners who have offered grants to further develop our learning algorithm and platform. Osmosis has received numerous educational technology awards including the ‘E-Learning Award’ and the ‘Overall Award’ at the 2015 QS-Wharton Re-imagine Education Awards, where Osmosis was vetted by education leaders at Apple and Google, and took 1st place at the Milken-Penn GSE Education Business Plan Competition. Most intriguing have been applications of Osmosis in ongoing clinical education, an approach currently being used by Kaiser Permanente.

Dissemination: Describe your scale up plan including your target market and 2-4 milestones for the coming year. What are the biggest hurdles to dissemination and how do you plan to overcome these obstacles? (400 words)

We have three major milestones for the next year. First, we’re aiming to have 40 medical schools officially use Osmosis for curriculum delivery (currently there are 18). Second, we’ll expand our content offering to include thousands of curated video-based resources in addition to creating over a hundred new hot-topic videos in-house. Our current library of videos are the official medical teaching videos on Wikipedia (e.g. Zika virus, pneumonia, and jaundice), and garner over 200,000 views/month and hundreds of positive comments. We will have these videos translated into ten major languages to facilitate global dissemination. Third, we will make a concerted push into the nursing market by delivering NCLEX-specific content to US and Canadian nursing schools.

With regard to scaling, we’ve brought on the former Khan Academy Health and Medicine Team, and have built strategic partnerships with content providers including Medscape and The Doctor’s Channel, to offer reliable, up-to-date information for students and faculty. Our core focus is medical and nursing students, because improving education of these front-line healthcare workers has been shown to improve healthcare outcomes. According to the International Medical Education Directory (IMED), there are 2,400+ medical schools in the world with a total enrollment of ~500,000 medical students. According to the American Association of Colleges of Nursing (AACN), there are 1,900+ US and Canadian nursing schools - whose combined annual enrollment is 100,000 students (~350,000 students at any given time). With enrollment in health care training programs increasing year-on-year, these ~1m medical and nursing students are our target market over the next 12 months.

The biggest hurdle to overcome is fighting the inertia of the status quo medical education system. We have received incredible reception from medical students, but strive to make learning by Osmosis even easier to adopt. We have taken on this challenge by simply listening to our users! We’ve run informal focus group-style conversations, formal surveys, and even employed live chat on the site to engage with thousands of our users; and the platform has improved immensely. Moving forward, we’ll continue to work closely with future physicians to identify ways to make clinical learning more effective and research the impact that has on patient lives.

1 comment

Join the conversation:

Photo of David

Hello Vishal, 
You have selected a very clever name for your technology. What source of medical education content does Omosis use? Traditional textbook content? Experience-based learning? During my first year in medical school, students were told that they would learn 50,000 facts in the first year alone, and I wonder how much that number has grown over the years. Absorbing facts is key to the medical school experience in passing licensing examinations, but in clinical practice, observational skills and medical decision-making are very important, and perhaps more important than just factual knowledge.

During my first year in medical school an instructor made a startling comment, which I have not forgotten: 90% of medicine is practiced by rote memory and last bad mistake, which unfortunately can be common in the habit-based practices that are common to humans.

Do you work with heuristic models? Is your primary target medical students, or do you have designs to expand to all physicians? We find peer review to be an excellent educational model, as physicians are provided with an opportunity to study their case management in hindsight and adjust/correct bad habits. We have thousands of unique case reviews that can be used for teaching on a national scale, and I would be interested to speak with you further to see if there is an opportunity for Osmosis to use this type of teaching material. David  djadwin@colubmia-analytics.com