hamidghanbari | Hamid Ghanbari, MD, MPH

Atrial fibrillation clinic of the future- Dr. Tarakji

hamidghanbari — Wed, 20 Oct 2021 19:34:30 +0000

Atrial fibrillation clinic of the future- Dr. Tarakji

New care models for the management of AF

Dr. Ghanbari and Dr. Tarakji discuss a recently published manuscript that presents a new care model for following patients after AF ablation that uses a smartphone ECG coupled with a novel cloud-based platform

Back to All Posts

AI enabled ECG interpretation- Drs Noseworthy and Kashou

hamidghanbari — Wed, 20 Oct 2021 19:22:53 +0000

AI enabled ECG interpretation- Drs Noseworthy and Kashou

A comprehensive artificial intelligence–enabled electrocardiogram interpretation program

Interview with Dr. Peter Noseworthy and Dr. Anthony Kashou covering their article “A comprehensive artificial intelligence–enabled electrocardiogram interpretation program”

Back to All Posts

Clinical Utilization of Digital Health Technology- Elaine Wan

hamidghanbari — Fri, 15 Oct 2021 20:03:09 +0000

Clinical Utilization of Digital Health Technology- Elaine Wan

A Conversation with Dr Elaine Wan

Guest Dr. Elaine Wan joins host Dr. Hamid Ghanbari to discuss the HRS White Paper on Clinical Utilization of Digital Health Technology.

Back to All Posts

Machine Learning Resources

hamidghanbari — Fri, 15 Oct 2021 19:24:47 +0000

Machine Learning Resources

Best of free machine learning resources

Introduction to statistical learning

https://www.statlearning.com/

Cool videos on Machine learning

https://www.youtube.com/watch?v=0MNVhXEX9to&feature=youtu.be

Machine Learning Street Talk

https://www.youtube.com/channel/UCMLtBahI5DMrt0NPvDSoIRQ

Alfredo Canziani

https://www.youtube.com/channel/UCupQLyNchb9-2Z5lmUOIijw

Yanic Kilcher- ML paper reviews

https://www.youtube.com/c/YannicKilcher/videos

Deep Learning Berkley

https://www.youtube.com/playlist?list=PL_iWQOsE6TfVmKkQHucjPAoRtIJYt8a5

UC berkeley

Designing, Visualizing, and Understanding Deep Neural Networks.

UC Berkeley.
100% Free.
Spring 2021.

https://www.youtube.com/playlist?list=PLuv1FSpHurUevSXe_k0S7Onh6ruL-_NNh

Made with ML

https://madewithml.com/

Applied Machine Learning

https://youtube.com/watch?v=vcE9WGbi4QY&list=PL2UML_KCiC0UlY7iCQDSiGDMovaupqc83&ab_channel=VolodymyrKuleshov…

Intro to Deep Learning

https://youtube.com/playlist?list=PL_iWQOsE6TfVmKkQHucjPAoRtIJYt8a5A…

Full Stack Deep Learning

https://course.fullstackdeeplearning.com

CS 329S: Machine Learning Systems Design

https://stanford-cs329s.github.io/syllabus.html

Andy Lukyaneko- Good place for ML paper reviews

https://andlukyane.com

Applied Machine Learning.

Cornell Tech, 2020.

100% Free.

80 videos.

https://youtube.com/playlist?list=PL2UML_KCiC0UlY7iCQDSiGDMovaupqc83

Machine learning examples with code

https://towardsai.net/p/machine-learning/machine-learning-algorithms-for-beginners-with-python-code-examples-ml-19c6afd60daa

Machine learning glossary

https://ml-cheatsheet.readthedocs.io/en/latest/

Interpretable machines

alignmentforum.org

ML notes

https://theptrk.files.wordpress.com/2020/02/aicartoons_week_1_5.pdf

Trending machine learning papers

http://www.arxiv-sanity.com/

https://42papers.com/

https://papers.labml.ai/papers/daily

Collages for ML papers

https://mobile.twitter.com/hashtag/mlcollage

Microsoft ML for beginners

https://github.com/microsoft/ML-For-Beginners

ML Blog

https://eugeneyan.com/start-here/

Math

https://twitter.com/prasoonpratham/status/1428325796431777792?s=21

Deep learning drizzle

https://deep-learning-drizzle.github.io/

Open resources for imaging

https://www.frontiersin.org/articles/10.3389/fcvm.2020.00025/full#T7

Yan LeCunn Course on DL

https://cds.nyu.edu/deep-learning/

https://probability4datascience.com/

Back to All Posts

Components of Innovation Ecosystem

hamidghanbari — Wed, 02 Oct 2019 09:34:44 +0000

Components of Innovation Ecosystem

Design Principles for an Innovation Ecosystem

We leveraged the existing design principles for building an effective innovation ecosystem within and outside of healthcare. These institutions include Proctor and Gamble’s “open innovation” model, Agency for Healthcare Research and Quality, Institute for Healthcare Improvement, Bromford Labs, Kaiser Permanente, Cleveland Clinic, Mayo Clinic, Ascension Health, and Stanford Health Care. We adopted the following key components of an effective innovation challenge in the design of CVC Innovation Ecosystem:

1. Pace: A predictable rhythm for producing new knowledge

We will establish an innovation framework to move from creating and testing to scaling ideas.

We will utilize the 90-day sprints to execute the plan, do, study, act (PDSA) framework for each innovation project.

2. Staff with dedicated team:

The dedicated staff initially is comprised of a five-member team with dedicated effort ranging from 20%-100% dedicated to the innovation ecosystem.

Each innovation project will be staffed by a four-person teams: a “lead” who had primary responsibility for the 90-day project cycles; a program manager; and two members of the innovation core team as thought partners. External advisors, “helpers” or UM faculty members might contribute to an innovation project, but most projects will be resourced with the four-person team to avoid responsibility dilution.

3. Work out loud: Share your work and prototype for rapid feedback

We will use open source tools to allow sharing and exchange of ideas.

Innovation Mission Control: We will use open source software (ie Trello) to create a project outline and to do list that can be viewed by everyone. The monthly meetings between the core team and project team are open to everyone at UM who would like to attend (remotely).

We will disseminate the ideas and methods developed by the innovation academy through newsletters, publishing online and in peer-reviewed journal

4. Accountability:

Innovation core team will organize weekly team meetings with the project team to discuss the sticking points of each project and challenge each member to reach further.

The core team will meet every six to eight weeks to discuss the progress of each project with UM senior leaders and content area experts. The core team will seek independent and candor feedback on their progress, contacts for further exploration or testing, ideas for redirection if necessary, and links within the organization.

Each project will be evaluated based on a combination of the following form of accountability:

Results accountability: Did you deliver the predicted outcome? It is ideal for projects where there is a precedence and predictions are reliable.

Action accountability: Did you execute the plan well? Ideal when the predictions are less reliable.

Learning accountability: Did you follow a rigorous learning process? Ideal when predictions are less reliable.

We will also track the following innovation key performance indexes (KPI):

Reporting KPIs (innovation teams, ideas generated, experiments, assumptions tested and validated

Governance KPIs (how well in the innovation team is executing its objectives)

Global KPIs (help evaluate the overall performance of investment in the innovation program)

5. Education

Each project team will be assigned a tailored educational program designed to facilitate the success of the innovation project

6. Communicate organization-wide understanding of the innovation projects

Ensure that all innovation projects are aligned with the strategic goals of the institution.

Each innovation project has a standard charter, including the aim, rationale, background, and all intended deliverables for the innovation project. These charters will be posted electronically and accessible to all UM staff. This will help the operational engine adopt successful innovation projects.

The core team will facilitate the partnership between the project team and the operational team within our institution. This will help ensure the project success as well as organization wide adoption and implementation of successful projects.

7. Specific Plan and Budget for each project within the innovation portfolio

Each innovation project requires its own dedicated, separate, stand-alone plan.

The innovation team will manage a portfolio of innovation projects that are different stages of development.

8. Run a disciplined experiment

Each innovation project is evaluated rigorously using the scientific method.

We will create a hypothesis of record that can be tested for each innovation project. The hypothesis of record is composed of set of conjectures about cause and effect relationships between actions, outcomes, and subsequent outcomes.

If the predicted and actual trends are similar, the cause and effect relationship is validated. Otherwise the hypothesis of record is revised and upgraded.

Each innovation project is evaluated rigorously using the scientific method.

Back to All Posts

Innovation road map

hamidghanbari — Tue, 17 Sep 2019 21:09:58 +0000

Innovation road map

Innovation Roadmap at University of Michigan.

Although United States leads the world in health care spending, it lags behind its peers in many quality measures. To ensure a sustainable future for health care, it needs to be proactive, predictive, hyper-personalized, digical (seamless interplay of digital and physical), decentralized, continuous, integrated, people-powered and value based.

The transition for reimbursement from volume to value is inevitable, and disciplined innovation can drive it forward. Organizations that succeed in adopting to this landscape need to draw from diverse perspectives, including patients, families, caregivers and payers and continually innovate to deliver high value care.

Few healthcare organizations perform this in an orderly and reliable way. Innovations in healthcare typically occur in silos lead by faculty and staff champions armed with specific expertise who are uniquely positioned to innovate within their environment. These projects rely on luck and incredible efforts by innovation champions to succeed. Therefore, many innovative projects remain captive in the minds of the employees and ideas that are developed are not the consistent with the highest strategic priorities of their organizations. For innovation to succeed, there is a need for new organizational structures combined with substantial investment to ensure innovation occurs consistently and at scale.

The University of Michigan Frankel Cardiovascular Center (CVC) Value Innovation Program is an example of such successful initiative that leverages informational technology to engage the community within the cardiovascular center and develop innovation projects that maximize value within our institution. The program is supported by a generous foundation grant which allows for hiring a dedicated staff. The dedicated staff will develop a core modular educational curriculum that can be tailored based on the needs of the innovation project (innovation academy). The dedicated team will also help execute and implement the selected projects over the year following the innovation challenge.

What is Value in Health Care?

Value in healthcare should be centered around the needs of patients. Value is defined by healthcare outcomes achieved per dollars spent (value=outcomes/costs). Therefore, value in healthcare is defined by results (health outcomes) and not inputs (volume, process of care used). Health outcomes for each patient are condition-specific and no single outcome captures the results of care. When measuring outcomes, we will utilize measures that assess health status achieved or retained, patient/provider experience, process of recovery, and sustainability of health. Cost refers to the total costs of the full cycle of care for the patient’s medical condition, not the cost of individual services. We will utilize time-driven activity based costing to assess the cost of each innovation project.

What is innovation?

There is significant emphasis on ideas in healthcare and that is often mistaken for innovation. Innovation represents the implementation of new or significantly improved products, services or processes. The ultimate measure of success for innovation is the development of new products with a sustainable and profitable business model. A business model is sustainable if the products deliver value to customers (the product is desirable by customers) and are profitable (some value is captured by the institution).

We use a rigorous definition of innovation that emphasizes the ultimate goal of innovation which is creation of value for patients.

We will define healthcare innovation as the implementation of new or altered products, services, processes, systems, policies, organizational structures, or business models for the purpose of creating value for customers and financial returns for the institution.

Innovation= invention (ideas) X commercialization (create value in the world)

Back to All Posts

The Hard Truth About Innovation Cultures

hamidghanbari — Sun, 30 Dec 2018 12:55:37 +0000

The Hard Truth About Innovation Cultures

From The Hard Truth About Innovative Cultures

The most common characteristics of innovative cultures are believed to be:

Tolerance to failure
Willingness to experiment
Psychological safety
Highly collaborative and non hierarchical

But these cultures are not all fun and games.

The easy-to-like behaviors that get so much attention are only one side of the coin. They must be counterbalanced by some tougher and frankly less fun behaviors. Innovative cultures are paradoxical and need to be carefully managed. These are the important paradoxes that need to be managed:

1. Tolerance for Failure but No Tolerance for Incompetence

2. Willingness to Experiment but Highly Disciplined

3. Psychologically Safe but Brutally Candid

4. Collaboration but with Individual Accountability

5. Flat but Strong Leadership

If you want your organization to strike the delicate balance required, then you as a leader must demonstrate the ability to strike that balance yourself.

Back to All Posts

Why predictions about the future of AI are wrong

hamidghanbari — Tue, 25 Sep 2018 15:52:43 +0000

Why predictions about the future of AI are wrong

I came across this great conversation with Rodney Brooks on econtalk on why many people are wrong about the future of AI with his accompanying MIT technology article. He argues that the following seven reasons lead to wrong predictions about the future of AI:

Overestimating and underestimating (Amara’s Law: We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run)
Imagining magic (If it is far enough away from the technology we have and understand today, then we do not know its limitations. And if it becomes indistinguishable from magic, anything one says about it is no longer falsifiable. Nothing in the universe is without limit.)
Performance vs competence (People hear that some robot or some AI system has performed some task. They then generalize from that performance to a competence that a person performing the same task could be expected to have. And they apply that generalization to the robot or AI system)
Suitcase words (When people hear that machine learning is making great strides in some new domain, they tend to use as a mental model the way in which a person would learn that new domain. However, machine learning is very brittle, and it requires lots of preparation by human researchers or engineers, special-purpose coding, special-purpose sets of training data, and a custom learning structure for each new problem domain.)
Exponentials (When people are suffering from exponentialism, they may think that the exponentials they use to justify an argument are going to continue apace. But Moore’s Law and other seemingly exponential laws can fail because they were not truly exponential in the first place.)
Hollywood scenarios (The plot for many Hollywood science fiction movies is that the world is just as it is today, except for one new twist. They ignore the fact that if we are able to eventually build such smart devices, the world will have changed significantly by then. We will not suddenly be surprised by the existence of such super-intelligences. They will evolve technologically over time, and our world will come to be populated by many other intelligences, and we will have lots of experience already.)
Speed of deployment (Capital costs keep physical hardware around for a long time, even when there are high-tech aspects to it, and even when it has an existential mission. Almost all innovations in robotics and AI take far, far, longer to be really widely deployed than people in the field and outside the field imagine.)

Back to All Posts

Data as labor

hamidghanbari — Fri, 21 Sep 2018 17:29:46 +0000

Data as labor

What if people paid for their data- Economist

Data labor: where you think about data provided by humans as a form of labor which powers artificial intelligence. But most individuals lack the bargaining power to negotiate a fair value for their data. The history of labour offers a hint about how things could evolve. Historically, if wages rose to acceptable levels, it was mostly due to unions. Similarly,“data-labor unions”, organisations that serve as gatekeepers of people’s data. They will negotiate rates, monitor members’ data work and ensure the quality of their digital output, for instance by keeping reputation scores. Unions could funnel specialist data work to their members and even organise strikes, for instance by blocking access to exert influence on a company employing its members’ data. Similarly, data unions could be conduits channelling members’ data contributions, all while tracking them and billing AI firms that benefit from them. Current obstacles to this include 1) lack of right legal framework 2) technology to keep track of data flows needs to be become better 3) development of class consciousness by people as data workers. It is likely that active forms of data labor (ex image labelling) and will be compensated better than passive forms of data labor.

Back to All Posts

Fears, uncertainties and doubts

hamidghanbari — Tue, 19 Jun 2018 12:11:45 +0000

Fears, uncertainties and doubts

There are many reasons why people do not live a creative life but fears, uncertainties and doubts (FUDS) rank very high on the list.

I have learned through my career that FUDS can be disabling. But you can also harness their energy to transform negative ideas into a positive and creative reality. You cannot ignore them. You have to acknowledge them and then focus aggressively on what you want to replace them with.

Psychologist, Peter Lecky, found that there were two powerful levers for changing and overcoming beliefs (it is always a relief to remember that everyone struggles with fear even the most creative and successful people among us):

1. The belief that one is capable of doing one’s share by exerting a certain amount of independence (agency). The belief that you are the “hero” in your story.

2. The belief that there is something inside us that makes us equal in talent and ability to the rest of the world (creative confidence)

Practical tips for dealing with FUDS

Tick Tock Exercise

This is a powerful exercise designed to help you overcome FUDS.

1. Zero in on and write down those negative thoughts that are preventing you from realizing your goals. (Tick)

2. Sit quietly and examine the negatives. Learn how you are irrationally twisting things and blowing them out of proportion.

3. Substitute objective positive thought for subjective negative one. (Tock)

Tick	Tock
My idea is so stupid	This is all or nothing thinking. The idea does not have to be blockbuster. If I was in leadership wouldn’t i want all the ideas to come to me
I’ll never be able to do it	Just do a little bit at a time and get started

This fear setting exercise can help in all aspects of life but it is very important in creative life when FUDS shows its ugly head. This is very similar to what Tim Ferris talks about in his TED talk. He has a slightly different approach. He advocates a three page exercise. On the first page you define all the worst case scenarios (Define), what you could do to prevent/reduce the effects of worst case scenarios (Prevent), and what you could do to repair the damage if it occurs (Repair). On the second page you write down the benefits of an attempt or partial success. Lastly and perhaps the most important part is to think about the price of inaction (emotionally, physically, financially, etc). This can be a very powerful method if employed consistently when faced with FUDS.

https://www.ted.com/talks/tim_ferriss_why_you_should_define_your_fears_instead_of_your_goals

I often hear people saying that they are not the creative type. I used to believe that about myself for a long time. I never thought I was the creative type. But with time, this slowly changed. I realized that entrenched beliefs about the world can distort our perception of reality. There are a variety of ways to change the belief system of those who do not think they are creative. Here are a couple of proven methods:

1. Self-Affirmation: get in the habit of remembering your successes, good qualities, and characteristics. Forget your failures. Success breeds success. what matters is the successful attempt. Keep log and update it regularly. These successes can be all arenas of life: home, work, family, …

2. Creative affirmation: Human beings act, feel and perform in accordance to what they imagine to be true about themselves and environment. Start by writing a general statement (ie I am creative) . take a few moments and write several variations of this statement about creativity. Then take one and expand on it.

I think it is very helpful to have some useful tools and methods to deal with FUDS wherever you face them. So next time you face them you know what to do.

References:

Cracking Creativity: The Secrets of Creative Genius

Back to All Posts