Enabling value-based care

The problem we are addressing is a huge global problem of increasing healthcare costs in the US. Access to Healthcare should be a human right, yet it is commoditized to the extent that processes are set up to slow the process if not block them completely.

Prior authorization lies at the crux of the problem as this process has severe cost and time implications. The current process is too slow, manual, and riddled with errors due to the manual nature of the process. It takes 3x the time it should take for the approval of certain life-saving drugs and procedures. It also eats away at the precious time of doctors that could have been otherwise spent in treating patients.

The process also leads to distrust between payers and providers as the providers constantly feel that the payers deny the requests because they do not want to reimburse. However, payers claim that the process helps in managing costs. The process is subjective and requires a deep understanding of the patient's conditions and biomarkers. Hence, it is difficult to bring consensus between providers and payers.

Our solution is a Prior Authorization web platform for health systems (providers) and health plans (payers) to bring objectivity to the process by removing the redundant and manual aspects of the process. We use machine learning to detect bad patient outcomes in case the prior authorization requests are not completed in the stipulated time, thus bringing providers and payers to the common ground. Prior authorization is also a wedge to the larger domain of utilization management for health plans (payers) and health systems (providers), which is a gateway to making healthcare cheaper for everyone.

Our unique value proposition includes a combination of our technology, provider partnerships, and unique business model. We have filed a patent for our machine-learning disease detection and progression platform. Within a year of inception, we have partnered with some of the largest healthcare organizations like the Mayo Clinic, Joslin Diabetes Center (the largest diabetes institute globally), and Beth Israel Deaconess Medical Center. Finally, we have defined a hybrid business model for payers and providers that provides the flexibility of a risk-sharing payment model (for Medicare and Medicaid patients that have higher volumes of prior authorization requests) and a volume-based payment model (for corporate patients with fewer prior authorization requests).

At the macro level, the solution is designed to make prior authorization streamlined. The current process of prior authorization is quite manual and takes a lot of time. We help by automating a significant part of the process and by using machine learning to make data-backed recommendations for improving the process.

At the micro level, our solution will make the process of receiving care for patients and their caregivers less stressful process. More specifically, we help in increasing the chances of getting reimbursed by insurance and in reducing the turnaround time for approval by collating all the data points needed. We also help in triaging care to patients so that everyone receives the care they deserve irrespective of their race, gender, ethnicity, or socioeconomic status.

For doctors, we help in automating a lot of tasks that are redundant and provide them with actionable summaries for each patient, thus letting them focus on the actual care.

Last, we also work with payers and providers to reduce the friction between them. This leads to building more trust in the system which helps in expediting the claim approval process.

My introduction to healthcare was through my dad's diabetes and cardiovascular health. I was his primary caregiver and saw the struggles first-hand. First, there were not clear enough guidelines to follow for keeping his glucose levels in check and continuous glucose monitoring (a good solution for keeping a real-time check on the glucose levels) was unaffordable for us. The doctors had limited time to consult and never really spent enough time to understand what is going wrong. Instead, they used to give very generic advice. I would not blame them as I know they have very limited time to treat patients.

I have also worked with United Nations Development Program (UNDP) on a project aimed to make migration safer for migrants in Africa and central Asia. It helped me appreciate the need and affordability of healthcare for all. During this project, I interviewed more than 100 migrants and developed an algorithm to recommend safer pathways of migration. Our algorithms also helped health workers prioritize interventions for the migrants based on the urgency and needs of migrants.

Through my organization, basys.ai, we have collaborated with the doctors at Mayo Clinic, Joslin Diabetes Center, Indian Healthcare Service (IHS), and Beth Israel Deaconess Medical Center to help in delivering better care through the use of technology. I would particularly like to underscore our work with IHS, where we have come up with algorithms to identify social determinants of health to reduce biases in healthcare.

Our work and mission at basys.ai to make healthcare affordable and accessible have been spotlighted by Harvard Kennedy School (HKS). I am fortunate to be one of the sixteen Cheng fellows for the Social Innovation Change Initiative at HKS. This provides us with a great network of mentors, public leaders, and peers to lean on and learn from their collective experiences.

In the past few years, I have volunteered to help organizations working on women's safety like SafeCity and Pingaa to help them with the development of algorithms for reducing gender-based violence and crimes. These organizations provide a great support ecosystem.

Navigating bureaucracy: Finally, based on what I have seen in healthcare, there is a lot of ingrained bureaucracy. One of the ways of navigating it is through understanding the system and proposing a solution that is not only social but also impacts the key performance indices that healthcare organizations care about. At basys.ai, we realize and appreciate that in order to realize the true potential of the impact we want to bring about, we would have to be aware and active of such nuances.

100+ doctors, 1k+ other clinical staff, and 1M+ patients

A platform for spotlighting the ideas we care about: Solve would be a great problem to not only spotlight the ideas but to also help us in recruiting individuals passionate about making healthcare accessible.

Strategic partnerships: It would also help in forging more strategic partnerships with organizations that care about the problems we care about. Besides, it would give us access to an amazing network of peers, mentors, and former awardees.

Technical expertise: Given we are a technology organization and MIT is probably the best place to scout for technology talent, we would like to tap the resources MIT has to offer.  We are already working with Professor Ghassemi, who has been extremely brilliant and kind in supporting our initiative as an advisor.

Staying on top of legal processes: We would also like to get some help with legal matters as this is something we have a limited understanding of. It takes up significant time from our schedule and we would love to piggyback off of the mentors' expertise to navigate this complex process.

Codeveloped with clinicians: We are extremely aware and conscious of the involvement of clinicians in setting up basic frameworks for improving patient outcomes and access. So, we have developed our platform with the clinicians including a consortium of doctors who played a pivotal role in establishing the initial partnerships.

Adoption by major healthcare organizations: We are grateful that our platform has been adopted by some of the largest healthcare organizations like Mayo Clinic, Joslin, and Beth Israel Deaconess Medical Center.

Support by healthcare and data science experts: We have been supported by our advisors in healthcare and data science like James Roosevelt, Jr., Marzyeh Ghassemi, and Leo Anthony Celi.

IP: We have filed a patent for our machine learning platform for improving patient outcomes. We have also developed an easy-to-use platform for doctors so that they can focus on treating patients.

Using data to determine social determinants of health and drive action: We use geocoding while maintaining patient privacy to determine social determinants of health and flag the potential biases to doctors. Based on our work so far, we have seen that nudges can help pivot healthcare to be fairer.

Impact goals for the next year: 

Demonstrate results with the existing partners.

1. Automation of the prior authorization pipeline and reduce costs by 22%

2. Reduce prior authorization volume by 10%

3. Increase accessibility to 10% more population

Link to the first whitepaper released on our work with Joslin on reducing costs while treating diabetic retinopathy. We have to release more whitepapers on reducing costs and increasing accessibility for cardiovascular diseases, surgery, and cancer.

Impact goals for the next five years: 

Scale the results nationally, to new geographies and audiences.

1. Show the results can scale nationally across 500 mid-sized providers, 50 large providers, 30 mid-sized payers, and 5 large payers

2. Demonstrate material improvement to accessibility of healthcare to underserved populations. We are working with Indian Healthcare Service to define the annual targets for the same.

How will I achieve them?

The model of engagement with payers and providers is that we help these organizations in improving patient outcomes and reduce costs by automating manual processes like prior authorization. We charge payers and our product is free for the providers.

For a more detailed business plan, you may also check out this deck.

We measure the progress toward impact goals using the following key performance indices:

Patient outcomes: We use machine learning to detect pathologies like heart failure earlier and recommend intervention strategies. We would soon be releasing the work we did with Beth Israel Deaconess Medical Center and Joslin Diabetes Center.

Cheaper healthcare: As an organizational commitment, we are going to semi-annually release our work and its cost benefits to healthcare. Our first research with Joslin on improving patient outcomes and reducing cost (by 22%) for patients with diabetic retinopathy was recently released.

More inclusive healthcare: In our work with Indian Healthcare Service (IHS) and Joslin Diabetes Center, the work is supposed to triage care in low-resource settings to facilitate better distribution of limited available resources. We are currently working on quantifying the result of our first pilot with IHS.

Input:

People: Co-founders, who are data scientists and healthcare professionals + software and data science team + sales team + Chief Medical Officer

Knowledge: data science and healthcare

Financial Resources: $3M for 2 years

Partnerships:  Health Systems (Providers) and Health Plans (payers)

Activities:

Released the results of phase 1 of the partnership with Joslin on improving patient outcomes and reducing procedural costs

Pilot with BIDMC to paid partnership

New customer and data acquisition: Mayo Clinic and Point32 Health

Output (so far):

Patient outcome: Better Glycated hemoglobin/HbA1c (reduction of 1.6 percentage points) for patients with diabetes

Time saved: More time saved for doctors (~40%)

Better information flow and reduced friction between providers and payers, resulting in reduced turnaround time for processes like prior authorization (reduction by one-third) 

Outcomes:

A better lifestyle for patients and reduced anxiety for patients and caregivers

More time saved for doctors (which could be used to treat more patients or more time could be spent with each patient in delivering more personalized and better care) 

More cost-saving for payers and healthcare in general, which can be translated to cheaper healthcare for patients

Impact: 

Individual level: Better patient outcomes and non-inflated impact of diseases, unlike the status-quo

Aggregate level: Better and equitable access to healthcare

System level: More emphasis on research and less operational burden

Our core technology is two-fold:

1. Simple automation as low-hanging fruits: A lot of healthcare practices require simple innovation like rule-based engines and basic API-based communication to reduce manual and repetitive tasks. This in itself alleviates a lot of burden off the clinicians. We use as many simple and transparent approaches as we can.

2. Machine Learning for complex tasks: Given the advances in data science and the vastness of data with health systems, there is a huge scope for using data to stay on top of care delivery. We work with some of the amazing researchers in machine learning and healthcare to define product pipelines for healthcare.

We have filed a patent for our proprietary platform and machine learning algorithms for detecting the progression of diabetes and making personalized recommendations. Our product launch was covered by Forbes and Boston Business Journal.

Our team of 10 members represents a diverse range of nationalities (4), races, ethnicities, and socioeconomic profiles, creating a rich and inclusive workplace culture. We all respect and celebrate each others' diverse backgrounds.

Diversity is also one of the five pillars of our startup among integrity, passion, accountability, and collaboration. I am proud of my team's commitment and the way we support each other. We would want to be more gender-diverse though as we are 7 male team members currently.

As we expand, it would not be easy to maintain the same level of diversity. We have to constantly keep reminding ourselves that all the effort made so far will go in vain if we do not stick to our commitment of maintaining an inclusive environment for everyone.

We charge health plans (payers) on a hybrid model:

1. For Medicare and Medicaid populations, we offer a risk-sharing subscription-based model on a per-member-per-month (pmpm) basis.

2. For employer insurance, we charge on the volume of usage (number of prior authorizations processed)

The reasoning behind the hybrid approach is that given the high volume of prior authorization requests for Medicare and Medicaid patients, a risk-sharing model makes more economic and actuarial sense. For employers, there are not so many prior authorization requests and a volume-based approach provides more transparency to the health plans on what they are spending for.

We are getting revenue from our customers like Joslin Diabetes Center for improving their patient outcomes and operational efficiency.

We have two revenue streams:

1. Health Plans (Payers): Going forward, we would be scaling on health plans as they are the source of reimbursements

2. Health Systems (Providers): Although providers are constrained in their capacity to pay, they have a high potential in improving patient outcomes. For us, they also provide the strategic moat for selling into the health plans.

1. We receive annual recurring revenue ($172k) from Joslin Diabetes Center. We have received investments from Norrsken Impact Foundation and an angel investor who is the founder of the largest Covid-19 testing lab in Sweden.

2. Mayo Clinic is investing in us and they are our strategic partners. We have access to their datasets for training our machine-learning algorithms.

3. We are running a paid pilot with Beth Israel Deaconess Medical Center and soon going to close a pilot with Point32 Health.

4. Besides, we have a strong pipeline of customers including both mid and large payers e.g. Kaiser, BCBS MA, Elevance Health (Anthem), and Emblem. They are interested in the data coming out of our clinical partnerships.