Joint replacement surgery is one of the most common in the US with more than 1 million Americans undergoing the operation each year. Recent evidence suggests that surgical outcomes for this common operation and many others can improve when patients undergo pre-habilitation (or ‘PreHab’) in the weeks leading up to surgery. For example, a PreHab program might include exercises to strengthen the muscles around the joint that will be replaced, an improved nutrition regimen, or practicing mindfulness training to help manage pain.

To help patients improve surgical outcomes with PreHab Cal alum Manish Shah (CS ’05) has launched a new startup called PeerWell which uses machine learning tools to create personalized conditioning plans for patients. (Earlier in his career, Shah co-founded the data/marketing company RapLeaf, which later became LiveRamp, along with fellow Cal alum Auren Hoffman (IEOR ’96). The company was later sold for $310 million.)

There is one particular group where surgery is more common than average, and where the outcome can have a huge impact on their careers — pro athletes. PeerWell has been reaching out to the pros to invite them to try the PreHab program. NFL stars Arian Foster, Glover Quin and Duane Brown have all publicly acknowledged PeerWell’s role in helping them to achieve positive results with their own surgeries.

In fact, Foster was so impressed that he became an investor in the company.

PeerWell has been focusing on pro athletes to learn how they prepare for surgeries, so that they can get back to playing shape as quickly as possible. But Shah has bigger plans for PeerWell besides just helping pro athletes get back on to the field. The goal is to bring the program to the millions of people that undergo surgery each year who are not pro athletes but have just as much reason to get back to their lives after an operation. In addition to helping the 1 million joint replacement surgery patients each year, PeerWell wants to expand its application to work for almost any type of medical procedure.

To learn more about how PeerWell works and plans for the future, we followed up with Manish to ask a few questions about his new venture.

SCET: What made you interested in focusing on improving surgical outcomes?

Manish Shah: I’ve spent over three years researching healthcare to understand which factors impact the quality of care patients receive. I’ve also been interested in understanding what makes the biggest difference in terms of reducing costs to look for ways that healthcare services can be more accessible to people in need. Costs are a huge impediment to people getting cared for the way you would expect in 2016 in the US. In our exploration of healthcare and the specific needs stakeholders (i.e. patients, doctors, hospitals, payers etc.) in the system have, we found that by a large margin, the place to improve things is around surgical outcomes.

We often hear about chronic disease, and it dominates discussions in healthcare reform. What we learned from doctors and hospitals is that this is not where improvements should be made. While chronic disease spending accounts for 80% of costs, the vast majority of that is around end-of-life and high-touch intervention (e.g. surgeries). In fact, in the US studies show that the cost of one surgery is on average the same as thirty-three years of costs associated with diabetes.

SCET: How are you using data / machine learning / AI to optimize the PreHab program?

MS:  There are a few different ways we are using AI in our PreHab program:

1. Since each person has a different amount of time for PreHab before their surgery, we can update which activities they are assigned to match the time they have available. For example, some people won’t have time for losing weight and instead should focus on quitting smoking or blood sugar control.
2. We adapt the program in-place as we learn from people. We collect various data points during PreHab from sensors and clinical assessments we’ve incorporated into the experience. The assessments measure things like pain tolerance, depression, anxiety, etc. We use these data points to learn where PreHab can have the biggest impact for this person and assign daily activities based on that prediction.
3. PeerWell is a peer-based system so the system learns collectively what works for people with different profiles. We can identify people that are most similar to the current person and see what has had an impact on that cohort at a given stage in the program. We can also look historically to see how others have done pre/post surgery to identify which activities yield the best outcome.
4. We use data to predict time to recovery, success of being discharged back to home, and likelihood of success for having a surgery done at a surgery center versus a hospital.

SCET: One unique thing about PeerWell seems to be that you match individuals with others who are going through the same process.  What made you think of this? Does this help improve PreHab success rates?

MS:  This is key. We use data from “peers” to adjust the program as people are going through it. But, we also connect patients with others to share experiences and stories with each other. This helps them alleviate some of the anxiety prior to the procedure because they hear from someone like them that just had the same experience.

This is a behavior we found many patients were doing already. Patients actively seek out other patients to learn from their experience. Additionally, since surgery is such a big event in people’s lives, they want to talk about it — even long after it is over and done with. So patients have a strong urge to share their own experiences and to hear those of other patients. PeerWell provides that social layer to make this exchange possible, and also incorporates it into the program to help motivate people to do more.

Doctors can even use this feature to enlist their patients as new advocates and mentors to help new patients go through PreHab and rehab in a more social way.

SCET: Do you track surgical outcomes (and relate them to PreHab data somehow?) 

MS:  We are optimizing for a few clinical outcomes that impact patient experience, health outcomes, and cost:

1. How long does the patient stay in a hospital before going home?
2. When they leave the hospital, can they go straight home, or do they need to go to an in-patient rehab facility?

The longer the patient stays “in the system” the longer their recovery and more prone to infections, slow rehab, etc.

With our data and our upcoming rehab program, we will also be tracking mean time to recovery based on range of motion detection and other physical assessments we are building into the platform. This is entirely new data to healthcare. We will be able to show patients, doctors, and hospitals how long it takes for a patient to walk again after surgery.

SCET: You mention that you plan to expand PeerWell to all 50 million surgical procedures done in the US annually. Is this focused on knee and hip replacement only, or are there other procedures where this kind of program could make a big difference? What is the roll out plan?

MS:  We are starting out with prehab for orthopedics, specifically with total joint replacement surgery. This is due to healthcare policy changes that were put in place this year. There are one million of these procedures done each year in the US and growing fast (about 15% CAGR). PreHab is actually applicable across the board in healthcare for major medical procedures.

We’ve been approached to use our PreHab program for heart patients and cancer patients. The objective is to make PreHab the standard of care with any major medical procedure, because it helps patients get better results.

SCET:  You’ve already started (and sold) a company successfully since leaving Cal. What were the biggest lessons you learned that you are now applying with PeerWell?

MS:  There are a lot but a few to note here:

1. Data is important and builds a long-term, durable defense to the business. We had many companies try to copy LiveRamp but failed because they could not compete against our data. PeerWell is being built from the ground up as a data company. We know how to create and apply leverage with the data we collect. We know exactly where the data should be applied to enable decision support. This makes us invaluable to our customers and we make it easier and easier for them to lean on us for data that we are the best at collecting.

2. Sales-cycles are long.  It’s important to be prepared ahead of time and work to reduce any time you can from the sales process. Don’t ever be your own bottleneck in an enterprise sales process.

3. LiveRamp created a new category in digital marketing, “CRM Retargeting”. It’s obvious now, but wasn’t back then. PreHab is a similar opportunity in healthcare. Data and policy changes now make it possible, and PeerWell is in the pole position.

SCET: Do you have any other general advice for entrepreneurship students at Cal who are just starting up?

1. It matters a lot who you work with and start your company with — especially the first 10 people of the company. This will make or break the company. Teams that don’t work well together or are not committed to each other will not last through the gauntlet a startup has to fight through in its first couple of years. This includes investors you add to your extended team.
2. Don’t burn cash. Spend every dollar as if it’s your last.
3. Don’t be afraid to take on something big. Even if its a new field for you or doesn’t fit the pattern of what you’ve been doing. It’s often the people that can translate between domains that find a new and better way of doing things.
4. Don’t pay attention to people on the sidelines. Talk to and learn from people who are involved in building an organization. They are much harder to meet with, but the value per minute is inordinately bigger.
5. Exercise often, sleep well, be nice, love what you do and who you work with. The rest will take care of itself.