Clinic Price Check

Joanne Rodrigues, the principal investigator on this project, is an experienced data scientist with master degrees in mathematics (London School of Economics), political science (U.C. Berkeley) and demography (U.C. Berkeley) and a bachelor’s degree in international economics (Georgetown University). She has a forthcoming book in Addison-Wesley’s Data and Analytics Series, Product Analytics: Applied Data Science Techniques for Actionable Consumer Insights and six years of experience applying machine learning/statistical algorithms to derive business insights (in healthcare, gaming). She managed analytics teams at Sony and led MeYou Health’s data science efforts. For this proposed work, she built the data pipeline, the machine learning algorithms to collate pricing data and the backend for the web application for Clinic Price Check.

Due to COVID-19, the United States will face a dramatically altered healthcare landscape, but with an essentially unchanged system for insurance or cost-sharing. Many Americans will need healthcare, but will have lost their employment and in many cases their insurance at the same time. Many underinsured and uninsured individuals will see their savings wiped out, face garnished wages or bankruptcy. 

Our solution is price transparency of American health providers, including list (‘sticker prices’), cash-pay and Charity Care rates, which will allow consumers to minimize their cost exposure while also putting downward pressure on providers with inappropriate cost structures, which will be critical as we attempt to overcome this pandemic. Many Americans avoid visiting the doctor, due to fear of the financial repercussions. Helping consumers access the cost information prior to receiving care can mitigate unnecessary spread of COVID-19 and save lives. 

The problem that we address is the lack of price transparency in the American health system. The proposed project could have a high-impact, because the healthcare industry has largely eschewed efforts at greater price transparency, with deleterious effects for American consumers. According to the American Cancer Association, 56% of Americans suffer from hardships related to the cost of healthcare. In 2019, 67% of bankruptcies were due to medical costs. American's healthcare costs are some of the highest in the world, averaging two and a half times other developed countries.

Due to COVID-19 over the next six months, the U.S. will face a dramatically altered healthcare landscape, but with an essentially unchanged system for insurance and cost-sharing. Many individuals will have their savings wiped out, and have lost their employment and insurance. Uninsured and underinsured Americans are wary of receiving testing and/or treatment from health providers because of the cost of care, resulting in further spread of the virus. According to a CNN interview of a doctor, one of the most jarring experiences in the pandemic was a patient that eventually died from COVID-19’s worrying in his last minutes about who would pay for his care.

Our solution providers consumers with prices for medical services at local health providers, particularly list ('sticker prices'), cash-pay (opt-out of insurance), and Charity Care (below a certain income threshold) prices. 

On January 1st, 2019, Centers for Medicare and Medicaid Services (CMS) rule (CMS-1696-F) required hospitals to post Chargemasters, a list of all goods and services with prices, to their websites. The core innovation is to use machine learning to extract, cleanse and collate Chargemaster and are pricing data into an aggregated, easy-to-compare format to help consumers make more informed decisions about their healthcare spending.  In comparison to a human expert, the machine learning methods pioneered in this project had a much higher rate of accuracy and are easily scalable. 

We offer a number of novel elements to help consumers comparison shop for health services:

1. Easy comparison of the pricing of medical services at local providers. We have over 100 times the provider density compared to other services like MDSave and ClearHealthCosts.

2. Estimation of cash-pay rates and Charity Care rates. Cash-pay and Charity Care rates are often substantially cheaper than commercial insurer negotiated rates for health services.

3. Comparison of quality information such as CMS quality ratings, national awards of excellence and technological sophistication.

Two groups of healthcare consumers will benefit from this project, the uninsured and underinsured, particularly those in High Deductible Health Plans (HDHP) members, or about 60% of the American population under 65. 

Many uninsured or low-income patients do not realize that they qualify for substantial Charity Care discounts. Health providers are at no obligation to share this information with them. Billions of dollars in health provider Charity Care goes unused yearly, even though it could potentially allow many consumers to avoid deleterious financial consequences like wage garnishment and bankruptcy. 

Many underinsured individuals in High Deductible Health Plans (HDHP) can save by paying cash-pay rates (or the rates for those who do not have insurance or opt-out of using their insurance). Cash-pay rates are on average 20% less in California, but can be as high as 95% less than insurer negotiated rates at some health providers. One user's insurance adjusted rate to see a cardiologist was $435. His provider, had a 55% self-pay discount, the cash-pay rate for the same service was $250, saving himself $185.

The majority of Americans suffer from hardships due to the cost of healthcare. In 2019, almost 70% of bankruptcies were due to medical costs. Our project relates to the dimension of "Elevating Issues" because we are fostering awareness of the lack of price transparency in healthcare. We are confronting this issue directly by using machine learning to recreate health providers pricing structures. By providing consumers with this valuable information, our organization hopes to increase competition, put downward pressure on providers with inappropriate pricing structures and prevent individuals from facing wage garnishment and bankruptcy due to their inability to afford care. 

The reason I built ClinicPriceCheck was because of my negative experience with health provider billing in the past. When my child got sick about two years ago, we had a High Deductible Healthcare Plan (HDHP) and opted to take her to an urgent care facility. Months later we received an $800 bill for a flu test that was administered during the visit. After calling the hospital multiple times, I found out if I had paid cash-pay rate, I could have gotten a 30%+ discount on the procedure.

I'm an experienced data scientist who has a passion for solving real-world (i.e. fuzzy) problems with advanced analytics. After I heard the hospital Chargemasters would be released, I knew that I could collate this data into a comparable format. From calling hospitals to negotiate rates, I knew simple patterns for how hospitals and other health providers set prices. I also knew that consumers being able to compare cash-pay rates to insurer negotiated rates would lead to much greater price transparency and massive consumer savings.

As a data scientist, I spent years working for large companies building addictive products that tacitly change consumer behavior toward company KPIs. After realizing that this sometimes negative affects consumers, I decided that I wanted to focus on using data for good. 

In health technology, insurers and large provider systems employ teams of data scientists and analysts focused on increasing revenue and lowering costs. I wanted to give consumers the same access to information to help them make more informed decisions about their healthcare spending.

I’m coming to this problem, not as an outsider, but as a mom with children in High Deductible Health Plan(HDHP) in one of the most expensive regions for medical care in the nation. My passion to solve this problem stems from facing it myself. The United States desperately needs a solution to high health costs. Healthcare price transparency is the first step. 

My background as a data scientist, with master's degrees in mathematics, demography and political science, helped me deliver on this project. 

While collating health provider prices might seems straightforward, it's decidedly not. Health providers do their best to eschew price transparency. There are three main challenges: 

• Data collection is time-consuming and technically challenging.  My background as a data scientist helped me quickly collect the data. 
• Application of machine learning is not straightforward. One must be able to make this data comparable between providers. My masters in math gave me the technical skills to solve this problem. 
• Making complex data consumer friendly. Having an undergraduate and masters in political science/economics helped me understand healthcare policy, how to make this data useful to consumers and improved my ability to write about it.

In addition, having experience in industry, working for start-ups in the healthcare space, has helped me understand how to grow the business. 

To build this product, I had to teach myself web development. For years I had wanted to start my own company, but had failed to build a web product. I had gotten stuck with learning a complex task like web development while dealing with the complexities of building a business. However, this time, I set my expectations lower and gave myself the time to learn. I think being patient with myself helped me to overcome my initial inability to learn Python web development. I started out building each component of Clinic Price Check one-by-one, like user profiles or the pricing tables. 

Over six months, I taught myself Python web development and built a production quality website. When the site when live, there were initially lots of production bugs, but I worked assiduously until I figured out why each bug was occurring. Clinic Price Check is interactive, has a basic and membership site, a number of pricing modals and advanced features like downvoting incorrect prices and personalized user features. When it was featured on NBC Bay Area, we got 2,500 users over two days and had only two production bugs. 

When I started this project, I was a stay-at-home mom of two kids. My daughter was two years old and my son was one year old. While taking care of both children full-time, I taught my self web development and used my skills as a data scientist to extract, clean and collate the data for California deployed Clinic Price Check. Clinic Price Check is actually the first website that I've built. Clinic Price Check with over a million prices for 5,000+ services at 400+ hospitals in California was built and deployed with less than $2,000 dollars. NBC Bay Area was impressed with the quality and usefulness of the site that they aired four times on their morning and nightly programs.

Clinic Price Check is a social venture and our goal is to help underinsured and uninsured individuals save money on healthcare. We are staying true to our mission. After shelter-in-place is lifted, we scheduled to hold over twenty library seminars in the San Fransisco Bay Area to listen to individuals healthcare woes and build new free features on the site to meet their needs. 

Our competitors, Castlight health, Clear Health Costs, and Healthcare Bluebook, are focused on using either claims data or shared bills to create price transparency. However, these companies have not been very effective because of the sheer number of prices in the system. A conservative estimate of the number of prices is 1.25 billion prices (15 insurer groups* 15,000 services * 5,500 hospitals). This is 4X’s as great as the number of people in the U.S. 

Clinic Price Check takes a different approach in two key ways: 

1. Use of Chargemaster data and other pricing data for services

2. Focus on different patient populations, particularly cash-pay and the uninsured

This project relies on recent release of health provider Chargemasters, per the CMS disclosure rules. We then use machine learning to extract, clean and collate Chargemaster and other pricing data into a consumer-friendly, comparative format. By removing the complexity of insurers, we are able to eliminate a large amount of pricing variation and offer our users comparable baselines, i.e. the cost if they opt-out of insurance or qualify for Charity Care. 

By fostering greater price transparency, even consumers in Cadillac health plans have the ability to compare cash-pay prices to their insurer negotiated prices, which in the long-run will reduce costs for commercial health plans.

The activities that will drive change for healthcare consumers is having a easy-to-use, real-time pricing application that will allow consumers to estimate healthcare costs at local providers, choosing the lowest-cost, highest-quality healthcare provider. 

Comparison shopping will start with cash-pay consumers. This process has already started in cities with greater hospital competition like Los Angeles. Cash-pay consumers used to pay list prices, but cash-pay rates at many LA hospitals have fallen to the Medicare Reimbursement rate, due to the Afordable Care Ave and consumers shopping around. The Medicare Reimbursement rate is the only price in the healthcare system related to the cost of providing care. Currently, consumers in commercial health plans at the same Los Angeles hospitals pay 5-20 times the Medicare Reimbursement rate for the same service. 

In regions, where there has been greater health provider concentration, like in the Bay Area, we assume that better price comparison among providers will lead to similar dynamics in the cash-pay market, with some providers dropping their cash-pay rates to the Medicare reimbursement rate. 

After it becomes common knowledge that consumers can cash-pay at lower rates for health services. Informed health consumers in Cadillac health plans will see the lower cash-pay rates and start paying cash. This will grow the cash-pay market and force more health providers to offer competitive cash-pay rates. A growing cash-pay market will in turn create pressure for commercial insurer rates to fall because more employers will try to become self-insured to take advantage of the lower cash-rates. With a declining employer market, insurers will be forced to lower commercial rates. 

Over the last four months that we have been operational, we've helped over 12,000 users search health prices in California. Currently, we're in the process of expanding to the entire United States. In one year, we're hoping to reach a 500,000 people, by adding pricing data for the entire US, specialized COVID-19 pricing tools to help consumers determine if they qualify for government or financial assistance and with a marketing campaign to increase awareness. 

The size of the uninsured and underinsured population in the US is about 161 million people. We hope to serve the majority of these users in the next five years, through either our consumer or enterprise product.

Our goals are in the next year are:

• Display prices for the top 5,000 medical services at the vast majority of healthcare providers in all states

• Create quality video content on health plan and health provider pricing topics

• Grow our user base from 12,000 to 500,000 users

• Document over $10,000 of user healthcare savings due to the website

  
  

Our goals for the next five years:

• Add insurer negotiated rates if they are released

• Release an enterprise product with health plan details like deductibles, co-pays and out-of-pocket maximums

• Grow the user base to over 10 million users 

• Become a profitable social venture

There are a number of key barriers that would prevent us from accomplishing our goals of complete price transparency:

1. Centers for Medicare and Medicaid (CMS) revokes the Prospective Payment Services (PPS) rules to release Chargemasters CMS-1696-F. Even if these rules are not revoked, but not enforced, providers might not comply with the guidelines. About 5% of providers are currently not in compliance with CMS-1696-F. 

2. The difficulty of educating consumers on how health provider pricing works. Through internal surveys, we have estimated that approximately  90% of health consumers do not understand health provider pricing. 

3. Similar to pharmacy pricing, a minority of hospitals offer substantially lower cash-pay rates, then the rates they charge commercially insured patients. Providers make it clear in their written policies that patients must not have insurance to get their substantially lower cash-pay rates. It's unclear if consumers with insurance can cash-pay and if health providers have any obligation to share cash prices with them if they ask. 

4. Hospital final bills are often bundled meaning that services are grouped together without informing patients. Bundling makes estimating medical costs difficult for consumers and for us, because there are hundreds of thousands of medical services and supplies and each provider bundles these differently. 

This is how we plan to overcome these barriers: 

1. If disclosure rules are reversed, we will develop a method to estimate the annual provider price increases, which are generally relatively consistent. 

2. The proposed effort will work on ways to make pricing untethered to the abstract and archaic current billing system, by linking pricing to diagnoses to help consumers better understand healthcare pricing.  

3. We worked with NBC Bay Area to call the California Managed Care Department to determine if consumers with health insurance can pay lower cash-pay rates. In a similar way to determining California's policies, we are planning to contact Managed Care Departments in each state to determine the answer to this particular question. 

4. We will develop a method to address health provider bundling, including accessing the feasibility of using predictive models to predict bundling and subsequent pricing of those service bundles by different providers.

Clinic Price Check currently has a few partners that have aided with business development and development of the product:

• Tech Futures Group (a California small business administration program), which has helped us with development of our accounting, IP and business strategy

• University of California Berkeley SkyDeck, the Berkeley Institute of Data Science, HAAS Start-up Squad and University of California alumni and founders networks

• Georgetown Tech Alliance and Georgetown Entrepreneurial Alliance

Clinic Price Check is a social venture and will always offer price comparison for most health services at local providers for free. We will charge for an enterprise version and premium consumer version, which will integrate plan specifics like deductibles, copays, coinsurance and out-of-pocket maximums and AI pricing tools.

Unlike anything else on the market, this novel application allows consumers to compare prices for medical procedures at local health providers. There are two customer groups that we are targeting: (1) uninsured and underinsured consumers directly and (2) employer groups, particularly self-insured employers.

The total addressable market is the cash-pay healthcare market in the United States (U.S.), which would be around $495 billion annually. Clinic Price Check will monetize the product as a low-cost employer and consumer add-on (charging a $1 per employee per month). Clinic Price Check would charge consumers the same amount charged to employer groups (without the initial start-up fees). The full Service Available Market for consumers and employer groups is $1.886 billion.The Service Available Market is not completely obtainable because some employers and consumers have high quality insurance plans or may not be interested in our service for other reasons. 

If we are able to get the majority of employees in High Deductible Health Plans, a portion of uninsured/underinsured consumers who employers do not join our service and some government employees, the Serviceable Obtainable Market is $847 million.

In the long-term, Clinic Price Check will fund our mission of greater healthcare price transparency through the development of an enterprise version, which includes insurer negotiated rates, plan details and potential integration with Health Saving Accounts (HSAs). 

Clinic Price Check, the website, is production quality(has been thoroughly bug-tested) and currently scaleable to millions of users. The next step for commercialization is to improve the underlying matching algorithm and complete the data extraction, cleaning and collation for the other forty-nine states. (We estimate this will cost $20,000 to complete the process to cover the entire United States.)

Once Clinic Price Check covers the entire U.S.United States, the next step is getting press for the site and improving site SEO, through thousands of entry points for each billable service. (For SEO/marketing campaigns, we estimate it would cost $20,000.) Once a sizable user base is secured, the search for ad partners will commence to support the scalability of the application to tens of millions of users. 

Concurrently while growing the free version of the site, there will be a push to find early adopters in the employer benefits space, to try our product on a free-trial basis for the first three months. One strategy for gaining employer groups is the bottom-up approach or having participants sign-up with their employer email. Once we have reached some density in smaller employers, the final stage is to convince large self-insured employer groups to try our service.

Clinic Price Check is a health technology start-up company, which currently only offers a free version, and, as such, has not generated any revenues to date. The initial funding, about $10,000, was provided by the founder and covered development of the product/algorithms for California, legal fees for two utility patents, business taxes and other startup expenses.

We are applying for government research grants to fund the development of this project. At the time of this application, we have not received a response on grant funding applications. 

Our expenses for 2020 will be about $40,000. 

It will cost us $20,000 to finish the data collection process. 

It would cost us $20,000 to run a marketing Google Adwords campaign, hire a consultant to help us with SEO and develop social media/video content to help consumers better understand health provider pricing. 

The full development of an intelligent pricing system would cost about $250,000 to fund hiring of two full-time data scientists to aid PI in development of this system. 

I'm applying to the Elevate Prize because I need help to grow and scale this project. 

• The project can touch a large number of people (about 60% of the American public directly), but need to grow the public's awareness. We hope that the Elevate Prize can help grow awareness for the project. 
• We need funding to cover the costs of project development, such as expanding to other states and building an intelligent pricing agent to help consumers predict health costs. The Elevate Prize can help us raise either grant funding or institutional investor funding.
• We need technical support and to partner with leading AI researchers to improve to make the healthcare pricing information more accessible to consumers. The Elevate Prize can put us on the radar of leading researchers, who might offer some of their time and support to the project. 
• We need support of corporate partners to build an enterprise version to become a sustainable social venture. The Elevate Prize can help us build some of those partnerships 

Clinic Price Check has a number of partnership goals:

• Branding/marketing our free consumer product

• Educating consumers about cash-pay

• Creating, selling and piloting our enterprise product

• Determining the legality around cash-pay and financial assistance if a patient has insurance

• Securing funding or grant money to continue development of the product

Clinic Price Check would like to partner with a few different groups:

• Partner with MIT Machine Learning/Artificial Intelligence facility and institutes such as the MIT Statistics and Data Science Center and Data, Systems and Society. They can help us develop our intelligent pricing agent to help consumers connect symptoms or diagnoses with health provider pricing.

• Partner with Elevate Community Companies or Foundation Partners. They can help us determine what aspects of our product are most interesting to employer groups and help us launch an initial enterprise product pilot.

• Partner with MIT Solve network of Social Entrepreneurship Companies and Organizations. We can learn from their experiences and form a shared network and support system.