Friday, November 29, 2013

Simple Math with Scary Implications for Hospitals: What Value-Based Purchasing means for hospitals’ bottom lines

I spend a lot of my time exploring the drivers of top and bottom-line growth for payers. As a physician, I had a nagging curiosity to learn about what the economics look like for my (part-time) workplace. So I explored one of the biggest changes in hospital reimbursement reform, value-based purchasing (VBP).

Value-based purchasing is a reimbursement structure that is intended to reward improved clinical processes, outcomes, and patient satisfaction. In many cases, it is perceived as the inverse: a financial penalty for below average hospital performance. The following is some overly-simplified math than demonstrates how VBP works and more frighteningly, the implications of VBP for hospitals. This rudimentary analysis may serve as a useful backdrop for entrepreneurs trying to sell their software into hospitals, doctors considering how they can add value to their hospital, or patients who want to get empowered about the role of their opinion.

First, let’s start with how hospitals were paid before VBP was put in place. Medicare paid hospitals per discharge based on a diagnosis-related group (DRG) specific for a disease state. More elaborate admissions like heart surgery were reimbursed more than more benign admissions like a cellulitis.

With the introduction of VBP, hospitals will have their reimbursement go up or down depending on their performance in clinical processes, clinical outcomes, and patient satisfaction. The performance in each of these domains gets incorporated into a total performance score (TPS) for each hospital each year by Medicare. Depending on a hospital’s TPS compared to that of the national median TPS, a given hospital will have the reimbursement for each discharge systematically increased or decreased by some multiple (VBP Incentive Multiplier).

To incentivize better performance, Medicare is reducing the baseline reimbursement for each DRG by 1% starting in 2013. That reduction will worsen to 2% by 2017.

The combination of reimbursement based on the VBP Incentive Multiplier and the DRG Percent Reduction over the next few years results in the following overly-simplified formula:

VBP Incentive MultiplierHospital A =

1 + [Percent Reduction in DRG x (TPSHospital A/TPSNational Avg) - Percent Reduction in DRG]

This is where the scary part begins. Let’s fast forward to 2017 when the reduction in reimbursement for each discharge is at the full 2%. Let’s also assume that a given Hospital A is performing at an annual TPS of 25 while the national average is 50. Plugging our numbers into the above formula, the VBP Incentive Multiplier = 1 + [2% x 25/50 - 2%] = 1 + [0.02 x 0.5 - 0.02] = 1 + [0.01 - 0.02] = 1 - 0.01 = 0.99.

So in 2018, Hospital A will have each of its discharged diagnoses reimbursed at 99% of what the DRG reimbursement rate is for that year. If, for example, Hospital A has an operating budget of $100 mil, and if all of the revenue comes from clinical care (vs research funding), then Hospital A’s below-average performance would have just resulted in $100 mil x 0.99 = $99 mil actual operating budget, or a $1,000,000 penalty. With profit margins for highly profitable hospitals in the 2-3% range, a 1% hit on operating budget can be as much as a 50% decrease in the profit margin. The board of a hospital would not be happy with such numbers, the CEO may be out of a job, and more importantly, there may be less money to reinvest in expanding the hospital’s ability to serve vulnerable populations.

Let’s take a look at a slightly less gloomy scenario. I’ll return to 2013 for this example where the DRG reduction is just at 1%. Let’s assume that in 2013, Hospital A was performing above average with a TPS of 75 while the national average was 50. Plugging our numbers into the above formula again would yield a VBP Incentive Multiplier = 1 + [1% x 75/50 - 1%] = 1 + [.01x 1.50 - .01] = 1 + [.015-.01] = 1 + .005 = 1.005.

So, for 2014, Hospital A will have each of its discharges reimbursed at 1.005 times the baseline DRG reimbursement rate for that year. Assuming the same operating budget of $100 mil for Hospital A and assuming that all of its revenue comes from clinical care (vs research funding), then Hospital A’s above average performance would have turned its $100mil x 1.005 into $100,500,000 or $500,000 increase in operating budget. The resultant potential for a 25% increase in profit margin would make for a happy board, an employed CEO, and potentially reinvestment into interesting programming for needy patients.

But there’s a catch! I didn’t take into account the cost of what it would take to improve a hospital’s performance. If the hospital starts and remains above the median at the same level, it will continue to enjoy higher margins. But what about hospitals that need to improve? With improvement, in theory, the goal of VBP, underperforming hospitals have a very steep uphill battle.

For example, investments in improving clinical processes, outcomes, and patient satisfaction could include software, training, and evaluation costs, among others. Using EMR costs as a proxy for the range of cost for software, the cheaper EMRs cost about $4k per provider per year. If hospital A employes 100 providers, that's $400k per year, not to mention set up fees. My calculations are obviously over-simplified. But the point is that hospitals will have a difficult time achieving a positive ROI in the newly emerging reimbursement landscape. Even for hospitals that outperform the rest, their net gain from improving outcomes may be minimal at best.

There are many limitations to this analysis. I didn’t include the additional burden for hospitals due to the impact of efficiency metric that is due to roll out in 2015 and 30-day readmission penalties, among others.

 Although well intended, VBP has many unforeseen consequences that may disproportionally and negatively impact hospitals that serve disadvantaged patients. These unforeseen consequences and my rudimentary math exercise above are examples of why hospitals and the healthcare system as a whole may need to focus (even faster) on what happens before and after the admission.

Wednesday, November 27, 2013

Health Information Exchanges for Startups: Leapfrogging the glacial pace of interoperability among providers

There is an emerging transition from fee-for-service to capitated and episode based reimbursement models. With the fastest growing cost driver in healthcare being post-acute care, there is a lot of attention being paid to how to mitigate the increasing spending. One approach to curb the growing cost of post-acute care is to improve interoperability among provider organizations.

A large financial driver to support interoperability is the need for hospitals to meet the requirement for meaningful use (MU). Among those requirements is the need for hospitals to communicate with post-acute providers.

The main mechanism for interoperability between the myriad provider organizations is the health information exchange (HIE). HIEs allow for providers to communicate by creating a common roadway for data to travel. There are emerging national standards for the layout of those roadways, their on-ramps and off-ramps, the types of vehicles that drive on them, and the type of cargo those vehicles can carry.

The focus of these evolving standards, appropriately, is on the provider organizations. Yet there remains much difficulty for providers to communicate with each other. And it is especially difficult for startups to break into that dialogue. Without access to communicate health information with providers, it is challenging for startups to test their interventions for value.

Adding to the difficulty of accessing providers and payers is the frothiness of the digital health startup community. Many fledgling entrepreneurs spend too little time performing the necessary review of analogs and antilogs to explore whether there are others trying to solve the same problem in the same way.

Even when there is some differentiating factor between startups, the incremental difference between startups is drowned out by too little attention span and too high expectations for performance by payers. For example, many startups in the post-acute care sound something like: “care coordination platform to drive down admissions.” Startups need a way for payers to better hear the type of care coordination platform, by whom it is used, and how they drive down admissions.

What if instead of aiming to be interoperable with providers, startups focused on being interoperable with each other? Since HIEs have already created standards for the way data is transmitted and in what form, startups can leap frog the EMR community and revolutionize the use and value derived from HIE and interoperability standards.

Interoperability among startups would create a widget ecosystem where each startup had a small but unique value proposition that was built on a common, best-practice care coordination scaffold. The standardized data architecture would make it much easier for provider and payer organizations to tap into the fabric of collaborating startups. The value of that fabric would grow with each new startup that wove in its data. And the new insights derived from the transformation of multiple strings of data into one robust quilt would reinforce the unique value of each startup contributing to the startup HIE.

The process for creating a startup HIE would include utilizing standards for the way data is input and output. There would need to be agreement on the mechanism of transmitting the data, whether it was through a local HIE or directly between startups. The value of the synergistic relationship would then have to be tested and conveyed to the customer in the form of offering more robust data and analytics or helping to better meet some sort of reporting requirements. Finally, the collaborating startups should establish and share best practices with their community of collaborators in order to eliminate waste in the network and increase the network value for all participants.

The interoperability between startups would allow each one to focus on differentiating themselves by focusing on addressing direct customer needs and reinforcing their verticals through focused product development rather than struggling on common horizontal problems like finding out how to communicate with their customers.

Perhaps there is a role for startup Incubators to be the leaders, facilitators and moderators of these types of collaborations. What if there was a digital health startup incubator HIE?! The incubator that hosted a startup HIE would draw pre-qualified startup applicants, attract investors seeking curated investments, and provider organizations seeking connected startups.

One limitation of this approach is the perceived risk of sharing data with competitors. This risk can be mitigated by a strong incubator community that sets strict ethical rules and the option for startups to be able to remove themselves from the Startup HIE.

I look forward to seeing if startups that collaborate have a higher likelihood of delivering better value to their customers and if they experience more growth then non-collaborating companies.

Thursday, November 21, 2013

The Structure of an Innovation Academy within a Medical Residency Program

We are at the beginning of a historic era for innovation in health care delivery in the US due to the convergence of payment reform and the proliferation of mobile technology. Academic medical centers (AMCs) have the potential to be leaders in this era of delivery reform, but most have yet to display a commitment to delivery innovation on par with their commitment to basic research. [Ackerly et al] This discrepancy is not due to lack of talent or innovative spirit in AMCs, but rather because of a paucity of training in designing and implementing end-user validated interventions and a lack of established pathways for career advancement in clinical innovation outside of research, among other barriers.

A group of medical residents in the Boston Combined Residency Program (BCRP) have come together to explore how changes in the medical training experience can help cultivate leaders in efficient, patient-centered, and commercially sustainable healthcare delivery innovation. Below is a draft overview of the themes and structure of an Innovation Academy for a medical residency training program.

What is an Innovation Academy?
The innovation academy is a career development, networking, and idea incubation program for creating healthcare delivery innovations by trainees in medical residencies.

Why an Innovation Academy?
The innovation academy is needed in order to equip future healthcare leaders with skills necessary to usher AMCs and health systems more broadly through the immense change that is happening because of health care reform. The missions of most AMCs include thought leadership, creation of generalizable knowledge, training future leaders in healthcare, and providing care to patients. Until recently, the business models of AMCs supported these missions with volume-driven reimbursement for patient care and NIH-funded research as the largest revenue generators for AMCs. With shrinking NIH budgets and hospitals facing 10-15% decreases in reimbursement over the next several years, the traditional missions of AMCs may no longer be aligned with their current business models.

Furthermore, the skills being taught to trainees today, namely clinical and research skills, are only a fraction of the skills that will be requisite for clinicians to thrive or even survive in the delivery systems of the future. The innovation academy can help equip residents with the skills and network that will enable them to effectively contribute to achievement of the Triple Aim and to find or create jobs as clinician-innovators when they complete their training.

The Pillars of the Innovation Academy
To be aligned with the spirit of innovation, the Innovation Academy should adhere to a core principle of design thinking which is creating an experience with the end user in mind. In that vein, prescribing a fixed set of “pillars” can limit the necessary out-of-the-box thinking that may be needed for residents to learn how to create truly innovative system redesigns. So the pillars should be used as a guide rather a fixed construct.

The pillars include technology, policy, entrepreneurship, system redesign, and quality improvement. Each of these realms can overlap with one another. But each pillar also encompasses sufficient unique skill sets and career paths that warrant their own designation.

The Curriculum
The innovation academy curriculum should be based on the currency of clinical innovation which is value-added to patients. That value may be direct or indirect, but that value must have immediate effect. So the goal of the innovation academy curriculum should be to enable residents to create a deliverable that is validated by some end user. Given the time constraints on residents to obtain clinical training, the expectation for deliverables should be modest. Examples of deliverables may include:

-Technology: prototype software

-Policy: written legislation

-Entrepreneurship: creation of a business plan

-System Redesign: completion of a management project for the hospital

-QI: completion of a microsystem improvement

Although the incentive structure of the innovation academy should be based on value-adding deliverables, scholarly activity is still important for dissemination of knowledge as well as reinforcing and solidifying lessons learned through writing. To enable maximum flexibility and still encourage writing, the innovation academy should require all participating residents to submit at least 2 blog posts per year. The blog would enable an informal peer-review process without the time-consuming obligations of writing, submission, and revision for a formal peer-reviewed publication.

In addition to learning by doing, the innovation academy would provide expert speakers and mentors to provide structured bursts of wisdom through noon time talks and afternoon skills sessions. We propose including 1 residency-wide lunch time talk by experts each quarter.

Speakers should reflect the skills necessary to excel in each of the pillars of the innovation academy. Some of those skills should include:


-articulating value of personal skills and project (elevator pitch)

-using mentors



-Lean Startup Thinking

-Design Thinking


In addition to group didactic learning opportunities, we strongly support the role of mentors in identifying individual goals. Coaches can assist with achievement of personal life goals, “here and now” career goals, and longer term career arc goals. Residents will outgrow mentors over time, so having a robust mentorship program is essential. Mentors should be internal and external to the home institution. Mentors should also be internal and external to the healthcare industry altogether. Furthermore, mentors can and should be peers as well as faculty-level. The goal of using faculty-level mentors is to help the resident answer the questions that they now they don’t know. The goal of peer-mentors is to help think through what they don’t know that they don’t know.

The orientation to the innovation academy should occur during the first few weeks of residency. The goal of the orientation is to expose residents to the need for and purpose of the innovation academy, sample careers, and key concepts/skills. Examples of career paths and examples of how to find a job “doing innovation” include Dr. Clay Ackerly (Assistant Director of Population Health at Partners Health Care, Dr. Daniel Stein (Director Medical and Clinical Services of Walmart), Dr. Patrick Conway (Director of Centers for Medicare and Medicaid Innovation - CMMI), Dr. Joshua Sharfstein (Former Deputy Director of FDA, Current Health Director of Maryland), Dr. Rushika Fernandapulle (Founder, CEO Iora Health), Dr. Wen Dombrowski (Chief Medical Information Officer VNA Health Group).

Intern Year
The goal of the innovation academy for interns is introduce them to projects and people that represent various approaches to achieving their vision for impacting health. Another goal is to help them find mentors. Additionally, the innovation academy can help interns start exploring interesting projects through their Keystone experience.

Junior Year
The goal of the innovation academy for juniors is to identify the approach the resident will take to achieve the vision they clarified during intern year. With strong mentor guidance, residents should be able to leverage the internal and external network of the innovation academy to identify a project to work on as a senior during the 3 month academic development block (ADB). 1-2 months of “research” elective can be used as a dry run working with a team that may be the home of the innovation project during senior year.

Senior Year
The goal of the innovation academy for seniors is to provide the mentorship and incubation of project ideas during their ADB. By the end of senior year, residents should have a deliverable completed. Part of creating the deliverable should be creating the connections and demonstration of capacity to create or find a job in clinical innovation.

The goal of the innovation academy for recent graduates is to serve as a job pipeline for clinicians with hard skills and experience in the various pillars of innovation. The job-pipeline would be aligned with formal or informal pathways to organizations that are leading health delivery innovation and redesign including CMMI, Insitute for Healthcare Improvement (IHI), Management and Policy Fellowship at MGH, Kauffman Fellowship in entrepreneurship, Clinician-in-residence at a startup incubator such as Startup Health, among others.

Executing an innovation academy requires leadership at three levels. First, senior program leadership has to fully support the vision, mission, and culture of the innovation academy. Second, resident leaders need to own the innovation academy, it’s goals, and it’s activities, supported by faculty mentors. And third, resident leaders including current and future chief residents should be involved to preserve institutional memory and serve as a bridge and/or buffer between residents and faculty.

Future Work
Many interesting avenues of future work are possible around the innovation academy and innovation in medical training. One particularly interesting project would be to create a “Myer’s Briggs” personality test to predict the type of clinical innovation career that best reflects the resident’s strengths and interests. This could be done by retrospectively looking at successful clinician innovators and project to when they were interns. Furthermore, the efficacy of such a program will need to be evaluated. But before studying this approach, we will first prototype!

For more information and public discussion of this topic, please come to the panel on "Hacking Medical Training" at SXSW in March 2014.