The Costs of Accelerating EHR Adoption

Many studies have demonstrated that Electronic Health Records (EHRs) can improve quality and reduce costs through coordinated delivery of the right care at the right time. The escalating cost of healthcare and the downturn in the economy are the perfect storm to create real urgency for implementing Healthcare Information Technology.

Many clinicians cannot afford EHR implementation. Stark safe harbors help physicians affiliated with hospital systems but do not help unaffiliated clinicians.

An early priority for the Obama administration should be decisive, rapid action to accelerate the adoption of EHRs via broadened Medicare/Medicaid incentives to implement and use Certification Commission on Healthcare Information Technology (CCHIT) certified products. Federal funds are needed to subsidize implementation teams and locally credible EHR champions who inspire and motivate providers at the grass roots level in each State. In order to receive funding, States should have to create EHR adoption services that effectively and efficiently deploy EHRs to achieve low failure rates and meet quality/safety goals. Funds should include direct payment, low interest loans, tax credits, pay for performance incentives, and penalties for delayed adoption. Grant funding, however, is probably not an effective vehicle, since it doesn't give Federal/State governments enough control, nor is it usually focused on sustainability.

How much is needed? Our Massachusetts experience suggests that approximately $350 million is needed to complete the rollout of EHRs in our state - about $50,000 per practicing unaffiliated clinician.

Here are the breakdowns of community EHR implementation costs at BIDMC/BIDPO, the Massachusetts eHealth Collaborative, and the New York Department of Health and Hygiene EHR project.

Costs per licensed user	BIDMC	MAEHC	NYC
Software	5,998	10,800	4,500
Hardware	10,561	17,783	15,000
People	29,641	17,660	16,000
Total	46,200	46,243	35,500

Assumptions:
1) Software costs include only the direct licensing costs for EHR and non-EHR software. MAeHC software costs are higher because several different EHRs were implemented, creating more complexity.
2) The NYC costs do not include non-EHR software
3) Hardware includes practice-level and central-site hardware.
4) People includes direct services from staff, whether vendor-provided or sponsor-provided.
5) These costs are for implementation only. The average annual per physician support costs are roughly $5,500 per user for BIDMC and $6,500 per user for MAeHC.


Further detail:
1) BIDMC includes 300 docs. MAeHC includes 575 docs. NYC includes 1,200 docs.
2) The people costs are not directly comparable, because neither MAeHC nor NYC have accounted for the entire provider-side of the costs of hardware integration. For example, with MAeHC, vendors designed and the hospitals implemented the local ASP environments, but we do not know the labor cost at the hospitals. With the BIDMC project all costs are explicit because the ASP environment was outsourced. The NYC practices are purchasing hardware on their own, so we do not know the exact costs.
3) BIDMC will get some scale benefit once the number of implementations grows. The per user people costs include the design and build of the central site and the cost of the Project Management Office.
4) BIDMC actual hardware costs will probably be higher based on the implementations to date because the practices are purchasing more equipment than original budgeted ­(i.e. more printers, laptops, and tablets for support staff)

As a country, we have enough experience with live implementations to know what needs to be done to implement EHRs and the cost of doing it. The time for grants and experimentation has passed. To borrow a marketing slogan, the time is right to "Just Do it" by providing financial incentives.

An Epidemic of Overtreatment

This blog entry was co-authored with Rich Parker MD, Assistant Professor, Internal Medicine, Healthcare Associates.

Healthcare costs in the US are approaching 17% of the GDP and may be as high as 20% in the next few years.

What is causing the US to have the highest cost and lowest value for the healthcare dollar? Simple - it's overtreatment.

Overtreatment takes many forms - from over ordering expensive diagnostic tests to the prescribing of expensive and sometimes unneeded therapeutics.

There are many reasons for this. Here are just a few:

1. Incentives are misaligned. Healthcare reimbursement in the US pays for quantity, not quality. This means that clinicians benefit from performing more procedures, hospitals benefit from more diagnostic testing, and the pharmaceutical industry benefits from adoption of new name brand drugs. If you do not believe this to be the case, spend a day in an ambulatory care clinic or a hospital and see what goes on. Ask any resident, fellow or attending how many tests and treatments are unneeded. We believe that paying for wellness or paying for outcomes will solve this piece of the overtreatment puzzle. If doctors and hospitals had to live within a budget, diagnostic and treatment strategies would change quickly and become less expensive for all of us with equally good clinical outcomes.


2. We've attended many gatherings where parents discuss brand name powerful antibiotics and recommend that they become the first line drug for treatment of anything their children complain about. "Don't accept Amoxicillin, go for the Augmentin or Cipro". John's daughter is 16 and has not ever taken an antibiotic in her life. She's had a few viruses, but no virus is cured by antibiotics. Overtreatment of the pediatric population with powerful antibiotics creates resistant organisms that make children sicker and create a dependency on ever more powerful antibiotics. The problem with adults is equally severe. Watch the evening news and within an hour you'll hear about a dozen brand name pharmaceuticals treating diseases you've never heard of, but may now suspect you have. The United States is the only country in the world that allows “direct to consumer” advertising. We believe this advertising should be regulated to solve this piece of the overtreatment puzzle. Those advertising dollars end up coming out of your pocket too!

3. Some patients are not willing to accept risk or shared decisionmaking with their doctors. They want to begin the evaluation of back pain with an MRI instead of trying a course of gentle exercise and pain meds. Many issues do not have a clean or simple diagnosis. Eat right, exercise, avoid caffeine/nicotine, and let the body heal itself. For many conditions, rest and time cure the problem. Although the healthcare systems of Canada and the UK have their problems, the fact that access to expensive diagnostics is limited enables patients and their doctors to work together on simpler evaluations and therapies as a first step. We need to change the cultural expectation that expensive tests are "first line".

4. As a country the US eats poorly, avoids exercise, drinks an infusion of lattes, and then wants to take a pill to make all the lifestyle diseases go away. Lifestyle issues should be treated with lifestyle changes, not pharmaceuticals or nutraceuticals. Our own experience convinced us of this. John gave up the lattes, the super-sized meals, and began daily exercise 7 years ago. Since that time, all his lifestyle diseases have disappeared.

5. Overtreatment begets overtreatment. If a lifestyle disease is treated with pharmaceuticals, it's likely that those medications will cause side effects. The symptoms of side effects lead to further diagnostic testing and more pharmaceuticals are often the result. We know several patients who are on medications for hypertension due to overeating, H2 blockers due to excess caffeine/nicotine consumption, and several medications to treat the side effects of their initial medications. Two or three medications can fast become ten. We've suggested taking a medication holiday with appropriate clinical supervision, redesigning their diets, and beginning daily exercise. The answer we often hear is that taking all those pills, having all those tests, and visiting their clinician often is easier than changing their lifestyle.

6. Today on the local radio station, an attorney asked the question "have you ever had a bad outcome or misdiagnosis? I've been holding doctors accountable for 30 years. Call me and we'll get you the cash settlement you deserve." There are bad doctors. There are doctors who are unskilled at surgery or provide very non-standard care. However, most clinicians are trying to do the right thing. Medicine is not an exact science. It's based on experience and probabilities. This means that even the best clinician will miss a rare disease or an atypical presentation of a common disease. As a country, we need to realize that delayed or misdiagnosis will occur despite best efforts and accept a low level of imperfect outcomes instead of forcing every doctor to overtreat every patient in the pursuit of 100% certainty. Both patients and doctors together must accept some degree of uncertainty or we will continue to bankrupt our system.

Our economy has lost its competitive edge because our healthcare costs have ballooned to extreme levels due to misaligned incentives, overzealous pharmaceutical marketing, expectations of high cost testing/therapeutics, excessive administrative costs and complications due to overprescribing and fear of litigation.

The diagnosis of overtreatment is simple. The therapies are complex. We've proposed a few fixes above and will continue to write this topic in blogs to come.

Interoperability Advice for the New Administration

As policymakers consider ways to reduce healthcare cost and improve quality, I'm often asked about the current readiness of standards and interoperability.

I believe that standards are no longer the rate limiting step.

On November 12, I presented an overview of standards readiness to Secretary Leavitt and AHIC. The video is available online

My presentation begins at 1 hour and 8 minutes. Anyone wanting to view it can just use real player to advance to that point.

You'll see that as a country, we have finished:

2006 - Personal Health Records, Laboratories, Biosurveillance
2007 - Medications, Quality, Clinical Summaries
2008 - Medical devices, Referrals, Family History/Genome, Secure messaging, Public Health Reporting, Immunizations

In 2009, we'll complete Newborn screening, Clinical Trials/Research and close a few minor gaps

All the stakeholders (vendors, government, academic, pharma, labs, payers, providers, patients) have agreed on the needed standards by consensus. Secretary Leavitt has Recognized all the 2006 and 2007 standards and will be Accepting the 2008 standards on January 8, 2009. Recognition means that the standards are required for use by all Federal agencies. Acceptance means that a year of testing begins and Recognition will follow.

Thus, there is no need to wait for the standards. Vendors are beginning to implement these standards and the Certification Commission on Health Information Technology is beginning to require them.

If standards are not the issue, what about security and privacy? As readers of my blog know, I am passionate about the need to protect confidentiality.

I believe that security is no longer the rate limiting step.

The standards for security were finished in 2007. They are available online and have been fully incorporated into all the HITSP interoperability specifications including all the needed security standards to support encryption, authentication, authorization, audit trials, non-repudiability, and patient consent.

These security standards can enforce any local privacy policies - from something basic like HIPAA to something complex like the Massachusetts approach to opt-in consent at the institutional level.

It is true that the US has very heterogeneous privacy policies in states and localities that pre-empt HIPAA, but that is not a security or technology issue.

What about architecture?

I think that we've done enough pilots and experiments to know what architecture we need.

The US already has a functional architecture for e-Prescribing including retrieval of comprehensive medication history. The US already has a functional architecture for exchange of lab results among providers, patients and commercial labs.

What's missing is a clinical summary exchange that ensures care coordination among providers of care and patients. I've written about a simple, internet-based, service oriented architecture that can securely exchange structured healthcare data between stakeholders. This can approach can be used to
a. Send / push / route hospital data to appropriate parties
b. Send / push / route visit and other data in support of referral consultation
c. Send / push / route visit and other data for standardized quality reporting
d. Send / push / route data for patient health records (PHRs)

Note that none of these transactions creates new privacy issues. Every one of them is currently required by good medical practice or by law, and are performed on paper today.

Thus, interoperability is implementable today with harmonized standards, appropriate security, and a service oriented architecture using the internet.

Now we need incentives to implement it.

Data exchange is a public good in many ways, so it will be challenging to fund purely based on local stakeholder contributions. There is a need for Federal leadership and funding to mandate very specific transactions on a defined implementation timetable. We should accelerate adoption through the same approach the US is using for e-Prescribing: regulation to create mandates and incentives to create urgency, followed by penalties for late implementation.

Experience has taught me that it's best to automate existing processes rather than trying to simultaneously change process and add technology. The approach I've presented above is a good short term solution. In the long term, let's hope that patients become a steward for their own data via PHRs or establish a "medical home" - a primary care giver who coordinates all their care. The architecture could easily evolve such that every entity which provides care has to push the data into a "medical home" EHR in a standardized fashion.

Engage with Grace

Today, you'll find that many healthcare bloggers have devoted their blog to the post below. My wife and I have have completed the exercise. In my case, I do not want end of life care in a hospital. When I cease to be me through brain injury or diminished mental capacity, I do not want to be supported. I want my ashes scattered from the top of Mt. Scowden on the Tioga Crest.

This guest post was written by Alexandra Drane and the Engage With Grace team.

We make choices throughout our lives - where we want to live, what types of activities will fill our days, with whom we spend our time. These choices are often a balance between our desires and our means, but at the end of the day, they are decisions made with intent. But when it comes to how we want to be treated at the end our lives, often we don't express our intent or tell our loved ones about it. This has real consequences. 73% of Americans would prefer to die at home, but up to 50% die in hospital. More than 80% of Californians say their loved ones know exactly or have a good idea of what their wishes would be if they were in a persistent coma, but only 50% say they've talked to them about their preferences. But our end of life experiences are about a lot more than statistics. They're about all of us.

So the first thing we need to do is start talking. Engage With Grace: The One Slide Project was designed with one simple goal: to help get the conversation about end of life experience started. The idea is simple: Create a tool to help get people talking. One Slide, with just five questions on it. Five questions designed to help get us talking with each other, with our loved ones, about our preferences. And we're asking people to share this One Slide wherever and whenever they can ... at a presentation, at dinner, at their book club. Just One Slide, just five questions.

Let's start a global discussion that, until now, most of us haven't had.

Here is what we are asking you: Download The One Slide and share it at any opportunity with colleagues, family, friends. Think of the slide as currency and donate just two minutes whenever you can. Commit to being able to answer these five questions about end of life experience for yourself, and for your loved ones. Then commit to helping others do the same. Get this conversation started.

Let's start a viral movement driven by the change we as individuals can effect...and the incredibly positive impact we could have collectively. Help ensure that all of us - and the people we care for - can end our lives in the same purposeful way we live them. Just One Slide, just one goal. Think of the enormous difference we can make together.

To learn more please go to www.engagewithgrace.org.

Outliers

It's too cold for rock climbing and kayaking but not yet cold enough for ice climbing or skiing, so I'm using the time to read through the books on my nightstand and on my Kindle.

I just finished Outliers by Malcolm Gladwell (author of The Tipping Point) , a well written thought piece on what really produces the movers and shakers in each generation.

Darwin would have loved the heady questions raised by this book. Is success more about nature or nurture? What's more important the Intelligence quotient or Emotional quotient. Is the Stanford-Binet test a useful measure of your likelihood to succeed? Can Harvard really tell the differences among 3000 valedictorians with perfect SATs applying to college?

Here's my own story put in the context of the book.

When I was 12 years old (1974), my parents went to law school and I spent my free time after school scouring surplus stores in Southern California. Sunny Trading Company on Torrance Boulevard was my treasure trove. For 10 cents I could buy NAND gates, Shift Registers, and LM555 digital timing chips. Reading through National Semiconductor product catalogs and the entire contents of our local library's Dewey Decimal 620-622, I learned digital logic, analog to digital conversion, and the basics of microprocessor design.

Then, in 1975, a major breakthrough. The Popular Electronics January issue announced the Altair 8800, which made home computing possible and I devoted myself to learning about "personal computers".

I spent my high school years programming in numerous languages from Assembler to Fortran to Cobol to BASIC. I used minicomputers, microcomputers, and mainframe computers. In 1978, I designed the software and hardware for my first experimental medical device - a computer capable of gathering visual and audio evoked potentials then performing signal averaging and fast fourier transforms in real time.

All of this was possible because I lived in Southern California in the mid 1970's where surplus stores had cheap integrated circuits and because my local library gave me access to great books about emerging technology. I truly believe that this foundational portion of my career was more about time and place than me.

From high school I went to Stanford, started a software firm and began the parallel life of medicine and technology that leads to the present.

In Outliers, Gladwell points out that Bill Gates, Steve Jobs, Larry Ellison, Scott McNealy and Bill Joy - the leaders behind our largest technology companies - were all born in 1955. They were 20 when the Popular Electronics issue was published, just completing college (or dropping out of it). The were at the beginning of their careers but without families, a mortgage or an established job in a traditional technology firm. They were at the right place at the right time to ride the wave of emerging technologies.

Of course they were smart, but numerous people are smarter. Gladwell concludes that once you are "smart enough" then culture, circumstance, timing, and luck are key differentiators for success.

Our lives are complex paths with daily choices that lead to success or failure. I know that I could have ended up in a dozen different careers, lifestyles, and economic strata. However, as Outliers suggests, the world around me shaped my outcome and I can really link my Harvard faculty position to my parents choice of living near an electronics surplus store in the 1970's.

Outliers is worth reading to understand the external context which shaped some of the most successful people in our generation.

Reducing IT budgets

I oversee the budgets of several technology organizations, all of which are under pressure to react to the faltering economy. At BIDMC, capital budgets are constrained, operating budgets are being tightly managed, and no staff reductions are planned. At Harvard Medical School, senior leaders are seeking a 10% reduction in the school's operating budget. How can IT organizations approach operating budget reductions?

1. Engage all your staff. They can identify operational inefficiencies, redundancy, and savings opportunities. Any budget reduction results in rumors, speculation, and fear of job loss. Engaging your staff in the budget process empowers them, informs them, and reduces their worry.

2. Find the low hanging fruit - vacancies, travel/training, consulting fees, food/entertainment, and other "nice to have" expenditures are the first place to start any budget reductions. I passionately support training, but when faced with budget cuts, most staff would elect to support salaries and reduce training.

3. Identify service reductions - all IT projects are a function of scope, resources, and timing. Reducing the scope of service and determining what projects to cancel is an important part of budget reductions. One challenge is that organizations often have short memories. If you offer a budget reduction linked to a service reduction, you may find that the budget reduction is happily accepted but the service reduction is forgotten in a few weeks. In fact, many departments throughout the organization will suggest that budget reductions are possible if more automation and technology is added to their work processes. Reducing service at a time when customers need more service may not be the optimal approach, which I will discuss further below.

4. Extend timelines - assuming that resources are diminished and scope is already reduced, the last lever a CIO has is to extend the timelines of new projects. Instead of delivering new software this year, delay it to next year. Existing staff can take on more projects only if they have a longer time to do them.

5. Accept risk - Our job in IT is to ensure stability, reliability, and security. 99.99% uptime requires multiple redundant data centers, but there is not a precise cookbook as to how this should be done. I have implemented 2 data centers a few miles apart and not a grid of worldwide data centers. Why? Because risk = likelihood of a bad outcome * the impact of that outcome. When creating budgets, I decided that the likelihood of a regional disaster which destroys the IT capability of the entire Boston region is small. The likelihood of a single data center fire, flood or explosion is measurable, so I chose to mitigate that risk. In times of budget stress, a re-evaluation of risk is appropriate. Can network, server, storage and desktop components be kept for a year or two beyond their usual lifetimes? Can maintenance contracts be reduced or eliminated and mitigated by having spare components handy? Just as with service reductions, the strategy of increasing risk to reduce costs must be widely communicated, so when a failure occurs, everyone understands it was a risk accepted as a result of budget reductions.

In general, I do not recommend fighting budget reductions with overly dramatic stories of doom and gloom. That is not professional. Instead, CIOs should provide senior management with a list of services and a list of risks, then decide collaboratively what to do. This ensures that the CIO and IT is seen as an enabler and team player rather than a cause of the budget problem.

Having been through numerous budget reduction experiences over the past decade, I have witnessed the paradoxical effect that IT budgets are sometimes increased when organizational budgets are decreased. Savvy administrators know that economic downturns provide the urgency to re-engineer processes and accomplish politically difficult strategic changes. A short investment in automation can lead to long term reductions in operating costs. Thus, this downturn may be the opportunity to eliminate paper, streamline labor intensive manual methods, and consolidate/centralize for economies of scale.

Over the next 60 days, I will work with Harvard Medical School administration on all these issues and report back as to what we collectively decided to do.

Cool Technology of the Week

This week I had the opportunity to carry the ultimate geek utility belt - an iPhone 3G, a Blackberry Bold (AT&T) and a Blackberry Storm (Verizon). I tested all three devices in parallel for emailing, web browsing, and usability. Here are my first impressions:

The Bold has a bright 2.8-inch glass LCD screen with 480 x 320 resolution (same as the iPhone 3G) and is very easy to read. I started using bifocals last week and I really appreciated the Bold screen as compared to my Curve. The increased graphic density makes a big difference for the web browser since it's now possible to render an entire page on one screen. An easy to use zoom feature and scroll bars make browsing much more "desktop like" than previous Blackberries. The HSDPA 3G wireless over AT&T is fast and the coverage in Boston is quite good. The WiFi linked easily to my hospital infrastructure. The Bluetooth bonded seamlessly with my Prius. The only downsides - the battery life of the Bold is about a day and half, compared to 3 days on my Blackberry Curve. I could not find a way to turn off the 3G and use EDGE only. The QWERTY keypad is very similar to the 8800 series, without the generous key spacing of the Curve. I thought this would be a problem but after a few minutes of typing, I adapted and was able to type at my usual 60 words per minute without errors.

The Storm has a bright 3.25 inch glass touch screen with 480 x 360 resolution and tactile feedback. The entire screen is a mouse button that depresses but since your finger only contacts a small portion of the screen at time, your brain believes that the screen is a push button under your finger. The on screen keyboard is a full QWERTY keyboard when the device is tilted horizontally. When you hold the Storm vertically, it displays a "SureType" keyboard - the same doubled up keys found on the Blackberry Pearl. They question that I'm sure everyone is asking - was the typing experience on the Storm better than the iPhone 3G? The answer for me, is yes. My brain wants a key to click when I push it. Even though I really like the engineering of the iPhone screen, I have not achieved better than about 80% accuracy at 60 words per minute. With the Storm in full QWERTY mode, I easily adapted to the touch screen for typing. "The Quick Brown Fox Jumped Over the Lazy Dogs" actually appeared as typed, time after time.

In general, I found the Blackberry Storm easy to use although the multiple buttons on the case made navigation less obvious and elegant than the iPhone. The Storm has EVDO but no WiFi. For me, the WiFi is not a major need since most airports charge for WiFi and I prefer wide area networks. I've heard others complain about the lack of WiFi.

I'm sure many will ask - is the Storm an iPhone killer? I think this question is just as inappropriate as is the iPhone 3G a Blackberry killer?

They are different devices with different purposes. The Blackberry Storm is an email device that includes a great screen and a virtual keyboard with tactile feedback. The iPhone is a small computer running numerous lifestyle applications and an email client that is not optimized for high volume email use. They are both excellent devices for their respective niches.

What am I going to use? As with every technology this question needs to be answered with a specific set of requirements in mind. I need to email 21 hours a day while walking, commuting (only at stoplights), while in meetings, and while in airports. Although I really like the idea of touch screens and believe the next generation of handheld computers will focus on touch screen user interfaces, I still prefer the Bold's keyboard which is easier to use, faster, and more accurate for on the go typing than the Storm or the iPhone.

Next week I'll handoff these devices to my staff for testing and we'll see how they work for others with different requirements.