BIDMC CIO discusses innovations at the hospital and moving to the cloud.
Hospitals and health systems host a lot of patient data and innovations. Some choose to build the systems which hold the data, while others outsource their electronic health records (EHR). Either way, with the wealth of data, executives are tasked with the decision of keeping the data on their server or migrating to the cloud.
Manu Tandon, MBA, is the senior vice president and chief information officer (CIO) at Beth Israel Deaconess Medical Center (BIDMC), which is an academic medical center affiliated with Harvard Medical School. An executive at BIDMC for five years, he previously served as the CIO for the executive office of the U.S. Department of Health and Human Services in the Commonwealth of Massachusetts. Tandon is an engineer by training and has a master’s in business administration from Harvard Kennedy School.
I spoke with Tandon about the innovations in BIDMC’s pipeline, the hospital’s homegrown electronic health record (EHR) and the decision to migrate to the cloud.
Editor’s note: This interview has been lightly edited for length, style and clarity.
Manu Tandon: We have a homegrown EHR system. We are probably one of the last remaining large academic medical centers using a homegrown EHR system. The Center for IT Exploration is tasked with looking at technology innovations outside of healthcare to see if they can be brought into the healthcare sector. More specifically, if they can be brought in to help BIDMC’s patients, providers and staff. And if we find something that we think will work, then we integrate it into our EHR system’s workflow.
Manu Tandon: One example is when you need a surgery, you need to have a consent form from the patient that allows the provider to perform it. These consent forms come in various sort of shapes and forms and can come from different places, since most of our patients come from different parts of the state or even the country. We were finding that some surgeries were either delayed, postponed or cancelled because there wasn’t a consent form.
We thought that we could solve this with machine learning, so, the Center for IT Exploration worked to develop this machine-learning solution that now detects whether a consent form is completed for surgery and if it's not, then it flags it. That has helped in saving time for nurses looking for consent services and forms so that they can focus more on the nursing aspects of their jobs.
Similarly, there is something called a history and physical form, which is basically a patient's history that is needed before surgery can happen. That's also hard to sometimes locate and can result in delays in surgery if it's not there. We have a machine-learning solution to identify them in advance and save time for our nurses who were otherwise looking through faxes and paperwork to find them.
We also noticed that our midweek inpatient census was peaking during Wednesdays and Thursdays, which is actually very common in large hospitals where the number of patients in the inpatient centers sort of peaks in the middle of the week. Most of it is not in our control, but some of it is related to surgeries that are optional and scheduled a certain way. We use machine learning to help us figure out if we could schedule our surgeries in a different way to not have such a high patient census in the middle of the week, which puts pressure on the hospital and on patient flow. We used machine learning to identify and adjust schedules for a few surgeons in our orthopedic joint surgery program that resulted in a lower mid-week census attributed to that program.
And then we looked at patients who come for certain oncology treatment. These patients have to go through different stations and parts of the hospital. This one is not live yet, but we are looking for the Center of IT Exploration to use Bluetooth to locate where the patients are.
We are usually looking at machine learning, artificial intelligence, natural language processing, location-based services, home monitoring solutions and voice.
Manu Tandon: Location-based services are going to become very important. A lot of our hospital is spread across multiple campuses. There’s a lot of movement of patients, physicians and staff. And patient and physician flow are a big topic of interest for us.
For example, if before a physician is entering a patient room, we can bring up the EHR of the patient on their smartphone, or their smart device, before they entered the room, or conversely, we can inform the patient on the TV screen in inside their room of who's about to see them. Or if we can use location-based services to prepare operating rooms to be efficient as a patient is being carted in advance of them reaching the room.
Those are all ways we can shave off minutes from the workflow and patient progression, which can all add up very quickly and can lead to decreasing of delays.
Manu Tandon: Our EHR was developed in the ‘70s and started with a grant from NIH to two of our physicians who wanted to build a patient-centric EHR, which was very innovative at the time. They started by basically building a master patient index with about 500,000 patients, which is how many we had at the time. Then they kept building clinical systems, lab systems, ancillary systems and ordering systems around it, and we sort of just kept on that path.
Around 2000, the system became browser-enabled — we were actually one of the first to have a browser-based EHR system, which means you can access it from any location on any device. Then we launched the part of the Open Notes program, which is a program under which patients can read the notes written about them on the patient portal. We were part of the group that was a pioneer for that initiative — and I think it has just kept evolving since then.
It compares favorably in terms of provider satisfaction scores and we also compare its functionality to the leading EHR vendor functionalities, and it compares on that account as well.
We have now recently merged, and we are now looking at this question of what our long-term EHR strategy is. But so far, this has been how it has evolved.
Manu Tandon: I think in-house systems allow you to control the workflow. Our system is highly tailored to our needs. We are constantly evolving our system based on the needs of our own conditions. It has been highly customized and was always built with a focus towards usability. Folks tell me that our system is highly usable, even folks that have experience with other EHRs.
It does come with its own with its challenges, though. You have to keep up with regulatory needs. And every once in a while, there's some state or federal mandate that we need to work on.
Most of the country is using outsourced EHR solutions, though, and I think we are the exception. For those that have already made the transformation, I can’t see them going back. It's so hard to put EHRs into place that I can't see people making an effort to going back. But there are definitely some advantages of having your own system in terms of being able to customize it and manage the workflow to your needs.
Manu Tandon: Dr. Halamka runs the center and he's looking at a system-wide strategy and IT innovations at the system level.
We’re more focused on the operations of the operational side of the hospital. So sometimes, tools that we will develop at the hospital level, Dr. Halamka might take that and see if they have applicability at the system level, or even at the national level. He also works with several international agencies to try to see if they're applicable there.
On the other hand, when he finds something that is valuable from somewhere else, he will introduce that to us. It's a two-way lane and we work well on that. Dr. Halamka’s group has a different focus than us, but we are complementary to each other.
Manu Tandon: The data center is where we host our EHR and all our other assets. It’s where we host our enterprise resource planning, PACS systems, file servers, research, infrastructure and exchange email systems. But we have been migrating assets from our data center to a public cloud — Amazon Web Services (AWS).
Manu Tandon: Multiple reasons. For one, the location that the data center is hosted in is a rented space that we have to vacate in about four years.
We also think that there are some advantages for us to go to the public cloud. We have done an extensive design work to mitigate any security concerns with the public cloud. We also found that going on the public cloud allowed us to take advantage of innovative solutions much faster. And actually, we leverage that public cloud model for our Center for IT Exploration quite gainfully.
Manu Tandon: The storage requirements are growing at a rate which is much faster than we have ever seen before — especially imaging requirements are growing exponentially. In the last two years, we have doubled the need for storage that took us almost 10 years to double before that.
So, I think cost is a big driver and people can ask themselves, if the hospital should be in the data center business. Of course, they will be data centers for certain things. But there is the possibility of getting flexibility and elasticity by going to the cloud. That’s a question people people will ask themselves. Migrating to the public cloud does require careful design and consideration from a security perspective and is something that is a different skill set that has to be developed.
We developed it internally using training and support. But in general, the long-term costs of running and maintaining a data center, especially as the cost of healthcare delivery is rising at a higher rate than the rate at which the revenues are growing, I think that's a strong consideration that folks will have to think through.
Manu Tandon: Patient engagement with the healthcare system. The way we engage, today's brick and mortar type of model, I think that'll change. With the advent of technology, I think
it will become more consumer oriented.
Providers would hopefully benefit in the long run with technologies that would allow them to increase time with patients and not have to have too much burden on the EHR side. I think EHRs have to mature quite a bit to get there and there is a huge opportunity for improvement.
Healthcare is already a complex field operation. More process engineering and productivity improvement ideas from other industries or even from the healthcare industry would need to continue to be adopted, so that we can reduce inefficiencies and also reduce costs in the system.
Manu Tandon: Workflow is king. It's not enough to have a cool new technology. Unless we think through the people and process aspect and the workflow aspect of the adoption or the implementation of the technology, it will never scale. So, thinking about scale is really important.
Get the best insights in digital health directly to your inbox.
Hear from More Executive Voices