A call for healthcare systems engineering
BOSTON--“Healthcare is a mess” and the industry sorely needs more systems engineering to bolster patient and provider safety, said Jeanne Huddleston, MD, medical director of the health systems engineering program at the Center for the Science of Health Care Delivery, at the Big Data Healthcare Analytics Forum on Nov. 21.
Huddleston, who received an industrial engineering degree later in life, had looked at the implementation of guidelines and checklists in the healthcare setting and said, “I couldn’t believe the lack of data behind the things we do.”
Hospital-acquired infections account for $9.8 billion each year, with projected spending at $4.5 trillion by 2019. “How in the world do we take all the data we have and create something meaningful when healthcare can’t get the right thing to the right patient 54 percent of the time?”
Healthcare systems engineering is an answer, as it’s a data-driven approach to how care is delivered. “We have the pieces and parts, all the medical sciences, but we’re not putting them together in the right way,” Huddleston said.
She cited four problem domains to tackle with systems engineering: capacity of resource management; safety (both employee and patient); workload and treatment optimization; and process efficiency and reliable care delivery. Part of this translates to locating the right equipment to the right patient at the right time, lab test tracking and smarter appointment setting, for example.
Hotels use predictive analytics to fill beds and credit card companies spot red flags before consumers do, she said. “Businesses have been re-engineering themselves for the 21st century. Imagine if healthcare did, too.”
For example, at Mayo Clinic, Huddleston’s Center for the Science of Health Care Delivery analyzed neurology appointments, including unfilled appointments, overbooks and days to the next available appointment. Every week they send out the data, which some groups use to adjust appointment times based on type. The groups that have used it have seen results. “We could have amazing data with amazing results, but unless we can get providers to use it, it’s useless.”
The process of coming up with a surgery time for spine surgery, for example, often is decided without consulting data. At Mayo, they came up with an interface in which patients pick a day for the surgery that isn’t already overscheduled. “We started with surgeons and patients, and only used those variables. If it doesn’t pass the sniff test for surgeons, we don’t use it.”
When it was implemented, they saw results: a cost reduction from 3 to 2 rooms, “fewer no-hitters,” decreased weekend stays of Medicare patients, less use of the operating room outside of prime times and increased revenue. The model also takes into account whether or not the patients will need skilled care and their anticipated length of stays. “These are the types of things we don’t think about when [the physician] and Ms. Smith are in the room.”
Systems engineering and process design also are critical to ensure physicians are directed to the patient who needs it most at a given time. Many years ago, Mayo Clinic performed a mortality review of 9,000 deaths in Rochester and found that the No. 1 cause of death was nurses not recognizing the quickly deteriorating condition of patients.
They applied healthcare system engineering and found some inherent failures, including failure to re-assess the patient’s clinical condition, that the physician wasn’t reviewing nurses’ notes and that care providers were overburdened with too many complex things to do.
Mayo Clinic started collecting data from nurses on how worried they were about each patient at the end of their shift. With a 95 percent response rate, they collected 31,000 nurse data points. “We hit a nerve, we hit something important.”
The data yielded a positive predictive value over 40 percent, and now prospective trials are underway to integrate this into clinical decision support. “We’re developing visual ways for make it easy for physicians to know if patients need to be looked at,” Huddleston said. “Docs needs to come back to the bedside.”