Implementing a new data system for quality improvement or patient safety, such as an incident reporting system, can come with a wide range of challenges and potential pitfalls.
Missteps along the way can delay the rollout or undermine the performance of an expensive and critical tool.
Any data system can affect a wide range of hospital operations, interacting with and possibly hindering the operation of other systems, as well as clinical and business operations. Those far-reaching effects can be underestimated, giving hospital leaders a false sense of the scope of the project they are undertaking.
Successful implementation begins with hospital leaders identifying an accurate perspective on the work ahead, says Michael Rudomin, vice president of supply chain operations for Vizient, a company based in Irving, TX, that offers healthcare performance improvement assistance. That includes a realistic view of how big the project will be and how long it will take, he says.
“Every implementation I’ve been involved with ended up being longer and more complicated than anyone had planned. That has a lot of implications in a hospital on everything you do, on the clinical and business sides,” Rudomin says. “In one hospital I just recently worked with, the CFO knew that because of all the different processes we were about to implement, even at the very end we would still need to produce patient bills to send to the insurance companies. Patient bills, of course, are the end product of hundreds of transactions that occur upstream ... and those processes had been interrupted by some bumps in the road we encountered with introducing the new system.”
One System Affects Others
In anticipation of that interruption in billing, the CFO planned to have an additional three to four months of cash on hand at the end of implementation.
Those funds were intended to help cover the shortfall in income while the billing system caught up with the delays induced by the system introduction, Rudomin says. But it still was not enough.
That hospital’s experience involved billing and revenue, but Rudomin says the same sort of cascading problem can occur with quality and safety data, resulting in lags and inabilities to extract information needed for compliance reports and ongoing quality initiatives.
“The initial estimate of how they would handle that period of missing data was made in good faith, but their problem was based on the assumption that everything would go reasonably well,” Rudomin says. “We tested this system six ways from Sunday and thought we had identified everything that might possibly be a problem,” he continues. “Yet when it was time to flip the switch ... certain things didn’t work. No one could identify why they weren’t working because they were working 10 minutes ago in the test environment.”
Command Center Responds
Fortunately, the hospital had set up a 24/7 command center to respond to problems once the system went live. The command center included representatives from the vendor, consultants, hospital administrators, and clinicians. For the first three weeks after going live, command center team members met three times a day on all three shifts to look for problems and respond to any complaints.
When the bar code scanners for the “4 East” section stopped working, no one could scan patients’ identification bracelets; thus, patients could not receive their meds.
Leaders in the command center talked staff through the situation, going immediately to Plan B for manual identification. There still was a delay in providing medication, but without the command center’s ready response, it could have been much longer, Rudomin explains.
He notes the same command center concept can be used on a large scale, or on a small scale for system implementations that involve only one department.
No matter what is planned for a data system, it is critical to anticipate other problems arising. Leaders must be prepared for these issues, including the stress it puts on everyone to set up the system properly so they can do their jobs.
“It’s really quite a heavy lift. I don’t think I’ve ever worked with a hospital that went into it understanding that fully,” Rudomin observes. “Everyone talks to their colleagues and hears the war stories of what has happened somewhere else, and you incorporate that into your plan. But, inevitably, something will come up ... that throw a monkey wrench into your plans.”
Define Quality Improvement
New data systems may be implemented with the goal of improving quality of care. However, Rudomin says it is important to define exactly what that means. An ambiguous goal of improved quality may help obtain funding and approval for the new system. Still, in practice, the implementation must consider what that means specifically.
“You’ve implemented a new system, and now you have the ability to produce a lot more information, more quickly and organized in a way that is more useful than before,” Rudomin says. “But you’re only dealing with that side of the ledger. The other side of the ledger is the human side.”
If the people involved with the data system have been providing improper care or inputting data incorrectly, the new system has just provided them a faster, more elegant way to do the wrong thing, Rudomin explains.
Addressing the human side of the equation, through a review of the care processes that generate the data, should be a component of any quality improvement-driven data system implementation, he says.
Quality leaders also must remember that the introduction of any data system will standardize how some processes are handled. This can upset operations because many will have to switch from their old processes, which may have involved workarounds that made sense to them, to the new method. The hospital will have the power to require that behavioral change with staff who are employed or primarily work at that facility, Rudomin notes.
But the change in process can be particularly challenging for nonemployed healthcare providers who work at multiple sites. “At some point, physicians may decide that to be effective in seeing their patients, they are going to do it how they want to do it,” Rudomin warns. “You have to pay equal attention to who is interacting with the system, and how well or poorly they interacted with it before. If they interacted with it poorly before, it’s unlikely that will improve with a new system unless you address the problem.”
Determine Data Storage
Introducing a new system also may force an assessment of how much historical data a facility will store and make available, Rudomin says.
This can be a business decision based on the cost of data storage. Quality leaders may have to accept they cannot keep access to an unlimited amount of clinical data.
“Especially when your new system is a different software or process, not an evolution of a product you already have, you may have to establish a cutoff point. Keeping 10 years of clinical data online is just not cost-effective,” Rudomin says.
“Most hospitals will use the two-year rule, which means that everything from the past two years will be transferred to the new system and kept in active storage,” he continues. “Everything prior to that will require talking to the IT folks to obtain it from the old system.”
Still, hospitals often find two years worth of data is insufficient. Someone may need to see the record from a patient’s visit five years ago.
The standard process for requesting that access from the IT department could be too cumbersome and take too long. Many hospitals develop an expedited request process for such situations, Rudomin notes.
“Hospitals underestimate that need a lot. They think that if a file hasn’t been used in five years, then they don’t need to keep it in active storage,” he says. “But there must be a provision for the exceptions when that file is needed right away.”
Staggered Training Imperfect
Training employees on a new system also brings the possibility of missteps and oversights, Rudomin says. The timing of training is one potential problem area. When many employees must be trained on a new system, it becomes necessary to stagger the education so that groups go through the process as the system is implemented. “That means that the people who were the first to be trained on the system are the farthest away from the implementation date. When the system goes live, it has been a good while since they learned how to use it,” Rudomin explains. “If you trained people six months ago ... before it goes live, that training is worthless. You flip the switch to go live, and these people realize instantaneously that they have forgotten how to use it.”
That necessitates retraining, which comes with a cost. While they are going through retraining, staff are not fully functional in the new system, either less productive or nonproductive.
“No matter how well you train people, hospitals underestimate how detrimental that training period will be to the hospital’s productivity. Some hospitals will bring in additional human resources to fill the gaps in productivity while their employees are in training,” Rudomin explains. “You may have 50% of a department in training, but patients are still coming in, and there is work to be done.”
When implementing systems, hospital leaders may acknowledge productivity will slow down, but it can be managed. Rudomin says productivity slows much more than leaders like to believe.
“Where you can do something to minimize the effect with additional staff, do that. But please remember that you are stretching these people very thin and staff burnout will start to show up after a couple months,” Rudomin says. “People will be training half the time, then working the other but trying to catch up on the work that backed up when they were training. They start working overtime ... after a couple months, they are completely burned out.”
Rudomin also cautions against an overly rosy estimate of how a new data system will improve quality of care or patient safety. Once a system is operational, the actual impact may not be what was expected.
“Some will find that the system is a great tool and allows them to be more productive and effective. Other folks will find that it is a step backward for them,” Rudomin notes. “It takes them longer, not because they don’t know how to use the system but because the design of the system requires them to jump through more hoops to do the same thing they were able to do before.”
That can lead to a significant morale problem. In severe cases, employees will start leaving the hospital because the new data system was so burdensome that they had become miserable.
A common refrain is the new system may have optimized one department’s work, but it wasn’t designed to optimize all departments, Rudomin says.
Hospital leaders must prepare to deal with this outcome, even if it cannot be entirely avoided. That might mean staffing up in anticipation of some departures or offering bonuses to offset the frustration of employees, Rudomin suggests.
For the implementation of any data system, large or small, it is vital to manage expectations and follow the “rule of twos,” Rudomin says.
“The rule of twos says that it will take twice as long to get half as far as you thought you were going to be at a certain time,” he says. “If hospitals are willing to accept that and factor it into their expectations, they are going to have a better outcome.”
- Michael Rudomin, Vice President, Supply Chain Operations, Vizient, Irving, TX. Phone: (972) 830-0000.