Hi, my name is Mark Zaghi. I'm a fourth year Harvard MD MBA student. I'm going to talk to you today about OR efficiency, identifying causes of delayed first starts. Operating room delays lead to poor utilization of both labor and resources, as well as leading to increased healthcare costs. The highly complex and multifactorial nature of the process often makes it difficult to isolate individual causes of delays, and attempts are often subjective or biased. Thus, our multidisciplinary team aimed to use objective EHR data to identify causes of delays. Specifically, we extracted timestamp data from EPIC on all neurosurgery first case starts at Mass General Hospital for one year. To focus on more meaningful delays, we only included delays which were greater than 15 minutes from the time the patient was scheduled to enter the operating room. Our team also shadowed the process from the patient's perspective, from hospital arrival to entering the OR, as well as validating timestamps and assessing the various tasks the stakeholders encountered along the way. Here is the process map of the EPIC time variables we used. You can see that the patient preparation and room preparation are occurring in parallel, while the following timestamps after the patient enters the room are listed in serial. We identified three main segments and listed their respective key personnel that play a majority of the role in processing the patient through these timestamps. In the prep for surgery segment, for example, this would include patient positioning, shaving the head, and brain lab registration. While the induction segment and prep for surgery segment are shown in serial, it is important to note that there can be a significant amount of parallel processing that occurs between these two. Overall, we found that 27% of our first starts were delayed. Keeping in mind that only delays greater than 15 minutes were included, our median delay was 25 minutes, and we had a total delay time of 157 hours for the year. Here, you can see the variability amongst various surgeons. Although the subspecialty of each surgeon is not shown for anonymity's sake, there was no association found between subspecialty and likelihood of the patient being delayed into the OR. On delayed cases, the time segment that was significantly different appeared to be the pre-op complete to in-room segment of 86 plus or minus 41 minutes for delayed cases versus 69 plus or minus 33 minutes for non-delayed cases with a p-value of less than 0.01. Regarding patient arrival, patients are normally instructed to arrive two hours prior to the start of their procedure. Excluding our academic day, where the procedures are scheduled for later in the morning, we found that our delayed patients were arriving at 5.56 a.m. on average, or 11 minutes later than expected. It's easy to attribute this to the patient's arriving time being delayed, but our observation experience painted a very different picture. Interestingly, when we observed this process, we found that upon arriving to the hospital lobby, patients are not allowed to go to the Center for Perioperative Care, or CPC, until 5.30 a.m. for security purposes. At that point, there's a significant bottleneck which leads to about an approximately 40-person line outside the CPC extending to the elevator bay. For the first half hour, there are only one to two receptionists, which take approximately two minutes per patient to process this line. Once the patient finally gets to the receptionist, the patient arrival and department arrival timestamps are both checked simultaneously. The patient is then taken to a CPC nurse a few minutes later, at which point pre-op start is checked, and then they are transferred to the neurosurgery PACU after the patient changes into their gown and reviews a short checklist. At the PACU, the nursing, neurosurgery, and anesthesia teams all see the patient. We found out there was significant confusion regarding when pre-op complete should be checked, and found that often it was when the PACU nurse ran through his or her checklist, but prior to anesthesia or neurosurgery completing their respective preoperative work. Rarely, we even saw that this timestamp was checked after the patient entered the room. To help determine why these delays occurred, we had one of our nursing managers circulate around the ORs and ask for reasons why delays occurred when they did. Notably, we saw no Hawthorne effect in this month compared to our aggregate data. The most common cause of delay was patient preparation. Breaking that apart revealed that lines, both venous and arterial, tend to be common causes of these delays. In summary, we used objective data to identify causes of delayed OR first starts, which included delays in patients being checked in, patient preparation in the pre-op setting, as well as some confusion regarding a variable used for data collection. Our potential solutions include exploring direct transfers to the neurosurgery PACU, shifting lines into the OR to allow for more parallel processing, as well as standardizing and explaining variables for data collection. Future directions include applying these solutions, which has been delayed due to COVID-19, expanding beyond first starts into second and third surgeries, including turnover time, expanding to the other surgical departments and other institutions. Lastly, more of a note than a future direction, it's important that these process improvement initiatives allow all stakeholders' voices to be heard to facilitate everyone's collective buy-in towards improvement. With that, I want to thank all the members of our multidisciplinary team and give my sincere gratitude to Dr. Brian Naheed for his outstanding mentorship and guidance. Thank you.

operating room delays

resource utilization

electronic health record data

neurosurgery cases

workflow optimization