Small Patients, Small Errors, Big Impact
January / February 2009
Small Patients, Small Errors, Big Impact
A Six Sigma project in a community hospital emergency department found and corrected the root cause of inappropriate orders that were undermining pediatricians’ confidence in the quality of care.
How would you react if you learned that local pediatricians were steering patients away from your hospital’s emergency department and sending them to a competitor because they lacked confidence in the quality of care in your emergency department? Memorial Hermann Baptist Hospital (MHBH) in Beaumont, Texas, confronted that disturbing scenario in 2006. The business impact was disturbing enough, but the patient safety implications were even more unsettling. As clinical nurse specialist for pediatric services at a sister institution, Children’s Memorial Hermann Hospital (CMHH), I was asked to lead a quality improvement project that employed Six Sigma methods to address and correct the problem quickly and permanently.
Pediatricians were concerned about cases in which ED physicians gave inappropriate care orders for patients destined for admission. The investigation found that the errors were not reaching the young patients because pediatric unit nurses, recognizing inappropriate orders, would call the children’s pediatricians, who then issued new and appropriate orders. Although the children received the appropriate care, the confidence of their pediatricians in the ED was shaken.
The Six Sigma project, assisted by the Solutions consulting practice of GE Healthcare, led to process changes that significantly reduced inappropriate orders from the ED. It also restored pediatricians’ faith in the ED and helped build better relationships between ED and pediatric nurses and physicians. Strict adherence to the rigorous discipline of Six Sigma was critical to the project’s success.
Methodical, Measurable
I first addressed the care of pediatric patients in community EDs back in 2006 when I was completing an application to attend the Virginia Commonwealth University (VCU) Executive Fellowship in Patient Safety. Having reviewed numerous pediatric cases as a legal nurse consultant, I had seen many patient safety cases involving children originate in community EDs. For some time, I had been looking for a way to marry my clinical knowledge with information I had acquired as a legal nurse consultant.
In my role with Children’s Memorial Hermann, I was often called upon by directors throughout the 15-hospital Memorial Hermann system to help troubleshoot pediatric issues. Soon after applying to the VCU program, I received a request from the pediatric director at MHBH to assist with issues affecting pediatric patients in the ED at her hospital.
MHBH is a 250-bed acute-care hospital located in Beaumont, a city of 100,000 in southeast Texas. It houses a 17-bed pediatric inpatient unit with three intermediate-care beds. The issues in the ED were the same ones I had listed in my VCU proposal.
The hospital’s pediatric director said the community pediatricians were quite pleased with the care given by the nursing staff on the inpatient pediatric unit, but they were not comfortable with the care given in the ED. The pediatric unit census remained high throughout the week when the pediatricians directly admitted their patients. However, on the weekends, when the pediatricians’ offices were closed, the inpatient unit census dropped.
The pediatricians readily admitted to steering patients away from the ED as much as possible because they did not feel the care there was optimal. Their specific complaints included orders for fluid boluses not based upon weight, fluid boluses using fluids containing glucose, orders for inappropriate antibiotic and inappropriate dosing of antibiotics.
The VCU Executive Fellowship for Patient Safety accepted me and approved my proposal to work on a project identical to the MHBH case as my graduation project. Early in the VCU program I was selected in my job to be trained as a Six Sigma Green Belt. As I progressed through Six Sigma training, I noted several flaws in my VCU proposal — mainly the lack of measurement.
The more I learned in Six Sigma training, the clearer it became that if the ED project were to succeed, it would have to be completed in a highly methodical and measurable manner. With the help of my Six Sigma Master Black Belt, I redesigned the proposal to use Six Sigma methodology to guide the safety project at MHBH. Here is how the project progressed through five phases of the Six Sigma method, commonly known as DMAIC: define, measure, analyze, improve, control.
The Define Phase
The problem was defined as pediatricians’ displeasure with the care given to their patients in the ED at MHBH. This displeasure caused them to send their patients to the competitor’s ED, even though they were pleased with the pediatric inpatient care at MHBH. The pediatricians were the primary customers for the project, and patients and their families were secondary customers. Because the primary focus of the project was patient safety, it was aligned administratively with the chief patient care officer for MHBH, who provided approval and support.
The ED project’s goal was to decrease the number of variances in care by 75% within 6 months. Variances were defined as events in the ED where the care provided or ordered by the ED physician differed or conflicted with what the patient’s pediatrician would have found acceptable for the patient.
The Measure Phase
To determine the exact nature of the problems with pediatric patients in the ED, we pulled and analyzed all variances (incident reports) for 1 calendar year. Pediatric patients were defined as non-maternity patients aged 17 years or younger. We identified 18 variances among pediatric patients, and of these, 11 were specifically related to patient care issues in the ED. Those issues fell into six categories:
- Inappropriate medications
- Inappropriate medication dosing
- Inappropriate fluid type
- Inappropriate fluid dosing
- Missing diagnostics
- Incomplete nurse-to-nurse report
We reported this analysis to the pediatric nursing director and to the medical director for the MHBH pediatric unit. They agreed that these issues were among those reported as concerns by various pediatricians, but they felt there might be other issues as well. They added six more categories of possible adverse occurrences:
- Missing orders
- Inappropriate treatments
- Inappropriate triage
- Failure to assess or reassess
- Failure to follow up
- Other events
Once the list of issues was acknowledged, the task was to identify a group of patients to examine. A list of all pediatric patients for the 2005 calendar year aged 17 or younger revealed that there had been 8,963 pediatric visits. This large number of visits made it a challenge to sort the data into usable information. And it was apparent that the types of errors the pediatricians were reporting did not occur in all pediatric patients seen in the ED, but only in those patients with higher acuity levels. Categorizing nearly 9,000 patients seemed impossible, but I worked with my Six Sigma Master Black Belt to devise a sort plan.
The patients had 3,221 distinct diagnoses. After assigning an acuity ranking to each diagnosis, we sorted the children into groups. After sorting, the sample size was reduced to 1,969 patients. We randomized these patients and searched 60 records for the 12 concerns identified by the pediatric nursing and medical directors. We found only two errors.
What did this mean? Were the issues in the ED not real but only perceived? The pediatric nursing and medical directors did not think so. In fact, they reported seeing another error in the ED only days before. Their assurance led to the conviction that something was amiss in the way the data had been sorted.
A return to the original variances for a closer look revealed the misstep. We had randomly examined charts from all pediatric ED patients of higher acuity, including those treated and released. But all of the variances had occurred in patients who had been admitted to the pediatric unit from the ED. In addition, all the patients for whom variances had been written were less than 6 years of age.
We then went back and reviewed the charts of all the patients from the original sample of 1,969 who had been admitted. That reduced the sample size to 222 patients, and we reviewed 149 of those charts.
The Analyze Phase
The 149 charts we reviewed contained 77 defects or errors. In terms of Six Sigma measures, this represented a defect per million opportunities (DPMO) of 80,800 with a yield of 92.6% — meaning that among orders written for pediatric patient admissions, seven of each 100 would contain a defect. Analysis revealed that the 77 errors fell into six categories:
- Inappropriate medication
- Inappropriate medication dosing
- Inappropriate fluid type
- Inappropriate fluid dosing
- Missing diagnostics
- Other events
The number-one concern driving the project forward was fear that a child would be harmed by our care. We had no evidence that any harm had occurred, but we were concerned that it could. In analyzing the data, we found several things. First, we were pleased to find that the errors had not reached our patients. Second, we found that the errors were written into the admitting orders and did not represent the actual care given in the ED. Upon examination, the care given to children in the ED was found to be appropriate. It was the process of admitting pediatric patients that led to the problem areas.
The ED physicians were writing the admission orders. The children were then sent to the pediatric unit. Upon examination of the admission orders, the pediatric nurses, recognizing the inappropriateness of the orders, were then calling the patients’ pediatricians who issued new, appropriate orders for the children. The fallout from this, however, was that pediatricians thought the ED had no idea how to care for children.
The Improve Phase
We first presented the results of the data analysis to the ED physicians. Their initial reaction was predictable. They were not pleased with someone critiquing their care. However, once they realized that the emphasis was focused on the process and not on individuals, they became more engaged. The ED physicians readily admitted that they had no idea how to care for children as inpatients and were open to looking for solutions jointly with the pediatricians. Their suggested solutions included a specific order set for the admission of children, as well as education about pediatrics for ED nurses.
Once the ED physicians had seen the data, we devised a plan to share it with the pediatricians and win their support to improve the process of admitting pediatric patients from the ED. We presented the information at a quarterly pediatric medical staff meeting. The pediatricians were also pleased to learn that the errors had never reached their patients; they were willing to help develop process improvements. They agreed with the proposed solutions from the ED physicians.
The pediatric nursing director then developed a draft of the pediatric admission orders; the pediatric medical director reviewed it and made changes. Next, the two directors invited several pediatricians to review the orders at a lunch meeting. All invited pediatricians attended the meeting, which began with a review of the data and the processes currently in place. The ED physicians had issued a slight challenge that the meeting hosts shared with the group: They doubted the pediatricians would be able to agree upon one standardized set of orders.
The pediatricians took the challenge and finalized a complete order set in one hour. This set of orders was then sent to all community pediatricians for their approval. They accepted unanimously — and so did the ED physicians when asked for their input.
A pilot using the order set was run for 3 weeks. Twenty charts were reviewed after the pilot, and only one defect was found. The pilot was then extended. After 3 months of use, 75 charts were reviewed, and eight defects were found. That translated into a DPMO of 15,238. The yield increased to 98.5%, representing a 79% improvement. The ED physicians used the pediatric order sets 97% of the time.
The Control Phase
After 3 more months of order set usage, we reviewed 82 more charts. The ED physicians were 100% compliant in using the standard pediatric order set. There were six defects, translating into a DPMO of 10,452. The yield increased to 99%, for an overall project improvement of 86% (see Figure 1).
|
This information was reported to all involved: the ED physicians, the pediatricians and hospital administration. All were extremely pleased with the results. A part of the control phase was to institute a plan to maintain the improvement. This included handing the project over to the process owners. The pediatric director who had initiated the project agreed to take ownership.
Going forward, the plan included monitoring of order set usage and continued random chart reviews on a quarterly basis. The data from these reviews will be reported at the ED medical staff meetings monthly. Any drop-off in the order set usage rate or increase in defects — if it happens — will be addressed jointly by the pediatric and ED medical directors, at the initiation of the pediatric nursing director.
Conclusions
Use of the Six Sigma methodology made the project a success. If the original proposal had been carried through, the project would have been entirely different, and likely unsuccessful. Because Six Sigma required careful definition and investigation of the factors involved, the project accurately identified the root causes of the problem.
Several months after the second measurement of results, the pediatricians had made no further complaints to the pediatric nursing director about the ED. The pediatric nursing staff expressed contentment with the new order set, and they were pleased that they no longer had to call the pediatricians each time an admission came up from the ED. The ED physicians expressed satisfaction with the results of the project.
In addition, collaborative relationships were forged. The nursing staff and physicians in the ED consult the pediatric unit and pediatricians much more often. Because the pediatric nurses now perceive the ED staff as partners in quality care, they are much more willing to assist the ED nursing staff, go to the ED to start intravenous lines, or make suggestions for possible interventions. This collaboration is an additional benefit that further promotes patient safety for pediatric patients at MHBH.
Susan Engleman is a clinical nurse specialist and pediatric nurse practitioner who is currently employed as director, System Children’s Services, with Children’s Memorial Hermann Hospital in Houston, Texas. Engleman has many years of pediatric nursing experience in numerous roles practicing in a variety of settings including pediatric intensive care, intermediate care, acute care, and home care. In 2006, Engleman completed the Executive Fellowship in Patient Safety through the Virginia Commonwealth University in Richmond, Virginia. She may be contacted at susan.engleman@memorialhermann.org.