Reuters Health Information (2014-01-17): Does compliance with process measures impact quality of care?
Does compliance with process measures impact quality of care?
Last Updated: 2014-01-17 17:39:42 -0500 (Reuters Health)
NEW YORK (Reuters Health) - Canadian researchers saw fewer postoperative complications when quality scores for process measures were higher - and the main "driver" of higher quality was early ambulation, they say.
One expert not involved in the study, however, thinks "early ambulation" was a proxy for other patient characteristics rather than a protective factor.
The study, by Dr. Simon Bergman and colleagues at the Jewish General Hospital in Montreal, used data from the hospital's National Surgical Quality Improvement Program (NSQIP) database on 273 adults (mean age, 64) who had elective major abdominal surgery and hospital stays of at least two days.
Hospital-based experts distilled more than 60 quality improvement parameters down to 10 that could be reviewed using the institution's electronic medical record system: use of prophylactic antibiotics, postoperative euglycemia, venous thromboembolism prophylaxis, central venous line care, urinary catheter documentation, postoperative ambulation, presence of a current medication list, pressure ulcer risk assessment, oral intake documentation, and the use of a surgical safety checklist.
They calculated a quality score for each patient by dividing the number of quality indicators charted by the number for which each patient was eligible.
The average Charlson Comorbidity Index (CCI) was 3.4 and the mean hospital stay was eight days. Almost two-thirds of the operations were colorectal resections, and 10.3% were hepatobiliary procedures.
As reported online in Surgery, 82 patients (30%) had at least one complication. The most common were surgical site infections (occurring in 14.7%) and bleeding requiring transfusion (in 12.8%). Four patients (1.5%) died within 30 days of surgery.
The mean percentage of quality indicators in the charts was 65.9%. Oral intake was documented in only 9.6% of records, while prophylactic antibiotics were documented in 95.6%.
Most indicators were present in more than 50% of cases, however, and overall patient quality scores averaged 67.2%.
Poisson regression showed that as quality scores rose, the rate of postoperative complications fell (p = 0.011).
But the authors say the only process indicator significantly associated with fewer complications was documentation of early postoperative ambulation, which was charted in 61.5% of patients. Of that group, 21.4% suffered a complication. For 38.5% of patients, there was no note in the chart about ambulation, and 43.8% of those patients had at least one complication (p<0.001).
Dr. John D. Birkmeyer of the University of Michigan, a surgeon with an extensive bibliography in the area of outcomes research, told Reuters Health by phone that the results "are way too good to be true."
"A randomized clinical trial of early ambulation or even a constellation of perioperative processes is unlikely to duplicate the results found in this paper," Dr. Birkmeyer said.
He believes early ambulation was probably a proxy for patient factors rather than a cause of fewer complications and that the results could have been confounded by two factors.
"Patients who experienced early complications become noncompliant with perioperative pathways," he said. "Nurses likely are busy with higher priority clinical needs to chart many of these processes."
He added, "Preoperative patient frailty, and particularly mobility, would affect the link between early ambulation and postoperative outcomes. A patient who is frail and immobile before surgery is unlikely to do well with early ambulation."
The authors noted that higher CCIs and wound classifications were tied to more complications in this cohort (with p values of 0.002 and <0.001, respectively).
"Large multicenter outcomes registries such as NSQIP, are powerful tools, but it is important to consider the impact of selection bias and confounding factors when analyzing the data," Dr. Birkmeyer said.
The corresponding author declined to comment.