Audit and Feedback Urinary Catheter Duration

The ubiquitous urinary catheter is associated with 80% of hospital‐acquired urinary tract infections (UTIs)1estimated to number one million annuallyaccounting for 40% of all nosocomial infections.2, 3 The clinical consequences of these catheter‐associated urinary tract infections (CAUTIs) are substantial and include prolonged hospital stay, bacteremia, and death.4 Despite the known risks of CAUTIs, it is estimated that 25% of hospitalized patients receive urinary catheters and that inappropriate urinary catheter use is widespread.5 Among catheterized patients, catheter duration is the most important modifiable risk factor for CAUTI. The excess risk of any bacteriuria accrues at a rate of 5% per catheter‐day beyond the first 48 hours of catheterization.6, 7 The reduction of catheter‐days for a given patient is an important component of quality improvement efforts to reduce CAUTIs. Unfortunately, as of a 2005 survey, most hospitals do not systematically track urinary catheter insertions and removals.8

The above concerns are highlighted for surgical patients among whom indwelling urinary catheter use is particularly high. In a 2001 sample of Medicare beneficiaries, 85% of major surgical patients had perioperative urinary catheters.9 In this population, postoperative catheter duration exceeded 48 hours in nearly 50%, despite concern that the risks of infection offset the benefits of continued catheterization after 24 hours to 48 hours postoperatively.4, 7, 10 Patients with catheters greater than 2 postoperative days had a 21% increased likelihood of in‐hospital UTI, increased 30‐day mortality, and decreased odds of discharge to home.9

To address the risk of CAUTI associated with excess urinary catheter days, The Centers for Medicare and Medicaid Services' (CMS) Surgical Care Improvement Project (SCIP) added catheter removal on postoperative day 1 or 2 to its process measure set beginning in October 2009. SCIP is 1 of several high‐profile surgical quality improvement programs that employs performance audit and feedback of patient‐level process or outcome measures to address deficiencies in surgical care.11, 12 In addition, audit and feedback of CAUTI rates has been used to successfully reduce CAUTIs in medical‐surgical patients.13 The goal of our study was to audit patient‐level postoperative urinary catheter duration and measure the impact of its feedback to nursing staff on postoperative catheter duration, CAUTI rates, and nurse's attitudes about CAUTI prevention.

Methods

Results

During the study period there were a total of 1657 surgeries on the 2 study units during the baseline and follow‐up periods. After exclusions for urologic or gynecologic surgery (271), no indwelling catheter for surgery (505), or first surgery of 2 or more during the hospitalization (31), there were 846 eligible surgeries (51%).

Table 1 describes the population for the baseline and follow‐up periods for orthopedic and general surgery patients. Within each unit, the surgical populations were comparable during the baseline and follow‐up periods with the exception that the mean length of stay for eligible general surgery patients was significantly shorter in the baseline period as compared to the follow‐up period (6.6 vs. 8.5 days, P = 0.02). Cases on the orthopedic surgery unit were predominantly knee, hip and spine surgeries (85.9%), while those on the general surgery unit were predominantly gut and other gastrointestinal (GI) procedures (80.3%).

BaselineFollow‐Up Comparison

Table 3 describes the measures of urinary catheter use for each surgical population for both data collection periods. On both units, measures of catheter duration were improved following the education and feedback intervention. For the orthopedic unit, mean postoperative catheter duration was reduced from 1.7 to 1.4 days (P = 0.01) and the proportion of patients with catheter removal before day 3 was increased from 86% to 92% (P = 0.04). For the general surgery unit, mean postoperative catheter duration was reduced from 2.6 to 2.2 days (P = 0.01) and the proportion of patients with catheter removal before day 3 was increased from 56% to 63% (P = 0.14). When the general surgery measures were adjusted to account for the difference in length of stay between the 2 time periods, the odds of meeting the performance measure at follow‐up compared to baseline increased from unadjusted odds of 1.38 (P = 0.14) to adjusted odds of 1.69 (P = 0.02).

Table 3.Measures of Indwelling Urinary Catheter Duration and CAUTI Rates

Measure	Orthopedic Surgery			General Surgery
	Baseline, n = 206	Follow‐Up, n = 290	P Value	Baseline, n = 167	Follow‐Up, n = 183	P Value
Abbreviations: CAUTI, catheter‐associated urinary tract infection; ns, not sufficient; SD, standard deviation.
Postoperative catheter duration in days (mean, SD)	1.70 (1.24)	1.44 (0.85)	0.01	2.64 (1.85)	2.19 (1.40)	0.01
Postoperative catheter duration performance measure (%)	86	92	0.04	56	63	0.14
Total catheter days	350	418		441	401
Catheter days/1000 hospital days	423	394	ns*	398	259	s*
Catheter‐associated UTIs	3	0		3	3
Catheter‐associated UTI rate (infections/1000 device days)	8.6	0	ns*	6.8	7.5	ns*

Figure 1a and b are histograms of the frequency of cases having a given postoperative catheter duration in days. The dark bars show the baseline distribution and the light bars show the follow‐up distribution. Although the number of patients in the follow‐up period is greater than for the baseline period for the orthopedic surgery cohort, the images are instructive. In both groups of patients, but most notably in the general surgery population, the reduction in catheter measures resulted from a left shift in the frequency distribution, both for the longer duration outliers (removing the tail of each plot) and for the shorter duration catheters (3 days), increasing the proportion of catheters removed on postoperative day 1.

Figure 1

Discussion

This preobservational and postobservational study found that audit and feedback of patient‐level postoperative urinary catheter duration delivered in the context of an educational intervention and brainstorming session was temporally associated with clinically meaningful reductions in urinary catheter duration. In so doing, we demonstrated the feasibility of collecting patient level urinary catheter duration, and delivering it in a manner that had utility for frontline staff. Our results are consistent with the quality improvement literature which demonstrates that audit and feedback is a successful quality improvement strategy in many contexts and may be as good as more complex interventions at increasing adherence to the performance of process measures for surgical infection prevention.14 Two large national programs, the VA National Surgical Quality Improvement Program (NSQIP) and SCIP, use audit and feedback as the backbone of their large‐scale quality improvement strategies with promising results.11, 12

Given that the 2 study units were so different in practice patterns regarding urinary catheter management and nursing‐identified barriers to evidence‐based care, this work suggests that urinary catheter management may pose unique challenges for different clinical areas and provides a caution that one‐size‐fits‐all interventions for the rationalization of urinary catheter management and reduction of CAUTIs may be of limited effectiveness in the absence of local tailoring. As such, audit and feedback is well‐suited to this purpose as more proscriptive quality improvement strategies may meet with a variety of implementation challenges.

The impact of our intervention on CAUTI rates was not significant. There are several possible explanations for this finding. First, the study was not powered to detect a difference in CAUTI rates given the low infection rate at our institution. However, we cannot exclude the possibility that a reduction in mean catheter duration of one‐third to one‐half of a day is insufficient to impact CAUTI rates in postoperative patients, particularly when many of the follow‐up patients still had postoperative catheter duration 2 daysthe timeframe beyond which bacterial colonization of the catheter begins. While both study units had similar increases in postoperative catheter duration, the UTI rate was only decreased in the orthopedic surgery group which had much higher rates of postoperative catheter duration 2 days at baseline.

In recent years, there has been a renewed focus on the eradication of hospital‐acquired infections prompted by intense interest from the public, federal and state legislators, and others.15, 16 The CMS has recently used the revamping of the Inpatient Prospective Payment System (IPPS) as an opportunity to align financial incentives so that reimbursements for claims with certain hospital‐acquired conditions, including CAUTIs, will be reduced to that of the reimbursement of the same claim without the presence of the complication.17 This move is just one of several strategies to motivate hospitals and clinicians to address the pervasive problem of hospital‐acquired infection.

As urinary catheters are intimately linked to hospital‐acquired UTI, a focus of reduction efforts on catheter use is appropriate. The National Quality Forum (NQF) endorsed a postoperative catheter duration quality measure which was incorporated by the CMS's SCIP in late 2009.18 As a result, every hospital in the country that performs surgery and participates in the Medicare program is now tasked with determining patient‐level urinary catheter duration for selected surgical patients. This move represents a departure from current recommendations from the Centers for Disease Control and Prevention (CDC) and its National Healthcare Safety Network19, 20 which endorse the measurement of a catheter utilization ratio (urinary catheter days/patient days) for patient care units, but does not endorse any patient‐level utilization measures. In this instance, the use of patient‐level data may be better suited to quality improvement interventions such as audit and feedback because of its clinical relevance to frontline providers. However, it may also increase the data collection burden on hospitals. Notably, the measurement of postoperative catheter duration in this study was semiautomated using queries of an EHR. Such an approach can significantly reduce the data collection burden for this process measure and is consistent with national initiatives to integrate EHRs with quality improvement initiatives going forward.21

Our study has several limitations. This study took place in the year following the announcement of a high profile Medicare rule change regarding payment for hospital‐acquired harms. Certainly, the uncontrolled prestudy and poststudy design does not allow for the assessment of the impact of our intervention independent of this context. We are unable, therefore, to attribute the observed reduction solely to the intervention. Additionally, we did not follow postoperative urinary catheter duration beyond the immediate follow‐up period. It is anticipated that the impact of an audit and feedback intervention may diminish over time without a mechanism for repeated feedback. Certainly the sustainability of such repeated feedback in a single institution would be improved with an appropriately configured EHR.

In addition, we have reliable data on catheter reinsertions only from the follow‐up period. While the rates of reinsertions we recorded (0.7% on orthopedic surgery and 2.7% on general surgery) were lower than expected based on the literature,22 we are unable to determine if our intervention led to increases in postoperative urinary retention.

This study was limited to 2 surgical units of a single academic medical center and therefore the urinary catheter utilization patterns may not be representative of other patient populations at other institutions. However, the urinary catheter patterns were comparable to those identified in our prior work in a national sample of Medicare patients undergoing elective surgery.9

Finally, the field of CAUTI prevention has evolved rapidly since this study was performed. In particular, the surveillance definition of CAUTI was altered twice by the CDC in December of 2008 and March of 2009. In addition, the Infectious Diseases Society of America issued a new definition of CAUTI in February of 2010.23 All of these changes highlight the difference between asymptomatic bateriuria (ASB) and symptomatic CAUTI. However, the surveillance definition in use at the time of this study did not make this distinction. Therefore, we are unable to comment on the relative occurrence of ASB versus symptomatic CAUTI under the new definitions.

Rational urinary catheter use is a central component of CAUTI prevention efforts.24 We describe the use of patient‐level urinary catheter use in an audit and feedback intervention to frontline staff that was associated with reductions in urinary catheter duration. To do so, we employed a methodology for tracking urinary catheter use patterns that can provide important data for infection preventionists and frontline providers in efforts to improve urinary output management. This promising approach merits further study as an adjunct to current efforts to rationalize urinary catheter utilization and reduce CAUTIs. In the current environment, having the right data is a powerful aide for ongoing performance improvement.

The ubiquitous urinary catheter is associated with 80% of hospital‐acquired urinary tract infections (UTIs)1estimated to number one million annuallyaccounting for 40% of all nosocomial infections.2, 3 The clinical consequences of these catheter‐associated urinary tract infections (CAUTIs) are substantial and include prolonged hospital stay, bacteremia, and death.4 Despite the known risks of CAUTIs, it is estimated that 25% of hospitalized patients receive urinary catheters and that inappropriate urinary catheter use is widespread.5 Among catheterized patients, catheter duration is the most important modifiable risk factor for CAUTI. The excess risk of any bacteriuria accrues at a rate of 5% per catheter‐day beyond the first 48 hours of catheterization.6, 7 The reduction of catheter‐days for a given patient is an important component of quality improvement efforts to reduce CAUTIs. Unfortunately, as of a 2005 survey, most hospitals do not systematically track urinary catheter insertions and removals.8

The above concerns are highlighted for surgical patients among whom indwelling urinary catheter use is particularly high. In a 2001 sample of Medicare beneficiaries, 85% of major surgical patients had perioperative urinary catheters.9 In this population, postoperative catheter duration exceeded 48 hours in nearly 50%, despite concern that the risks of infection offset the benefits of continued catheterization after 24 hours to 48 hours postoperatively.4, 7, 10 Patients with catheters greater than 2 postoperative days had a 21% increased likelihood of in‐hospital UTI, increased 30‐day mortality, and decreased odds of discharge to home.9

To address the risk of CAUTI associated with excess urinary catheter days, The Centers for Medicare and Medicaid Services' (CMS) Surgical Care Improvement Project (SCIP) added catheter removal on postoperative day 1 or 2 to its process measure set beginning in October 2009. SCIP is 1 of several high‐profile surgical quality improvement programs that employs performance audit and feedback of patient‐level process or outcome measures to address deficiencies in surgical care.11, 12 In addition, audit and feedback of CAUTI rates has been used to successfully reduce CAUTIs in medical‐surgical patients.13 The goal of our study was to audit patient‐level postoperative urinary catheter duration and measure the impact of its feedback to nursing staff on postoperative catheter duration, CAUTI rates, and nurse's attitudes about CAUTI prevention.

Methods

Results

During the study period there were a total of 1657 surgeries on the 2 study units during the baseline and follow‐up periods. After exclusions for urologic or gynecologic surgery (271), no indwelling catheter for surgery (505), or first surgery of 2 or more during the hospitalization (31), there were 846 eligible surgeries (51%).

Table 1 describes the population for the baseline and follow‐up periods for orthopedic and general surgery patients. Within each unit, the surgical populations were comparable during the baseline and follow‐up periods with the exception that the mean length of stay for eligible general surgery patients was significantly shorter in the baseline period as compared to the follow‐up period (6.6 vs. 8.5 days, P = 0.02). Cases on the orthopedic surgery unit were predominantly knee, hip and spine surgeries (85.9%), while those on the general surgery unit were predominantly gut and other gastrointestinal (GI) procedures (80.3%).

BaselineFollow‐Up Comparison

Table 3 describes the measures of urinary catheter use for each surgical population for both data collection periods. On both units, measures of catheter duration were improved following the education and feedback intervention. For the orthopedic unit, mean postoperative catheter duration was reduced from 1.7 to 1.4 days (P = 0.01) and the proportion of patients with catheter removal before day 3 was increased from 86% to 92% (P = 0.04). For the general surgery unit, mean postoperative catheter duration was reduced from 2.6 to 2.2 days (P = 0.01) and the proportion of patients with catheter removal before day 3 was increased from 56% to 63% (P = 0.14). When the general surgery measures were adjusted to account for the difference in length of stay between the 2 time periods, the odds of meeting the performance measure at follow‐up compared to baseline increased from unadjusted odds of 1.38 (P = 0.14) to adjusted odds of 1.69 (P = 0.02).

Table 3.Measures of Indwelling Urinary Catheter Duration and CAUTI Rates

Measure	Orthopedic Surgery			General Surgery
	Baseline, n = 206	Follow‐Up, n = 290	P Value	Baseline, n = 167	Follow‐Up, n = 183	P Value
Abbreviations: CAUTI, catheter‐associated urinary tract infection; ns, not sufficient; SD, standard deviation.
Postoperative catheter duration in days (mean, SD)	1.70 (1.24)	1.44 (0.85)	0.01	2.64 (1.85)	2.19 (1.40)	0.01
Postoperative catheter duration performance measure (%)	86	92	0.04	56	63	0.14
Total catheter days	350	418		441	401
Catheter days/1000 hospital days	423	394	ns*	398	259	s*
Catheter‐associated UTIs	3	0		3	3
Catheter‐associated UTI rate (infections/1000 device days)	8.6	0	ns*	6.8	7.5	ns*

Figure 1a and b are histograms of the frequency of cases having a given postoperative catheter duration in days. The dark bars show the baseline distribution and the light bars show the follow‐up distribution. Although the number of patients in the follow‐up period is greater than for the baseline period for the orthopedic surgery cohort, the images are instructive. In both groups of patients, but most notably in the general surgery population, the reduction in catheter measures resulted from a left shift in the frequency distribution, both for the longer duration outliers (removing the tail of each plot) and for the shorter duration catheters (3 days), increasing the proportion of catheters removed on postoperative day 1.

Figure 1

Discussion

This preobservational and postobservational study found that audit and feedback of patient‐level postoperative urinary catheter duration delivered in the context of an educational intervention and brainstorming session was temporally associated with clinically meaningful reductions in urinary catheter duration. In so doing, we demonstrated the feasibility of collecting patient level urinary catheter duration, and delivering it in a manner that had utility for frontline staff. Our results are consistent with the quality improvement literature which demonstrates that audit and feedback is a successful quality improvement strategy in many contexts and may be as good as more complex interventions at increasing adherence to the performance of process measures for surgical infection prevention.14 Two large national programs, the VA National Surgical Quality Improvement Program (NSQIP) and SCIP, use audit and feedback as the backbone of their large‐scale quality improvement strategies with promising results.11, 12

Given that the 2 study units were so different in practice patterns regarding urinary catheter management and nursing‐identified barriers to evidence‐based care, this work suggests that urinary catheter management may pose unique challenges for different clinical areas and provides a caution that one‐size‐fits‐all interventions for the rationalization of urinary catheter management and reduction of CAUTIs may be of limited effectiveness in the absence of local tailoring. As such, audit and feedback is well‐suited to this purpose as more proscriptive quality improvement strategies may meet with a variety of implementation challenges.

The impact of our intervention on CAUTI rates was not significant. There are several possible explanations for this finding. First, the study was not powered to detect a difference in CAUTI rates given the low infection rate at our institution. However, we cannot exclude the possibility that a reduction in mean catheter duration of one‐third to one‐half of a day is insufficient to impact CAUTI rates in postoperative patients, particularly when many of the follow‐up patients still had postoperative catheter duration 2 daysthe timeframe beyond which bacterial colonization of the catheter begins. While both study units had similar increases in postoperative catheter duration, the UTI rate was only decreased in the orthopedic surgery group which had much higher rates of postoperative catheter duration 2 days at baseline.

In recent years, there has been a renewed focus on the eradication of hospital‐acquired infections prompted by intense interest from the public, federal and state legislators, and others.15, 16 The CMS has recently used the revamping of the Inpatient Prospective Payment System (IPPS) as an opportunity to align financial incentives so that reimbursements for claims with certain hospital‐acquired conditions, including CAUTIs, will be reduced to that of the reimbursement of the same claim without the presence of the complication.17 This move is just one of several strategies to motivate hospitals and clinicians to address the pervasive problem of hospital‐acquired infection.

As urinary catheters are intimately linked to hospital‐acquired UTI, a focus of reduction efforts on catheter use is appropriate. The National Quality Forum (NQF) endorsed a postoperative catheter duration quality measure which was incorporated by the CMS's SCIP in late 2009.18 As a result, every hospital in the country that performs surgery and participates in the Medicare program is now tasked with determining patient‐level urinary catheter duration for selected surgical patients. This move represents a departure from current recommendations from the Centers for Disease Control and Prevention (CDC) and its National Healthcare Safety Network19, 20 which endorse the measurement of a catheter utilization ratio (urinary catheter days/patient days) for patient care units, but does not endorse any patient‐level utilization measures. In this instance, the use of patient‐level data may be better suited to quality improvement interventions such as audit and feedback because of its clinical relevance to frontline providers. However, it may also increase the data collection burden on hospitals. Notably, the measurement of postoperative catheter duration in this study was semiautomated using queries of an EHR. Such an approach can significantly reduce the data collection burden for this process measure and is consistent with national initiatives to integrate EHRs with quality improvement initiatives going forward.21

Our study has several limitations. This study took place in the year following the announcement of a high profile Medicare rule change regarding payment for hospital‐acquired harms. Certainly, the uncontrolled prestudy and poststudy design does not allow for the assessment of the impact of our intervention independent of this context. We are unable, therefore, to attribute the observed reduction solely to the intervention. Additionally, we did not follow postoperative urinary catheter duration beyond the immediate follow‐up period. It is anticipated that the impact of an audit and feedback intervention may diminish over time without a mechanism for repeated feedback. Certainly the sustainability of such repeated feedback in a single institution would be improved with an appropriately configured EHR.

In addition, we have reliable data on catheter reinsertions only from the follow‐up period. While the rates of reinsertions we recorded (0.7% on orthopedic surgery and 2.7% on general surgery) were lower than expected based on the literature,22 we are unable to determine if our intervention led to increases in postoperative urinary retention.

This study was limited to 2 surgical units of a single academic medical center and therefore the urinary catheter utilization patterns may not be representative of other patient populations at other institutions. However, the urinary catheter patterns were comparable to those identified in our prior work in a national sample of Medicare patients undergoing elective surgery.9

Finally, the field of CAUTI prevention has evolved rapidly since this study was performed. In particular, the surveillance definition of CAUTI was altered twice by the CDC in December of 2008 and March of 2009. In addition, the Infectious Diseases Society of America issued a new definition of CAUTI in February of 2010.23 All of these changes highlight the difference between asymptomatic bateriuria (ASB) and symptomatic CAUTI. However, the surveillance definition in use at the time of this study did not make this distinction. Therefore, we are unable to comment on the relative occurrence of ASB versus symptomatic CAUTI under the new definitions.

Rational urinary catheter use is a central component of CAUTI prevention efforts.24 We describe the use of patient‐level urinary catheter use in an audit and feedback intervention to frontline staff that was associated with reductions in urinary catheter duration. To do so, we employed a methodology for tracking urinary catheter use patterns that can provide important data for infection preventionists and frontline providers in efforts to improve urinary output management. This promising approach merits further study as an adjunct to current efforts to rationalize urinary catheter utilization and reduce CAUTIs. In the current environment, having the right data is a powerful aide for ongoing performance improvement.

The ubiquitous urinary catheter is associated with 80% of hospital‐acquired urinary tract infections (UTIs)1estimated to number one million annuallyaccounting for 40% of all nosocomial infections.2, 3 The clinical consequences of these catheter‐associated urinary tract infections (CAUTIs) are substantial and include prolonged hospital stay, bacteremia, and death.4 Despite the known risks of CAUTIs, it is estimated that 25% of hospitalized patients receive urinary catheters and that inappropriate urinary catheter use is widespread.5 Among catheterized patients, catheter duration is the most important modifiable risk factor for CAUTI. The excess risk of any bacteriuria accrues at a rate of 5% per catheter‐day beyond the first 48 hours of catheterization.6, 7 The reduction of catheter‐days for a given patient is an important component of quality improvement efforts to reduce CAUTIs. Unfortunately, as of a 2005 survey, most hospitals do not systematically track urinary catheter insertions and removals.8

The above concerns are highlighted for surgical patients among whom indwelling urinary catheter use is particularly high. In a 2001 sample of Medicare beneficiaries, 85% of major surgical patients had perioperative urinary catheters.9 In this population, postoperative catheter duration exceeded 48 hours in nearly 50%, despite concern that the risks of infection offset the benefits of continued catheterization after 24 hours to 48 hours postoperatively.4, 7, 10 Patients with catheters greater than 2 postoperative days had a 21% increased likelihood of in‐hospital UTI, increased 30‐day mortality, and decreased odds of discharge to home.9

To address the risk of CAUTI associated with excess urinary catheter days, The Centers for Medicare and Medicaid Services' (CMS) Surgical Care Improvement Project (SCIP) added catheter removal on postoperative day 1 or 2 to its process measure set beginning in October 2009. SCIP is 1 of several high‐profile surgical quality improvement programs that employs performance audit and feedback of patient‐level process or outcome measures to address deficiencies in surgical care.11, 12 In addition, audit and feedback of CAUTI rates has been used to successfully reduce CAUTIs in medical‐surgical patients.13 The goal of our study was to audit patient‐level postoperative urinary catheter duration and measure the impact of its feedback to nursing staff on postoperative catheter duration, CAUTI rates, and nurse's attitudes about CAUTI prevention.

Methods

Results

During the study period there were a total of 1657 surgeries on the 2 study units during the baseline and follow‐up periods. After exclusions for urologic or gynecologic surgery (271), no indwelling catheter for surgery (505), or first surgery of 2 or more during the hospitalization (31), there were 846 eligible surgeries (51%).

Table 1 describes the population for the baseline and follow‐up periods for orthopedic and general surgery patients. Within each unit, the surgical populations were comparable during the baseline and follow‐up periods with the exception that the mean length of stay for eligible general surgery patients was significantly shorter in the baseline period as compared to the follow‐up period (6.6 vs. 8.5 days, P = 0.02). Cases on the orthopedic surgery unit were predominantly knee, hip and spine surgeries (85.9%), while those on the general surgery unit were predominantly gut and other gastrointestinal (GI) procedures (80.3%).

BaselineFollow‐Up Comparison

Table 3 describes the measures of urinary catheter use for each surgical population for both data collection periods. On both units, measures of catheter duration were improved following the education and feedback intervention. For the orthopedic unit, mean postoperative catheter duration was reduced from 1.7 to 1.4 days (P = 0.01) and the proportion of patients with catheter removal before day 3 was increased from 86% to 92% (P = 0.04). For the general surgery unit, mean postoperative catheter duration was reduced from 2.6 to 2.2 days (P = 0.01) and the proportion of patients with catheter removal before day 3 was increased from 56% to 63% (P = 0.14). When the general surgery measures were adjusted to account for the difference in length of stay between the 2 time periods, the odds of meeting the performance measure at follow‐up compared to baseline increased from unadjusted odds of 1.38 (P = 0.14) to adjusted odds of 1.69 (P = 0.02).

Table 3.Measures of Indwelling Urinary Catheter Duration and CAUTI Rates

Measure	Orthopedic Surgery			General Surgery
	Baseline, n = 206	Follow‐Up, n = 290	P Value	Baseline, n = 167	Follow‐Up, n = 183	P Value
Abbreviations: CAUTI, catheter‐associated urinary tract infection; ns, not sufficient; SD, standard deviation.
Postoperative catheter duration in days (mean, SD)	1.70 (1.24)	1.44 (0.85)	0.01	2.64 (1.85)	2.19 (1.40)	0.01
Postoperative catheter duration performance measure (%)	86	92	0.04	56	63	0.14
Total catheter days	350	418		441	401
Catheter days/1000 hospital days	423	394	ns*	398	259	s*
Catheter‐associated UTIs	3	0		3	3
Catheter‐associated UTI rate (infections/1000 device days)	8.6	0	ns*	6.8	7.5	ns*

Figure 1a and b are histograms of the frequency of cases having a given postoperative catheter duration in days. The dark bars show the baseline distribution and the light bars show the follow‐up distribution. Although the number of patients in the follow‐up period is greater than for the baseline period for the orthopedic surgery cohort, the images are instructive. In both groups of patients, but most notably in the general surgery population, the reduction in catheter measures resulted from a left shift in the frequency distribution, both for the longer duration outliers (removing the tail of each plot) and for the shorter duration catheters (3 days), increasing the proportion of catheters removed on postoperative day 1.

Figure 1

Discussion

This preobservational and postobservational study found that audit and feedback of patient‐level postoperative urinary catheter duration delivered in the context of an educational intervention and brainstorming session was temporally associated with clinically meaningful reductions in urinary catheter duration. In so doing, we demonstrated the feasibility of collecting patient level urinary catheter duration, and delivering it in a manner that had utility for frontline staff. Our results are consistent with the quality improvement literature which demonstrates that audit and feedback is a successful quality improvement strategy in many contexts and may be as good as more complex interventions at increasing adherence to the performance of process measures for surgical infection prevention.14 Two large national programs, the VA National Surgical Quality Improvement Program (NSQIP) and SCIP, use audit and feedback as the backbone of their large‐scale quality improvement strategies with promising results.11, 12

Given that the 2 study units were so different in practice patterns regarding urinary catheter management and nursing‐identified barriers to evidence‐based care, this work suggests that urinary catheter management may pose unique challenges for different clinical areas and provides a caution that one‐size‐fits‐all interventions for the rationalization of urinary catheter management and reduction of CAUTIs may be of limited effectiveness in the absence of local tailoring. As such, audit and feedback is well‐suited to this purpose as more proscriptive quality improvement strategies may meet with a variety of implementation challenges.

The impact of our intervention on CAUTI rates was not significant. There are several possible explanations for this finding. First, the study was not powered to detect a difference in CAUTI rates given the low infection rate at our institution. However, we cannot exclude the possibility that a reduction in mean catheter duration of one‐third to one‐half of a day is insufficient to impact CAUTI rates in postoperative patients, particularly when many of the follow‐up patients still had postoperative catheter duration 2 daysthe timeframe beyond which bacterial colonization of the catheter begins. While both study units had similar increases in postoperative catheter duration, the UTI rate was only decreased in the orthopedic surgery group which had much higher rates of postoperative catheter duration 2 days at baseline.

In recent years, there has been a renewed focus on the eradication of hospital‐acquired infections prompted by intense interest from the public, federal and state legislators, and others.15, 16 The CMS has recently used the revamping of the Inpatient Prospective Payment System (IPPS) as an opportunity to align financial incentives so that reimbursements for claims with certain hospital‐acquired conditions, including CAUTIs, will be reduced to that of the reimbursement of the same claim without the presence of the complication.17 This move is just one of several strategies to motivate hospitals and clinicians to address the pervasive problem of hospital‐acquired infection.

As urinary catheters are intimately linked to hospital‐acquired UTI, a focus of reduction efforts on catheter use is appropriate. The National Quality Forum (NQF) endorsed a postoperative catheter duration quality measure which was incorporated by the CMS's SCIP in late 2009.18 As a result, every hospital in the country that performs surgery and participates in the Medicare program is now tasked with determining patient‐level urinary catheter duration for selected surgical patients. This move represents a departure from current recommendations from the Centers for Disease Control and Prevention (CDC) and its National Healthcare Safety Network19, 20 which endorse the measurement of a catheter utilization ratio (urinary catheter days/patient days) for patient care units, but does not endorse any patient‐level utilization measures. In this instance, the use of patient‐level data may be better suited to quality improvement interventions such as audit and feedback because of its clinical relevance to frontline providers. However, it may also increase the data collection burden on hospitals. Notably, the measurement of postoperative catheter duration in this study was semiautomated using queries of an EHR. Such an approach can significantly reduce the data collection burden for this process measure and is consistent with national initiatives to integrate EHRs with quality improvement initiatives going forward.21

Our study has several limitations. This study took place in the year following the announcement of a high profile Medicare rule change regarding payment for hospital‐acquired harms. Certainly, the uncontrolled prestudy and poststudy design does not allow for the assessment of the impact of our intervention independent of this context. We are unable, therefore, to attribute the observed reduction solely to the intervention. Additionally, we did not follow postoperative urinary catheter duration beyond the immediate follow‐up period. It is anticipated that the impact of an audit and feedback intervention may diminish over time without a mechanism for repeated feedback. Certainly the sustainability of such repeated feedback in a single institution would be improved with an appropriately configured EHR.

In addition, we have reliable data on catheter reinsertions only from the follow‐up period. While the rates of reinsertions we recorded (0.7% on orthopedic surgery and 2.7% on general surgery) were lower than expected based on the literature,22 we are unable to determine if our intervention led to increases in postoperative urinary retention.

This study was limited to 2 surgical units of a single academic medical center and therefore the urinary catheter utilization patterns may not be representative of other patient populations at other institutions. However, the urinary catheter patterns were comparable to those identified in our prior work in a national sample of Medicare patients undergoing elective surgery.9

Finally, the field of CAUTI prevention has evolved rapidly since this study was performed. In particular, the surveillance definition of CAUTI was altered twice by the CDC in December of 2008 and March of 2009. In addition, the Infectious Diseases Society of America issued a new definition of CAUTI in February of 2010.23 All of these changes highlight the difference between asymptomatic bateriuria (ASB) and symptomatic CAUTI. However, the surveillance definition in use at the time of this study did not make this distinction. Therefore, we are unable to comment on the relative occurrence of ASB versus symptomatic CAUTI under the new definitions.

Rational urinary catheter use is a central component of CAUTI prevention efforts.24 We describe the use of patient‐level urinary catheter use in an audit and feedback intervention to frontline staff that was associated with reductions in urinary catheter duration. To do so, we employed a methodology for tracking urinary catheter use patterns that can provide important data for infection preventionists and frontline providers in efforts to improve urinary output management. This promising approach merits further study as an adjunct to current efforts to rationalize urinary catheter utilization and reduce CAUTIs. In the current environment, having the right data is a powerful aide for ongoing performance improvement.