Healthcare-associated infections (HAIs) are the most frequent adverse event in healthcare delivery worldwide, affecting up to 10% of patients on general medical-surgical wards and up to 30% of patients in intensive care units.¹ Annually, HAIs cost billions of dollars and cause an estimated 1.7 million infections and 99,000 associated deaths in the United States. Furthermore, 25% of all hospitalized patients have indwelling catheters,^{Reference Advani and Fakih2-Reference Gokula, Hickner and Smith4} and device-related HAIs, such as catheter-associated urinary tract infections (CAUTIs), are more likely to be caused by antibiotic-resistant bacteria.^{Reference Weiner-Lastinger, Abner and Edwards5}

Reduction and prevention of HAIs, especially CAUTIs, which can account for up to one-third of all HAIs, are leading priorities for hospitals, regulators, and payors.^{Reference Advani and Fakih2,Reference Wright, Kharasch, Beaumont, Peterson and Robicsek6,Reference Fakih, Gould and Trautner7} However, the US Center for Disease Control and Prevention’s National Health Safety Network (CDC-NHSN) surveillance definition of CAUTI has limited clinical correlation; it does not fully reflect noninfectious harm related to the catheter.^{Reference Advani and Fakih2,Reference Fakih, Gould and Trautner7,8} Therefore, experts continue to debate and refine HAI surveillance metrics to best capture both catheter-related infectious harm and catheter-related noninfectious (CRNH) harm such as urethral trauma, immobility, pain, bleeding, and leakage. In fact, indwelling urinary catheters (IUCs) usually cause more noninfectious than infectious harm.^{Reference Advani and Fakih2,Reference Hollingsworth, Rogers and Krein3,Reference Fakih, Gould and Trautner7}

The commonly used standardized infection ratio (SIR) does not fully capture CRNH or the impact of prevention efforts in all settings.^{Reference Advani and Fakih2,Reference Hollingsworth, Rogers and Krein3,Reference Fakih, Gould and Trautner7} Alternatively, device utilization rates and ratios (DURs) do not reflect differences in other factors that may describe levels of device use.⁸ Furthermore, DURs lose comparability over time and across settings, and they can mask truly effective interventions by selecting for a higher-risk group of catheterized patients.⁸ Experts, including the CDC-NHSN, now support the use of the standardized utilization ratio (SUR).⁸ However, a literature search revealed only 1 small, short-duration, single-center study that used the SUR as a metric.^{Reference Mueller9} Similarly, experts have criticized the quality of descriptions of interventions as “remarkably poor.”^{Reference Hoffmann, Glasziou and Boutron10} To address this, they encourage using the Template for Intervention Description and Replication (TIDieR) guidelines.^{Reference Hoffmann, Glasziou and Boutron10} However, reports using the TIDieR format for CAUTI interventions are rare.

We implemented a multidimensional intervention capable of reducing IUC use and CAUTI across a 5-hospital healthcare system, totaling 1,056 acute-care beds. The intervention and outcomes are presented using the TIDieR guideline and the most recently recommended quality measure, namely the SUR.

Methods

A before-and-after study was conducted at a 5-hospital healthcare system, totaling 1,692 acute-care beds from September 1, 2017, to October 1, 2019. Participants included all adult inpatients. The intervention occurred from August 15, 2018, to September 14, 2018, and consisted of 3 major features. The first feature was an interactive educational campaign comprising one-on-one engagements between infection preventionists and frontline nurses and providers. This campaign was combined with an Olympic-style competition among units rewarding overall participation, device utilization, clandestinely monitored hand hygiene, and CAUTI rates. Both were collectively referred to as the ‘DeCATHlon’ (Supplemental Table 1 online). The second feature was making sure device alternatives, in the form of female external urinary collection devices and male custom-fitted condom catheters, were available to all units, and practitioners were trained in the indications and capabilities of the noninvasive devices. The third feature consisted of increasing the urinalysis reflex to culture threshold from >5 to ≥10 WBCs, and electronic prompts or computerized decision support (CDS) for ordering urine cultures for patients with intrauterine contraceptives (IUCs), and for encouraging device alternatives and catheter removal (Fig. 1). In addition to the 3 main components, weekly unit-specific device utilization reports were sent to each ward or patient care location. Monthly, quality department representatives discussed unit-level DURs with managers, who then reviewed patient-level device use at daily informal meetings with physicians and advanced practice providers. Executives, including the chief medical officer, also provided feedback to units and individual providers, especially those considered to be performance outliers. Supplemental Table 1 (online) details the intervention based on the TIDIER checklist.

Fig. 1. Electronic prompts or clinical decision support for ordering urine cultures on patients with indwelling urinary catheter, and for encouraging device alternatives and catheter removal.

The significance of differences between pre intervention and postintervention metrics was determined using the R statistical package. We performed 2 tests: the Student t test, which is optimized for normally distributed data, and the Kolmogorov-Smirnov test, which makes no assumptions about data distribution.

Results

For the 2-year study period, the combined average daily census for all hospitals was 87% of capacity, the combined average length of stay was 6.8 days, the average infection prevention staff (IPS) to inpatient bed ratio was 1 IPS per 175 beds, and the average hand hygiene compliance rate was 74%. We noted no significant changes in length of stay, occupancy, staffing levels or hand hygiene compliance during the study period. In the postintervention period, a left ventricular-assist device (LVAD) service was launched. One year after the intervention, CAUTI rates, the SIR, device days, the SUR, and antibiotic prescriptions initiated for inpatient urinary tract infections (UTIs) significantly decreased. Orders for device alternatives and device-alternative days increased. Orders for IUCs significantly decreased (Supplemental Fig. 1 online), and orders using decision-support prompts for IUCs and urine cultures also significantly increased (Table 1). All major quality measures significantly improved in the postintervention period, and the results achieved significance, whether the parametric t test for normally distributed data or the nonparametric Kolmogorov-Smirnov test was used. Similarly, infection-related complications were reduced, regardless of denominator used, infection per device day, or infection per patient day (Table 1).^{Reference Wright, Kharasch, Beaumont, Peterson and Robicsek6} These improvements were comparable across all facilities and were proportional to the facility size.

Table 1. Preintervention and Postintervention Results

Note. SD, standard deviation; SIR, standardized infection ratio; CAUTI, catheter-associated urinary tract infection; CDS, computerized decision support; UTI, urinary tract infection.

^a September 1, 2017, to September 14, 2018.

^b September 15, 2018, to September 30, 2019.