A Pediatric-Specific View on Improved Nurse-Sensitive Indicators Related to Increased BSN or Higher Rates Over Time

Abstract

Background: 

Nurses obtaining higher levels of education has been recommended nationally for more than a decade to support improved patient outcomes.

Local Problem: 

Organizational strategies were implemented to achieve a highly educated workforce at the project site over 14 years. However, there was no evaluation of relationship with increasing education levels and pediatric patient outcomes.

Methods: 

A quality improvement project explored the relationship of Bachelor of Science in Nursing (BSN) or higher prepared nurses with 9 different nurse-sensitive indicators (NSIs). Educational preparation and NSI data were retrieved retrospectively from hospital databases from up to 14 years.

Results: 

Moderate to moderately strong correlations were noted between the proportion of BSN or higher prepared nurses and 2 NSIs (central line–associated bloodstream infections, r = −0.55 and surgical site infections, r = −0.71). Four of the 9 indicators were approaching moderate correlation.

Conclusions: 

Increasing the proportion of highly educated nurses at the project site correlated with improvements in select NSI outcomes.

The Institute of Medicine's (IOM's) Future of Nursing report 2010 challenged nurses to pursue higher levels of education.¹ This recommendation resulted in increasing rates of nurses in the workforce with a bachelor of science in nursing (BSN) or higher degree. While the correlation of BSN or higher degrees with patient outcomes has been reported, limited literature has addressed nursing education at the bachelor's level related to pediatric patient care outcomes. An urban, academic pediatric specialty hospital studied the correlation between BSN or higher educational preparation and quality of nursing care as measured by nurse-sensitive indicators in the pediatric population over time.

Overall, 57% of nurses in the United States hold a BSN or higher degree.² In 2019, the National Council of State Boards of Nursing reported 49.2% of US-educated, first-time NCLEX testers as BSN prepared compared with 39.3% in 2010.³ Nurses in the workforce are entering practice with higher levels of education or returning to school to meet the expectations of employers. In 2020, of nurses graduating with a BSN degree, 46.0% were registered nurses (RNs) returning to school for a BSN degree.⁴ In addition, there is a nationwide trend of increasing BSN percentages from 2010 to 2018 across all states except for Alaska, Nevada, Rhode Island, and South Dakota.⁵

The urban, academic pediatric specialty hospital accepted the IOM's challenge more than a decade ago to make a committed effort to increase the educational level of nurses within the organization. Since 2009, the hospital successfully increased the year-over-year rate of BSN or higher prepared nurses on an annual basis through vision and goal setting, recruitment strategies, financial support, and leadership role modeling/accountability strategies.⁶ BSN or higher degree rates increased from 60% in 2006 to 85.9% in 2019 at the project facility during project timeframe.

Although the project facility had experienced improved BSN or higher degree rates over 14 years, the specific impact on patient outcomes in the pediatric population was not evaluated. The benefits of nurses having BSN or higher educational preparation have demonstrated improved patient outcomes in the adult setting including decreases in mortality and heart failure mortality, higher rate of survival to discharge after cardiac arrest, lower length of stay, readmission rates, pressure injury, failure to rescue, postoperative deep vein thrombosis, and pulmonary embolism.^7–11 However, there is limited literature on the relationship of nursing education to pediatric patient outcomes. One study was identified examining improved patient outcomes, which demonstrated fewer complications after pediatric cardiac surgery with higher levels of nursing education and experience.¹²

Because of the lack of literature describing the relationship of nursing education to pediatric patient outcomes, the project team developed a quality improvement (QI) project to measure the relationship of patient outcomes to increases in BSN rates over 14 years. For the purposes of this project, the team explored the relationship between BSN or higher education among pediatric nurses at this site and selected nurse-sensitive indicators from fiscal year (FY) 2006 through FY2019. The aim was to use these findings to support future strategic initiatives to promote the continuing advanced education among nurses and to better support the impact of nursing education on patient outcomes.

METHODS

Approach

A QI project, consisting of a retrospective analysis of patient outcome and nursing education preparation data over time, was conducted at a 313-bed, tertiary, freestanding, urban pediatric hospital located in the Mideast. The hospital is affiliated with a large university and is a teaching facility with more than 30 school of nursing partnerships. The project was approved by the internal Quality Review Committee as QI.

Data collection

Nurse-sensitive indicators for this QI project were identified and included medication administration errors, patient falls, peripheral intravenous infiltrations and extravasations (PIVIEs), and pressure injuries, which are reported and tracked through a voluntary, electronic patient safety occurrence reporting system (see Supplemental Digital Content, Table 1 available at: https://links.lww.com/JNCQ/A964). The events are internally coded for severity based on the statewide Patient Safety Reporting System Harm Score Taxonomy Scale.¹³

Medication errors, patient falls, and PIVIEs were defined internally by the organization's quality department, in part, due to the requirements of the state reporting definitions and harm taxonomy, which includes no harm events that reach the patient. The harm taxonomy score is assigned by internal patient safety staff trained on the state reporting requirements. The organization also made the decision to continue to use internal definitions for these indicators over the years as pediatric national benchmarks were not clearly defined at the time the data capture began. By including no harm and harm events that reach the patient in the internal definitions, it allowed the facility to easily track and trend events and the QI efforts over the years.

Medication errors were internally defined as any error that reached the patient during the administration process. High alert medication errors are a subset of the administration errors defined as medications that have a heightened risk of causing significant patient harm and require a separate double-check prior to administration. The medications included as high alerts are heparin (excluding prefilled flushes), parenteral nutrition and intravenous fat emulsion, chemotherapy, insulin (intravenous route only), and continuous narcotic infusions. Patient falls and PIVIEs were internally defined as any event that reached the patient and required additional monitoring or intervention.

Pressure injuries are staged by certified wound, ostomy, and continence nursing staff using the National Pressure Injury Advisory Panel definitions.¹⁴ Hospital-acquired injuries were included if identified as a mucosal, stage 1 through stage 4, unstageable, or a deep-tissue pressure injury. Hospital-acquired infection data, central line–associated bloodstream infection (CLABSI), catheter-associated urinary tract infection (CAUTI), and surgical site infection (SSI) were confirmed using active surveillance based on the National Healthcare Safety Network (NHSN) definitions by internal infection control staff.¹⁵ The non–risk-adjusted mortality percentage is internally tracked by quality department staff and identified as any inpatient (excludes patients expiring in the emergency department) who expires within the facility as the numerator over the number of inpatient discharges as the denominator. Nurse-sensitive indicator data were retrieved from existing hospital databases and were reported quarterly per FY.

The project team collected BSN or higher degree rates from all RNs working in the hospital who reported through patient care services under the chief nursing officer. The rates included bedside nurses of inpatient units (both acute care and critical care), emergency department, perioperative areas, leadership, ancillary support services including dialysis and infusion center, cardiac services, clinical informatics, care coordination, vascular access, training and education, transport, and radiology. BSN or higher percentages were retrieved from existing hospital databases through human resource files and internal databases used for Press Ganey National Database of Nursing Quality Indicators (NDNQI) reporting. BSN or higher percentages were annualized from FY2006 through FY2014 due to the lack of quarterly reporting. Actualized BSN rates were reported quarterly from FY2015 through FY2019. Data were collected in aggregate by overall hospital rates only. No identifiers for any nursing staff were included.

Data analysis

Although this was a QI project, the project team developed a general hypothesis $r_{x{y}_i}$ = 0, where x is the proportion of staff with a BSN or higher degree and y_i the variable metrics of interest (nurse-sensitive indicators). The general alternate hypothesis is as follows: $r_{x{y}_i}$ is not equal to 0, where x is the proportion of staff with a BSN or higher degree and y_i the variable metrics of interest. The outcome data (error rates) were calculated on a per 1000 basis to ensure comparison because the researchers were interested in comparisons over time in relation to the percentage of BSN or higher educated nurses. Correlations were based on quarterly rates and calculated to examine the relationship between BSN or higher percentage and error rates across the entire organization.

RESULTS

Fourteen years of data were collected retrospectively starting in FY2006 and ending in FY2019. Over this time, the percentage of BSN or higher degree rates increased from 60% in FY2006 to 85.91% at the end of FY2019 (range, 60%-86.72%). Nurse-sensitive indicators for CLABSI, medication errors, falls, and pressure injuries were reported quarterly for FY2006 through FY2019. Other indicators and patient outcomes had variable years of data due to limitations with initiating of reporting (see Supplemental Digital Content, Table 2, available at: https://links.lww.com/JNCQ/A965).

Correlations between the proportion of BSN or higher prepared nurses and patient outcomes were analyzed across the metrics. CLABSIs and SSIs had correlations in the data set defined as moderate to moderately strong. The project team analyzed the largest data set, which included 56 quarters of data. At that sample size, any correlation over 0.25 would be below the traditional P value at .05 cutoff. For smaller data sets of 4 years or 16 quarters, any correlation over 0.48 would be less than the traditional P value of .05 cutoff. As BSN rates increased, CLABSIs decreased (r = −0.55) and SSIs decreased (r = −0.71) (Figure 1). Additional outcomes approaching moderate correlations were mortality (r = −0.46) (Figure 2), PIVIEs (r = −0.47), falls (r = −0.34), and CAUTIs (r = 0.36). There were weak correlations observed with medication administration errors (r = 0.05), medication high alert errors (r = 0.14), and pressure injuries (r = −0.10).

Figure 1.: Correlation between BSN or higher education,^a CLABSI rates,^b and SSI rates.^{c a}BSN or higher education data retrieved from internal databases at the project facility. ^bCLABSI rates retrieved from an internal database; infections identified using active surveillance based on the National Healthcare Safety Network definitions.^{15 c}SSI rates retrieved from an internal database; infections identified using active surveillance based on the National Healthcare Safety Network definitions.¹⁵ BSN indicates bachelor of science in nursing; CLABSI, central line–associated bloodstream infection; SSI, surgical site infection; FY, fiscal year.

Figure 2.: BSN or higher education proportion^a and mortality^b. ^aBSN or higher education data retrieved from internal databases at the project facility. ^bMortality data retrieved from internal dataset at the project facility. BSN indicates Bachelor of Science in Nursing: FY, fiscal year.

DISCUSSION

In 2010, the IOM released its recommendation to have 80% of nurses have a BSN or higher degree by the year 2020.¹ There is a growing body of research supporting the findings by Aiken et al,^16,17 who demonstrate a correlation between level of education of staff with improved patient mortality and failure to rescue rates. One study showed that “each 10% increase in the proportion of nurses with higher degrees decreased the risk of mortality and of failure to rescue by a factor of 0.95, or by 5%, after controlling for patient and hospital characteristics.”^16(p1620) National organizations support having a highly educated workforce through a position statement issued in 2010 by the Tri-Council for Nursing.¹⁸ However, there is limited research reviewing the relationship of nurse education levels and patient care outcomes in pediatric populations.¹²

This QI project consisted of a retrospective review of internal quality data that was suggestive of correlations between selected nurse-sensitive indicators and nurse education levels over time in a pediatric facility. The results of the project displayed a moderate to moderately strong relationship between the increase in the percentage of BSN or higher prepared staff to the decrease in the CLABSI and SSI rates. The indicators of mortality, PIVIEs, falls, and CAUTIs showed weaker correlations but were approaching moderate levels.

Overall, these results suggest that as we increased the education level of staff to BSN or higher degree, we showed positive impact to specific patient care outcome measures similar to the results published in adult care settings.^6–10 Of note, CLABSI rates appear to have remained relatively flat but this does not illustrate the changes that have occurred in our patient complexity, with a higher volume of critical care patients. In addition to hospital infection rates, the standardized infection ratio (SIR) is utilized to compare our infection measures to account for patient risk factors. The SSI data reported in this project cover less than 14 years due to the data not being captured in our internal databases in the same manner as other quality measures until FY2016. The project team included the SSI data in the analysis as 16 quarters of data. The intent of the team was not to argue causality but to identify a potential relationship to nursing education and therefore these data were included in the analysis.

Weak correlations were found for overall medication and high alert medication administration error rates. The absence of a correlation with education was not surprising, as medication administration errors have been an area of focus for the institution with implemented strategies for improvement. Through online physician order entry and positive patient identification using barcode-assisted medication administration technology, the facility has been able to maintain low error rates over time. There has been variation in the high alert medication error rates due to changes to the medications included in the definition, an updated administration process for staff, and denominator calculation modifications, all of which may have contributed to the weaker correlation.

The pressure injury indicator also demonstrated weak correlation to education level. Contributing factors include definition updates, changing from monthly surveillance data to a rate of all identified injuries, as well as internal identification including adopting medical record charting triggers and reporting of injuries. At the end of FY2017, 2 new wound care nurses were hired and created robust processes to identify injuries including surveillance rounds on high-risk patients, which has become a focus in pediatric care.

LIMITATIONS

Limitations to the project include other factors that could have influenced improvement in patient outcomes including improved nurse staffing to patient ratios in most units over time, increased nursing specialty certification rates, and improvements to the hospital working environment for staff satisfaction and retention, which were not controlled for in this QI project. Over the 14-year period, there were changes in the quality and infection prevention staff who reviewed and determined whether hospital-acquired conditions or infections occurred in addition to changes to definitions of some of the indicators. In addition, the use of a voluntary reporting system to capture patient adverse events may have not inclusively captured all events.

The mortality percentages consisted of all-cause, inpatient mortality and not risk-adjusted based on patient comorbidities or acuity. Throughout the time frame reviewed, the facility had multidisciplinary performance improvement teams to monitor and create action plans to improve nurse-sensitive indicator outcomes to meet internal or national benchmarks. The teams have focused on evidence-based best practice initiatives and implementation of patient care bundles, which are audited monthly across the hospital. The organization actively participates in multiple patient safety organizations and collaboratives including pediatric-specific networks focusing on decreasing patient harm events. Magnet designation for the institution was earned in 2012 and has since been maintained. Thus, the findings in this QI project are suggestive of a relationship between education levels and selected patient outcomes in the pediatric population at this study site but not casual in nature.

CONCLUSION

Pediatric health care organizations strive to deliver high-quality care to their children and families. In addition, Magnet organizations must monitor and demonstrate outperformance in nurse-sensitive clinical indicators. Through increased proportions of BSN or higher prepared nurses at the project site, there were improvements in select nurse-sensitive indicator outcomes over time. Because of the limited published pediatric literature exploring this relationship, these findings may provide some insight into suggestive relationships between education and pediatric patient outcomes. Future research and QI studies are needed to fully examine the relationship of BSN or higher education on pediatric patient outcomes.