Opportunities for improvement can be identified in numerous ways: bedside staff input, audits of medical records, occurrence of a serious adverse event, deviations from usual performance, and comparison to peers. Consistent measurement of quality indicators is an efficient way to recognize gaps between actual and ideal performance. Data is the driving force behind problem identification.

Visual tools that display repeated measures of an outcome over time (e.g., control charts) facilitate recognition of problematic areas (Langley et al., 2009). NDNQI graphs allow nurses to track their quality indicators over time, with the added benefit of comparison to peers. NDNQI comparison data features
percentiles, which are calculated from a rank ordering of units. For example, NDNQI analysts list the catheter-associated urinary tract infection (CAUTI) rates on adult critical care units in non-teaching hospitals from lowest to highest. The 75th percentile is the level at which 75% of these critical care units have lower CAUTI rates. In Figure 2, the 75th percentile is represented by the top of the gray bar. The critical care unit in this example often has CAUTI rates above the 75th percentile, which indicates a definite need for improvement.

NDNQI percentiles provide normative standards for judging quality. The nurse manager of a unit with injury falls rates consistently above the 50th percentile (the median) knows that over half of the unit’s peers have fewer injury falls. Until 2008, NDNQI reports provided the mean instead of the median (50th per centile) for most indicators. Many hospitals used the mean (average) to identify problems. However, the mean is inflated when a few units have unusually high rates of adverse events. In such situations, the median is a more accurate reflection of “average” performance (Polit & Hungler, 1999). Therefore, the median is preferred over the mean when evaluating patient outcomes.

The NDNQI median represents “middle-of-the-road” quality. The 10th, 25th, 75th, and 90th percentiles in NDNQI reports provide the distribution of performance and can help identify the degree of a problem. A unit with hospital-acquired pressure ulcers (HAPUs) above the 75th percentile should recognize that 75% of their peers have better HAPU rates. Similarly, a pediatric unit whose average number of pain assessments is at the 10th percentile would understand that 90% of peer units provide more frequent pain assessments.
Such reports should definitely spur these units to pursue the remaining QI steps described in this chapter.

Once a problem is identified, a “process owner”, someone who will be responsible for the entire improvement effort, should be selected. While support from hospital executives is necessary, the owner may be a manager or clinician who can lead the improvement (Berwick, Godfrey, & Roessner, 1990). The owner works with an interprofessional team to further define the problem, implement improvements, and measure progress. The team can be newly formed or an existing group (e.g., quality council) and can be unit-based or hospital-wide depending on the scope of the problem. Representatives from multiple professions (nursing, medicine, pharmacy, physical therapy, hospital resource management, etc.) and from multiple levels of the organization (unit managers, bedside nurses, executives) bring unique views and solutions to the problem. Hospitals often form and empower interprofessional outcome-specific teams (such as a pressure ulcer prevention team) to continually design, test, and implement changes.

Before making any changes, the team should clearly define their purpose and examine possible causes of the problem. Setting goals and “drilling down” to the root of the problem are the next steps.

Goals keep a team focused and provide a ruler for measuring progress. Useful goals are quantifiable and specify the indicator, units, and timeframe. For example:

NDNQI reports provide ready-made goals. Percentiles tell you what qualifies as average, above average, and excellent quality (see Table 1). An acceptable mid-way goal is to reduce negative outcome rates to below the median. At the median (the middle), 50% of similar units have higher rates and 50% have lower rates. If a unit’s total falls per 1,000 patient days had been above the 75th percentile, achieving fall rates below the median would represent a considerable improvement. However, the median does not necessarily represent ideal care. NDNQI percentiles are based on actual practice, with some hospitals attaining the desired goals of zero—e.g., for fall rates—not because the value is above average and is above the 50th percentile. Being at the median shows your quality of nursing care is “typical” or “average” compared to your peers.

For negative outcomes, above average quality is seen in the 25th percentile. A unit with catheter-associated urinary tract infections at or below the 25th percentile is in the top 25% of peers. Excellent care, or best practice, is reflected in the 10th percentile. A unit with total falls at or below the 10th percentile is in the top 10% of peers. For positive structure or process indicators such as RN hours per patient day, percent of
RNs with certification, or pediatric pain assessment, the 90th percentile shows best practice.

To design effective improvements, an examination of the problem and its causes is needed. The team should invest some time drilling down to the problem’s core. In some cases, a problem’s causes and solutions may be quite obvious. In others, a little digging may reveal an unexpected process at work.

Information from multiple sources paints a clearer picture of the problem. The team can examine medical records, review patient satisfaction surveys, and interview or observe direct-care providers.

Health care has borrowed numerous analysis tools from the manufacturing industry (Graban, 2009; Ursprung & Gray, 2010). Tools such as root cause analysis and
process mapping can reveal the sources of poor outcomes. For example, in root cause analysis, teams diagram the details of a specific case, then repeatedly ask, “Why did this happen?” until the underlying causes are exposed (Ursprung & Gray, 2010). Several authors have provided excellent guides for conducting effective root cause analyses (Pham et al., 2010; Ursprung & Gray, 2010).

NDNQI reports also help teams drill down. By examining NDNQI structure and process indicators, teams can identify factors contributing to undesirable patient outcomes. If a unit’s total hours of nursing care per patient day is below the 25th percentile, or their use of agency personnel is higher than 90% of their peers, improved staffing may be part of the solution.

For outcomes such as falls and pressure ulcers, nursing care process indicators may reveal more about the problem. Figure 3 illustrates how data from NDNQI reports can help teams drill down on risk assessment and prevention interventions. On this unit, 16% of patients were not assessed for fall risk. Although 95% of the patients at risk had a prevention protocol in place, the unit’s high fall rate suggests the protocol may need to be improved. The team should explore: