Zrelak, Patricia A. PhD, RN, NEA-BC, CNRN; Utter, Garth H. MD, MSc; Sadeghi, Banafsheh PhD, MD; Cuny, Joanne MBA, RN; Baron, Ruth BScN, RN; Romano, Patrick S. MD, MPH
THE AGENCY for Healthcare Research and Quality (AHRQ) patient safety indicators (PSI) are a set of hospitalization quality indicators that take advantage of readily available administrative data.1 The PSIs identify 18 potentially preventable adverse events and complications that patients may experience through contact with the health care system. Events identified by the PSIs collectively impart a picture of patient safety within a hospital and may represent occurrences amenable to system-level change. Medical records identified as having a PSI of interest may then be reviewed for opportunities for improvement, such as issues related to nursing care.
The aim of this article is to evaluate opportunities for improving nursing care related to the AHRQ PSIs identified through detailed review of a national sample of medical records of patients identified as having at least 1 event of interest, as part of the AHRQ PSI validation pilot project and a related collaboration with the University HealthSystem Consortium (UHC). Information about the accuracy (criterion validity) of the PSIs, based on the same review process, has been published elsewhere.2–5
This study was approved by the federal Office of Management and Budget and by the institutional review board at the University of California, Davis. University HealthSystem Consortium provided permission to use data submitted by member organizations. Data collection procedures complied with the Health Insurance Portability and Accountability Act Privacy Rule provisions for disclosure of protected health information without subject authorization to a public health authority [45 CFR 164.512(b)].
This was a retrospective, multifacility, cross-sectional chart abstraction study.
Two groups of hospitals participated. The first group consisted of 47 nonfederal hospitals, representing 29 states and various hospital types, that responded to a call for volunteers through the AHRQ Quality Indicator technical support listserve. These hospitals reviewed records that met criteria for PSI 13 (postoperative sepsis) or PSI 7 (selected infections due to medical care; now known as central line–related bloodstream infection or CLRBSI).4,5 The second group included 34 academic medical centers affiliated with UHC, identified through the UHC listserve. Their staff reviewed medical records meeting criteria for PSI 12 (postoperative deep vein thrombosis or pulmonary embolus or DVT/PE); PSI 11 (postoperative respiratory failure); and/or PSI 3 (decubitus ulcer; now known as pressure ulcer).2,3
Evidence-based data abstraction instruments and corresponding guidelines, available at http://qualityindicators.AHRQ.gov/validationpilot.aspx, were developed by study investigators based on a literature review and input from national experts. Group 1 instruments were pretested at hospitals in Sacramento, California. Final abstraction tools included questions about demographic characteristics, verification of the reported events, patient-level risk factors, preventive measures, evaluation and treatment, and patient outcomes. Instruments used by group 2 were developed similarly, except a national steering committee was formed for each PSI, and pretesting occurred within UHC.
Group 1 hospitals applied a modified version of the AHRQ PSI Windows software version 3.1 (March 12, 2007), provided by the authors, to extract concurrent probability samples of records that met criteria for each PSIs under study between October 1, 2005, to March 31, 2007. Sampling targets included 240 cases nationally and 30 cases per hospital for each PSI. The overall number of abstracted records often fell short of the target because some hospitals had no qualifying or retrievable cases. For group 2, the UHC coordinating center, using the public use version of the same AHRQ software, prepared the sample list for each study hospital with targets ranging between 30 and 60 cases per PSI per facility.
For both groups, if after answering an initial set of questions, the patient did not appear to qualify for the targeted PSI event, the abstractor was allowed to exit the abstraction process. Criteria for each PSI are defined in the AHRQ PSI Technical Specifications (currently available at http://www.qualityindicators.AHRQ.gov/Modules/PSI_TechSpec.aspx).1 Abstractors for each group were trained via Web-based teleconferences. Group 1 abstractors used a paper tool, which was submitted to study staff for data entry. Group 2 entered their data into an online database. Ongoing support was provided by e-mail, listserve, and telephone.
SAS, version 9.1 (Cary, North Carolina), was used to perform descriptive analyses. Proportions were calculated for categorical data and averages for continuous data. Only records found to have the PSI of event of interest were included in the results.
Participating hospitals reviewed a total of 2293 PSI events. Cases per indicator are as follows: CLABSI (n = 130), sepsis (n = 67), respiratory failure (n = 548), DVT/PE (n = 314), and hospital-acquired pressure ulcers (HAPU) (n = 1234).2–5
Central line–related blood stream infection
At the time of the study, this indicator included all vascular device types, although most confirmed events were related to central venous catheters (CVC) (n = 72), followed by peripheral intravenous (n = 13) and arterial lines (n = 3). Peripherally inserted central catheters represented 30% of the infected nontunneled CVCs. For short-term nontunneled CVCs, the internal jugular site was responsible for 34% of the events followed by the subclavian (25%) and femoral (11%) sites. Femoral catheters had the shortest mean time to infection (5.7 ± 3.4 days) compared to internal jugular (10.0 ± 6.1 days) and subclavian lines (11.7 ± 8.4). Abstractors were frequently unable to find documentation of line type (peripheral vs central) (n = 24), CVC insertion site (n = 12), and CVC insertion or removal dates (n = 12). Total parenteral nutrition was used in 29 (36%) cases. Abstractors found no documentation regarding the presence or absence of symptoms in 73 cases.
Postoperative deep vein thrombosis/pulmonary embolus
This PSI targets surgical cases that develop a postoperative DVT or PE. Only 42% of the surgical cases received American College of Chest Physicians guideline–directed venous thromboembolism prophylaxis before or on the day of surgery.6 The elapsed time from documentation of clinical signs of DVT and/or PE (eg, leg swelling, tenderness, or a palpable venous cord for DVT; new dyspnea, chest pain, or oxygen desaturation for PE) to the diagnosis or diagnostic evaluation was greater than 1 day in 21% of DVT and 17% of PE cases. The initiation of treatment was delayed, defined as greater than 1 day from diagnosis, in 20% of DVT and 17% of PE cases.
This PSI targets previously uninfected patients who develop sepsis after an elective surgical procedure. In a large number of cases (n = 28; 17%), the abstractor was unable to identify bacteremia, septicemia, sepsis, or systemic inflammatory response syndrome from the medical record and in some cases, may have incorrectly assumed that sepsis was absent based on “negative blood cultures” alone (according to their abstraction notes).
When hair removal is necessary, clipping and depilatory creams have been found to result in fewer surgical site infections when compared to razors.7,8 Among the 67 confirmed cases, only 12 cases had documented hair removal, and razors were used in 3 cases (25%).
Preoperative intravenous (IV) antimicrobial prophylaxis should be given within 1 hour of incision to maximize tissue concentration (2 hours are allowed for vancomycin and fluoroquinolones).9,10 In all but 3 cases in which antibiotics were ordered, administration of the antibiotic occurred within 1 hour of the operative start time. In each of these cases, antibiotics were administered within 15-minutes of the 1-hour standard. Pneumonia (54%), CLABSI (46%), and urinary tract infections (12%) were the most common infections (not mutually exclusive) associated with the septic episode. Optimal antimicrobial prophylaxis (defined as antimicrobials administered within 1 hour of surgery and stopped within 24 hours of incision time) occurred in 77% in those with urinary tract infections, 83% of those with surgical site infections, 88% of those with CLRBSI, and 100% of those with pneumonia.11,12 Perioperative hypothermia (temperature < 96.8°F) has been associated with increased risk of infection and sepsis.8,10 Among confirmed cases of sepsis, 7 of 31 and 11 of 32 had a temperature lower than 96.8°F in the operating room and postanesthesia care unit, respectively. In 11.8% of cases, the abstractor was unable to find a recorded intraoperative temperature, and in 17.6% of cases, the abstractor reported use of a different temperature scale between perioperative care settings.
Postoperative respiratory failure
This PSI is intended to capture cases of respiratory failure following an elective operation. General anesthesia was associated with 787 of the 834 procedures (94.4%) performed in the 548 records with at least 1 episode of postoperative respiratory failure. More than three-fourths (80.2%) of study patients developed respiratory failure within the first week after surgery, with 22.5% reintubated on the same day as surgery. The most common reasons for postoperative reintubation were hypoxemia, hypoventilation, and/or hypercapnia (69.8%), followed by airway protection (20.5%).
Of 834 procedures evaluated, explicit documentation of preoperative oxygen saturation was missing in 287 cases (34.4%), although 80.6% of cases had an American Society of Anesthesiologists physical status classification of III or IV (indicating severe systemic disease). In 323 procedures (38.7%), there was no documentation of a PaO2 < 60 mm Hg or a PaCO2 > 45 mm Hg at any time in the postoperative period. For patients who had a postoperative PaO2 < 60 mm Hg, the median time from the end of the surgical procedure to the first postoperative event was 50 hours (range; 10 minutes to 42.1 days). For patients who presented postoperatively with a PaCO2 > 45 mm Hg, the median time to the first postoperative event was 20 hours (range, 7 minutes to 42.1 days). An order or protocol for breathing exercises was not apparent in 31% of cases.
This measure targets HAPU; it excludes patients with a length of stay of 4 days or less, patients admitted from nursing homes and other acute care facilities, and patients who are particularly susceptible to pressure ulcers due to paralytic conditions. Of the 1234 confirmed HAPU cases, we found that 96.5% had a skin assessment and 80.9% had PU risk assessment completed on the date of or the date following admission. There were 152 HAPU cases with an emergency department (ED) length of stay greater than 24 hours, of which 80.9% had documentation of a first skin assessment completed prior to ED departure. For cases admitted to a preoperative holding area for greater than 24 hours prior to admission, 87.2% had a first skin assessment before departure to another inpatient unit. Twenty-five percent of HAPU cases were not assessed for pressure ulcer risk within the 2 days prior to the diagnosis.
Only 3.6% (median hospital performance) of HAPU cases included documentation of the location, dimensions, exudate, and site when the ulcer was first recognized, and only 3.5% (median hospital performance) of cases had these characteristics documented at discharge. Staging, which is important for deciding on nursing interventions and treatment, was initially documented in 64.3% of HAPU cases (including ulcers documented as stages 1–4 or unstageable). Stage 3, 4, and unstageable ulcers represented 20.1% of HAPU cases on diagnosis. Only 13.3% of new HAPUs were stage 1 when first recognized. Physicians documented the presence of a HAPU in less 50% of all cases. Staging was documented in 58% (496/850) of HAPU cases at discharge.
Height or weight was missing from 20.6% of HAPU cases, despite a median length of stay of 26 days (range, 5–319). Excluding ambulatory patients, those with a Braden score greater than 18 (low risk), or those with a documented contraindication to turning, less than half (49.1%) of bed-bound patients were turned or repositioned at least every 2 hours in the 2 days prior to being diagnosed with a HAPU, although most (61.5%) were located in an intensive care unit on the day of diagnosis.
Potential nursing opportunities to improve care relevant to the AHRQ PSIs are summarized in the Table. Common trends across indicators were lack of documentation of an accurate and timely patient assessment, lack of adherence to national standards of care related to prophylaxis or prevention, and delays in recognition and intervention.
Table-a. Summary of ...Image Tools
Accurate patient assessments are essential not only for providing safe nursing care but also for effective and efficient team collaboration. Documentation of such is an integral part of safe and appropriate practice, reflecting both the judgment and critical thinking that nurses are expected to provide in acute care settings. Improved documentation of CLRBSI, HAPU, and DVT events present on admission is not only essential to developing the plan of care but can directly impact observed PSI rates and reimbursement under the Centers for Medicare & Medicaid Services (CMS) Deficit Reduction Act.13 Many of these areas of care are good candidates for improving interdisciplinary rounding and communication, incorporation into daily tracking tools, and rapid cycle change. Some of measures examined in the study are reflected in the Institute for Healthcare Improvement's (IHI) bundles associated with the prevention of hospital-acquired events.9
Table-b. Summary of ...Image Tools
Unlike other national CLRBSI measures (such as The National Healthcare Safety Network and National Database for Nursing Quality Indicators), PSI 7 has the advantage of using readily available administrative data and is not limited to specific hospital units.12,14 Our review of PSI 7 cases demonstrated opportunities to improve care, starting with CVC site selection. National guidelines, such as those reflected in the IHI bundle for decreasing CLABSI call for the preferential use of the subclavian vein over other insertion sites for short-term CVC placement, to reduce the likelihood of infection, although these recommendations need to be weighed against the risk of other complications in individual cases.15–18 Although physicians typically choose catheter insertion sites, registered nurses are able to advocate for optimal site placement and are primarily responsible for postinsertion CVC line and site care, which are important in prevention of CLRBSI. For example, the internal jugular site was linked to 34% of CLRBSI events in this study, perhaps because it is harder to dress, there is more movement at the site, and it is closer to sputum and oral secretions. Femoral catheters, linked to 11% of infections in our study, are at highest risk for contamination, as evidenced by the shorter time from insertion to infection (5.7 ± 3.4 days), and therefore require extra-vigilance and should only be used when other sites are unavailable with a documented justification.16,17 Peripherally inserted central catheter lines, linked to 30% of the CLRBSI in our study, are routinely placed and managed by nurses.
In many cases, the abstractor was unable to find documentation of important information such as CVC type (peripheral vs central), insertion site, insertion or removal dates, and signs and symptoms of bloodstream infection. This lack of documentation not only impedes nursing care but also complicates surveillance efforts and affects other team members who rely on nursing documentation and assessment. For example, one of the recommended practices associated with decreased incidence of CLRBSI is daily documentation of CVC necessity, which is unlikely if the catheter type, site, and associated dates are not routinely documented.9 About one-third (36%) of CLRBSI were associated with lines used for total parenteral nutrition. Prior studies report that infections related to CVCs used for total parenteral nutrition can be avoided or greatly reduced through proper insertion and line care along with strict control of blood glucose levels, suggesting that further review of these cases may offer opportunities for improvement.19
Deep vein thrombosis and PE are among the most frequent preventable causes of hospital death in the United States.20 Nurses often provide clinical leadership in hospital initiatives to improve prophylaxis (present in only 42% of eligible cases). They also are vital team members in the timely recognition and treatment of events (missing in 17%-21% of cases). For example, nurses may participate in local audits of current practice, aid in the identification of barriers, promote methods to remind clinicians including nurses to assess patients for DVT/PE risk and the adequacy of prophylaxis, develop documentation aids, and refine local policy to further improve practice. Use of the AHRQ DVT/PE PSI has the potential to help identify areas of concern, monitor the impact of changes in practice, and provide benchmarks for comparisons.
Postoperative sepsis is relevant to perioperative and postoperative nursing practice. Identified areas for improvement include several Surgical Care Improvement Project and IHI Surgical Site Infection bundle measures such as using alternatives to razors for hair removal, administering prophylactic antimicrobials within 1 hour of incision, and maintaining normothermia.21 In addition to helping to prevent surgical site infection and sepsis, avoiding perioperative hypothermia is associated with less bleeding, faster recovery, quicker metabolism of anesthesia, and a lower likelihood of prolonged postoperative ventilation. It is therefore of interest to numerous regulatory and oversight agencies as well as the Association of periOperative Registered Nurses.21,22 Specific opportunities to promote normothermia include improved surveillance and use of the same temperature scale across the continuum of care (Celsius or Fahrenheit). Hypothermia or hyperthermia, when present with at least 1 other symptom of systemic inflammatory response syndrome, may be an early warning sign for sepsis, especially in high-risk settings.
As with CLABSI, nursing documentation of the signs and symptoms of infection or sepsis were often missing. Early recognition is essential for improving outcomes from sepsis, and to a lesser extent other infections, which are predicated on early administration of best practice therapy. A handful of hospitals have implemented automatic sepsis alerts within their electronic documentation systems to aid with earlier sepsis recognition and treatment. These systems are based on combinations of changes in clinical criteria such as decreased mental status, decreased capillary refill, and altered respiratory rate, blood pressure, urine output, and temperature. Even more hospitals have implemented formal sepsis code response teams.23
Importantly, many of the leading reasons for sepsis (urinary tract infection 12%, pneumonia 54%, CLABSI 46%) are amenable to preventive nursing interventions. For example, nursing interventions to prevent urinary tract infection include use of aseptic technique during catheter insertion, daily review of catheter necessity with prompt removal of catheters when no longer necessary, avoidance of unnecessary urinary catheters, and keeping the collection bag below the level of the bladder at all times including during patient transport. For postoperative pneumonia, preventive nursing actions include proper positioning and turning, judicious pain control, limited use of nasogastric tubes, and good oral hygiene.11 Many interventions to prevent infection, such as those related to aseptic technique, historically have not been documented in the medical record. With increasing emphasis on preventing complications, hospitals have increasingly adapted checklist and flow sheets to help identify opportunities for improvement and measure compliance to national guidelines.
Explicit documentation of preoperative oxygen saturation was missing in 287 cases (34.4%) of postoperative respiratory failure, although 80.6% of cases had an American Society of Anesthesiologists Physical Status Classification of III or IV and more than 94% received general anesthesia. Because of the rapidity with which changes in respiratory status can occur, nurses serve a particularly important role in detecting signs of incipient respiratory failure, such as increased work of breathing, hypoventilation, and hypoxia. In fact, the leading reason for reintubation was hypoxemia, hypoventilation, and/or hypercapnia (69.8%), which are most often correctable with earlier detection and intervention. On the basis of the high percentage of cases reintubated within the first day and week (22.5% and 80.2% respectively), closer monitoring of study patients in the early postoperative period and more consistent protocols for postoperative lung expansion exercises may have prevented at least some cases of postoperative respiratory failure. Nurse-led rapid response teams have demonstrated some promising but inconsistent results in preventing poor outcomes in patients with postoperative respiratory failure.24
Of the PSIs, HAPU is most closely related to care provided by nurses. National guidelines call for skin and pressure ulcer risk assessment to be performed at admission and at regular intervals thereafter, with high-risk patients being reassessed at least daily.25 Initial assessments must start at the time of entry into the health care system including holding areas (missing in 12.8%) and the ED (missing in 19.1%). The lack of documentation of skin and pressure ulcer risk assessments, PU ulcer characteristics (defined as location, dimensions, exudate, and site, all of which were documented in only 3.6% of cases at PU diagnosis and 3.5% at discharge), and basic nursing interventions such as frequent repositioning of high-risk patients, suggest that at least some pressure ulcers may have been preventable. Because only 13.3% of HAPU were stage 1 when first documented, there are opportunities to recognize PU events earlier and possibly to prevent progression to deeper stages. Basic hygiene, skin integrity, and pressure ulcer treatment involve independent nursing actions that do not require physician orders. For example, patients with low and/or high body mass indices are at increased risk for pressure ulcer and undernutrition. Height and weight, therefore, should be part of the nursing admission assessment, but these parameters were missing in nearly 21% of cases. Turning or repositioning bed-bound patients at least every 2 hours is an important nursing intervention, not documented in about half of all high-risk patients, even though 61.5% were receiving intensive care on the day of diagnosis.
Not all data elements of interest were available via chart review (eg, hand washing and aseptic technique in cases of CLRBSI) and time constraints limited our ability to assess other potentially interesting processes of care (eg, frequency of oral care and head of bed position in postoperative respiratory failure). We had no mechanism for assessing the reliability or validity of data collection, as we did not have direct access to source documents and were unable to confirm that documented actions were actually performed. Absence of documentation related to the PSI events was not limited to nurses, but reflects on the entire health care team. Sampling was limited to cases that were flagged positive and therefore we had no comparison group with which to identify differences in nursing processes of care.
Nurses are critical to the prevention and early recognition of potential complications and adverse events. Quantifying the impact of nursing on quality-related events is becoming increasingly important. The AHRQ PSIs represent an inexpensive data source that may be used to further evaluate the work of nurses; however, additional research is required. Systematic review of cases flagged by the PSIs demonstrates opportunities for improvement in nursing care and nursing documentation and supports the importance of nurse-led system-level changes. Patient safety indicators rates may be used to evaluate the impact of such changes in processes of care and organizational function.
2. Utter GH, Cuny J, Sama P, et al. Detection of postoperative respiratory failure: how predictive is the AHRQ patient safety indicator? J Am Coll Surg. 2010;211(3):347–354.
3. White RH, Sadeghi B, Tancredi DJ, et al. How valid is the ICD-9-CM based AHRQ patient safety indicator for postoperative venous thromboembolism? Med Care. 2009;47(12):1237–1243.
4. Cevasco M, Borzecki AM, Chen Q, et al. Positive predictive value of the AHRQ patient safety indicator “postoperative sepsis”: implications for practice and policy. J Am Coll Surg. 2011;212(6):954–961.
5. Zrelak PA, Sadeghi B, Utter GH, et al. Positive predictive value of the AHRQ patient safety indicator for infections due to central venous catheters (“selected infections due to medical care”). J Healthc Qual. 2011;33(2):29–36.
6. Geerts WH, Bergqvist D, Pineo GF, et al. Prevention of venous thromboembolism: American college of chest physicians evidence-based clinical practice guidelines. 8th ed. Chest. 2008;133(6)(suppl):381S–453S.
8. QualityNet. The surgical care improvement project (SCIP). http://www.qualitynet.org
. Published 2010. Accessed January 6, 2010.
9. Institute for Healthcare Improvement (IHI). http://www.ihi.org/ihi
. Published 2010. Accessed January 6, 2010.
10. Association of periOperative Registered Nurses (AORN). Perioperative Standards and Recommended Practices. Denver, CO: AORN Inc; 2011.
11. O'Keefe-McCarthy S, Santiago C, Lau G. Ventilator-associated pneumonia bundled strategies: an evidence-based practice. Worldviews Evid Based Nurs. 2008;5(4):193–204.
12. The Joint Commission. Implementation guide for the NQF endorsed nursing sensitive care measure set, 2009. http://www.jointcommission.org
. Accessed March 26, 2010.
13. Department of Health and Human Services. Medicare program; proposed changes to the hospital inpatient prospective payment systems and fiscal year 2009 rates; proposed changes to disclosure of physician ownership in hospitals and physician self-referral rules; proposed collection of information regarding financial relationships between hospitals and physicians. Part II 42 CFR parts 411, 412, 413, 422, and 489. 2008;73.
14. Centers for Disease Control and Prevention. National healthcare safety network (NHSN). http://www.cdc.gov/nhsn/about.html
. Published 2011. Accessed February 11, 2011.
15. Marschall J, Mermel LA, Classen D, et al. Strategies to prevent central line-associated bloodstream infections in acute care hospitals. Infect Control Hosp Epidemiol. 2008;29(suppl 1):S22–S30.
16. Goetz AM, Wagener MM, Miller JM, Muder RR. Risk of infection due to central venous catheters: effect of site of placement and catheter type. Infect Control Hosp Epidemiol. 1998;19(11):842–845.
17. Merrer J, De Jonghe B, Golliot F, et al. Complications of femoral and subclavian venous catheterization in critically ill patients: a randomized controlled trial. JAMA. 2001;286(6):700–707.
18. Institute for Healthcare Improvement. How to Guide: Prevent Central Line-Associated Bloodstream Infections. Cambridge, MA: Institute for Healthcare Improvement; 2011.
19. Dimick JB, Swoboda S, Talamini MA, Pelz RK, Hendrix CW, Lipsett PA. Risk of colonization of central venous catheters: catheters for total parenteral nutrition vs other catheters. Am J Crit Care. 2003;12(4):328–335.
20. National Quality Forum (NQF). National quality forum (NQF). Safe practices for better healthcare—2010 update: a consensus report; 2010.
21. The Medicare Quality Improvement Community (MedQIC). Surgical care improvement project (SCIP). http://www.qualitynet.org
. Published 2011. Accessed February 11, 2011.
22. Centers for Medicare and Medicaid Services (CMS) and The Joint Commission (TJC). Specifications Manual for National Hospital Inpatient Quality Measures. Baltimore, MD: CMS and TJC; 2011.
24. Massey D, Aitken LM, Chaboyer W. Literature review: do rapid response systems reduce the incidence of major adverse events in the deteriorating ward patient? J Clin Nurs. 2010;19(23/24):3260–3273.
25. The National Pressure Ulcer Advisory Panel (NPUAP). Pressure Ulcer Prevention Points. Washington, DC: NPUAP; 2010.
administrative data; adverse event; central line–related bloodstream infection; deep vein thrombosis; hospital-acquired conditions; nursing-sensitive indicator; patient safety indicator; postoperative respiratory failure; pressure ulcer; quality improvement; sepsis
© 2012 Lippincott Williams & Wilkins, Inc.