Cascade iatrogenesis refers to a cascade of complications that may result from a seemingly benign initial insult, such as that brought about as a result of treatment or hospitalization. If the initial insult is not recognized or managed properly, the after effects can snowball into a larger adverse event. To illustrate, when an older patient with postoperative pain is oversedated, a decline in respiratory function may occur, that if not recognized by providers, can result in atelectasis and even respiratory depression. Alternatively, untreated pain can result in the patient being unable to engage in early ambulation or active breathing exercises, leading to lower lung volumes, atelectasis, and respiratory depression. In either case, the unrecognized atelectasis and respiratory depression can ultimately lead to respiratory failure requiring intubation. Cascading further, intubation, if not managed properly, can culminate in ventilator-associated pneumonia and even sepsis and death (Thornlow, Anderson, & Oddone, 2009). Cascading complications are not uncommon and can result in further functional decline due to the illness itself, deconditioning, or the adverse effects of treatment (Callahan, Thomas, Goldhirsch, & Leipzig, 2002; Quinlan & Rudolph, 2011). Even early complications, those that occur within the first day after surgery, increase risk for death (Silber et al., 2005).
Postoperative respiratory failure is common and served as the exemplar of cascade iatrogenesis for the current study, with the initial events identified as oversedation, fluid overload, and atelectasis. According to national estimates, older adults (age ≥65) are four times more likely than younger adults (ages 18 to 44) and 60% more likely than middle-aged adults (ages 45 to 64) to experience postoperative respiratory failure following elective surgery; individuals older than 85 experience the highest rates (Agency for Healthcare Research and Quality [AHRQ], 2010; Thornlow, 2009). Surgical patients with postoperative respiratory failure had 3.74 times higher odds of death within 90 days compared to those without postoperative respiratory failure (Encinosa & Hellinger, 2008).
Identifying patients at risk is a first step in preventing complications. Arozullah, Khuri, Henderson, and Daley (2001) and other investigators (Smetana, Lawrence, & Cornell, 2006) developed risk indices for predicting postoperative pulmonary complications after non-cardiothoracic surgery, and these multifactorial risk tools were incorporated into The American College of Physicians guidelines (Qaseem et al., 2006). These risk calculation tools assess patient risk factors (e.g., age, comorbidities) and surgical risk factors (e.g., type of surgery, duration of anesthesia); however, they do not consider the role of postoperative nursing care, which would include postoperative monitoring and assessment. Further, these tools ignore the points at which a cascading event of iatrogenesis is likely to begin and progress—little to no attention is given to the frequency of postoperative monitoring and assessment, adequacy of pain relief, or the level of nursing surveillance, all of which may affect morbidity and mortality (Thornlow et al., 2009).
In this pilot study, early events in cascade iatrogenesis are described using the example of postoperative respiratory failure. The aim was (a) to determine whether the selected trigger events were present and recorded in the patients’ charts, and (b) to examine the relationship between these initial trigger event(s) and the subsequent occurrence of postoperative respiratory failure. Not only were events that precipitated postoperative respiratory failure examined, but the nursing interventions that may have prevented, mitigated, or exacerbated risk were described. Given that this was a pilot study, the goal for analyses was to detect trends to guide development for a future larger study.
We used a cascade iatrogenesis framework (Thornlow et al., 2009) to guide our study of the risk factors and nursing care that preceded the development of postoperative respiratory failure in hospitalized older adults (Figure). Developed from a literature review and two existing frameworks (Mitchell, Ferketich, & Jennings, 1998; Picone et al., 2008; Titler et al., 2006), the conceptual model incorporates factors associated with older adults’ risk for developing the outcome, postoperative respiratory failure, including unit characteristics, events that may trigger the initial cascade, and the nursing care variables that may prevent, mitigate, or exacerbate these risks. The model includes unit characteristics, such as nursing staff levels, which have been shown to affect pulmonary outcomes (Beck et al., 2013; Brennan, Daly, & Jones, 2013); nursing care processes, such as nursing surveillance and pain management, which have been demonstrated to affect clinical outcomes (Kutney-Lee, Lake, & Aiken, 2009; Rakel & Herr, 2004; Shea, Brooks, Dayhoff, & Keck, 2002); and trigger events that may provoke respiratory depression, and moreover, events that if not recognized, may result in postoperative respiratory failure. As noted in the model, trigger events may include episodes of oversedation, atelectasis, fluid overload, and even delirium, each of which may predispose patients to poorer outcomes. Patient characteristics, including age and comorbidities, along with surgical procedural factors, such as type of surgery and length of anesthesia time, have been associated with increased risk for developing postoperative respiratory failure (Arozullah et al., 2001; Smetana et al., 2006) and thus are included in the model (Thornlow et al., 2009). The model guided the variables selected to study.
Cascade iatrogenesis framework for postoperative respiratory failure. Source: Thornlow, D., Anderson, R., & Oddone, E. (2009). Cascade iatrogenesis: Factors leading to the development of adverse events in hospitalized older adults. International Journal of Nursing Studies, 46(11), 1528–1535. Reprinted with permission.
In this retrospective, descriptive, case-control pilot study, medical record data were abstracted to examine trigger events in more detail and compare risk, nursing care, and outcomes of cases that developed postoperative respiratory failure to controls who did not. The study was approved by the participating university’s Institutional Review Board. All data were protected on a secured, password-protected server, and in a secured encrypted folder within that server.
Setting, Sample, and Procedures for Matching
A 924-bed tertiary and quaternary teaching hospital located in the southeastern United States was the setting for this pilot study. Records of patients ages 65 and older who underwent an elective surgical procedure and were discharged from the hospital in fiscal year 2008 with a secondary diagnosis of postoperative respiratory failure were retrospectively reviewed. Postoperative respiratory failure was defined using the Patient Safety Indicator definition by the AHRQ (2009), which stipulates that the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) (World Health Organization, 2013) codes 518.81 or 518.84 (designating acute respiratory failure) appear as secondary diagnoses in the discharge records of elective surgical patients. All cases meeting the inclusion criteria in the full year prior to the study (N = 28) were included. Patients in the same age range who underwent the same procedure in the same year, but who did not develop postoperative respiratory failure during their hospitalization, served as controls. An analyst who was not part of the study team randomly selected 28 control records matching cases to controls based on their age, year of surgery, type of surgical procedure (i.e., ICD-9-CM code), and American Society of Anesthesiologists (ASA, n.d.) score. The ASA preoperative classification score is a subjective assessment of a patient’s overall health status that is based on five classes, ranging from 1 (completely healthy), to 5 (a moribund patient who is not expected to live 24 hours with or without surgery). The ASA classification score was originally designed to estimate overall mortality risk in surgical patients, yet also has been shown to predict postoperative cardiovascular and pulmonary complications (Smetana, 2006). In this study, the ASA score was used as a measure of patient complexity preoperatively.
Study Variables and Record Review Procedures
Table 1 describes the variable definitions, data abstracted, and data sources. Data available electronically were retrieved from the hospital’s clinical data repository, the Decision Support Repository (DSR); data unavailable electronically within the DSR were abstracted from the patient’s hard copy medical record. All abstracted data were entered and stored on a secured server. Patient characteristics (demographic data); the type, frequency, and timing of potential trigger events that preceded postoperative respiratory failure; and the unit on which the event occurred were recorded. Recognizing the importance of the temporality of the data, the timing and sequence of trigger events were recorded. For cases, trigger events were recorded, noting when they occurred, from the time of admission to the postanesthesia care unit (PACU), or similar unit, to the time of postoperative respiratory failure (intubation). For matched controls, the trigger events and when they occurred during a postoperative period of the same length were recorded. The type, frequency, and timing of nursing care provided during hospitalization for both cases and controls were also recorded.
Construction of Variables
Descriptive statistics were calculated for the type, frequency, and timing of potential trigger events and the nursing care that preceded development of postoperative respiratory failure. Cases and controls were compared using independent t tests for continuous variables and chi-square analysis for categorical variables (Table 2, Table 3, and Table 4) rather than the more robust odds ratios and logistic regression because of the small sample.
Patient Demographic and Procedural Variables
Unit and Nursing Care Variables
Trigger Events and Patient Outcomes
The results are organized using the concepts in the Cascade Iatrogenesis Model (Thornlow et al., 2009) (Figure).
No differences were found in the characteristics that were used for matching cases (Table 2); however, cases had more documented secondary conditions than controls. Because it was unclear whether these secondary codes were comorbidities that were present on admission or complications that developed during hospitalization, the charts were re-reviewed to identify and document conditions that would conceivably have been present on admission. These were coronary artery disease (e.g., previous myocardial infarction, hypertension) and lung disease (i.e., chronic obstructive pulmonary disease, lung cancer, history of tobacco use), diabetes mellitus, and documented obesity. Cases were compared to controls for these particular characteristics with no significant differences found for the presence of coronary artery disease, lung disease, or diabetes. Controls were more likely to be obese than cases.
No statistically significant differences were found between the two groups by type of anesthesia; however, the cases differed significantly in the total minutes of anesthesia, exceeding controls by an average 33 minutes (p = 0.009). Cases returned to the operating room (OR) significantly more often than their matched counterparts (Table 2). Reasons for return to the OR were complications such as bleeding or tracheostomy placement following prolonged intubation for respiratory failure.
This study was conducted in one hospital and most patients were treated on one of three floors (i.e., neurosurgery, general surgery, or orthopedics). Cases were admitted more often to intensive care units (ICUs) following surgery than their matched counterparts, but this finding was not significant. For both cases and controls, more trigger events occurred on the floor than in the ICU (Table 3). Cases, or patients who developed postoperative respiratory failure, were transferred more often than their matched counterparts and thus resided on more types of units during their hospital stay.
Cases and controls used patient-controlled analgesia (PCA) at similar rates, with three case patients and one control patient requiring naloxone (Narcan®) administration (Table 3) for oversedation. Cases were less likely to ambulate early and required more calls to rapid response and code teams. These patients were subsequently transferred to higher levels of care more frequently than patients who did not develop postoperative respiratory failure (Table 3).
Because nurse staffing data were attributed to nursing units, not individual patients, the type of unit where the patient resided was used as a proxy variable for nurse staffing levels. Nurse staffing in a step-down unit is higher than on a floor (generally 1 nurse to 3 patients in step-down and 1 nurse to 5 patients on the floor), yet less intensive than nurse staffing in a critical care unit (generally 1 nurse to 1 or 2 patients). No significant differences were found between cases and controls in the use of step-down care, yet cases were cared for more often in the ICU (n = 12) than on the floor (n = 3), whereas the reverse was true for controls: More controls were cared for on the floor (n = 9) than in the ICU (n = 6).
Oversedation, delirium, atelectasis, aspiration, and fluid overload were found to be trigger events (Table 4). On average, cases experienced a trigger event that precipitated transfer to a higher level of care on Day 5 of their hospitalization; cases experienced certain trigger events at significantly higher rates than controls. For example, atelectasis occurred more often in cases who developed postoperative respiratory failure than controls who did not, with the odds of documented atelectasis almost two times greater in cases than controls. Patients who developed postoperative respiratory failure experienced more episodes of fluid overload and pulmonary edema than matched controls and these patients received more doses of furosemide (Lasix®).
Although the groups did not differ in the frequency of episodes of oversedation, three cases and two controls, or 9% of the sampled patients, required intervention for oversedation. Three cases and one control required naloxone, whereas one control had his PCA discontinued.
Significant differences were noted in the rates of delirium between cases and controls. All cases who exhibited signs of delirium (n = 10) or who became oversedated (n = 3) required intubation, whereas none of the controls who exhibited signs of confusion or delirium (n = 2) or became oversedated (n = 1) required intubation. Of note, all 10 patients who developed delirium in the case group also aspirated; yet, in the control group, neither of the two patients who developed confusion or delirium aspirated. Overall, cases aspirated more often than controls.
Patients who developed postoperative respiratory failure demonstrated higher rates of pneumonia than controls. Four of the 13 patients who developed pneumonia (13 cases, 0 controls) died during their hospital stay. Ultimately, patients who developed postoperative respiratory failure had longer lengths of stay than controls, were less likely to be discharged home, and were more likely to die than patients who did not develop postoperative respiratory failure.
Most of the variables were able to be abstracted from either the electronic or hard copy charts, yet it was difficult to tease out from the medical records the actual nursing care provided. Nursing care was documented according to established postoperative protocols and free-text notes were written, yet finding clinically relevant nursing care variables that were readily available and could be abstracted posed challenges. These challenges are addressed below.
Trends in the data that would guide further exploration in a larger study were noted. Primarily, evidence of cascade iatrogenesis was found with some catastrophic outcomes. For example, both cases and controls experienced similar rates of oversedation, yet case patients were more likely to experience aspiration following episodes of oversedation than their matched counterparts. Evidence was found to substantiate a cascade: of the 11 cases who aspirated, five subsequently developed pneumonia, with three of these pneumonias attributed to the aspiration event (classified as aspiration pneumonia). One of these three patients died from aspiration pneumonia. Recognizing patients who are at risk for trigger events and taking steps to prevent or mitigate their occurrence remain key nursing responsibilities that may prevent this tragic cascade.
Evidence of trigger events (e.g., oversedation, delirium, atelectasis, fluid overload) was found in the current study. More than 7% of all sampled patients (3 cases and 1 control or 4/56 patients) required intervention for oversedation (e.g., naloxone administration), suggesting that nursing care related to pain management remains somewhat challenging in the older adult postoperative population. This finding is supported by a retrospective cohort study of 8,855 patients ages 16 and older who received short-term opioid agents for pain control (Cepeda et al., 2003), in which investigators found that patients ages 71 to 80 had a 5.4 times greater risk and those older than 80 had an 8.7 times greater risk of respiratory depression than patients ages 16 to 45. Further, of those patients who were oversedated, none of the control patients aspirated, whereas two of the case patients aspirated, perhaps indicating earlier detection or mitigation of oversedation by nurses caring for the control group. This hypothesis deserves attention in future studies. Moreover, although the literature is replete with assessment tools for pain ratings, sedation levels, and respiratory status and guidelines for completing them before and after opioid agent administration (Hjermstad et al., 2011; Smith, 2007), Zeitz and McCutcheon (2006) found that postoperative vital signs were collected based on traditional policies rather than patient risk profile. More evidence is needed to establish age- and risk-based protocols to guide nurses to assess patients at higher risk for oversedation more frequently than those patients with lower risk for oversedation.
In the current study, no significant differences were found in the rates of delirium between cases and controls. However, six of 10 case patients with delirium aspirated, whereas neither of the two control patients with delirium aspirated. Aspiration posed a clear risk for complications in this study. Both cohorts experienced similar rates of oversedation and delirium, yet no control patients aspirated. It is possible that earlier detection or intervention for oversedation and delirium by nurses prevented aspiration and the subsequent cascade toward postoperative respiratory failure in the control patients. This hypothesis warrants further investigation—recognizing patients who are at risk for aspiration, such as those who are confused, oversedated, or lethargic, and taking precautions to prevent aspiration, remain key nursing responsibilities. Perhaps nurses caring for the control patients who exhibited signs of delirium in this study referred those patients for swallowing studies, managed their pain differently, or intervened in other ways to reduce their risk for aspiration. Based on these findings, it is posited that aspiration is an important trigger event that should be added to the conceptual model (Figure). Further, nursing care for early detection and mitigation of oversedation and delirium is a fruitful topic for further research, and nursing care strategies for this population remain an important area for further study. Moreover, delirium may further exacerbate inability to engage in early ambulation and deep breathing exercises and may complicate pain management strategies. Prior research also suggests that delirium and cognitive impairment, a precursor to delirium, predict adverse outcomes in hospitalized older adults (McCusker, Kakuma, & Abrahamowitcz, 2002).
The odds of documented atelectasis were almost five times greater in cases than controls, suggesting that it is a significant factor in postoperative respiratory failure. This finding is difficult to interpret because only half of controls had documented chest x-rays following surgery or during their hospitalization, whereas all cases had documented chest x-rays during their hospital stay. Therefore, it is possible that control patients would demonstrate incidental findings of atelectasis on their chest x-rays had such x-rays been procured. Other explanations for these differences may suggest that early ambulation and other interventions by nurses may have prevented atelectasis and its sequelae in the control patients.
Older adults are particularly sensitive to over- or under-hydration due to age-related changes in renal function, thus there is a need for careful fluid balance in the postoperative period. In this study, patients who developed postoperative respiratory failure experienced more episodes of fluid overload and pulmonary edema than their matched controls and received more doses of furosemide. Patients who received furosemide in the control group received the dose earlier in the course of their hospitalization, but whether this difference is merely related to their shorter length of stay or whether nurses intervened earlier to prevent or mitigate further deterioration from fluid overload is unclear and warrants further investigation.
For both cases and controls, more trigger events occurred on the floor than in the ICU (Table 3). ICUs generally staff with a higher nurse-per-patient ratio than floors, yet the floor and ICU comprise differing populations. Because these trends were not examined between floors, or between ICUs, it is difficult to draw conclusions from these data, thus limiting the use of the proxy variable for nurse staffing, the “unit on which the patient resided.” In prior research, higher nurse staffing levels have been associated with lower rates of postoperative pulmonary complications, although findings are mixed (Beck et al., 2013; Brennan et al., 2013). Although Needleman, Buerhaus, Mattke, Stewart, and Zelevinsky (2002) found no relationship between nurse staffing levels and pulmonary failure among major surgery patients, other investigators reported that higher nurse staffing levels were associated with lower rates of postoperative pulmonary complications, including pneumonia (Cho, Ketefian, Barkauskas, & Smith, 2003; Kovner, Jones, Zhan, Gergen, & Basu, 2002; Mark, Harless, McCue, & Xu, 2004; Twigg, Geelhoed, Bremner, & Duffield, 2013). Unruh (2003) reported that higher rates of adverse events, including rates of atelectasis and pneumonia, were associated with higher unit acuity levels. In future studies, comparing trends between similar types of units with varying levels of nurse staffing, or using patientlevel nurse staffing variables, may proffer evidence that nurse staffing ratios correlate with trigger events and rates of postoperative respiratory failure.
Patients who developed postoperative respiratory failure were transferred more often than their matched counterparts and thus resided on more units during their hospital stay. The timeline for these transfers was not ascertained (i.e., whether these transfers occurred prior to or after the patient developed respiratory distress). Future studies should examine this timeline. Titler et al. (2006) found that the clinical outcomes of nosocomial infection, adverse occurrences, patient falls, medication errors, and discharge disposition were significantly associated with the number of units per hospitalization—as the number of units increased from one unit, the odds of acquiring at least one nosocomial infection, one fall, or one adverse event also increased. This suggests that factors other than staffing, such as familiarity with the patients and potential gaps in patient handoffs, are possible explanations for the ability of nurses to recognize at-risk patients or those experiencing signs of deterioration. Number of units is a proxy variable for issues related to multiple handoffs in which key patient information may not be relayed to the receiving provider. Ascertaining the timeline for these transfers and handoffs will illuminate the role handoffs may play in postoperative respiratory failure. Nursing education (Aiken et al., 2012; Blegen, Goode, Park, Vaughn, & Spetz, 2013) and experience (Gillespie, Chaboyer, Wallis, & Werder, 2011; Tourangeau, Giovannetti, Tu, & Wood, 2002) also have been linked to the quality of care. These factors may be important for nursing surveillance and suggest the need to more carefully examine the context of care for its effect on cascade iatrogenesis. In addition, unfamiliarity of the units by patients and families may place patients at risk for untoward outcomes, such as may be the case due to unfamiliarity or discomfort communicating with new providers.
Nursing surveillance is a process through which nurses monitor, evaluate, and act on emerging indicators of a patient’s change in status. The components of this process include ongoing observation and assessment, recognition, interpretation of clinical data, and clinical decision making (Kutney-Lee et al., 2009). Observational studies suggest that deterioration of patients on general medical and surgical wards is often preceded by changes in physiological observations that are recorded by clinical staff 6 to 24 hours prior to a serious adverse event. In the current study, nurses placed more calls to rapid response teams for cases than controls, and these case patients were subsequently transferred to higher levels of care. Events that may have triggered the calls to these teams were identified, and although the trigger events mirrored early warning signs (Cei, Bartolomei, & Mumoli, 2009; Cuthbertson, Boroujerdi, McKie, Aucott, & Prescott, 2007; Gardner-Thorpe, Love, Wrightson, Walsh, & Keeling, 2006), specific physiological data or when these warning signs became apparent were unable to be denoted. It can be supposed that more frequent surveillance by nurses and thus earlier intervention for warning signs of deterioration in the control group may have prevented the need to call rapid response teams and hence reduced subsequent transfers to higher levels of care. More research is needed to delineate the decision making that occurs along critical points in the cascade toward postoperative respiratory failure. Detection and intervention are closely tied with nursing care (Friese & Aiken, 2008). The inability to recognize and intervene during the early signs of deterioration leads to late ICU admission, excess mortality, and increased hospital costs (Cuthbertson et al., 2007).
And finally, patients who developed postoperative respiratory failure were less likely to ambulate early. As noted, these patients developed higher levels of atelectasis than their matched controls. Perhaps nurses caring for control patients recognized the protective effect of early ambulation and ensured that their patients ambulated early and often, or encouraged alternate means for lung aeration, which may have prevented atelectasis. These findings support other evidence about the importance of early ambulation in the postoperative setting (Shea et al., 2002). Further research is needed to understand the barriers to early ambulation.
In summary, all 28 case patients required intubation for their postoperative respiratory failure, whether their postoperative respiratory failure was related to atelectasis, pulmonary edema, aspiration, or some other trigger event. Almost half of the patients who required intubation (N = 28) subsequently developed pneumonia (n = 13) and four of these 13 patients died during their hospital stay. In total, patients who developed postoperative respiratory failure had longer lengths of stays, were less likely to be discharged home, and were more likely to die than patients who did not develop postoperative respiratory failure.
Although the sample was small for this study, it was adequate to test procedures and examine trends in the data to support a larger study. Patients were matched as intended by age, ASA score, and year and type of surgery, yet the majority of patients were classified as ASA score 3, signaling severe systemic disease. The matching factors were chosen based on prior evidence that these factors predict postoperative risk (Arozullah et al., 2001; Smetana, 2006; Smetana et al., 2006). The ASA score was used as a summary measure of preoperative comorbidity. The ASA preoperative classification score, originally designed to estimate overall mortality risk in surgical patients, also has been shown to predict postoperative cardiovascular and pulmonary complications. Based on general health status, patients who are graded higher than class 2 in the 5-class ASA system have a two- to three-fold increased risk of postoperative pulmonary complications compared to those graded class 2 or lower (Smetana, 2006). Borzecki et al. (2012) found that 56% of patients with postoperative respiratory failure had an ASA level higher than 2. In the current study, most patients had ASA scores of 3 or 4, thus potentially limiting the effectiveness of matching cases to controls using this variable. Future matching should consider not only ASA level, but also certain types of comorbidities present on admission (e.g., pre-existing lung disease). Although no significant differences were found between cases and controls for the presence of coronary artery disease, lung disease, or diabetes, controls were more likely to be obese than cases. In future studies, it is recommended to include only patients who receive general anesthesia, thus excluding patients who receive other types of anesthesia (e.g., regional or spinal). All types of anesthesia were included in this pilot study so that orthopedic surgical patients who are at risk for pulmonary complications due to their potential difficulties mobilizing after surgery could be examined.
It was difficult to gather from the medical records the actual level of nursing surveillance provided. Future work must focus on identifying measurable, clinically relevant nursing variables that can be abstracted from the medical record and that can be used to analyze the association among nursing care, trigger events, and postoperative respiratory failure in hospitalized older adults. “Nursing dose” may be a more effective measure of the intensity of nursing care (Manojlovich & Sidani, 2008; Sidani, Manojlovich, & Covell, 2010), although staffing data were recorded per nursing unit, rather than the patient level of analysis in the current study, making any nurse dose calculation difficult. For this reason, the type of unit where the patient resided was used as a proxy variable for nurse staffing. Where available, other nursing variables shown to influence patient outcomes, such as education and experience, could be collected in future studies. Determining whether patients received respiratory therapy during the postoperative period may also be warranted. And finally, a larger sample from multiple settings would enable a more robust multivariate analysis. Observation via case study could also shed light on the actual nursing care provided during the postoperative period.
The current study used a cascade iatrogenesis framework (Thornlow et al., 2009) to examine risk factors that address the relationship between risk, nursing care, and postoperative respiratory failure. Using this model, events were identified that not only may trigger the cascade toward postoperative respiratory failure in hospitalized older adults, but that also may be prevented through nursing care. In this pilot study, patients who developed postoperative respiratory failure were more likely to experience trigger events such as atelectasis and aspiration. Patients who developed postoperative respiratory failure experienced more calls to rapid response or code teams, more transfers to higher levels of care, longer lengths of stay, and more deaths than matched controls.
A cascade implies a series of events. It is hypothesized that certain events or conditions may trigger respiratory depression in hospitalized older adults, and if these trigger events are not prevented or mitigated, the cascade toward postoperative respiratory failure would begin. Understanding how small events lead to larger untoward outcomes requires additional research to identify and interrupt key points in the cascade. Recognizing patients who are at risk for trigger events and taking steps to prevent or mitigate their occurrence remain key nursing responsibilities that can prevent small events from cascading into larger untoward outcomes. Much of the actual care provided by nurses was nearly invisible in the medical records. Further evidence is needed to support that nursing care and surveillance, beyond simply nurse staffing levels and staff composition, contribute to patient outcomes. Procuring this evidence may require using additional measures or undertaking direct observational studies to not only identify factors that contribute to the development of postoperative respiratory failure in hospitalized older adults, but that also will enable clinicians to assess high-risk older adult patients and develop interventions to prevent this complication.
- Agency for Healthcare Research and Quality. (2009). Patient safety indicators overview. Retrieved from http://www.qualityindicators.ahrq.gov/Modules/psi_resources.aspx
- Agency for Healthcare Research and Quality. (2010). National healthcare quality report. Retrieved from http://www.ahrq.gov/research/findings/nhqrdr/nhqr10/pdf/nhqr10.pdf
- Aiken, L.H., Cimiotti, J.P., Sloane, D.M., Smith, H.L., Flynn, L. & Neff, D.F. (2012). Effects of nurse staffing and nurse education on patient deaths in hospitals with different nurse work environments. Journal of Nursing Administration, 42(10 Suppl.), S10–S16. doi:10.1097/01.NNA.0000420390.87789.67 [CrossRef]
- American Society of Anesthesiologists. (n.d.). ASA physical status classification system. Retrieved from http://www.asahq.org/Home/For-Members/Clinical-Information/ASA-Physical-Status-Classification-System
- Arozullah, A.M., Khuri, S.F, Henderson, W.G. & Daley, J. (2001). Development and validation of a multifactorial risk index for predicting postoperative pneumonia after major noncardiac surgery. Annals of Internal Medicine, 135, 847–857. doi:10.7326/0003-4819-135-10-200111200-00005 [CrossRef]
- Beck, S.L., Weiss, M.E., Ryan-Wenger, N., Donaldson, N.E., Aydin, C., Towsley, G.L. & Gardner, W. (2013). Measuring nurses’ impact on health care quality: Progress, challenges, and future directions. Medical Care, 51(4 Suppl. 2), S15–S22 doi:10.1097/MLR.0b013e3182802e8b [CrossRef]
- Blegen, M.A., Goode, C.J., Park, S.H., Vaughn, T. & Spetz, J. (2013). Baccalaureate education in nursing and patient outcomes. Journal of Nursing Administration, 43, 89–94 doi:10.1097/NNA.0b013e31827f2028 [CrossRef]
- Borzecki, M., Cevasco, M., Chen, Q., Shin, M., Itani, K.M. & Rosen, A.K. (2012). Improving identification of postoperative respiratory failure missed by the Patient Safety Indicator algorithm. American Journal of Medical Quality, 28, 315–323 doi:10.1177/1062860612468482 [CrossRef]
- Brennan, C.W., Daly, B.J. & Jones, K.R. (2013). State of the science: The relationship between nurse staffing and patient outcomes. Western Journal of Nursing Research, 35, 760–794 doi:10.1177/0193945913476577 [CrossRef]
- Callahan, E.H., Thomas, D.C., Goldhirsch, S.L. & Leipzig, R. (2002). Geriatric hospital medicine. Medical Clinics of North America, 86, 707–729. doi:10.1016/S0025-7125(02)00014-7 [CrossRef]
- Cei, M., Bartolomei, C. & Mumoli, N. (2009). In-hospital mortality and morbidity of elderly medical patients can be predicted at admission by the Modified Early Warning Score: A prospective study. International Journal of Clinical Practice, 63, 591–595 doi:10.1111/j.1742-1241.2008.01986.x [CrossRef]
- Cepeda, M.S., Farrar, J.T., Baumgarten, M., Boston, R., Carr, D.B. & Strom, B.L. (2003). Side effects of opioids during short-term administration: Effect of age, gender, and race. Clinical Pharmacology and Therapeutics, 74, 102–112. doi:10.1016/S0009-9236(03)00152-8 [CrossRef]
- Cho, S.H., Ketefian, S., Barkauskas, V.H. & Smith, D.G. (2003). The effects of nurse staffing on adverse events, morbidity, mortality, and medical costs. Nursing Research, 52, 71–79. doi:10.1097/00006199-200303000-00003 [CrossRef]
- Cuthbertson, B.H., Boroujerdi, M., McKie, L., Aucott, L. & Prescott, G. (2007). Can physiological variables and early warning scoring systems allow early recognition of the deteriorating surgical patient?Critical Care Medicine, 35, 402–409. doi:10.1097/01.CCM.0000254826.10520.87 [CrossRef]
- Encinosa, W.E. & Hellinger, F.J. (2008). The impact of medical errors on ninety-day costs and outcomes: An examination of surgical patients. Health Services Research, 43, 2067–2085 doi:10.1111/j.1475-6773.2008.00882.x [CrossRef]
- Friese, C.R. & Aiken, L.H. (2008). Failure to rescue in the surgical oncology population: Implications for nursing and quality improvement. Oncology Nursing Forum, 35, 779–785 doi:10.1188/08.ONF.779-785 [CrossRef]
- Gardner-Thorpe, J., Love, N., Wrightson, J., Walsh, S. & Keeling, N. (2006). The value of Modified Early Warning Score (MEWS) in surgical in-patients: A prospective observational study. Annals of the Royal College of Surgeons of England, 88, 571–575. doi:10.1308/003588406X130615 [CrossRef]
- Gillespie, B.M., Chaboyer, W., Wallis, M. & Werder, H. (2011). Education and experience make a difference: Results of a predictor study. AORN Journal, 94, 78–90 doi:10.1016/j.aorn.2010.11.037 [CrossRef]
- Hjermstad, M.J., Fayers, P.M., Haugen, D.F., Caraceni, A., Hanks, G.W., Loge, J.H. & Kaasa, S. (2011). Studies comparing numerical rating scales, verbal rating scales, and visual analogue scales for assessment of pain intensity in adults: A systematic literature review. Journal of Pain and Symptom Management, 41, 1073–1093 doi:10.1016/j.jpainsymman.2010.08.016 [CrossRef]
- Kovner, C., Jones, C., Zhan, C., Gergen, P.J. & Basu, J. (2002). Nurse staffing and postsurgical adverse events: An analysis of administrative data from a sample of U.S. hospitals, 1990–1996. Health Services Research, 37, 611–629. doi:10.1111/1475-6773.00040 [CrossRef]
- Kutney-Lee, A., Lake, E.T. & Aiken, L.H. (2009). Development of the Hospital Nurse Surveillance Capacity Profile. Research in Nursing & Health, 32, 217–228 doi:10.1002/nur.20316 [CrossRef]
- Manojlovich, M. & Sidani, S. (2008). Nurse dose: What’s in a concept?Research in Nursing & Health, 31, 310–319 doi:10.1002/nur.20265 [CrossRef]
- Mark, B.A., Harless, D.W., McCue, M. & Xu, Y. (2004). A longitudinal examination of hospital registered nurse staffing and quality of care. Health Services Research, 39, 279–300. doi:10.1111/j.1475-6773.2004.00228.x [CrossRef]
- McCusker, J., Kakuma, R. & Abrahamowitcz, M. (2002). Predictors of functional decline in hospitalized elderly patients: A systematic review. Journals of Gerontology. Series A, Biological Sciences and Medical Sciences, 57, M569–M577. doi:10.1093/gerona/57.9.M569 [CrossRef]
- Mitchell, P.H., Ferketich, S. & Jennings, B.M. (1998). Quality health outcomes model: American Academy of Nursing Expert Panel on Quality Health Care. Journal of Nursing Scholarship, 30, 43–46. doi:10.1111/j.1547-5069.1998.tb01234.x [CrossRef]
- Needleman, J., Buerhaus, P., Mattke, S., Stewart, M. & Zelevinsky, K. (2002). Nurse-staffing levels and the quality of care in hospitals. New England Journal of Medicine, 346, 1715–1722. doi:10.1056/NEJMsa012247 [CrossRef]
- Picone, D.M., Titler, M.G., Dochterman, J., Shever, L., Kim, T., Abramowitz, P. & Qin, R. (2008). Predictors of medication errors among elderly hospitalized patients. American Journal of Medical Quality, 23, 115–127 doi:10.1177/1062860607313143 [CrossRef]
- Qaseem, A., Snow, V., Fitterman, N., Hornbake, E.R., Lawrence, V.A., Smetana, G.W. & Weiss, K.B. (2006). Risk assessment for and strategies to reduce perioperative pulmonary complications for patients undergoing noncardiothoracic surgery: A guideline from the American College of Physicians. Annals of Internal Medicine, 144, 575–580. doi:10.7326/0003-4819-144-8-200604180-00008 [CrossRef]
- Quinlan, N. & Rudolph, J.L. (2011). Postoperative delirium and functional decline after noncardiac surgery. Journal of the American Geriatrics Society, 59(Suppl. 2), S301–S304 doi:10.1111/j.1532-5415.2011.03679.x [CrossRef]
- Rakel, B. & Herr, K. (2004). Assessment and treatment of postoperative pain in older adults. Journal of Perianesthesia Nursing, 19, 194–208. doi:10.1016/j.jopan.2004.03.005 [CrossRef]
- Shea, R.A., Brooks, J.A., Dayhoff, N.E. & Keck, J. (2002). Pain intensity and postoperative pulmonary complications among the elderly after abdominal surgery. Heart & Lung, 31, 440–449. doi:10.1067/mhl.2002.129449 [CrossRef]
- Sidani, S., Manojlovich, M. & Covell, C. (2010). Nurse dose: Validation and refinement of a concept. Research in Theory and Nursing Practice, 24, 159–171. doi:10.1891/1541-65184.108.40.206 [CrossRef]
- Silber, J.H., Rosenbaum, P.R., Trudeau, M.E., Chen, W., Zhang, X., Kelz, R.R. & Even-Shoshan, O. (2005). Changes in prognosis after the first postoperative complication. Medical Care, 43, 122–131. doi:10.1097/00005650-200502000-00005 [CrossRef]
- Smetana, G. (2006). Preoperative pulmonary evaluation: Identifying and reducing risks for pulmonary complications. Cleveland Clinic Journal of Medicine, 73(Suppl. 1), S36–S41. doi:10.3949/ccjm.73.Suppl_1.S36 [CrossRef]
- Smetana, G.W., Lawrence, V.A. & Cornell, J.E. (2006). Preoperative pulmonary risk stratification for noncardiothoracic surgery: Systematic review for the American College of Physicians. Annals of Internal Medicine, 144, 581–595. doi:10.7326/0003-4819-144-8-200604180-00009 [CrossRef]
- Smith, L.H. (2007). Opioid safety: Is your patient at risk for respiratory depression?Clinical Journal of Oncology Nursing, 11, 293–296. doi:10.1188/07.CJON.293-296 [CrossRef]
- Thornlow, D.K. (2009). Increased risk for patient safety incidents in hospitalized older adults. Medsurg Nursing, 18, 287–291.
- Thornlow, D.K., Anderson, R. & Oddone, E. (2009). Cascade iatrogenesis: Factors leading to the development of adverse events in hospitalized older adults. International Journal of Nursing Studies, 46, 1528–1535 doi:10.1016/j.ijnurstu.2009.06.015 [CrossRef]
- Titler, M., Dochterman, J., Xie, X.J., Kanak, M., Fei, Q., Picone, D.M. & Shever, L. (2006). Nursing interventions and other factors associated with discharge disposition in older patients after hip fractures. Nursing Research, 55, 231–242. doi:10.1097/00006199-200607000-00003 [CrossRef]
- Tourangeau, A.E., Giovannetti, P., Tu, J.V. & Wood, M. (2002). Nursing-related determinants of 30-day mortality for hospitalized patients. Canadian Journal of Nursing Research, 33(4), 71–88.
- Twigg, D.E., Geelhoed, E.A., Bremner, A.P. & Duffield, C. (2013). The economic benefits of increased levels of nursing care in the hospital setting. Journal of Advanced Nursing, 10, 2253–2261 doi:10.1111/jan.12109 [CrossRef]
- Unruh, L. (2003). Licensed nurse staffing and adverse events in hospitals. Medical Care, 41, 142–152. doi:10.1097/00005650-200301000-00016 [CrossRef]
- World Health Organization. (2013). International classification of diseases, ninth revision, clinical modification. Retrieved from http://www.cdc.gov/nchs/icd/icd9cm.htm
- Zeitz, K. & McCutcheon, H. (2006). Observations and vital signs: Ritual or vital for the monitoring of postoperative patients?Applied Nursing Research, 19, 204–211. doi:10.1016/j.apnr.2005.09.005 [CrossRef]
Construction of Variables
|Variable||Operational Definition||Data Source|
|Age||Age in years on day admitted to hospital||Discharge abstract|
|Race||White; Black/African American; Asian; Hispanic; Other||Discharge abstract|
|Sex||Male or female||Discharge abstract|
|Comorbid conditions||Number of ICD-9-CM secondary diagnoses documented before admission to hospital||Discharge abstract|
|ASA score||Calculated preoperatively (scores 1 to 5)||Operative record|
|Procedure||ICD-9-CM primary procedure code further categorized into surgery type (e.g., orthopedic, abdominal, thoracic) for reference in tables||Operative record|
|Anesthesia type||Type administered (e.g., general, local)||Operative record|
|Return to OR||Return to OR during hospitalization||Operative record|
|Length of anesthesia||Start time to stop time||Operative record|
|Nurse staffing||Type of unit on which patient resided is a proxy variable for nurse staffing (i.e., nurse staffing differs per unit [e.g., nurse staffing is higher in a step-down unit than on a floor, but less intensive than in a critical care unit])
Type of unit admitted post surgery
Type of unit trigger event occurred on
|Number of units||The sum of the number of units on which treatment was provided over course of hospitalization||Medical record|
|Pain management||Pain medication administered via patient-controlled analgesia (PCA)||Nursing/medication administration record|
Call to rapid response or code team
Transfer to higher level of care (e.g., step-down to ICU)
|Trigger event||Oversedation—sedation requiring naloxone (Narcan®) administration; adjustment to PCA||Medical record|
|Atelectasis—collapse of lung documented by chest x-ray||Medical record|
|Aspiration—entry of secretions or foreign material into the trachea and lungs, documented by chest x-ray or in medical record||Medical record|
|Fluid overload—hypervolemia requiring administration of furosemide (Lasix®) or other diuretic or pulmonary edema on chest x-ray||Chest x-ray|
|Delirium—documented acute decline in cognition||Medical record|
|Respiratory failure||Respiratory failure (AHRQ, 2009) recorded as secondary diagnosis (ICD-9-CM 518.81/518.84)||Discharge abstract|
Patient Demographic and Procedural Variables
|Cases (N = 28)||Controls (N = 28)|
|Variable||n (%)||n (%)||Statistical Test||p Value|
|Age in years (mean, SD)||71.96 (5.27)||72.57 (5.20)||t = 0.434||0.333|
|ASA score (mean, SD)||3 (0.459)||3 (0.459)||χ2 = 0.000||0.500|
| 2||2 (7)||2 (7)|
| 3||24 (86)||24 (86)|
| 4||2 (7)||2 (7)|
|Type of surgery||χ2 = 0.411||0.469|
|Abdominal||8 (29)||9 (32)|
|Esophageal||5 (18)||5 (18)|
|Orthopedic||8 (29)||6 (21)|
|Other||7 (25)||8 (28)|
|Sex||χ2 = 0.299||0.392|
| Male||16 (57)||18 (64)|
| Female||12 (43)||10 (36)|
| Race||χ2 = 3.798||0.289|
| White||18 (64)||20 (71)|
| Black||8 (29)||5 (18)|
| Other||2 (7)||3 (11)|
|Comorbidities (mean, SD)||13.64 (2.78)||10.15 (3.8)||t = 3.897||0.000|
| Pre-existing CADa||21 (75)||23 (82)||χ2= 0.891||0.251|
| Pre-existing lung diseaseb||5 (18)||11 (39)||χ2= 3.489||0.058|
| Pre-existing diabetes mellitus||4 (14)||6 (21)||χ2= 0.582||0.251|
| Documented obesity||5 (18)||13 (46)||χ2= 5.728||0.012|
|Type of anesthesia||χ2 = 2.166||0.353|
| General||22 (79)||21 (75)|
| Other (epidural, regional, local)||6 (21)||7 (25)|
|Total anesthesia time (mean, SD) (minutes)||619.04 (266.69)||450.86 (233.28)||t = 2.449||0.009|
|Return to OR||17 (61)||1 (4)||χ2 = 20.959||0.000|
Unit and Nursing Care Variables
|Case (N = 28)||Control (N = 28)|
|Variable||n (%)||n (%)||Statistical Test||p Value|
|Type of unit admitted post PACU|
| Floor||6 (21)||12 (43)||χ2 = 3.518||0.086|
| Step-down||11 (39)||10 (36)|
| ICU||11 (39)||6 (21)|
|Number of units resided (mean, SD)||3.29 (1.462)||1.68 (0.819)||t = −5.075||0.000|
|Number of transfers (mean, SD)||4 (2.653)||0.96 (1.293)||t = −5.445||0.000|
|Day patient ambulated (mean, SD)||8.62 (13.29)||1.67 (1.167)||t = −2.556||0.007|
|PCA use||20 (71)||19 (68)||χ2 = 0.007||0.500|
|Naloxone (Narcan®) use||3 (11)||1 (4)||χ2 = 1.002||0.306|
|Type of unit event occurred on|
| Floor||9 (32)||16 (57)|
| Step-down||13 (46)||11 (39)||χ2 = 5.698||0.029|
| ICU||6 (21)||1 (4)|
|Rapid response team called||6 (21)||0||χ2 = 6.72||0.012|
|Code called||6 (21)||0||χ2 = 6.72||0.012|
|Transferred to higher care level||23 (82)||1 (4)||χ2 = 37.225||0.000|
Trigger Events and Patient Outcomes
|Cases (N = 28)||Controls (N = 28)|
|Variable||n (%)||n (%)||Statistical Test||p Value|
|Oversedation||3 (10.7)||1 (3.6)||χ2 = 0.229||0.315|
|Atelectasis||24 (86)||15 (54)||χ2 = 6.842||0.019|
|Aspiration||11 (39)||0||χ2 = 13.689||0.000|
|Fluid overloadb||16 (57)||4 (14)||χ2 = 11.200||0.001|
|Delirium||10 (36)||2 (7)||χ2 = 6.788||0.010|
|Pneumonia||13 (46)||0||χ2 = 16.93||0.000|
|Length of stay (days)||28.76 (21.72)||7.47 (4.69)||t = −4.505||0.000|
| Home||5 (18)||21 (75)||χ2 = 18.564||0.000|
| LTC, RC||14 (50)||5 (18)|
|Death||9 (32)||2 (7)||χ2 = 2.500||0.057|