Postoperative complications: Does intensive care unit staff nursing make a difference?

ISSUES IN ADMINISTRATION

Postoperative complications: Does intensive care unit staff nursing make a difference? Deborah Dang, RN, MS,a,b Mary E. Johantgen, PhD,b Peter J. Pronovost, MD,c,d,e Mollie W. Jenckes, MHSc,f and Eric B. Bass, MD, MPH,e,f Baltimore, Maryland

OBJECTIVE: The purpose of this study was to examine the association between intensive care unit nurse (ICU) staffing and the likelihood of complications for patients undergoing abdominal aortic surgery. DESIGN: The study is a retrospective review of hospital discharge data linked to data on ICU organizational characteristics. SETTING: Research took place in ICUs in non-federal, short-stay hospitals in Maryland. PATIENTS: Study included 2606 patients undergoing abdominal aortic surgery in Maryland between January 1994 and December 1996. OUTCOME MEASURES: Outcome measures included cardiac, respiratory, and other complications. RESULTS: Cardiac complications occurred in 13% of patients, respiratory complications occurred in 30%, and other complications occurred in 8% of patients. Multiple logistic regression revealed a statistically significant increased likelihood of respiratory complications (odds ratio [OR], 2.33; 95% confidence interval [CI], 1.50-3.60) in abdominal aortic surgery patients cared for in ICUs with lowversus high-intensity nurse staffing, an increased likelihood of cardiac complications (OR, 1.78; CI, 1.16-2.72) and other complications (OR, 1.74; CI, 1.15-2.63) in ICUs with medium- versus highintensity nurse staffing, after controlling for patient and organizational characteristics. CONCLUSIONS: Within the range of ICU nurse staffing levels present in Maryland hospitals, decreased nurse staffing was significantly associated with an increased risk of complications in patients undergoing abdominal aortic surgery. (Heart Lung® 2002;31:219-28.)

T

he current and projected nursing shortage has captured international attention of legislative bodies and professional and health care organizations. Strategies to manage the spiraling costs of health care in the United States have had significant downstream effects on the systems and

From aThe Johns Hopkins Hospital Department of Nursing, the b University of Maryland Graduate School of Nursing, cThe Johns Hopkins University School of Medicine Department of Anesthesiology/Critical Care and dDepartment of Surgery, eThe Johns Hopkins University School of Hygiene and Public Health Department of Health Policy and Management, and fThe Johns Hopkins University School of Medicine Department of Medicine, Baltimore, Maryland. Reprint requests: Deborah Dang, MS, RN, 600 N Wolfe St, Baltimore, MD 21287-1007. Copyright 2002, Mosby, Inc. All rights reserved. 0147-9563/2002/$35.00 ⫹ 0 2/1/122838 doi:10.1067/mhl.2002.122838

HEART & LUNG

VOL. 31, NO. 3

providers of health care, culminating in the present and predicted shortage of nurses. Hospitals, traditionally at the core of the health care system, have implemented a wide range of initiatives to maintain economic viability. Although these strategies have the potential to reduce costs and improve efficiencies, most have sought to reduce nurse staffing because registered nurses (RN) represent the single largest labor cost for hospitals.1,2 Because nursing services are central to the provision of hospital care, changes in levels of nurse staffing have profound implications for adversely affecting the quality of care and patient outcomes at a time when hospitals are seeing fewer but sicker patients. Moreover, many of these changes have been undertaken in the absence of empirical evidence of their effectiveness. This was confirmed in 1996 by the Institute of Med-

www.mosby.com/hrtlng

219

Postoperative complications

icine in its report on the adequacy of nurse staffing in hospitals and nursing homes.3 The Institute of Medicine committee concluded that “there is a serious paucity of recent research on the definitive effects of structural measures, such as specific staffing ratios, on the quality of patient care in terms of patient outcomes.”3 Although intensive care units (ICU) were not cited specifically, they should be studied because risk-adjusted outcomes vary widely among ICUs, ICUs represent units with high RN intensity, and care provided in these units represents a significant portion of health care costs.4,5 Previous studies conducted in the early 1990s included the proportion of RNs as an independent variable in examining patient outcomes.6-9 Typically, the outcome variables, as measures of quality of care, were mortality and length of stay. These studies suggested that a higher proportion of RN caregivers was associated with lower in-hospital mortality and shorter lengths of stay. However, hospital mortality may have low sensitivity for evaluating the full effect of nurse staffing because mortality is a relatively rare event and other organizational characteristics of hospitals are associated with mortality.10 Recent studies have sought to determine the effects of differences in nurse staffing on the risk of complications because they may be more sensitive to RN ratios.11-14 This is particularly important in ICUs, where the foundation of critical care nursing practice is early detection and prompt intervention in critically ill patients. Although some studies provide evidence of an association between the occurrence of complications and nurse staffing levels,11-13 few have examined this relationship in the ICU setting at the unit level where the effect of nurse staffing is most direct and more sensitive to the effects of variation in staffing level.11 One study examined the relationship between 6 adverse events thought to be sensitive to nursing care and the proportion of hours of care delivered by RNs in 42 nursing units in 1 hospital, 4 of which were critical care units.11 The ICU units had lower fall rates than units with a lower proportion of RNs, but higher rates of decubitus, infections, complaints, and medication errors. However the hospital-developed acuity system may not have been sensitive enough to adequately control for the higher acuity of the ICU patients. The relation of structural characteristics to patient outcomes in 38 ICUs in Maryland was previously examined.14 The study demonstrated that patients who had undergone abdominal aortic surgery and who were cared for in ICUs with lower nurse staffing was associated with a significant increase in ICU length of stay. A ratio of

220

www.mosby.com/hrtlng

Dang et al

1 nurse to 3 or more patients on the day shift was also associated with an increased likelihood of respiratory-related complications and contributed to an increased length of stay.15 In this analysis, only day shift staffing was examined in relation to complications. The present study extended the above analysis by including nurse staffing on all shifts in 38 ICUs while controlling for nursing unit structure variables that were not examined in previous analyses. The objective was to isolate the effects of nurse staffing on patient outcomes by examining the association between ICU nurse staffing and the likelihood of medical complications for patients undergoing abdominal aortic surgery, controlling for patient and other hospital characteristics.

METHODS This study represents a secondary analysis of hospital discharge data that were linked to a survey of organizational characteristics of ICUs. The target population of patients who had undergone abdominal aortic surgery was chosen because this surgery is a relatively common procedure performed in a variety of acute-care hospitals and typically requires care in an ICU after surgery.

Sample After approval by the institutional review board, the study population was drawn from inpatient hospital stays for all patients undergoing abdominal aortic surgery in Maryland between January 1994 and December 1996 (N ⫽ 2987). Data were obtained from the Uniform Health Discharge Data Set maintained by the Maryland Health Services Cost Review Commission (HSCRC). This is a nonconfidential data set containing information on all patients discharged from the 52 non-federal acute care hospitals in Maryland and includes patient-level clinical, demographic, and resource use information. Variables selected for this study included age, sex, race, nature of admission, vital status at discharge, hospital length of stay, ICU days, and codes from the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) for the primary discharge diagnosis, the principle procedure, as many as 14 secondary discharge diagnoses, and as many as 14 secondary procedures.16 The study population included patients age 30 years or older who were discharged from a Maryland hospital and who had a principal procedure code for abdominal aortic surgery (ICD-9-CM code 3844 for resection of abdominal aorta with replacement and ICD-9-CM code 3925 for aorto-iliac-femoral bypass).

MAY/JUNE 2002

HEART & LUNG

Dang et al

Postoperative complications

Table I Medical complications for abdominal aortic surgery patients Medical complications

Cardiac Acute myocardial infarction Cardiac arrest Cardiac complications after a procedure Any cardiac complication Respiratory Aspiration Pneumonia Pulmonary insufficiency after a procedure Mechanical ventilation ⬎96 h Tracheal reintubation Any respiratory complication Other Acute renal failure Septicemia Platelet transfusion Any other complication Any complication

ICD-9-CM codes

Number (%)

410* 4275 9971

72 (3) 33 (1) 291 (11) 341 (13)

507*, 9973 480*-487* 5184, 5185, 5188 9672 9604

372 (14) 137 (5) 307 (12) 184 (7) 379 (15) 787 (30)

584* 038* 9905

122 (5) 91 (4) 44 (2) 221 (8) 1032 (40)

ICD-9-CM; International Classification of Diseases, 9th Revision, Clinical Modification. *Includes all 4th and 5th digit codes.

Only 9 patients younger than 30 years of age were excluded because they had suffered an injury to a blood vessel (ICD-9-CM code 902).

Outcome measures A panel of 4 critical-care physicians independently identified ICD-9-CM procedure codes from the HSCRC database that were likely to represent important complications of abdominal aortic surgery. For this analysis, the focus was on medical complications that were independently associated with in-hospital mortality in the previous analysis14 and complications that have been shown to be nurse sensitive.11-13,17 The complications and associated ICD-9-CM codes are given in Table I. Even though discharge diagnosis codes do not distinguish between complications and comorbidities, the complications selected reflect acute problems that are likely to represent complications for this patient population. Complications were collapsed into 3 groups on the basis of their physiologic relationship (cardiac and respiratory) and to achieve a minimum number of independent variables (other complications). Cardiac complications included acute myocardial infarction (MI), cardiac arrest, and cardiac

HEART & LUNG

VOL. 31, NO. 3

complications after a procedure. Respiratory complications included aspiration, pneumonia, pulmonary insufficiency after a procedure, mechanical ventilation longer than 96 hours, and tracheal reintubation. Other complications included acute renal failure, septicemia, and platelet transfusion.

Nurse staffing Data was obtained on nurse staffing from the survey on ICU organizational characteristics. The survey was mailed to physician ICU directors who were asked to complete the survey in consultation with the nurse manager. Two questions measured nurse staffing on the day and night shifts: “what is the average nurse-to-patient ratio in the ICU during the daytime?” and “what is the average nurse-topatient ratio during the nighttime?” Possible responses were 1:1 to 1:2, 1:3 to 1:4, and greater than 1:4. The 2 nurse staffing variables were collapsed into a new variable coded as low-intensity staffing (1:3 or greater on the day and night shifts), mediumintensity staffing (1:3 or greater on either the day or night shift, but not both), and high-intensity staffing (1:2 or fewer on the day and night shifts).

www.mosby.com/hrtlng

221

Postoperative complications

Risk adjustment Adjustment for risk factors is critical in assessing outcomes of care in that it helps to account for patient and organizational characteristics that may increase the likelihood of poor outcomes. Adjustment for patient characteristics included demographics such as age, sex, and race. The RomanoCharlson comorbidity index was used to identify potentially important comorbid diseases.18-20 To adjust for severity of illness, patients were classified as having a ruptured or unruptured aorta (ICD-9-CM code 441.3), and the nature of admission field coded at admission was used to identify each case as elective, urgent, or emergent.21 Organizational characteristics were evaluated for their association with the 3 complication groups. These included hospital and ICU bed size, hospital and surgeon volume of abdominal aortic surgery, type of unit, full-time medical director and nurse manager, RN attendance at daily rounds, and use of written protocols and critical paths for abdominal aortic surgery patients.

Data collection This analysis used data on organizational characteristics from the previous survey.14 The survey included 32 items that were derived from an instrument to assess the organization of ICUs.22 Content areas included items related to ICU nurse and physician staffing, technology, and care processes. Five intensive care physicians independently reviewed the survey to ensure that the items reflected the intended domains. Inter-rater reliability, established through the use of 2 sets of intensivists, resulted in 100% and 97% agreement, respectively. During the original study period, 46 hospitals in the Maryland HSCRC database performed abdominal aortic surgery. The survey was mailed to physician ICU directors at these hospitals. After the initial letter explaining the survey, reminder letters were sent at 2 weeks, a second survey was mailed at 4 weeks, and telephone contacts were made at 6 weeks after the initial survey was mailed.

Statistical analysis A descriptive analysis was performed for hospital and ICU characteristics, patient characteristics, and complications. For ease of interpretation and because several variables had skewed distributions, race, hospital, and surgeon volume were dichotomously coded. Race was coded as white or other. Cut points for hospital and surgeon volume were obtained by examining the relationship between volume and mortality with a LOWESS smoothing

222

www.mosby.com/hrtlng

Dang et al

curve.23 Low volume was defined as less than 36 cases per year for hospitals and less than 8 cases per year for surgeons. Subsequently, a bivariate analysis was performed to evaluate the association among the independent variables, and between each independent variable and each type of complication. The following independent variables were highly correlated with each other: hospital and surgeon volume, hospital and ICU beds, and critical paths and protocols. To deal with the co-linear nature of these variables, variables in each pair that had the largest variance were excluded from the multivariate analysis. Multiple logistic regression assessed the contribution of conceptually important or significant independent variables to predicting the likelihood of cardiac, respiratory, or other complications. Each regression model contained a set of independent variables used to adjust for patient characteristics of age, race, sex, illness severity, and comorbid diseases. The final adjustment model for organizational characteristics included the number of cases of abdominal aortic surgery performed per year at the hospitals, the number of staffed ICU beds, and the use of critical paths for this patient population. Multilevel hierarchical modeling was used to adjust for the clustering of outcomes among patients within an ICU and for an ICU within a hospital so that an extreme effect in one ICU, for example, did not unduly influence the results.24 Although critical paths were not significant in the bivariate model and represented only 165 patients at 2 hospitals, this variable was retained as a result of its possible influence on variations in practice. Model fit was evaluated with the Hosmer-Lemeshow test.25 A multiple logistic regression analysis was conducted for each of the complications and for each of the 3 groups of complications: cardiac complications, respiratory complications, and other complications. A patient could have had 1 or more complications and therefore could be represented in any or all of the 3 major outcome variables.

RESULTS Hospital sample As summarized in Table II, 38 hospitals (83%) responded to the survey. The 7 nonresponding hospitals represented 381 patients (15%). However, there were no clinically meaningful differences between responding and nonresponding hospitals for patient or hospital characteristics, and these data were excluded from the analysis. The majority of hospitals had less than 400 beds

MAY/JUNE 2002

HEART & LUNG

Dang et al

Postoperative complications

Table II Characteristics of Maryland hospitals with ICUs caring for patients who have had abdominal aortic surgery, January 1994-December 1996 (N ⫽ 38) Characteristics

Number (%)

Hospital Bed size ⬍240 241-400 ⬎400 Hospital abdominal aortic surgery volume (mean cases/year) ⬎36 ⬍36 Surgeon abdominal aortic surgery volume (mean cases/year) ⬎8 ⬍8 ICU Staffed beds ⬍10 10-15 16-20 ⬎20 Type of patients Medical/surgical Surgical only Nurse staffing Low intensity Mixed intensity High intensity Medical director Full time Other Nurse Manager Full time Other Nurses attend daily rounds Yes No Written clinical protocols None ⬍25% 25%-50% ⬎50% Critical path for abdominal aortic surgery patients Yes No

14 (37) 17 (45) 7 (18) 6 (16) 32 (84) 14 (37) 24 (63) 11 (31) 13 (36) 9 (25) 3 (8) 33 (87) 5 (13) 6 (16) 8 (21) 24 (63) 16 (42) 22 (58) 30 (79) 8 (21) 18 (51) 17 (49) 14 (40) 5 (14) 4 (11) 12 (34) 2 (6) 32 (94)

ICU, Intensive Care Units

and performed less than 36 cases of abdominal aortic surgery per year. Most ICUs were staffed for 15 beds or less, had high-intensity nurse staffing, cared for a combination of medical and surgical

HEART & LUNG

VOL. 31, NO. 3

patients, and had full-time medical directors and nurse managers. Nurses’ attendance at daily rounds was evenly divided, and most ICUs used protocols. However only 6% had patients who had

www.mosby.com/hrtlng

223

Postoperative complications

Dang et al

Table III Characteristics of discharges undergoing abdominal aortic surgery in Maryland, January 1994-December 1996 (N ⫽ 2606) Characteristics

Number (%)

Age (years) ⬍60 60-70 71-85 ⬎85 Race White Black Other Gender Male Female Nature of admission Elective Urgent Emergent Type of aneurysm Unruptured Ruptured Comorbid diseases Chronic renal disease Chronic obstructive pulmonary disease Dementia Malignancy Mild diabetes mellitus Severe diabetes mellitus Mild liver disease Severe liver disease Old myocardial infarction Hospital length of stay ⬍12 12-20 ⬎20 Intensive care unit length of stay ⬍3 3-10 ⬎10

undergone abdominal aortic surgery on critical paths.

Patient sample Patient characteristics are summarized in Table III. The majority of patients were white men between 71 and 85 years of age, who were electively admitted with an unruptured aneurysm. Forty-seven percent of the patients had at least 1 of the comorbid diseases that are included in the Romano-Charleson

224

www.mosby.com/hrtlng

497 (19) 847 (33) 1185 (46) 77 (3) 2314 (89) 275 (11) 9 (0.3) 1779 (68) 827 (32) 1733 (67) 327 (13) 529 (20) 2382 (91) 223 (9) 98 (4) 267 (10) 17 (0.7) 108 (4) 310 (12) 84 (3) 19 (0.7) 1 (0) 324 (12) 1920 (74) 435 (17) 251 (10) 1393 (54) 1038 (40) 175 (7)

comorbidity index; the most frequent were old MI, mild diabetes mellitus, and chronic obstructive pulmonary disease. Seventy-four percent of the patients had a hospital length of stay less than 12 days (mean, 11 days), and the majority had an ICU stay of 10 days or less (mean, 4 days). The frequency of complications are presented in Table I. Approximately 40% of the patients had at least 1 complication, with respiratory complications being the most frequent (30%).

MAY/JUNE 2002

HEART & LUNG

Dang et al

Postoperative complications

Table IV Adjusted likelihood of postoperative complications for abdominal aortic surgery patients (N ⫽ 2606) Complications

Characteristics Structure

Nurse Staffing High intensity* Medium intensity Low intensity ICU staffed beds ⬍10* 10-15 16-20 ⬎20 Cases of abdominal aortic surgery/year ⬎36* ⬍36 Critical path for abdominal aortic surgery patients Yes* No Clinical Any comorbid disease† Patient Age† Race White* Other Gender Male* Female Severity Nature of admission Elective* Urgent Emergent Type of aneurysm Unruptured* Ruptured Hosmer-Lemeshow statistic‡

n

Cardiac (n ⴝ 341) Odds ratio (95%CI)

Respiratory (n ⴝ 787) Odds ratio (95%CI)

Other (n ⴝ 221) Odds ratio (95%CI)

1600 586 420

1.00 1.78 (1.16-2.72) 1.34 (0.82-2.17)

1.00 1.03 (0.78-1.38) 2.33 (1.50-3.60)

1.0 1.74 (1.15-2.63) 1.13 (0.73-1.75)

650 944 759 164

1.00 1.26 (0.82-1.94) 1.47 (0.72-3.00) 1.57 (1.01-2.44)

1.00 1.01 (0.72-1.41) 0.93 (0.54-1.60) 0.93 (0.21-4.19)

1.00 1.63 (1.10-2.43) 1.10 (0.59-2.05) 0.96 (0.47-1.95)

940 1666

1.00 0.70 (0.42-1.15)

1.00 1.82 (1.25-2.67)

1.00 0.87 (0.54-1.42)

165 2196

1.00 0.88 (0.56-1.40)

1.00 0.97 (0.21-4.47)

1.00 0.97 (0.58-1.62)

2036

1.04 (0.83-1.30)

1.14 (0.96-1.36)

1.23 (0.950-1.60)

2605

1.04 (1.02-1.05)

1.02 (1.01-1.03)

1.03 (1.01-1.04)

2314 284

1.00 1.05 (0.70-1.58)

1.00 1.02 (0.78-1.34)

1.00 1.23 (0.66-2.28)

1779 827

1.00 0.96 (0.78-1.17)

1.00 1.05 (0.86-1.29)

1733 327 528

1.00 1.09 (0.81-1.47) 1.22 (0.85-1.76)

1.00 1.15 (0.77-1.70) 1.31 (0.96-1.77)

2382 224

1.00 1.51 (0.98-2.35) ␹2 ⫽ 9.70, P ⫽ .29

1.00 97 (0.67-1.39) 1.0 1.88 (1.17-3.00) 1.82 (1.20-2.74)

1.00 1.00 5.26 (3.34-8.27) 5.29 (3.14-8.90) ␹2 ⫽ 12.15, P ⫽ .14 ␹2 ⫽ 7.40, P ⫽ .49

*Referent category. †Modeled as a continuous variable. ‡Hosmer-Lemeshow statistic evaluates the goodness-of-fit of regression models in predicting values on outcome variables that are consistent with the observed data. A nonsignificant P value indicates a good model fit.

Complications Table IV summarizes the results of the 3 logistic regression models. The Hosmer-Lemeshow

HEART & LUNG

VOL. 31, NO. 3

test reflected an adequate model fit for all models. After controlling for patient and organizational characteristics, nurse staffing showed sig-

www.mosby.com/hrtlng

225

Postoperative complications

nificant independent effects. Patients cared for on units with medium-intensity staffing were more likely to have cardiac complications (odds ratio [OR], 1.78; confidence interval [CI], 1.16-2.72) and other complications (OR, 1.74; CI, 1.15-2.63) than patients cared for on units with high-intensity staffing. Patients cared for on units with lowintensity staffing were more than twice as likely to have respiratory complications (OR, 2.33; CI, 1.503.60) than patients on units with high-intensity staffing. Regression analyses modeling individual complications (vs modeling the complication groups) were conducted to additionally understand the significant effects within cardiac, respiratory, and other complications. Patients cared for on units with medium-intensity staffing were more than twice as likely to develop cardiac complications after a procedure (OR, 2.10; CI, 1.26-3.50) than patients on units with high-intensity staffing. Within the respiratory complication group, patients were more than 5 times as likely to develop pulmonary insufficiency after surgery (OR, 5.11; CI, 2.89-9.04), and more than twice as likely to be mechanically ventilated after 96 hours (OR, 2.39; CI, 1.55-3.69) and reintubated (OR, 2.09; CI, 1.47-3.03) when cared for on units with low-intensity staffing compared with units with high-intensity staffing. Although ICU bed status was significant for cardiac and other complications, the pattern of significance was inconsistent and not easily interpretable. Interestingly, volume of cases was significantly related only to respiratory complications. Patients cared for in hospitals that performed less than 36 cases per year were 82% more likely to have a respiratory complication (OR, 1.82; CI, 1.25-2.67) than patients cared for in hospitals that did more than 36 cases per year. Modeling the individual complications, this effect was driven by pulmonary insufficiency after surgery (OR, 2.69; CI, 1.68-4.30) and reintubation (OR, 3.29; CI, 1.69-6.40). The significance of age, nature of admission, and type of aneurysm indicated moderate performance of the risk adjustment model and the importance of adjusting for patient risk. Patients admitted urgently or emergently were more than 80% as likely to have other complications compared with patients admitted electively. Patients admitted with a ruptured aneurysm were more than 5 times as likely to have a respiratory or other complication than patients without a ruptured aneurysm.

226

www.mosby.com/hrtlng

Dang et al

DISCUSSION This study assessed the association between ICU nurse staffing and the likelihood of cardiac, respiratory, and other complications in patients who had had abdominal aortic surgery. The intensity of nurse staffing was significantly associated with all 3 groups of complications. This study extends other recent studies examining nurse staffing and complications11-14 because it measured staffing at the unit level, included only RNs who were providing direct care, and included multiple facilities. Furthermore, the patient sample size was large, representing 2606 patients with a specific surgical procedure, 40% of whom had one or more complications. The significantly increased likelihood of cardiac, respiratory, and other complications in ICUs may reflect a difference in the level of monitoring by nurses or possibly an insufficient number of nurses to perform interventions such as pulmonary hygiene, an aspect of care for which nurses are responsible. The lack of association between complications and use of critical paths may indicate that critical paths have more of an effect on decreasing length of stay than on improving quality of care. As expected, the majority of patients (82%) were more likely to have respiratory complications if they were cared for in hospitals performing less than 36 cases of abdominal aortic surgery per year. The literature strongly supports a volume outcome relationship in the surgical population.26,27 More interesting was the finding that 64% of the patients were cared for in low-volume hospitals. Given that hospital volume is related to complications and that most patients have surgery at low-volume hospitals with higher complication rates, hospital referral strategies (such as those proposed by the Leapfrog Group28) or strategies to reduce complications at low-volume hospitals are imperative. Efforts to reduce the incidence of these complications could be the focus of quality improvement initiatives. Given these data, ICU nurses should take a prominent role in these efforts.

LIMITATIONS Although the sample of 38 ICUs was limited to 1 state, there is no indication that ICUs in this state differ from those in other states with regards to organizational characteristics.5 Although the response rate for the survey was 83%, the nurse manager may or may not have been involved in survey completion, and responses may represent perceptions or experiences over time. Specifically, the responses to items regarding nurse staffing were likely

MAY/JUNE 2002

HEART & LUNG

Dang et al

Postoperative complications

to represent the experiences of the ICU nurse managers or physician directors with the patterns of staffing over time. In ICUs there is likely to be less variation in nurse staffing compared with general care units. Therefore the nurse manager’s or physician director’s experiences can provide a reasonable approximation of the actual nurse-to-patient ratios. The use of secondary data analysis limited the selection of the patient outcomes measured in this study. The complications selected were those that were likely to be influenced by nursing interventions and the level of nursing surveillance, recognizing that nurses are not the only members of the health care team who provide care to patients. However the results are consistent with other studies that have found that the proportion of RNs is related to the occurrence of pneumonia, pulmonary compromise, and reintubation after surgery.12,13,15 Furthermore, in an ICU setting, nurses have a primary role in patient surveillance and care coordination within the health care team, and the effect of their presence and subsequent interventions on the risk of complications may be more pronounced. Several limitations with the use of discharges from administrative databases must be acknowledged. One is the reliability of coding of comorbid diseases and complications. Considering that a panel of critical-care physicians independently selected the complications, the bias in selecting complications likely to be seen in ICUs for this type of surgery should be minimal. Moreover, an analysis of the quality of the Maryland HSCRC database indicates a low error rate and no systematic differences in coding for comorbid diseases and the types of outcome variables used in this study.29 Another concern with the use of administrative databases is the inability to determine whether the complication occurred in the ICU or on the general surgical unit. Nevertheless, the complications selected for this study are likely to be seen in ICUs or to occur shortly after an ICU stay and would most likely reflect ICU care.

IMPLICATIONS FOR PRACTICE AND RESEARCH This study contributes to the growing evidence supporting the relationship between structural aspects of nursing care units and patient outcomes. Specifically, this study and others demonstrate that a lower proportion of RNs has consistently been associated with the development of respiratory complications.11-15 Nurse managers and administrators may be able to use this evidence to advocate

HEART & LUNG

VOL. 31, NO. 3

for increased non-RN support, such as the use of respiratory therapists to moderate the risk of patient’s developing respiratory complications in the ICU setting, and to inform their decision making related to allocation of scarce nurse resources. State and federal policy makers should take results such as these into account when weighing the costs, benefits, and risks of mandating minimum staffing levels. Decision making in the absence of systematic study of how variations in staffing levels affect patient outcomes may have unintended consequences for the quality of inpatient care. Future research is needed to investigate whether the relationship between nurse staffing has a direct influence on patient outcomes, as has been proposed in the Quality Heath Outcomes Model30 or if this relationship occurs through other process or contextual aspects of nursing units, such as the organization of nursing services and the professional practice environment.31 Studies employing more sensitive measures of nurse staffing such as nursing hours per patient day, obtained at the unit level, will strengthen the validity of such work and control for intrahospital and interhospital variations. The development of standardized databases at the state and national level will be important to establishing causal links between structural variables and patient outcomes in all settings where nurses deliver care. The pressures on hospitals to contain health care costs, the current and predicted nursing shortage, and the recent emphasis on patient safety and nurses’ role in preventing avoidable consequences to patients have made research in this area a high priority.

REFERENCES 1. Aiken LH, Fagin CM. Evaluating the consequences of hospital restructuring. Med Car 1997;35:OS1-OS4. 2. American Hospital Association. Hospital statistics. Chicago (IL): American Hospital Association; 1999. 3. Wunderlich GS, Sloan FA, Davis CK, editors. Nurse staffing in hospitals and nursing homes: is it adequate? Washington, DC: National Academy Press; 1996. 4. Shortell SM, Zimmerman JE, Rousseau DM, Gillies RR, Wagner DP, Draper EA, et al. The performance of intensive care units: does good management make a difference? Med Care 1994;32:508-25. 5. Groeger JS, Strosberg MA, Halpern NA, Russell CR, Kaye WE, Kalpalatha KG, et al. Descriptive analysis of critical care units in the United States. Crit Care Med 1992;20:846-83. 6. Aiken LH, Smith HL, Lake ET. Lower medicare mortality among a set of hospitals known for good nursing care. Med Care 1994;32:771-87. 7. Flood SD, Diers D. Nurse staffing, patient outcome, and cost. Nurs Manage 1988;19:34-43. 8. Hartz AJ, Krakauer H, Kuhn EM, Young M, Jacobsen SJ, Gay G, et al. Hospital characteristics and mortality rates. N Engl J Med 1989;321:1720-5.

www.mosby.com/hrtlng

227

Postoperative complications

Dang et al

9. Manhiem LM, Feinglass J, Shortell SM, Hughes EFX. Regional variation in medicare hospital mortality. Inquiry 1992; 29:55-66. 10. Mitchell PH, Shortell SM. Adverse outcomes and variation in organization of care delivery. Med Care 1997;35:NS19-NS32. 11. Blegen MA, Goode CJ, Reed L. Nurse staffing and patient outcomes. Nurs Res 1998;47:43-50. 12. Needleman J, Buerhaus PI, Mattke S, Stewart M, Zelevinsky K. Nurse staffing and patient outcomes in hospitals: final report. US Department of Health and Human Services; 2001. 13. Kovner C, Gergen PJ. Nurse staffing and adverse events following surgery in US hospitals. Image J Nurs Sch 1998;30: 315-21. 14. Pronovost PJ, Jenckes M, Dorman T, Garrett E, Breslow MJ, Rsoenfeld BA, et al. Organizational characteristics of ICUs related to outcomes of abdominal aortic surgery. JAMA 1999; 281:1310-7. 15. Pronovost PJ, Dang D, Dorman T, Lipsett PA, Garrett E, Jenckes M, et al. Intensive care nurse staffing associated with increased risk of complications after abdominal aortic surgery. Eff Clin Pract 2001;4:199-206. 16. Health Care Financing Administration. International Classification of Diseases, 9th Revision, Clinical Modification, Volumes 1-3. DHHS Publication No. (PHS) 80-1260. 17. Lichtig LK, Knauf RA, Milholland DK. Some impacts of nursing on acute care hospital outcomes. J Nurs Adm 1999;29: 25-33. 18. Romano PS, Roos LL, Jollis JG. Adapting a clinical comorbidity index for use with ICD-9-CM administrative data: differing perspectives. J Clin Epidemiol 1993;46:1075-9. 19. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method for classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987; 40:373-83.

20. Roos LL, Fisher ES, Sharp SM, Newhouse JP, Anderson G, Bubolz TA. Postsurgical mortality in Manitoba and New England. JAMA 1990;263:2453-8. 21. Dardik A, Burleyson GP, Bowman H, Gordan TA, Williams MA, Webb TH, et al. Surgical repair of ruptured abdominal aortic aneurysms in the state of Maryland: factors influencing outcome among 527 recent cases. J Vasc Sur 1998;28:413-21. 22. Shortell, SM, Rousseau DM, Gillies RR, Devers KJ, Simons TL. Organizational assessment of intensive care units (ICUs): construct development, reliability, and validity of the ICU nurse-physician questionnaire. Med Care 1991;29:709-26. 23. Gould WW. Linear splines and piecewise linear functions. J Am Stat Assoc 1979;74:829-36. 24. Paterson L, Goldstein H. New statistical methods for analysing social structures: an introduction to multilevel models. Br Edu Res J 1991;17:387-93. 25. Hosmer DW, Lemeshow S. Applied logistic regression. 2nd ed. New York: John Wiley & Sons; 2000. 26. Luft HS, Bunker JP, Enthoven AC. Should operations be regionalized? The empirical relationship between surgical volume and mortality. N Engl J Med 1979;301:1364-9. 27. Flood AB, Scott R, Ewy W. Does practice make perfect? The relation between hospital volume and outcomes for selected diagnostic categories. Med Care 1984;22:98-125. 28. The Leapfrog Group. Evidence-based hospital referral (EHR). Available at Leapfroggroup.org/FactSheet/EHR_FactSheet.PDF; Accessed 2000. 29. The Johns Hopkins Hospital report of clinical data quality review. Baltimore (MD): Deloitte, Haskins, & Sells; 1996. 30. Mitchell PH, Ferketich S, Jennings BM. Quality heath outcomes model. Image J Nurs Sch 1998;30:43-6. 31. Aiken LA, Sochalski J, Lake ET. Studying outcomes of organizational change in health services. Med Care 1997;35: NS6-NS18.

RECEIVE TABLES OF CONTENTS BY E-MAIL To receive the tables of contents by e-mail, sign up through our Web site at http://www.mosby.com/hrtlng Choose E-mail Noficication. Simply type your e-mail address in the box and click the Subscribe button. Alternatively, you may send an e-mail message to [email protected]. Leave the subject line blank and type the following as the body of your message: subscribe hrtlng_toc. You will receive an e-mail to confirm that you have been added to the mailing list. Note that table of contents e-mails will be sent out when a new issue is posted to the Web site.

228

www.mosby.com/hrtlng

MAY/JUNE 2002

HEART & LUNG