Abbreviations and Acronyms: ASA = American Society of Anesthesiologists; CHFO = continuous high-frequency oscillation; CPAP = continuous positive airway pressure; CPEP = continuous positive expiratory pressure; LOS = length of stay; OLE = oscillation and lung expansion; PPC = postoperative pulmonary complication; PSHA = post-surgical hospital admission.
Postoperative pulmonary complications (PPCs) result in significant morbidity, mortality, and healthcare costs. The incidence of “important” PPCs is 2% to 5% in the general surgical population. These consist of respiratory infection, respiratory failure, atelectasis, prolonged invasive mechanical ventilation, prolonged high-level respiratory support, and readmission to the ICU for pulmonary complications. However, additional patient factors and types of surgical procedures can increase the risks for PPCs. Pulmonary complications are especially common after thoracic and upper abdominal operations. Other risk factors include advanced age, higher American Society of Anesthesiologists (ASA) physical status classification scores, functional dependency, congestive heart failure, COPD, smoking history, and BMI ≥30 kg/m2. Of these, ASA score ≥2 and functional dependency have consistently been associated with a higher rate of significant PPCs, such as postoperative respiratory failure. Occurrence of postoperative respiratory failure requiring prolonged mechanical ventilation or re-intubation accounts for poor outcomes and high economic costs.1-5
Although PPCs are an incompletely understood multifactorial occurrence, atelectasis is recognized as a critical component.6 Evidence suggests that atelectasis is a common precursor of PPCs. Nearly all patients undergoing major operations experience some degree of transitory, clinically unimportant atelectasis. Most do not require aggressive therapy. However, in subgroups of high-risk individuals, treatment of atelectasis and mobilization of secretions can help prevent deterioration to postoperative respiratory failure.7
A variety of strategies have demonstrated therapeutic benefits reducing the risks for postoperative atelectasis. These include optimal analgesia, judicious use of respiratory-depressing medications, and selective use of nasogastric decompression. In patients at higher risk for complications, secretion mobilization techniques and lung expansion therapies might be indicated.6-9 Lung expansion devices that provide either oscillating airflow or positive expiratory pressure therapy have been used for more than 25 years.10-32 Those using continuous positive airway pressure (CPAP) have shown therapeutic benefits in high-risk hypoxemic patients.33,34 In a meta-analysis, Qaseem and colleagues34 found that lung expansion therapies reduced PPCs by >50% compared with no treatment. Delivering oscillating airflow with positive expiratory pressure can facilitate mucous clearance, promote lung expansion, and enhance aerosol delivery. This combination of oscillation and lung expansion (OLE) could represent an important adjunct to help reduce PPCs.
Preliminary evidence suggests use of OLE can be effective in postoperative patients prone to atelectasis and PPCs.35 The primary objective of this study was to determine if treatment with OLE reduces rate of pulmonary complications in high-risk postoperative patients compared with standard treatment alone.
METHODS
Study design
IRBs at each site approved the protocol. Informed consent was waived for stage I/retrospective cohort and obtained for stage II/prospective patients. This was a non-randomized pre-post intervention study (Fig. 1). Three academic facilities that had not previously used OLE conducted the study in patients undergoing open thoracic, aortic, or upper abdominal operations.
;)
Figure 1.:
Study design.
For the stage I retrospective cohort, CPT, ICD-9, and ICD-10 codes for thoracic, upper abdominal, and aortic procedures were queried to screen for 761 appropriate patients. From this group, patients were then randomly selected based on predefined entry criteria. Data were collected for patients who met inclusion criteria, for open thoracic, upper abdominal and aortic procedures. These patients received incentive spirometry but no OLE therapy. Data from 210 randomly selected patients were collected (70 patients from each site). At all sites, additional respiratory treatment was provided based on patient's clinical indication and attending physicians' preference. Stage I included patients randomly selected from a year before stage II. This allowed inclusion of patients from a seasonally consistent and contemporaneous period. The purpose was to determine the incidence of PPCs in patients receiving standard care. This group provided the control group for comparison with the stage II practice change cohort.
Stage II began after implementation of a practice change. Eligible patients received therapy with OLE in addition to standard care. Demographic, clinical, and outcomes data for 209 patients were collected. Outcomes were tracked for each day from post-surgical hospital admission (PSHA) through postoperative day 7. Total ICU length of stay (LOS), hospital LOS, and length of time on mechanical ventilation were determined. Re-admission after 30 days was captured, a timeline showing data procedures and collection points is available in the eTable 1.
Study participants
Post-thoracic, post-upper abdominal, and post-aortic surgical patients at least 18 years old and deemed high-risk (as defined by ASA class ≥3 or ASA class 2 and 1 or more of the following: current smoker or smoking history within past 6 months; history of COPD; documented obesity and/or BMI ≥30 kg/m2; aged 75 years or older) were eligible. Surgical procedures were opened with incisions at or above the umbilicus. Patients were excluded for contraindication to OLE therapy (eg untreated tension pneumothorax), organ transplantation, spinal operation involving a posterior approach, minimally invasive or a “scopic” procedure, and if they were on positive pressure ventilation at baseline.
Study device
The device used in the study was The MetaNeb System (Hill-Rom). It consists of a pneumatic compressor that delivers OLE therapy using both continuous high-frequency oscillation (CHFO) and continuous positive expiratory pressure (CPEP). It is recommended for mobilization of secretions, lung expansion, and treatment and prevention of pulmonary atelectasis. The device also provides supplemental oxygen when used with compressed oxygen and can deliver aerosol therapy during CPEP and/or CHFO.
Treatment regimen
Only patients in stage II received treatment with OLE in addition to standard care. For intubated patients, treatment regimen consisted of in-line therapy with OLE with the ventilator 6 times per day (q4h) in CHFO mode only. Non-intubated patients received therapy with OLE 4 times per day via a mouthpiece or mask, alternating between CHFO and CPEP mode. The target duration of each treatment was 10 minutes, for a minimum of 48 hours. The CHFO and CPEP treatment settings and duration were adjusted based on patient tolerance, at the discretion of healthcare providers.
Outcomes measures
The primary end point for the study was rate of significant PPCs. A significant PPC was defined as occurrence of 1 or more of the following within 7 days of the PSHA: prolonged invasive mechanical ventilation (>24 hours from PSHA); prolonged high-level respiratory support oxygen requirement of >40% or >5 L/min via nasal cannula (and above patient's baseline), and/or requirement for non-invasive ventilation or CPAP above patient's baseline requirement for more than 24 hours of PSHA; diagnosis of pneumonia (pneumonia was defined based on CDC criteria consisting of new lung infiltrate plus new onset of fever, purulent sputum, leukocytosis, and decline in oxygenation; the diagnosis of pneumonia was performed by directional sputum specimen or bronchoalveolar lavage, depending on institutional standard of practice); and readmission to the ICU (or transfer to an elevated level of care) for pulmonary complications.
Secondary end points included patient requirement for invasive mechanical ventilation for more than 48 hours, requirement for prolonged high-level respiratory support for more than 48 hours, total time on mechanical ventilation, time to extubation, ICU LOS, hospital LOS, and readmissions to the hospital, for any cause, within 30 days of discharge.
Statistical analyses
The primary analysis was performed using intent-to-treat methodology. Descriptive statistics were calculated for demographics, medical history, and primary and secondary end points. Time to event data (eg time to readmission) was calculated using unadjusted Cochran-Mantel Haenszel for survival data. The 95% CI was computed for the primary end point, incidence of significant PPCs. Incidence rates for various demographic subsets and operation types were also computed. Stage I and stage II incidence rates were compared to identify any statistical associations that can be attributed to use of OLE. Data are reported as mean ± SD and adjusted analyses used analysis of covariance (adjustment for age, operation duration, and ASA score).
RESULTS
Patients in stage I and II were similar with respect to baseline demographics and medical history (Table 1). The distribution of the primary operations performed, based on CPT coding, was similar between the groups (Table 2). However, stage II patients tended to be older by a mean of 3.7 years (p < 0.05). The proportion of patients with an ASA score of 4 or 5 was higher in stage II (24.9%) compared with stage I (17.6%; p < 0.05). Preoperative respiratory status was similar between the 2 stages, although more stage II patients were on CPAP before the operation compared with stage I patients (Table 3). There were more thoracic and aortic surgical procedures in stage II compared with stage I (p < 0.0001), and mean operation duration was higher in stage II period (p = 0.07) (Table 3).
Table 1: Patient Demographics and Illness Characteristics by Stage
Table 2: Number of Surgical Procedures by System
Table 3: Patient Preoperative Respiratory Status and Operation Information
In stage I, 22.9% of patients experienced ≥1 PPC compared with 15.8% in stage II, a 7.1% absolute reduction in PPCs (p = 0.06). After adjusting for age, ASA score, and operation duration, reduction in PPCs was statistically significant (Table 4).
Table 4: Postoperative Pulmonary Complications
The main contributor to PPC rate was prolonged mechanical ventilation after PSHA, but within 7 days of PSHA (Table 4). The portion of patients requiring prolonged (>24 hours) mechanical ventilation after PSHA was 12.9% in stage I and 8.1% in stage II. Patients requiring prolonged (>24 hours) high-level respiratory support decreased from 13.8% in stage I to 11.5% in stage II. In addition, there were 17 patients who had ≥2 PPCs in stage I and 14 patients in stage II. There was no difference in pneumonia or the number readmitted to the ICU or transferred to higher level of care between the 2 stages (Table 4). Of importance, patients with higher ASA scores had the largest reduction in PPCs (Fig. 2).
Figure 2.:
Postoperative pulmonary complication (PPC) rate by American Society of Anesthesiologists (ASA) class. *p = 0.02.
As a secondary end point, the number of patients requiring prolonged (>48 hours) mechanical ventilation in stage I vs stage II were 19 (9.0%) and 14 (6.7%), respectively. Similarly, the number of patients requiring prolonged (>48 hours) respiratory support was 21 in stage I and 19 in stage II. Differences were not statistically significant.
Mean reduction in hospital LOS was 1.62 days (p < 0.02) (Table 5). The decrease in hospital LOS was greater among patients who were intubated (stage I n = 53, stage II n = 60), with an observed mean reduction of 4.85 days (p < 0.05). A greater difference was observed after adjustments for age, ASA score, and operation duration by analysis of covariance.
Table 5: Mechanical Ventilation and Length of Stay End Points
The ICU LOS decreased by 2.04 days in stage II (Table 5). The reduction in ICU LOS was greater among patients who were intubated (stage I n = 53, stage II n = 60), with an observed mean reduction of 3.86 days. Differences were not statistically significant.
Mean reduction for time to initial extubation, from stage I to stage II, was 46.9 hours (p < 0.02) (Table 5). Similarly, mean reduction for total time on mechanical ventilation was 64.4 hours (p < 0.02). In addition, for intubated patients, there was an absolute PPC rate reduction of 33.1% in stage II patients.
Mortality rates were similar in the 2 stages. There was a total of 16 deaths in the study (9 in stage I and 7 in stage II). All stage II deaths were deemed unlikely/not related to the study device per the site principal investigators. Cause of subject deaths is available in eTable 2.
Number of patients readmitted to hospital, for any cause, within 30 days of discharge was 36 (17.1%) in stage I vs 28 (13.4%) in stage II.
Compliance to treatment regimen (stage II)
The study protocol called for each patient to receive OLE treatment approximately 4 hours from PSHA. More than 85% of enrolled patients received therapy within 6 hours. Study regimen called for patients to receive 4 or more OLE treatments within the first 40 hours of PSHA. Mean ± SD number of treatments was 5.0 ± 2.1, and 76.6% (160 of 209) of enrolled patients met the goal.
Adverse events
Two events were deemed possibly related to the device. The first was intolerance of OLE postoperative day 0. This event was deemed not serious. Therapy resumed 8 hours thereafter. The second event was pneumonia on postoperative day 3, deemed serious, but resolved by discharge.
DISCUSSION
This multicenter prospective trial evaluated integration of OLE into standard postoperative respiratory therapy in high-risk patients undergoing open thoracic, aortic, or upper abdominal operations.
The precise definition of PPC remains controversial. Reported incidence varies from 2% to 40%.6 In 2015, a European Perioperative Clinical Outcome task force recommended a definition of PPC including respiratory infection, respiratory failure, pleural effusion, atelectasis, pneumothorax, bronchospasm, aspiration pneumonitis, acute respiratory distress syndrome, pulmonary edema, and pulmonary embolism.36 In the current study, PPCs consisted of the following clinically relevant parameters: prolonged invasive mechanical ventilation, prolonged high-level respiratory support, requirement for non-invasive ventilation or CPAP above patient's baseline, diagnosis of pneumonia, and readmission to the ICU (or transfer to higher level of care) for pulmonary complications. Our stage I cohort had a 22.9% incidence of PPCs, comparable with previous studies. In addition, our definition of high-risk patients—incorporating types of surgical procedure, age, comorbidities, and ASA scores—is consistent with risk stratification models reported in the literature.5,37 Although the pathophysiology of PPCs is unclear and likely multifactorial, atelectasis and diminished mucous clearance constitute critical components.38
Aggressive pulmonary care in high-risk patients has potential to improve outcomes. However, although postoperative mobilization, chest physiotherapy, and oral hygiene bundles have been associated with reduced PPCs, evidence remains limited.38 In addition, practices, including multimodal analgesic regimen, judicious prescription of narcotics, selective use of nasogastric decompression, and secretion mobilization techniques, can provide clinical benefits; yet these modalities have not been evaluated systematically. Lung expansion therapy shows the most promising results in reducing PPCs.34
In the current healthcare climate, patients demand better outcomes, payers require value-based performance (eg decreased ICU LOS, hospital LOS, and reduced costs), and health agencies and advocacy groups mandate transparency and public reporting of quality and safety measures. Clinical providers and healthcare systems face escalating pressure to deliver high-quality care at greater value. The provision of value-based healthcare requires the best outcomes with fewest complications. The Center for Medicare and Medicaid, the Institute of Medicine, the Joint Commission, federal agencies, and industries require healthcare systems to measure and report surgical outcomes. Postoperative ventilation longer than 48 hours and hospital LOS after major operations represent important quality measures mandated by the American College of Surgeons NSQIP. As such, reducing PPCs, ventilator days, and hospital LOS constitute vital quality targets for healthcare systems to achieve best practices and to remain competitive.
In this study, we delivered lung expansion using OLE. In patients who remained intubated postoperatively, our data demonstrated that the number of patients who experience ≥1 PPC decreased from 69.8% to 36.7%. Importantly, the number of patients requiring prolonged mechanical ventilation decreased from 50.9% to 28.3%. Incidence of pneumonia (1.4% in each cohort), re-admission to higher level care (2.4% in each cohort), and 30-day mortality (3.9% overall) did not differ between groups. Incidence of these outcomes was low and our study was not adequately powered to examine these end points. We did not observe a reduction in ICU LOS. Although the median reduction was 2 days, 2 patients in the stage I cohort had long ICU stays (more than 40 days). This rendered the ICU outcomes nonstatistically significant. In addition, the inclusion of these 2 patients in the stage I cohort might have affected the hospital LOS analysis by disproportionally increasing the overall hospital LOS of the stage I group. Because of the concern about the impact of outliers, we conducted additional nonparametric (Wilcoxon) analysis, and the difference in hospital LOS remained statistically significant (p = 0.03). Although we did not examine financial data, reduction in PPCs might have provided cost savings. Cost of inpatient hospitalization has been reported at $2,300 per day.39 Cost of ICU stay is estimated at $10,800 per day for mechanically ventilated patients.40 Based on these estimates, savings from reduced hospital LOS in our study might be up to $10,800 per patient.
We acknowledge that daily sedation interruption, early mobilization, or other clinical practices might have changed during the course of the 31-month study period. However, except for adding OLE, none of the 3 institutions reported alterations in care for postoperative patients. We also acknowledge that, due to the non-randomized nature of our study, the 2 cohorts of patients might not be entirely comparable. However, the stage II cohort, who received OLE, were older, had higher ASA scores, longer operation duration, and included more thoracic and aortic procedures compared with stage I. As such, our results might underestimate the magnitude of effect for OLE. To account for seasonality differences, the stage I screening list was randomly sorted before patient selection. Operation dates for stage I ranged from December 2014 to April 2016, and stage II from October 2016 to July 2017. Overall, the distribution of operation dates was similar. Introduction of the OLE modality might have created inherent bias in stage II cohort of patients. Due to the before-after non-randomized design of the study, this confounder is perhaps unavoidable. That stated, our ICUs are staffed by multiple providers rotating on a weekly basis, the respiratory treatments were episodic, and only the principal investigators for each site had in-depth knowledge of the study; we believe these factors might have helped mitigate some of the confounding effects.
CONCLUSIONS
Our study demonstrated that patients who received treatment with OLE experienced fewer PPCs and had less time on mechanical ventilation. In addition, incorporating OLE resulted in a significant reduction in hospital LOS. Importantly, patients with higher ASA scores had the largest PPC reduction.
The results suggest the modality is feasible and can help achieve value-based quality care for surgical patients. In addition, other disease entities with a high likelihood of pulmonary complications should be considered for future OLE studies, including those admitted after blunt chest trauma, COPD exacerbation, cystic fibrosis, and pneumonia.
Author Contributions
Study conception and design: Liesching, Diette
Acquisition of data: Lei, Nahouraii, Frazer
Analysis and interpretation of data: Huynh, Liesching, Cereda, Diette
Drafting of manuscript: Huynh
Acknowledgment
The authors gratefully acknowledge the following individuals for their contributions in study development and/or conduct: Lynn Sullivan, RRT; Jana Hudcova, MD; Akmal Sarwar, MD; Elliot Servais, MD; Syed Quadri, MD; Edward Boroda, RRT; Gaurav Sachdev, MD; Kyle Cunningham, MD; Ken Hargett, MHA, RRT; Kim Lacy, RN; Margarete Pierce, RRT; Niraj Agarwalla, MD; Ronald Sing, DO; Bradley Thomas, MD; Michael Thomason, MD; William Miles, MD; John Green, MD; Tunisia Hunter, RRT; Julia Raddatz, MS; Leslie Sealey; Randy Johnston; Susan Evans, MD; Rita Brintzenhoff, MD; A Brit Christmas, MD; David Jacobs, MD.
REFERENCES
1. Gupta H, Gupta PK, Fang X, et al. Development and validation of a risk calculator predicting postoperative respiratory failure. Chest 2011;140:1207-1215.
2. Dimick JB, Chen SL, Taheri PA, et al. Hospital costs associated with surgical complications: a report from the private-sector National Surgical Quality Improvement Program. J Am Coll Surg 2004;199:531-537.
3. Zilberberg MD, Shorr AF. Prolonged acute mechanical ventilation and hospital bed utilization in 2020 in the United States: implications for budgets, plant and personnel planning. BMC Health Serv Res 2008;8:242.
4. Khuri SF, Henderson WG, DePalma RG, et al. Determinants of long-term survival after major surgery and the adverse effect of postoperative complications. Ann Surg 2005;242:326-341; discussion 341-323.
5. Agostini P, Cieslik H, Rathinam S, et al. Postoperative pulmonary complications following thoracic surgery: are there any modifiable risk factors? Thorax 2010;65:815-818.
6. Restrepo RD, Braverman J. Current challenges in the recognition, prevention and treatment of perioperative pulmonary atelectasis. Expert Rev Respir Med 2015;9:97-107.
7. Rama-Maceiras P. [Peri-operative atelectasis and alveolar recruitment manoeuvres]. Arch Bronconeumol 2010;46:317-324.
8. Hans GA, Sottiaux TM, Lamy ML, et al. Ventilatory management during routine general anaesthesia. Eur J Anaesthesiol 2009;26:1-8.
9. Sabate S, Mazo V, Canet J. Predicting postoperative pulmonary complications: implications for outcomes and costs. Curr Opin Anaesthesiol 2014;27:201-209.
10. Berlinski A, Willis JR. Albuterol delivery via intrapulmonary percussive ventilator and jet nebulizer in a pediatric ventilator model. Respir Care 2010;55:1699-1704.
11. Birnkrant DJ, Pope JF, Lewarski J, et al. Persistent pulmonary consolidation treated with intrapulmonary percussive ventilation: a preliminary report. Pediatr Pulmonol 1996;21:246-249.
12. Clini EM, Antoni FD, Vitacca M, et al. Intrapulmonary percussive ventilation in tracheostomized patients: a randomized controlled trial. Intensive Care Med 2006;32:1994-2001.
13. Dimassi S, Vargas F, Lyazidi A, et al. Intrapulmonary percussive ventilation superimposed on spontaneous breathing: a physiological study in patients at risk for extubation failure. Intensive Care Med 2011;37:1269-1276.
14. Fornasa E, Ajcevic M, Accardo A. Characterization of the mechanical behavior of intrapulmonary percussive ventilation. Physiol Meas 2013; 34:1583-1592.
15. Gatani T, Martucci N, La Rocca A, et al. Management of localized pneumothoraces after pulmonary resection with intrapulmonary percussive ventilation. Ann Thorac Surg 2010;90:1658-1661.
16. Ides K, Vos W, De Backer L, et al. Acute effects of intrapulmonary percussive ventilation in COPD patients assessed by using conventional outcome parameters and a novel computational fluid dynamics technique. Int J Chron Obstruct Pulmon Dis 2012;7:667-671.
17. Natale JE, Pfeifle J, Homnick DN. Comparison of intrapulmonary percussive ventilation and chest physiotherapy. A pilot study in patients with cystic fibrosis. Chest 1994;105:1789-1793.
18. Nava S, Barbarito N, Piaggi G, et al. Physiological response to intrapulmonary percussive ventilation in stable COPD patients. Respir Med 2006;100:1526-1533.
19. Newhouse PA, White F, Marks JH, et al. The intrapulmonary percussive ventilator and flutter device compared to standard chest physiotherapy in patients with cystic fibrosis. Clin Pediatr (Phila) 1998;37:427-432.
20. Nicolini A, Mollar E, Grecchi B, et al. Comparison of intermittent positive pressure breathing and temporary positive expiratory pressure in patients with severe chronic obstructive pulmonary disease. Arch Bronconeumol 2014;50:18-24.
21. Paneroni M, Clini E, Simonelli C, et al. Safety and efficacy of short-term intrapulmonary percussive ventilation in patients with bronchiectasis. Respir Care 2011;56:984-988.
22. Reper P, van Looy K. Chest physiotherapy using intrapulmonary percussive ventilation to treat persistent atelectasis in hypoxic patients after smoke inhalation. Burns 2013;39:192-193.
23. Reychler G, Keyeux A, Cremers C, et al. Comparison of lung deposition in two types of nebulization: intrapulmonary percussive ventilation vs jet nebulization. Chest 2004;125:502-508.
24. Riffard G, Buzenet J, Guerin C. Intrapulmonary percussive ventilation superimposed on conventional mechanical ventilation: comparison of volume controlled and pressure controlled modes. Respir Care 2014;59:1116-1122.
25. Toussaint M, De Win H, Steens M, et al. Effect of intrapulmonary percussive ventilation on mucus clearance in duchenne muscular dystrophy patients: a preliminary report. Respir Care 2003;48:940-947.
26. Toussaint M, Guillet MC, Paternotte S, et al. Intrapulmonary effects of setting parameters in portable intrapulmonary percussive ventilation devices. Respir Care 2012;57:735-742.
27. Tsuruta R, Kasaoka S, Okabayashi K, et al. Efficacy and safety of intrapulmonary percussive ventilation superimposed on conventional ventilation in obese patients with compression atelectasis. J Crit Care 2006;21:328-332.
28. Vargas F, Boyer A, Bui HN, et al. Effect of intrapulmonary percussive ventilation on expiratory flow limitation in chronic obstructive pulmonary disease patients. J Crit Care 2009; 24:212-219.
29. Yen Ha TK, Bui TD, Tran AT, et al. Atelectatic children treated with intrapulmonary percussive ventilation via a face mask: clinical trial and literature overview. Pediatr Int 2007;49:502-507.
30. Vargas F, Bui HN, Boyer A, et al. Intrapulmonary percussive ventilation in acute exacerbations of COPD patients with mild respiratory acidosis: a randomized controlled trial [ISRCTN17802078]. Crit Care 2005;9:R382-R389.
31. Nino G, McNally P, Miske LJ, et al. Use of intrapulmonary percussive ventilation (IPV) in the management of pulmonary complications of an infant with osteogenesis imperfecta. Pediatr Pulmonol 2009;44:1151-1154.
32. Testa A, Galeri S, Villafane JH, et al. Efficacy of short-term intrapulmonary percussive ventilation in patients with chronic obstructive pulmonary disease. Disabil Rehab 2015;37:899-903.
33. Branson RD. The scientific basis for postoperative respiratory care. Respir Care 2013;58:1974-1984.
34. Qaseem A, Snow V, Fitterman N, et al. Risk assessment for and strategies to reduce perioperative pulmonary complications for patients undergoing noncardiothoracic surgery: a guideline from the American College of Physicians. Ann Intern Med 2006;144:575-580.
35. Hsieh J, Prickett M, Willis N, et al. A comparison of positive pressure modalities in a respiratory therapist driven protocol for patients post anatomic lung resection. Respir Care 2014;59:OF2.
36. Jammer I, Wickboldt N, Sander M, et al. Standards for definitions and use of outcome measures for clinical effectiveness research in perioperative medicine: European Perioperative Clinical Outcome (EPCO) definitions: a statement from the ESA-ESICM joint taskforce on perioperative outcome measures. Eur J Anaesthesiol 2015;32:88-105.
37. Lupei MI, Chipman JG, Beilman GJ, et al. The association between ASA status and other risk stratification models on postoperative care unit outcomes. Anesth Analg 2014;118:989-994.
38. Miskovic A, Lumb AB. Postoperative pulmonary complications. Br J Anaesth 2017;118:317-334.
39. The Henry J Kaiser Family Foundation. Hospital adjusted expenses per inpatient day. Available at:
https://www.kff.org/health-costs/state-indicator/expenses-per-inpatient-day/. Accessed November 2017.
40. Dasta JA, McLaughlin TP, Mody SH, et al. Daily cost of an ICU care unit date: the contribution of mechanical ventilation. Crit Care Med 2005;33:1266-1271.
APPENDIX
eTable 1. Data Procedures and Collection
eTable 2. Causes of Patient Deaths
