|
|
||||||||
Ann Thorac Surg 2001;72:1855-1860
© 2001 The Society of Thoracic Surgeons
a Division of Cardiothoracic Surgery, The Oregon Clinic, Portland, Oregon, USA
b Medical Data Research Center, Providence Health Systems, Portland, Oregon, USA
c Regional Heart Data Services, Providence Health Systems, Portland, Oregon, USA
d Earle A. Chiles Research Center, Providence Health Systems, Portland, Oregon, USA
* Address reprint requests to Dr Handy, Division of Cardiothoracic Surgery, The Oregon Clinic, 507 NE 47th Ave, Portland, OR 97213, USA
e-mail: jhandy{at}orclinic.com
Presented at the Thirty-seventh Annual Meeting of The Society of Thoracic Surgeons, New Orleans, LA, Jan 2931, 2001.
| Abstract |
|---|
|
|
|---|
Methods. Five years of pulmonary resections, excluding lung biopsies, were analyzed from a prospective, computerized database. Readmission was defined as inpatient or emergency department admission within 90 days of operation. Search of 1,173,912 admissions to the Providence Health System in Oregon identified readmissions. Readmission analysis excluded operative deaths.
Results. A total of 374 patients underwent pulmonary resections, of whom 8 died (2.1%). Of 366 patients discharged, 69 (18.9%) were readmitted a total of 113 times: 42 had only one readmission, 16 had two readmissions, 7 had three readmissions, 2 had four readmissions, and 2 had five readmissions. Slightly more than half (51%) were readmitted as inpatients. Causes of the 113 readmissions included pulmonary (27%), postoperative infection (14%), cardiac (7%), and other (16%). Mean time to readmission was 32.5 ± 24.6 days. Inpatient readmission mean length of stay was 4.9 ± 3.4 days. Readmission to hospitals other than the hospital of the operation was as follows: first readmission, 15.9%; second readmission, 14.8%; third readmission, 36.3%; fourth readmission, 25%; fifth readmission, 0%. Analysis revealed only pneumonectomy as a risk for readmission. Twelve of 33 (36%) pneumonectomies were readmitted (p = 0.005). Of the 297 patients discharged after pulmonary resection and not requiring readmission, 12 (4%) died over the study interval, whereas 8 of 69 patients (11.6%) requiring readmission died.
Conclusions. Readmission after pulmonary resection is frequent and multiple readmissions are common. Causes are predominately pulmonary diagnoses and infections related to the operation. Pneumonectomy is a risk for readmission. An important portion of readmissions occurs outside the hospital of operation. The population requiring readmission after successfully undergoing pulmonary resection is at increased risk of subsequent mortality.
| Introduction |
|---|
|
|
|---|
Scant data exist regarding pulmonary operation readmission rates. Large series examining postoperative courses after pulmonary resection for malignancy including bronchoplasty [4] and pneumonectomy [5, 6] failed to note readmission data. Additionally, clinical series of lung resection extensively examining age [7], sex [8], diagnosis [9], and extent of procedure [10, 11] regarding postoperative outcomes did not mention hospital readmissions. Even so scrutinized a procedure as lung volume reduction operation has no readmission information [12]. Reviews of complications of pulmonary operations have not mentioned hospital readmission [13].
Cardiac operation readmission rates vary depending on the procedure studied and postoperative time frame selected. Hospital readmission after coronary artery bypass (CAB) has been noted to be 13.5% to 28% within 4 to 6 weeks postoperatively [14, 15] and approximately 25% when followed out to 6 months [16, 17]. Follow-up for as long as 4 years after cardiac transplantation has yielded a rehospitalization rate of 37% [18]. Other health care systempatient interactions, such as emergency department (ED) evaluations, are less well studied after cardiac operations. In the only study in the literature, postcardiac operation ED evaluations have had a readmission rate of 23% at 30 days [19].
The present study was undertaken to define the prevalence, patterns, and predisposing characteristics for hospital readmission after pulmonary operation.
| Material and methods |
|---|
|
|
|---|
Patients undergoing pulmonary resection were analyzed according to demographics, comorbidities, type and site of operative procedure, associated surgical procedures, diagnosis requiring or resulting from pulmonary resection, pathologic stage when diagnosis was malignancy, postoperative complications, operative mortality, length of hospital stay (LOS), and death after hospital discharge. Operative mortality was defined as death in the operating room or at any time during the initial hospitalization for pulmonary resection.
Hospital readmission was defined as inpatient admission or ED evaluation within 90 days of pulmonary resection. We chose a global approach in our definition of readmission, including both inpatient and ED evaluations, as both are expensive patienthealth system interactions. Pulmonary resection patients were queried against 1,173,912 admissions to the Providence Health Systems (PHS) during the same time to determine hospital readmission. The PHS within the Portland metro area consists of four hospitals (two tertiary and two community hospitals) and a comprehensive health insurance product. Readmission analysis excluded operative deaths.
Reasons for readmission were classified into nine categories according to admitting diagnosis. These categories (with noninclusive examples) include: pulmonary (pneumothorax, lung cancer, respiratory failure, pleural effusion, bronchopleural fistula, dyspnea), postoperative infection (fever, pneumonia, empyema, wound infection), cardiac (angina, myocardial infarction, atrial fibrillation, heart failure), gastrointestinal (GI; esophagitis, peptic ulcer, nausea and vomiting, impaction), deep venous thrombosis, neurologic (stroke), psychiatric (anxiety), urinary (retention), and other (nonlung malignancy, nonchest musculoskeletal problems, throat pain, otalgia, sinus pain).
The nonreadmitted and readmitted patients were compared with regard to demographics, comorbidities, type and site of operative procedure, associated surgical procedures, diagnosis requiring or resulting from pulmonary resection, pathologic stage when diagnosis was malignancy, postoperative complications, operative mortality, and LOS. Readmitted patients were further characterized regarding site of readmission (inpatient versus ED, hospital of operation versus not), time until readmission, reason for readmission, LOS, and number of readmissions.
Statistical analysis was carried out using a software package (SPSS 10.0, SPSS Inc, Chicago, IL). Frequencies of readmission and diagnoses were determined. The
2 test was used to compare categorical variables between the nonreadmitted and readmitted patients. Continuous variables were compared using the independent t test or analysis of variance; p less than 0.05 was considered significant.
| Results |
|---|
|
|
|---|
|
|
|
|
| Comment |
|---|
|
|
|---|
Little information exists in the literature regarding readmission after lung resection. McKenna and colleagues [20] reported 212 video-assisted lobectomies for lung cancer, with 2 patients (0.9%) requiring readmission in the 28.9-month mean follow-up. Wright and colleagues [21] noted no impact on 7-day readmission rates after implementation of a pulmonary lobectomy patient care pathway. The 7-day readmission for the entire patient population was 5 of 277 patients (1.8%).
Postoperative readmission is easily conceptualized as a complication. Our readmission rate of 18.9% after pulmonary resection makes readmission equal to pulmonary complications (19%) as the two most frequent postoperative problems in our 374 pulmonary resections. Our pulmonary resection readmission rate is consistent with that reported after cardiac operation. Recent series have demonstrated a postoperative readmission rate of 13.5% to 16% when followed to 30 days [14, 22]. Female sex [22] and diabetes [14, 22] were found to be risk factors for readmission, whereas age, other comorbidities, and decreased ventricular function were not [22]. Discharge before postoperative day 5 was found to lessen readmission after cardiac operation [14]. In an attempt to predict readmission after CAB among 4,835 patients in Israel, no reliable predictor was identified [23]. We identified no patient demographics, comorbidities, postoperative complications, or LOS as increasing readmission after pulmonary resection. However, we did ascertain that undergoing pneumonectomy significantly increased risk of readmission.
Readmission rates rise when patients are followed for longer time intervals. We chose 90 days to capture all postoperative readmissions. At 30-day follow-up after the cardiac operation, DAgostino and coworkers [14] reported a mean time to readmission of 11.7 ± 8.7 days. As our mean time to first readmission was 32.4 ± 24.6 days, many readmissions would have been missed with a postoperative time frame of 30 days. Conversely, we reasoned that after 90 days, readmission would unlikely to be secondary to the operation. In the previously mentioned Israeli study [23], a 24.1% rehospitalization rate was identified with a 3-month follow-up. Indeed, when CAB patients were followed for 6 months postoperatively in a 1985 study, readmission rate remained 24% [17], arguing in favor of use of a 90-day window to completely evaluate postsurgical hospital readmission.
Pulmonary problems and postoperative infection (fever, pneumonia, empyema, wound infection) dominated as causes of readmission after pulmonary resection (Fig 2). After cardiac operations, cardiac reasons predominate as causes of readmission, with wound (10.2%) and GI (8.0%) problems looming much larger than our series of lung resections [14]. In the Israeli study of 1,094 readmitted patients, cardiac reasons accounted for 54% of readmission while wound infections were responsible for 10.2% [23].
Of our pulmonary resection readmissions, 17% were to hospitals other than the surgery hospital; 48% to 61.1% of readmissions after cardiac operation were to hospitals other than the surgery hospital [14, 24]. This discrepancy may be related to differences in disease or postoperative problem patterns, referral patterns, or penetration of managed care and subsequent patient steering. Additionally, we demonstrated 44 of 113 readmissions experienced by 69 lung resection patients were multiple. Between 15% and 26% of recurrent, multiple readmissions were to hospitals besides the surgery hospital. No mention has been made of multiple readmissions in any cardiac studies.
Emergency department evaluations figure prominently after pulmonary resection as a factor in readmission. Our 90-day hospital readmission was evenly divided between inpatient stays and ED evaluations. Among 124 CAB patients at Yale within 6 weeks postoperatively, ED evaluations were found to be 9.7%, whereas hospital inpatient readmissions were 18.5% [24]. This low rate contrasts with a 30-day ED visit rate of 23.1% experienced by 425 patients at the Cleveland Clinic, one half of which were to hospitals other than the surgery hospital [19].
Finally, we unexpectedly discovered a postdischarge mortality of 11.6% in the readmitted and 4.0% in the nonreadmitted patients over the 5 years of study. Reasons for this higher death rate in the readmitted patients and the mode of death for both groups are unknown. In the large Israeli CAB study, 40 of 1,094 (3.7%) readmitted patients died during follow-up [23]. No data were given for the nonreadmitted patients.
Several real or potential flaws of this study bear mentioning. First, cause of readmission was based on categorizing the admitting diagnosis as opposed to the more accurate discharge diagnosis. Often, these diagnoses are not the same. We did not have computer access to discharge diagnoses. Secondly, we interrogated the computerized database of the PHS to identify readmissions. Thus, patients not readmitted to the PHS would be missed. Portland is unique in that health systems operate hospital systems coupled with health insurance products. It is unlikely that PHS patients cross out of PHS when experiencing postoperative problems. Third, complete preoperative physiologic evaluation (pulmonary function tests, 6-minute walk, arterial blood gases, etc) were not routinely entered into our thoracic operation database until 1997, thus these potentially important data were not available for evaluation of risk for readmission. Finally, we did not characterize the postdischarge mortality with regard to mean length of postoperative follow-up, or causes or patterns of death between the readmitted and nonreadmitted patients, making this information hard to interpret.
Future investigation of readmission after pulmonary resection data will focus on causes and patterns of the increased postdischarge death rate experienced by readmitted patients to guide effective prevention. Also, alteration of clinical practice is called for to decrease the readmission rate after pneumonectomy. These changes could include increased number and frequency of postoperative surgeon or nurse office visits, aggressive use of home health care beyond usual indications, or frequent phone contact to identify and handle problems on an outpatient basis.
Readmission to inpatient or ED care is frequent after pulmonary resection. Readmissions are often multiple, with a significant portion occurring outside the surgery hospital. Inpatient and ED readmissions occur with equal frequency. Postoperative pulmonary sequelae and infection predominate as reasons for readmission. Readmitted patients are at increased risk of postdischarge mortality. The sole predisposing risk factor for readmission is undergoing pneumonectomy.
| Discussion |
|---|
|
|
|---|
DR HANDY: One of the things I did not tell you was that the mean time to readmission was 32 ± 24 days. So these people are actually being readmitted late after their operation. It would seem unlikely that the cause of readmission is from an overly early discharge. In the cardiac literature, the fast-track, short-length-of-stay patient does not get readmitted, and we found that not to be the case here. A shorter hospitalization did not mean that the patient was likely to be readmitted. Conversely, a longer hospitalization was no guarantee that the patient would not be readmitted.
DR ROBERT J. CERFOLIO (Birmingham, AL): That is a perfect segue to my comments. We just presented a paper on fast-tracking pulmonary resections at the Western Thoracic Surgical Association and also found readmissions to be a problem. The patients would fly or drive back home and most of our readmissions were for "pneumonias." We found that with some patient educationthat is, by telling the patient, "Your X-ray is going to look abnormal afterwards"we could help decrease that incidence. Besides patient education at the time of discharge and talking with them, what other ways have you found to help decrease the incidence of readmissions? Readmission continues to be a problem for us, and I do not believe it is related to the fast track, but rather an education problem.
DR HANDY: That is a good question and I wish I had a good answer. Our information is novel and our observations are apparently unprecedented. So we are identifying a problem. The next step is to solve it.
DR MALCOLM M. DECAMP, JR (Cleveland, OH): I enjoyed your presentation very much. Given that pulmonary complications were clearly far and away the reason for readmission, I think you would need to look at the preoperative pulmonary status of your patients to see if it was the higher risk patient with lesser lung function that was re-presenting to their caregiver requiring readmission. And similarly, I was struck by the fact that there was a substantial number of nonanatomic resections given that most of the resections were for cancer, which makes me wonder whether these patients who were being readmitted perhaps had impaired lung function to begin with. I would appreciate your comments.
DR HANDY: That is a good question. One of the things that I wondered was this: Is the diagnosis of malignancy in and of itself, whether it be primary lung cancer or even more likely metastatic disease, a predisposition for readmission? The answer in this patient population was no, patients with the diagnosis of malignancy were not more likely to be readmitted. Our standard is just like everybody elses. We do anatomic resections, and one of the problems with our database, which is pretty extensive we have approximately 1,300 patientsis that we do not have good quantification globally of pulmonary function in this database.
DR PATRICK ROSS (Columbus, OH): I want to follow up on that. The mortality may be related to their preoperative pulmonary function as well as their stage. You are looking at a sicker population having a bigger operation and then saying that they did not live as long, and I am not sure, without differentiating and stratifying them preoperatively, that you can really justify that conclusion. I think that knowing their preoperative function and stage would be more important in contributing to our understanding of the difference between those two groups.
DR HANDY: If you are referring to the information where it seems as if readmitted patients have an increased subsequent mortality rate, unfortunately that was one of the weaknesses of the study. We have not yet characterized the length of follow-up. This was a 5-year study, but not a 5-year follow-up. Neither have we characterized the pattern or differences in the subsequent mortality of the readmitted versus nonreadmitted patients. That was one of the weaknesses in our study.
Thank you very much.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
J. Rubins, M. Unger, and G. L. Colice Follow-up and Surveillance of the Lung Cancer Patient Following Curative Intent Therapy: ACCP Evidence-Based Clinical Practice Guideline (2nd Edition) Chest, September 1, 2007; 132(3_suppl): 355S - 367S. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Shaw Genetics of postoperative complications following thoracic surgery. Seminars in Cardiothoracic and Vascular Anesthesia, December 1, 2006; 10(4): 327 - 345. [Abstract] [PDF] |
||||
![]() |
M. C. Engoren and C. M. Arslanian-Engoren Outcome After Tracheostomy for Respiratory Failure in the Elderly J Intensive Care Med, March 1, 2005; 20(2): 104 - 110. [Abstract] [PDF] |
||||
![]() |
G. Varela, J. L. Aranda, M. F. Jimenez, and N. Novoa Emergency hospital readmission after major lung resection: prevalence and related variables Eur. J. Cardiothorac. Surg., September 1, 2004; 26(3): 494 - 497. [Abstract] [Full Text] [PDF] |
||||
![]() |
G. L. Colice, J. Rubins, and M. Unger Follow-up and Surveillance of the Lung Cancer Patient Following Curative-Intent Therapy Chest, January 1, 2003; 123 (2009): 272S - 283S. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |