Ann Surg Oncol DOI 10.1245/s10434-014-3785-4
ORIGINAL ARTICLE – ENDOCRINE TUMORS
Assessing Safety and Outcomes in Outpatient versus Inpatient Thyroidectomy using the NSQIP: A Propensity Score Matched Analysis of 16,370 Patients Nima Khavanin, BS1, Alexei Mlodinow, BS1, John Y. S. Kim, MD, FACS1, Jon P. Ver Halen, MD, FACS2, Anuja K. Antony, MD, MPH, FACS3, and Sandeep Samant, MD, FACS4 1
Division of Plastic and Reconstructive Surgery, Northwestern University Feinberg School of Medicine, Chicago, IL; Division of Plastic and Reconstructive Surgery, Baptist Cancer Center-Vanderbilt Ingram Cancer Center, Memphis, TN; 3 Division of Plastic and Reconstructive Surgery, University of Illinois at Chicago, Chicago, IL; 4Department of Otolaryngology-Head and Neck Surgery, University of Tennessee Health Science Center, Memphis, TN 2
ABSTRACT Background. With increasing economic healthcare constraints and an evolving understanding of patient selection criteria and patient safety, outpatient thyroidectomy is now more frequently employed. However, robust statistical analyses evaluating outcomes and safety after outpatient thyroidectomy with matched comparisons to inpatient cohorts are lacking. Methods. The 2011-2012 NSQIP datasets were queried to identify all patients undergoing thyroidectomy. Inpatient and outpatient procedures cohorts were matched 1:1 using propensity score analysis to assess outcomes. Outcomes of interest included surgical and medical complications, reoperation, mortality, and readmission. Univariate and multivariate analyses were utilized to identify predictors of these events. Relative risk ratios were calculated for adverse events between inpatient and outpatient cohorts. Results. In total, 21,508 patients were identified to have undergone a thyroidectomy in 2011–2012. Inpatients and outpatients were matched 1:1 with respect to preoperative and operative characteristics, leaving 8,185 patients in each treatment arm. After matching, overall 30-day morbidity was rare with only 250 patients (1.53 %) experiencing any perioperative morbidity. 476 patients (2.91 %) were readmitted within 30-days of the operation. Both pre- and postmatching, inpatient thyroidectomy was associated with
Ó Society of Surgical Oncology 2014 First Received: 17 March 2014 J. P. Ver Halen, MD, FACS e-mail: [email protected]
increased risks of readmission, reoperation, and any complication. Conclusions. Based on this comprehensive populationbased study, outpatient thyroidectomy appears to be at least as safe as inpatient thyroidectomy. However, there are still differences in outcomes between inpatient and outpatient cohorts, despite statistical matching of preoperative and intraoperative variables. Future research needs to be spent identifying these as-of-yet unknown risk factors to resolve this discrepancy.
With increased focus on patient safety and an improved understanding of patient selection criteria, outpatient surgery is a viable option for many patients, enabling significant cost-savings in health care expenditures. However, the safety of outpatient surgery must be balanced with the potential for adverse events (AE).1–12 In keeping with this trend, thyroidectomy is now frequently undertaken as an outpatient (OP) procedure, with an ever-growing number of published studies reporting on its implementation since 2006.13–31 While there is a growing amount of peerreviewed research on the subject addressing procedurespecific outcomes (e.g., nerve injury, bleeding, airway compromise), there have not been any statistically rigorous, matched cohort analyses comparing outcomes between inpatient (IP) and OP cohorts. NSQIP is a nationally validated, risk-adjusted surgical outcomes database, aimed at measuring and improving the quality of care delivered to surgical patients. The database provides a robust cohort of patients from which a highpowered retrospective study can be performed, and it continues to be used to identify relevant clinical elements
N. Khavanin et al. TABLE 1 List of included CPT codes 60220 Total thyroid lobectomy, unilateral; with or without isthmusectomy 60225 Total thyroid lobectomy, unilateral; with contralateral subtotal lobectomy, including isthmusectomy 60240 Thyroidectomy, total or complete 60252 Thyroidectomy, total or subtotal for malignancy; with limited neck dissection 60254 Thyroidectomy, total or subtotal for malignancy; with radical neck dissection 60270 Thyroidectomy, including substernal thyroid; sternal split or transthoracic approach 60271 Thyroidectomy, including substernal thyroid; cervical approach
in various specialties.32–34 This study aims to compare outcomes for inpatient versus outpatient thyroidectomy using the NSQIP dataset, specifically evaluating the safety of outpatient thyroidectomy. METHODS Data Acquisition and Patient Selection Data collection methods for the American College of Surgeons National Surgical Quality Improvement Program (NSQIP) registry have previously been described in detail.35,36 All study aspects were approved by respective Institutional Review Boards. The 2011 and 2012 NSQIP registries were queried using the Current Procedural Terminology codes 60220, 60225, 60240, 60252, 60254, 60270, and 60271 to identify all patients undergoing thyroidectomy (Table 1). We excluded CPT codes 60210 (partial total lobectomy) and 60212 (partial total lobectomy with contralateral subtotal lobectomy), and 60260 (completion thyroidectomy) because we wanted to only include patients undergoing primary, total thyroidectomy for comparison. The authors felt that lobectomy and completion thyroidectomy would be biased towards OP and IP status, respectively. By convention, we track patients only using CPT codes, and not ICD-9 codes, as intention was to track thyroidectomy alone, irrespective of primary diagnosis. Outpatient procedures were identified by the inpatient/outpatient variable in the NSQIP registry. Outpatient status is defined by NSQIP as same day or 23 h or less in-hospital stay before discharge. Outcomes and Risk Adjustment Variables The primary outcomes of interest included 30-day adverse events.37 Adverse events included overall morbidity, surgical complications, medical complications, reoperations, unplanned readmission, and death within
30 days following the index procedure. Preoperative variables for risk adjustment included demographic data, medical comorbidities, operative time, American Society of Anesthesiologists (ASA) classification, and concurrent surgical procedures. Additionally, the sum of the relative value units (RVUs) was included to adjust for the added complexity and risk of concurrent procedures, as has been described previously.38 RVUs reflect the relative level of time, skill, training, and intensity required of a physician to provide a given service. RVUs therefore are a method for calculating the volume of work or effort expended by a physician in treating patients and are set by CMS in association with a physician advisory council. For a patient undergoing multiple procedures, the respective RVUs can be summed to estimate the total complexity of a given set of procedures (e.g., laryngectomy ? free flap reconstruction ? tracheostomy ? feeding tube in a patient undergoing thyroidectomy as one of multiple procedures). Given the heterogeneity of thyroidectomy codes used (e.g., lobectomy, total thyroidectomy, substernal thyroidectomy; Table 1), we matched RVU totals between cohorts in addition to comparing CPT codes. This matching has the direct effect of eliminating those procedures with concurrent procedures (e.g., laryngectomy, tracheal surgery), which are otherwise present in the inpatient (but not outpatient) cohort. Statistical Analysis Patients undergoing IP and OP thyroidectomies were propensity score matched to balance out differences between the patient populations. For each patient, the variables in Table 2 were used to model their probability for undergoing an outpatient procedure. Using these values, inpatients were paired with the nearest outpatient procedure without replacement in a 1:1 ratio. When a close match was not available, the case was eliminated. Propensity score matching was performed in SPSS (IBM Corp., Armonk, NY), and has been previously described in detail.39–42 Descriptive statistics and outcomes were analyzed in the matched cohorts using McNemar’s test for categorical variables and paired sample Student’s t tests for continuous variables, and in the unmatched cohort using v2 tests and unpaired Student’s t tests. Multiple logistic regression models were used to obtain unadjusted and adjusted odds ratios for the influence of an OP procedure on outcomes. Whenever possible the day of readmission for each patient was recorded and compared across the two cohorts. Detailed information regarding readmission was only available for the 2012 dataset. When available, International Classification of Disease 9th Edition revision codes were used to determine the reason for readmission.
Outpatient Versus Inpatient Thyroidectomy Outcomes TABLE 2 Patient demographics and clinical characteristics Full cohort
Matched cohort
% of patients Inpatients n = 9,817
p value Outpatients n = 11,691
% of patients Inpatients n = 8 185
p value Outpatients n = 8 185
Gender, (%) female
78.9 %
81.3 %
\0.001
80.5 %
80.7 %
0.695
Age (yr), mean ± SD
52.29 ± 14.8
51.13 ± 14.6
\0.001
51.99 ± 14.6
52.11 ± 14.5
0.579
Body mass index, mean ± SD
29.98 ± 7.8
29.94 ± 7.5
Diabetes
13.6 %
11.2 %
\0.001
0.681
30.14 ± 7.7
30.13 ± 7.5
0.897
13 %
12.8 %
0.796
Current smoker
13.9 %
14.9 %
0.037
14.1 %
14.1 %
0.964
Dyspnea
6.8 %
History of COPD
2.2 %
7.1 %
0.463
6.8 %
7.1 %
0.444
2.0 %
0.352
2.0 %
2.0 %
Congestive heart failure
0.912
0.4 %
0.1 %
\0.001
0.1 %
0.0 %
0.453
Hypertension
39.7 %
37.8 %
Disseminated cancer
1.2 %
0.4 %
0.006 \0.001
39.4 %
39.5 %
0.835
0.5 %
0.5 %
0.657
Steroid use
2.2 %
1.8 %
0.031
2.0 %
2.0 %
0.824
Bleeding disorders
1.6 %
1.1 %
\0.001
1.2 %
1.4 %
0.539
Chemotherapy within the past 30 days
0.5 %
0.3 %
0.164
0.2 %
0.1 %
0.754
Radiotherapy within the in past 90 days Prior operation within 30 days
0.2 % 0.8 %
0.2 % 0.4 %
0.608 0.470
0.1 % 0.3 %
0.1 % 0.2 %
0.453 0.607
68.9 %
73.1 %
70.6 %
70.7 %
\0.001
ASA classification 1 or 2
0.904
3, 4, or 5
31.1 %
26.9 %
29.4 %
29.3 %
Previous PCI or cardiac surgery
3.7 %
3.4 %
0.366
1.7 %
1.5 %
Previous stroke or TIA
3.2 %
3.0 %
0.494
1.4 %
1.2 %
0.125
Sum of RVUs, mean ± SD
18.71 ± 9.1
16.4 ± 5.9
\0.001
17.38 ± 6.8
17.36 ± 6.3
0.779
Operative time (min), mean ± SD
134.77 ± 76.6
106.67 ± 56.3
\0.001
119.82 ± 55.5
118.81 ± 59.5
0.052
0.456
RESULTS
Postoperative Outcomes
Cohort Characteristics
Overall 30-day morbidity was rare in both the matched and unmatched cohorts (Table 3). In both pre- and postmatched cohorts, outpatients tended to experience lower rates of postoperative complications than their inpatient counterparts. Relative risk values were calculated for inpatient versus outpatient procedures (Fig. 1). Values reflect ‘‘unadjusted’’ relative risk for readmission, any complication, reoperation, any surgical complication, and any medical complication, respectively. For each category of adverse event, the relative risk was significantly less for OP procedures. Within the matched cohorts, 71 outpatients (0.87 %) and 179 inpatients (2.19 %) experienced any postoperative complication (p \ 0.001; Table 3). Overall, 28 outpatients (0.34 %) and 45 inpatients (0.5 %) experienced a surgical complication within 30 days of thyroidectomy (p = 0.06); 43 outpatients (0.53 %) and 143 inpatients (1.75 %) experienced a medical complication within 30 days of thyroidectomy (p \ 0.001). It should be noted that while
In total 21,508 patients were identified who underwent thyroidectomy between 2011 and 2012. Overall, inpatients tended to be 1 year older with significantly higher rates of many comorbidities (Table 2). Of these, 16,370 patients (8,185 from each of the inpatient and outpatient cohorts) were matched and included in the analysis. After matching, the cohorts did not demonstrate any significant differences with respect to any preoperative variables, sum of procedural RVUs, and operative time (Table 2). With specific regard to RVU, pre-match differences (18.71 vs. 16.4, p \ 0.001) were controlled post-match (17.38 vs. 17.36, p = 0.779). Similar results were obtained for operative time and CPT code matching. In this manner, patients undergoing concomitant procedures that added additional complexity were either eliminated or matched to a similar patient, with the ultimate result of making the cohorts statistically indistinguishable.
N. Khavanin et al. TABLE 3 Patient outcomes following inpatient and outpatient procedures Full cohort
Matched cohort
% of patients Inpatients (%) n = 9,817
p value Outpatients (%) n = 11,691
% of patients Inpatients (%) n = 8,185
p value Outpatients (%) n = 8,185
Any 30-day complication
2.74
0.8
\0.001
2.19
0.87
\0.001
Overall surgical complication
0.63
0.29
\0.001
0.55
0.34
0.06
SSI
0.56
0.21
\0.001
0.50
0.24
0.01
Superficial SSI
0.34
0.21
0.063
0.33
0.23
0.302
Deep incisional SSI
0.14
0.01
\0.001
0.1
0.01
0.039
Organ/space SSI
0.08
0.00
0.002
0.07
0
0.031
Wound disruption
0.13
0.08
0.205
0.09
0.1
Overall medical complication
2.26
0.51
\0.001
1.75
0.53
Pneumonia
0.32
0.06
\0.001
0.26
0.07
0.006
Unplanned intubation
0.71
0.06
\0.001
0.55
0.09
\0.001
0.999 \0.001
VTE
0.23
0.06
\0.001
0.22
0.06
\0.001
Deep vein thrombosis
0.13
0.03
0.005
0.12
0.02
0.039
Pulmonary embolism
0.1
0.04
0.102
0.1
0.05
Failure to wean Progressive renal insufficiency
0.5 0
0 0
\0.001 –
0.38 0
0 0
\0.001 – 0.5
0.388
Acute renal failure
0.02
0
0.208
0.02
0
Urinary tract infection
0.39
0.27
0.117
0.35
0.24
0.253
Stroke
0.03
0.03
0.999
0.02
0.02
0.999
Coma
0
0
–
0.00
0.00
–
Peripheral nerve injury
0.02
0.03
0.8
0.02
0.02
0.999
Cardiac arrest
0.1
0.01
0.004
0.06
0.01
0.219
Myocardial infarction
0.08
0.01
0.014
0.05
0.01
0.375
Bleeding
0.5
0.02
\0.001
0.26
0.02
\0.001 0.004
Sepsis or septic shock
0.33
0.04
\0.001
0.22
0.05
Sepsis
0.24
0.04
\0.001
0.15
0.05
0.077
Septic shock
0.09
0.00
0.001
0.09
0
0.016
Readmission
3.19
2.23
\0.001
3.34
2.48
Reoperation
2.15
1.01
\0.001
1.97
1
\0.001
Death
0.15
0.01
\0.001
0.09
0
0.016
NSQIP has an extremely large sample size, and detailed variables collected, given the low proportions of patients experiencing an AE (many \1 %), the clinical relevance of these findings is questionable. Outcomes, such as SSI, wound disruption, pneumonia, respiratory failure/intubation, UTI, and bleeding, all have clinical relevance to thyroidectomy, but the significance of renal failure, MI, and CVA are debatable in this population. However, the study is intended to establish benchmarks for complication rates for IP and OP thyroidectomy and to assess the safety of OP thyroidectomy by making risk-adjusted comparisons with IP thyroidectomy. Of the 8,185 patients in each cohort, 161 inpatients (1.97 %) and 81 outpatients (1.00 %) underwent reoperation within 30 days (p \ 0.001). There were seven reported
0.001
mortalities in the inpatient cohort (Table 3). Overall in the matched cohorts, 203 (2.48 %) outpatients and 273 (3.34 %) inpatients were readmitted within 30 days of thyroidectomy (odds ratio [OR] = 0.74, 95 % confidence interval [CI] 0.613–0.886, p = 0.001). Data for timing of readmission were available for 156 IP and 125 OP. One outpatient was readmitted on the day of their operation, and across both cohorts, the readmission rate was greatest within the first 5 days postoperatively (34.6 % for IP, 44.8 % for OP, p = 0.048). Multiple Logistic Regression After adjusting for all preoperative variables and 30-day complications (Tables 1 and 2), outpatient procedural
Outpatient Versus Inpatient Thyroidectomy Outcomes FIG. 1 Thirty-day outcomes within matched cohorts. Relative risk values were calculated for inpatient versus outpatient procedures. Values reflect ‘‘unadjusted’’ relative risk for readmission, any complication, reoperation, any surgical complication, and any medical complication, respectively
Favors Outpatient Procedure Readmission
p = 0.001
Any Complication
p < 0.001
Reoperation
p < 0.001
Surgical Complication
p = 0.060
Medical Complication
p < 0.001
0.00
0.20
0.40
0.60
0.80
Favors Inpatient Procedure
1.00
1.20
1.40
1.60
1.80
2.00
Relative Risk (95% Cl) for an Outpatient Procedure within Matched Cohorts
status was associated with decreased odds of 30-day hospital readmission. Before adjusting for confounding variables, the unadjusted odds ratio for the influence of an outpatient procedure on readmission was 0.737 (p = 0.001). After serially adjusting for pre- and intraoperative variables and post-operative complications, the odds ratio remained significantly protective with a decreased likelihood of readmission (OR = 0.682, 95 % CI 0.511–0.909, p = 0.009; OR = 0.689, 95 % CI 0.498– 0.953, p = 0.024, respectively). Regression analysis was not performed for reoperation, mortality, surgical, or medical complications. Reasons for Readmission Detailed data on the reason for readmission was only available for patients undergoing thyroidectomy in 2012. The most commonly recorded reason for readmission across both cohorts was hypocalcemia/hypoparathyroidism; 32 of 132 inpatients (24.2 %) and 33 of 102 outpatients (32.3 %) recorded this AE, although this percentage was higher among outpatients (Table 4). Postoperative bleeding accounted for 8 inpatient and 6 outpatient readmissions, and reoperation following the identification of a malignant neoplasm of the thyroid accounted for 7 inpatient and 12 outpatient readmissions. Other recorded causes for readmission included respiratory issues, laryngeal nerve injury, VTE, and surgical site infection or wound disruption. Notably, 75 of 132 (56.8 %) inpatients and 49 of 102 (48.0 %) outpatients were readmitted for reasons unrelated to the principal surgical procedures. There were no significant differences in the reasons for readmission between IP and OP cohorts.
TABLE 4 Summary of reasons for readmission Reason for readmission
Inpatients (n = 132)*
Outpatients (n = 102)*
p value
Postoperative bleeding
8
6
0.955
Respiratory issue
2
1
1.000
Laryngeal nerve injury
1
0
1.000
Hypocalcemia/ hypoparathyroidism
32
33
0.22
7
12
0.12
Malignant neoplasm of thyroid gland VTE
2
1
1.000
Surgical site infection/wound disruption
5
0
0.126
Unrelated to principal surgery
75
49
0.229
* Reason for readmission available only for patients undergoing thyroidectomy in 2012
DISCUSSION In the absence of randomized, clinical trials evaluating the safety of outpatient thyroidectomy, our study maximizes the use of advanced statistical techniques to match inpatient and outpatient cohorts, thus providing comparable populations to assess AE and readmission rates. Before matching, inpatient and outpatient cohorts were significantly different with regard to gender, age, diabetes, active smoking status, congestive heart failure, hypertension, disseminated cancer, steroid use, bleeding disorder, ASA class, RVU total, and operative time. Post-matching, the two cohorts were not statistically different in any of these categories. While operative times were similar (p = 0.052), the absolute difference in means was only approximately 1 min, thus suggesting clinical irrelevance. Of great interest, matching only eliminated postoperative
N. Khavanin et al.
differences in overall surgical complications and cardiac complications (cardiac arrest and MI), whereas all other adverse event categories remained significantly increased in the inpatient cohort (Table 3; Fig. 1). One might contend that the increased incidence of specific medical complications (i.e., unplanned intubation, failure to wean from ventilator, septic shock, acute renal failure, bleeding requiring transfusion) may in fact be driving an increased utilization of IP status in this population. Other AEs, including VTE/DVT/PE and UTI, could be increased as a consequence of IP status. Regardless, the findings support the argument that OP thyroidectomy is at least as safe as IP, in properly selected patients. The future of outpatient thyroid surgery lies in two factors. The first, establishing safety outcomes comparable with traditional inpatient thyroidectomy, is described in this article. The second is proper patient selection.21,31,43–46 NSQIP tracks a number of previously established criteria, including age, ASA class, body mass index (BMI), presence of disseminated cancer, and ownership of a telephone. Of these factors, all were matched between cohorts, and a bias in favor of outpatient thyroidectomy remained (with respect to decreased odds of readmission and AE). These findings suggest that other, unidentified factors are involved. These include intraoperative surgical findings for which no adequate proxy is tracked in NSQIP. A number of specific complications merit note. Surgical site infection (the major recorded surgical complication) would not be affected by outpatient versus inpatient thyroidectomy, because it would most likely occur 5–7 days postoperatively. Also, NSQIP outcomes data are generic, and do not routinely track thyroid-specific complications, such as hypocalcemia or hematoma.28,47 As a surrogate, NSQIP does track the incidence of bleeding requiring transfusion (which in our cohort occurred much more frequently in inpatients; Table 3). Alternatively, one could instead track unplanned reoperation, which also was increased significantly in the IP cohort. These factors may bear importance as bleeding and hypocalcemia/hypothyroidism were major reasons for readmission (2012 data available only in Table 4), although the majority of reasons for readmission were listed as ‘‘other’’ (56.8 and 48 %). Statistical matching for procedural complexity via the use of RVU total and operative time has a number of advantages. This process allowed us to either match patients with adjunctive procedures (e.g., tracheal surgery, esophageal resection, laryngectomy) between cohorts or eliminate them from analysis (if an appropriately matched counterpart could not be found), thus removing any differences in the CPT codes between cohorts. This process further eliminated (or matched) patients who intraoperatively (based on events during the primary surgery)
required additional procedures (e.g., radical neck dissection, recurrent laryngeal nerve injury and repair, tracheal injury and repair), provided that surgeons reported and coded events appropriately. Conversely, there is no way to reliably track variables (including intraoperative or postoperative PTH levels), which could reasonably convert a given patient from OP to IP status. The closest surrogate for this information is the ‘‘reason for readmission’’ variable (Table 4), which reports equal numbers of readmission for bleeding, hypocalcemia, nerve injury, and so on. Our study has a number of limitations. NSQIP is only designed to detect if a given patient stayed in the hospital less than 23 h, neither what was initially planned, nor if a given patient was discharge the same day. Similarly, NSQIP cannot distinguish between same-day versus overnight (i.e., 23 h) observation. In this regard, the best indicator for a significant AE may be unplanned reoperation (e.g., postoperative bleeding, airway compromise), which also was significantly increased in the IP cohort. While the authors acknowledge this important limitation of the NSQIP database, they emphasize the importance of the current study’s findings. Further refinement of this variable in future iterations of NSQIP may be warranted. In addition, individual surgeon experience, another variable excluded from NSQIP, has been linked to adverse events.48 Finally, NSQIP does not specifically track the most critical complications for which monitoring is recommended following outpatient thyroidectomy (as noted above). The authors emphasize that the main outcome of interest in this study, namely comparison of readmission and incidence of adverse events after inpatient versus outpatient thyroidectomy, is not altered by this relative insensitivity to thyroidectomy-specific complications, in that both cohorts would be equally affected. Outpatient thyroidectomy appears to be safe, with rates of adverse events comparing well with other outpatient procedures. After matching, complications remain lower in the outpatient versus inpatient cohort, arguing that OP thyroidectomy is at least as safe as IP, in properly selected patients. As procedures continue to transition into outpatient settings and financial penalties associated with readmission become a reality, these findings will serve to optimize outpatient surgery utilization. Future research should aim towards tracking thyroidectomy-specific complications in NSQIP, and identifying additional patient selection criteria for outpatient thyroidectomy. FINANCIAL SUPPORT This particular research received no internal or external grant funding. CONFLICT OF INTEREST The authors report no relevant financial disclosures related to this current work.
Outpatient Versus Inpatient Thyroidectomy Outcomes ETHICAL APPROVAL De-identified patient information is freely available to all institutional members who comply with the American College of Surgeons National Surgical Quality Improvement Program (NSQIP) Data Use Agreement. The Data Use Agreement implements the protections afforded by the Health Insurance Portability and Accountability Act of 1996. DISCLAIMER The NSQIP and the hospitals participating in the NSQIP are the source of the data used herein; they have not been verified and are not responsible for the statistical validity of the data analysis, or the conclusions derived by the authors of this study.
REFERENCES 1. Hockenberry JM, Burgess JF Jr, Glasgow J, et al. Cost of readmission: can the Veterans Health Administration (VHA) experience inform national payment policy?. Med Care. 2013; 51(1):13–9. 2. Jencks SF, Williams MV, Coleman EA. Rehospitalizations among patients in the Medicare Fee-for-service program. N Engl J Med. 2009;360:1418–28. 3. Orszag PR, Emanuel EJ. Health care reform and cost control. N Engl J Med. 2010;363:601–3. 4. Medicare Payment Advisory Commission. A path to bundled payment around a rehospitalization. Report to the Congress: reforming the delivery system. Washington, DC, 2005:83–103. 5. Application of incentives to reduce avoidable readmissions to hospitals. Fed Regist. 2008;73:23673–7. 6. Centers for Medicare and Medicaid Services. Readmission Reduction Program. Web http://www.cms.gov/Medicare/Medicare-Feefor-Service-Payment/AcuteInpatientPPS/Readmissions-ReductionProgram.html (2013). Accessed 30 Jan 2013. 7. Patient Protection and Affordable Care Act Pub. L. No, 111-148 § 3025(a), 124 Stat. 119,408 (2010). 8. Hospital Readmissions Reduction Programs, 76 Fed. Reg. 51663 (Aug 18, 2011). 9. Fontanaros PB, McNutt RA. Revisiting hospital readmissions. JAMA. 2013;309(4):398–400. 10. McCarthy D, Johnson MB, Auget AM. Recasting readmissions by placing the hospital role in community context. JAMA. 2013;309(4):351–2. 11. Vaduganathan M, Bonow RO, Gheorghiade M. Thirty-day readmissions: the clock is ticking. JAMA. 2013;309(4):345–6. 12. Joynt KE, Jha AK. Characteristics of hospitals receiving penalties under the Hospital Readmissions Reduction Program. JAMA. 2013;309(4):342–3. 13. Snyder SK, Roberson CR, Cummings CC, Rajab MH. Local anesthesia with monitored anesthesia care vs general anesthesia in thyroidectomy: a randomized study. Arch Surg. 2006; 141(2):167–73. 14. Spanknebel K, Chabot JA, DiGiorgi M, Cheung K, Curty J, Allendorf J, LoGerfo P. Thyroidectomy using monitored local or conventional general anesthesia: an analysis of outpatient surgery, outcome and cost in 1,194 consecutive cases.World J Surg. 2006;30(5):813–24. 15. Terris DJ, Moister B, Seybt MW, Gourin CG, Chin E. Outpatient thyroid surgery is safe and desirable.Otolaryngol Head Neck Surg. 2007;136(4):556–9. 16. Inabnet WB, Shifrin A, Ahmed L, Sinha P. Safety of same day discharge in patients undergoing sutureless thyroidectomy: a comparison of local and general anesthesia. Thyroid. 2008;18(1):57–61. 17. Champault A, Vons C, Zilberman S, Labaille T, Brosseau S, Franco D. How to perform a thyroidectomy in an outpatient setting.Langenbecks Arch Surg. 2009;394(5):897–902.
18. Trottier DC, Barron P, Moonje V, Tadros S. Outpatient thyroid surgery: should patients be discharged on the day of their procedures?.Can J Surg. 2009;52(3):182–6. 19. Seybt MW, Terris DJ. Outpatient thyroidectomy: experience in over 200 patients. Laryngoscope. 2010;120(5):959–63. 20. Snyder SK, Hamid KS, Roberson CR, Rai SS, Bossen AC, Luh JH, Scherer EP, Song J. Outpatient thyroidectomy is safe and reasonable: experience with more than 1,000 planned outpatient procedures. J Am Coll Surg. 2010;210(5):575–82, 582–4. 21. Hessman C, Fields J, Schuman E. Outpatient thyroidectomy: is it safe and reasonable option?. Am J Surg. 2011;201(5):565–8. 22. Houlton JJ, Pechter W, Steward DL. PACU PTH facilitates safe outpatient total thyroidectomy. Otolaryngol Head Neck Surg. 2011;144(1):43–7. 23. Sklar M, Ali MJ, Solomon P. Outpatient thyroid surgery in a Toronto community hospital. J Otolaryngol Head Neck Surg. 2011;40(6):458–61. 24. Tuggle CT, Roman S, Udelsman R, Sosa JA. Same-day thyroidectomy: a review of practice patterns and outcomes for 1,168 procedures in New York State. Ann Surg Oncol. 2011;18(4):1035– 40. 25. Mazeh H, Khan Q, Schneider DF, Schaefer S, Sippel RS, Chen H. Same-day thyroidectomy program: Eligibility and safety evaluation. Surgery. 2012;152(6):1133–41. 26. Sahmkow SI, Audet N, Nadeau S, Camire’ M, Beaudoin D. Outpatient thyroidectomy: safety and patient’s satisfaction. J Otolaryngol Head Neck Surg. 2012;41(Suppl 1):S1–12. 27. Rutledge J, Siegel E, Belcher R, et al. Barriers to same-day discharge of patients undergoing total and completion thyroidectomy. Otolaryngol Head Neck Surg. 2014;150(5):770–4. 28. Dixon JL, Snyder SK, Lairmore TC, et al. A novel method for the management of post-thyroidectomy or parathyroidectomy hematoma: A single-institution experience after over 4,000 central neck operations. World J Surg. 2014;38(6):1262–7. 29. Stack BC Jr, Moore E, Spencer H, et al. Outpatient thyroid surgery data from the University Health System (UHC) Consortium. Otolaryngol Head Neck Surg. 2013;148(5):740–5. 30. Lang BH, Yih PC, Lo CY. A review of risk factors and timing for postoperative hematoma after thyroidectomy: is outpatient thyroidectomy really safe?. World J Surg. 2012;36(10):2497–502. 31. Terris DJ, Snyder S, Carneiro-Pla D, et al. American thyroid association statement on outpatient thyroidectomy. Thyroid. 2013;23(10):1193–202. 32. Mlodinow A, Ver Halen J, Rambachan A, Gaido J, Kim JY. Anemia is not a predictor of free flap failure: a review of NSQIP data. Microsurgery. 2013;33(6):432–8. 33. Kim BD, Ver Halen JP, Lim S, Kim JYS. Predictors of 61 unplanned readmission cases in microvascular free tissue transfer patients: Multi-institutional analysis of 774 patients. Microsurgery. 2014. doi:10.1002/micr.22230. 34. Khanavin N, Mlodinow A, Kim JYS, et al. Predictors of 30-day readmission after outpatient thyroidectomy. An analysis of the 2011 NSQIP dataset. Am J Otolaryngol. 2014;35:332–9. 35. Available at: site.acsnsqip.org/participant-use-data-file/. Accessed 14 Mar 2014. 36. Shiloach M, Frencher SK Jr, Steeger JE, et al. Toward robust information: data quality and inter-rater reliability in the American College of Surgeons National Surgical Quality Improvement Program. J Am Coll Surg. 2010;210:6–16. 37. Sellers MM, Merkow RP, Halverson A, et al. Validation of new readmission data in the American College of Surgeons National Surgical Quality Improvement Program. J Am Coll Surg. 2013;216(3):420–7. 38. Henderson WG, Daley J. Design and statistical methodology of the National Surgical Quality Improvement Program: why is it what it is?. Am J Surg. 2009;198(5 Suppl):S19–27.
N. Khavanin et al. 39. Austin PC. Some methods of propensity-score matching had superior performance to others: results of an empirical investigation and Monte Carlo simulations. Biom J. 2009;51(1):171–84. 40. Mioton LM, Buck DW 2nd, Gart MS, et al. A multivariate regression analysis of panniculectomy outcomes: does plastic surgery training matter? Plast Reconstr Surg. 2013;131(4):604e–12e. 41. Austin PC. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm Stat. 2011;10(2):150–61. 42. Thoemmes F. Propensity score matching in SPSS 2012. Available at: http://arxiv.org/ftp/arxiv/papers/1201/1201.6385.pdf. Accessed 12 Dec 2013. 43. Sun GH, DeMonner S, Davis MM. Epidemiological and economic trends in inpatient and outpatient thyroidectomy in the United States, 1996-2006. Thyroid. 2013;23(6):727–33.
44. Doran HE, England J, Palazzo F. Questionable safety of thyroid surgery with same day discharge. Ann R Coll Surg Engl. 2012;94:543–7. 45. Mowschenson PM, Hodin RA. Outpatient thyroid and parathyroid surgery: a prospective study of feasibility, safety, and costs. Surgery. 1995;118:1051–3; discussion 1053–4. 46. Materazzia G, Dionigic G, Bertia P, et al. One-day thyroid surgery: retrospective analysis of safety and patient satisfaction on a consecutive series of 1,571 cases over a three-year period. Eur Surg Res. 2007;39:182–8. 47. Doran HE, Palazzo F. Ambulatory thyroid surgery: do the risks overcome the benefits? Presse Med. 2014;43(3):291–6. 48. Stavrakis AI, Ituarte PH, Ko CY, Yeh MW. Surgeon volume as a predictor of outcomes in inpatient and outpatient endocrine surgery. Surgery. 2007;142(6):887–99.
ORIGINAL ARTICLE – ENDOCRINE TUMORS
Assessing Safety and Outcomes in Outpatient versus Inpatient Thyroidectomy using the NSQIP: A Propensity Score Matched Analysis of 16,370 Patients Nima Khavanin, BS1, Alexei Mlodinow, BS1, John Y. S. Kim, MD, FACS1, Jon P. Ver Halen, MD, FACS2, Anuja K. Antony, MD, MPH, FACS3, and Sandeep Samant, MD, FACS4 1
Division of Plastic and Reconstructive Surgery, Northwestern University Feinberg School of Medicine, Chicago, IL; Division of Plastic and Reconstructive Surgery, Baptist Cancer Center-Vanderbilt Ingram Cancer Center, Memphis, TN; 3 Division of Plastic and Reconstructive Surgery, University of Illinois at Chicago, Chicago, IL; 4Department of Otolaryngology-Head and Neck Surgery, University of Tennessee Health Science Center, Memphis, TN 2
ABSTRACT Background. With increasing economic healthcare constraints and an evolving understanding of patient selection criteria and patient safety, outpatient thyroidectomy is now more frequently employed. However, robust statistical analyses evaluating outcomes and safety after outpatient thyroidectomy with matched comparisons to inpatient cohorts are lacking. Methods. The 2011-2012 NSQIP datasets were queried to identify all patients undergoing thyroidectomy. Inpatient and outpatient procedures cohorts were matched 1:1 using propensity score analysis to assess outcomes. Outcomes of interest included surgical and medical complications, reoperation, mortality, and readmission. Univariate and multivariate analyses were utilized to identify predictors of these events. Relative risk ratios were calculated for adverse events between inpatient and outpatient cohorts. Results. In total, 21,508 patients were identified to have undergone a thyroidectomy in 2011–2012. Inpatients and outpatients were matched 1:1 with respect to preoperative and operative characteristics, leaving 8,185 patients in each treatment arm. After matching, overall 30-day morbidity was rare with only 250 patients (1.53 %) experiencing any perioperative morbidity. 476 patients (2.91 %) were readmitted within 30-days of the operation. Both pre- and postmatching, inpatient thyroidectomy was associated with
Ó Society of Surgical Oncology 2014 First Received: 17 March 2014 J. P. Ver Halen, MD, FACS e-mail: [email protected]
increased risks of readmission, reoperation, and any complication. Conclusions. Based on this comprehensive populationbased study, outpatient thyroidectomy appears to be at least as safe as inpatient thyroidectomy. However, there are still differences in outcomes between inpatient and outpatient cohorts, despite statistical matching of preoperative and intraoperative variables. Future research needs to be spent identifying these as-of-yet unknown risk factors to resolve this discrepancy.
With increased focus on patient safety and an improved understanding of patient selection criteria, outpatient surgery is a viable option for many patients, enabling significant cost-savings in health care expenditures. However, the safety of outpatient surgery must be balanced with the potential for adverse events (AE).1–12 In keeping with this trend, thyroidectomy is now frequently undertaken as an outpatient (OP) procedure, with an ever-growing number of published studies reporting on its implementation since 2006.13–31 While there is a growing amount of peerreviewed research on the subject addressing procedurespecific outcomes (e.g., nerve injury, bleeding, airway compromise), there have not been any statistically rigorous, matched cohort analyses comparing outcomes between inpatient (IP) and OP cohorts. NSQIP is a nationally validated, risk-adjusted surgical outcomes database, aimed at measuring and improving the quality of care delivered to surgical patients. The database provides a robust cohort of patients from which a highpowered retrospective study can be performed, and it continues to be used to identify relevant clinical elements
N. Khavanin et al. TABLE 1 List of included CPT codes 60220 Total thyroid lobectomy, unilateral; with or without isthmusectomy 60225 Total thyroid lobectomy, unilateral; with contralateral subtotal lobectomy, including isthmusectomy 60240 Thyroidectomy, total or complete 60252 Thyroidectomy, total or subtotal for malignancy; with limited neck dissection 60254 Thyroidectomy, total or subtotal for malignancy; with radical neck dissection 60270 Thyroidectomy, including substernal thyroid; sternal split or transthoracic approach 60271 Thyroidectomy, including substernal thyroid; cervical approach
in various specialties.32–34 This study aims to compare outcomes for inpatient versus outpatient thyroidectomy using the NSQIP dataset, specifically evaluating the safety of outpatient thyroidectomy. METHODS Data Acquisition and Patient Selection Data collection methods for the American College of Surgeons National Surgical Quality Improvement Program (NSQIP) registry have previously been described in detail.35,36 All study aspects were approved by respective Institutional Review Boards. The 2011 and 2012 NSQIP registries were queried using the Current Procedural Terminology codes 60220, 60225, 60240, 60252, 60254, 60270, and 60271 to identify all patients undergoing thyroidectomy (Table 1). We excluded CPT codes 60210 (partial total lobectomy) and 60212 (partial total lobectomy with contralateral subtotal lobectomy), and 60260 (completion thyroidectomy) because we wanted to only include patients undergoing primary, total thyroidectomy for comparison. The authors felt that lobectomy and completion thyroidectomy would be biased towards OP and IP status, respectively. By convention, we track patients only using CPT codes, and not ICD-9 codes, as intention was to track thyroidectomy alone, irrespective of primary diagnosis. Outpatient procedures were identified by the inpatient/outpatient variable in the NSQIP registry. Outpatient status is defined by NSQIP as same day or 23 h or less in-hospital stay before discharge. Outcomes and Risk Adjustment Variables The primary outcomes of interest included 30-day adverse events.37 Adverse events included overall morbidity, surgical complications, medical complications, reoperations, unplanned readmission, and death within
30 days following the index procedure. Preoperative variables for risk adjustment included demographic data, medical comorbidities, operative time, American Society of Anesthesiologists (ASA) classification, and concurrent surgical procedures. Additionally, the sum of the relative value units (RVUs) was included to adjust for the added complexity and risk of concurrent procedures, as has been described previously.38 RVUs reflect the relative level of time, skill, training, and intensity required of a physician to provide a given service. RVUs therefore are a method for calculating the volume of work or effort expended by a physician in treating patients and are set by CMS in association with a physician advisory council. For a patient undergoing multiple procedures, the respective RVUs can be summed to estimate the total complexity of a given set of procedures (e.g., laryngectomy ? free flap reconstruction ? tracheostomy ? feeding tube in a patient undergoing thyroidectomy as one of multiple procedures). Given the heterogeneity of thyroidectomy codes used (e.g., lobectomy, total thyroidectomy, substernal thyroidectomy; Table 1), we matched RVU totals between cohorts in addition to comparing CPT codes. This matching has the direct effect of eliminating those procedures with concurrent procedures (e.g., laryngectomy, tracheal surgery), which are otherwise present in the inpatient (but not outpatient) cohort. Statistical Analysis Patients undergoing IP and OP thyroidectomies were propensity score matched to balance out differences between the patient populations. For each patient, the variables in Table 2 were used to model their probability for undergoing an outpatient procedure. Using these values, inpatients were paired with the nearest outpatient procedure without replacement in a 1:1 ratio. When a close match was not available, the case was eliminated. Propensity score matching was performed in SPSS (IBM Corp., Armonk, NY), and has been previously described in detail.39–42 Descriptive statistics and outcomes were analyzed in the matched cohorts using McNemar’s test for categorical variables and paired sample Student’s t tests for continuous variables, and in the unmatched cohort using v2 tests and unpaired Student’s t tests. Multiple logistic regression models were used to obtain unadjusted and adjusted odds ratios for the influence of an OP procedure on outcomes. Whenever possible the day of readmission for each patient was recorded and compared across the two cohorts. Detailed information regarding readmission was only available for the 2012 dataset. When available, International Classification of Disease 9th Edition revision codes were used to determine the reason for readmission.
Outpatient Versus Inpatient Thyroidectomy Outcomes TABLE 2 Patient demographics and clinical characteristics Full cohort
Matched cohort
% of patients Inpatients n = 9,817
p value Outpatients n = 11,691
% of patients Inpatients n = 8 185
p value Outpatients n = 8 185
Gender, (%) female
78.9 %
81.3 %
\0.001
80.5 %
80.7 %
0.695
Age (yr), mean ± SD
52.29 ± 14.8
51.13 ± 14.6
\0.001
51.99 ± 14.6
52.11 ± 14.5
0.579
Body mass index, mean ± SD
29.98 ± 7.8
29.94 ± 7.5
Diabetes
13.6 %
11.2 %
\0.001
0.681
30.14 ± 7.7
30.13 ± 7.5
0.897
13 %
12.8 %
0.796
Current smoker
13.9 %
14.9 %
0.037
14.1 %
14.1 %
0.964
Dyspnea
6.8 %
History of COPD
2.2 %
7.1 %
0.463
6.8 %
7.1 %
0.444
2.0 %
0.352
2.0 %
2.0 %
Congestive heart failure
0.912
0.4 %
0.1 %
\0.001
0.1 %
0.0 %
0.453
Hypertension
39.7 %
37.8 %
Disseminated cancer
1.2 %
0.4 %
0.006 \0.001
39.4 %
39.5 %
0.835
0.5 %
0.5 %
0.657
Steroid use
2.2 %
1.8 %
0.031
2.0 %
2.0 %
0.824
Bleeding disorders
1.6 %
1.1 %
\0.001
1.2 %
1.4 %
0.539
Chemotherapy within the past 30 days
0.5 %
0.3 %
0.164
0.2 %
0.1 %
0.754
Radiotherapy within the in past 90 days Prior operation within 30 days
0.2 % 0.8 %
0.2 % 0.4 %
0.608 0.470
0.1 % 0.3 %
0.1 % 0.2 %
0.453 0.607
68.9 %
73.1 %
70.6 %
70.7 %
\0.001
ASA classification 1 or 2
0.904
3, 4, or 5
31.1 %
26.9 %
29.4 %
29.3 %
Previous PCI or cardiac surgery
3.7 %
3.4 %
0.366
1.7 %
1.5 %
Previous stroke or TIA
3.2 %
3.0 %
0.494
1.4 %
1.2 %
0.125
Sum of RVUs, mean ± SD
18.71 ± 9.1
16.4 ± 5.9
\0.001
17.38 ± 6.8
17.36 ± 6.3
0.779
Operative time (min), mean ± SD
134.77 ± 76.6
106.67 ± 56.3
\0.001
119.82 ± 55.5
118.81 ± 59.5
0.052
0.456
RESULTS
Postoperative Outcomes
Cohort Characteristics
Overall 30-day morbidity was rare in both the matched and unmatched cohorts (Table 3). In both pre- and postmatched cohorts, outpatients tended to experience lower rates of postoperative complications than their inpatient counterparts. Relative risk values were calculated for inpatient versus outpatient procedures (Fig. 1). Values reflect ‘‘unadjusted’’ relative risk for readmission, any complication, reoperation, any surgical complication, and any medical complication, respectively. For each category of adverse event, the relative risk was significantly less for OP procedures. Within the matched cohorts, 71 outpatients (0.87 %) and 179 inpatients (2.19 %) experienced any postoperative complication (p \ 0.001; Table 3). Overall, 28 outpatients (0.34 %) and 45 inpatients (0.5 %) experienced a surgical complication within 30 days of thyroidectomy (p = 0.06); 43 outpatients (0.53 %) and 143 inpatients (1.75 %) experienced a medical complication within 30 days of thyroidectomy (p \ 0.001). It should be noted that while
In total 21,508 patients were identified who underwent thyroidectomy between 2011 and 2012. Overall, inpatients tended to be 1 year older with significantly higher rates of many comorbidities (Table 2). Of these, 16,370 patients (8,185 from each of the inpatient and outpatient cohorts) were matched and included in the analysis. After matching, the cohorts did not demonstrate any significant differences with respect to any preoperative variables, sum of procedural RVUs, and operative time (Table 2). With specific regard to RVU, pre-match differences (18.71 vs. 16.4, p \ 0.001) were controlled post-match (17.38 vs. 17.36, p = 0.779). Similar results were obtained for operative time and CPT code matching. In this manner, patients undergoing concomitant procedures that added additional complexity were either eliminated or matched to a similar patient, with the ultimate result of making the cohorts statistically indistinguishable.
N. Khavanin et al. TABLE 3 Patient outcomes following inpatient and outpatient procedures Full cohort
Matched cohort
% of patients Inpatients (%) n = 9,817
p value Outpatients (%) n = 11,691
% of patients Inpatients (%) n = 8,185
p value Outpatients (%) n = 8,185
Any 30-day complication
2.74
0.8
\0.001
2.19
0.87
\0.001
Overall surgical complication
0.63
0.29
\0.001
0.55
0.34
0.06
SSI
0.56
0.21
\0.001
0.50
0.24
0.01
Superficial SSI
0.34
0.21
0.063
0.33
0.23
0.302
Deep incisional SSI
0.14
0.01
\0.001
0.1
0.01
0.039
Organ/space SSI
0.08
0.00
0.002
0.07
0
0.031
Wound disruption
0.13
0.08
0.205
0.09
0.1
Overall medical complication
2.26
0.51
\0.001
1.75
0.53
Pneumonia
0.32
0.06
\0.001
0.26
0.07
0.006
Unplanned intubation
0.71
0.06
\0.001
0.55
0.09
\0.001
0.999 \0.001
VTE
0.23
0.06
\0.001
0.22
0.06
\0.001
Deep vein thrombosis
0.13
0.03
0.005
0.12
0.02
0.039
Pulmonary embolism
0.1
0.04
0.102
0.1
0.05
Failure to wean Progressive renal insufficiency
0.5 0
0 0
\0.001 –
0.38 0
0 0
\0.001 – 0.5
0.388
Acute renal failure
0.02
0
0.208
0.02
0
Urinary tract infection
0.39
0.27
0.117
0.35
0.24
0.253
Stroke
0.03
0.03
0.999
0.02
0.02
0.999
Coma
0
0
–
0.00
0.00
–
Peripheral nerve injury
0.02
0.03
0.8
0.02
0.02
0.999
Cardiac arrest
0.1
0.01
0.004
0.06
0.01
0.219
Myocardial infarction
0.08
0.01
0.014
0.05
0.01
0.375
Bleeding
0.5
0.02
\0.001
0.26
0.02
\0.001 0.004
Sepsis or septic shock
0.33
0.04
\0.001
0.22
0.05
Sepsis
0.24
0.04
\0.001
0.15
0.05
0.077
Septic shock
0.09
0.00
0.001
0.09
0
0.016
Readmission
3.19
2.23
\0.001
3.34
2.48
Reoperation
2.15
1.01
\0.001
1.97
1
\0.001
Death
0.15
0.01
\0.001
0.09
0
0.016
NSQIP has an extremely large sample size, and detailed variables collected, given the low proportions of patients experiencing an AE (many \1 %), the clinical relevance of these findings is questionable. Outcomes, such as SSI, wound disruption, pneumonia, respiratory failure/intubation, UTI, and bleeding, all have clinical relevance to thyroidectomy, but the significance of renal failure, MI, and CVA are debatable in this population. However, the study is intended to establish benchmarks for complication rates for IP and OP thyroidectomy and to assess the safety of OP thyroidectomy by making risk-adjusted comparisons with IP thyroidectomy. Of the 8,185 patients in each cohort, 161 inpatients (1.97 %) and 81 outpatients (1.00 %) underwent reoperation within 30 days (p \ 0.001). There were seven reported
0.001
mortalities in the inpatient cohort (Table 3). Overall in the matched cohorts, 203 (2.48 %) outpatients and 273 (3.34 %) inpatients were readmitted within 30 days of thyroidectomy (odds ratio [OR] = 0.74, 95 % confidence interval [CI] 0.613–0.886, p = 0.001). Data for timing of readmission were available for 156 IP and 125 OP. One outpatient was readmitted on the day of their operation, and across both cohorts, the readmission rate was greatest within the first 5 days postoperatively (34.6 % for IP, 44.8 % for OP, p = 0.048). Multiple Logistic Regression After adjusting for all preoperative variables and 30-day complications (Tables 1 and 2), outpatient procedural
Outpatient Versus Inpatient Thyroidectomy Outcomes FIG. 1 Thirty-day outcomes within matched cohorts. Relative risk values were calculated for inpatient versus outpatient procedures. Values reflect ‘‘unadjusted’’ relative risk for readmission, any complication, reoperation, any surgical complication, and any medical complication, respectively
Favors Outpatient Procedure Readmission
p = 0.001
Any Complication
p < 0.001
Reoperation
p < 0.001
Surgical Complication
p = 0.060
Medical Complication
p < 0.001
0.00
0.20
0.40
0.60
0.80
Favors Inpatient Procedure
1.00
1.20
1.40
1.60
1.80
2.00
Relative Risk (95% Cl) for an Outpatient Procedure within Matched Cohorts
status was associated with decreased odds of 30-day hospital readmission. Before adjusting for confounding variables, the unadjusted odds ratio for the influence of an outpatient procedure on readmission was 0.737 (p = 0.001). After serially adjusting for pre- and intraoperative variables and post-operative complications, the odds ratio remained significantly protective with a decreased likelihood of readmission (OR = 0.682, 95 % CI 0.511–0.909, p = 0.009; OR = 0.689, 95 % CI 0.498– 0.953, p = 0.024, respectively). Regression analysis was not performed for reoperation, mortality, surgical, or medical complications. Reasons for Readmission Detailed data on the reason for readmission was only available for patients undergoing thyroidectomy in 2012. The most commonly recorded reason for readmission across both cohorts was hypocalcemia/hypoparathyroidism; 32 of 132 inpatients (24.2 %) and 33 of 102 outpatients (32.3 %) recorded this AE, although this percentage was higher among outpatients (Table 4). Postoperative bleeding accounted for 8 inpatient and 6 outpatient readmissions, and reoperation following the identification of a malignant neoplasm of the thyroid accounted for 7 inpatient and 12 outpatient readmissions. Other recorded causes for readmission included respiratory issues, laryngeal nerve injury, VTE, and surgical site infection or wound disruption. Notably, 75 of 132 (56.8 %) inpatients and 49 of 102 (48.0 %) outpatients were readmitted for reasons unrelated to the principal surgical procedures. There were no significant differences in the reasons for readmission between IP and OP cohorts.
TABLE 4 Summary of reasons for readmission Reason for readmission
Inpatients (n = 132)*
Outpatients (n = 102)*
p value
Postoperative bleeding
8
6
0.955
Respiratory issue
2
1
1.000
Laryngeal nerve injury
1
0
1.000
Hypocalcemia/ hypoparathyroidism
32
33
0.22
7
12
0.12
Malignant neoplasm of thyroid gland VTE
2
1
1.000
Surgical site infection/wound disruption
5
0
0.126
Unrelated to principal surgery
75
49
0.229
* Reason for readmission available only for patients undergoing thyroidectomy in 2012
DISCUSSION In the absence of randomized, clinical trials evaluating the safety of outpatient thyroidectomy, our study maximizes the use of advanced statistical techniques to match inpatient and outpatient cohorts, thus providing comparable populations to assess AE and readmission rates. Before matching, inpatient and outpatient cohorts were significantly different with regard to gender, age, diabetes, active smoking status, congestive heart failure, hypertension, disseminated cancer, steroid use, bleeding disorder, ASA class, RVU total, and operative time. Post-matching, the two cohorts were not statistically different in any of these categories. While operative times were similar (p = 0.052), the absolute difference in means was only approximately 1 min, thus suggesting clinical irrelevance. Of great interest, matching only eliminated postoperative
N. Khavanin et al.
differences in overall surgical complications and cardiac complications (cardiac arrest and MI), whereas all other adverse event categories remained significantly increased in the inpatient cohort (Table 3; Fig. 1). One might contend that the increased incidence of specific medical complications (i.e., unplanned intubation, failure to wean from ventilator, septic shock, acute renal failure, bleeding requiring transfusion) may in fact be driving an increased utilization of IP status in this population. Other AEs, including VTE/DVT/PE and UTI, could be increased as a consequence of IP status. Regardless, the findings support the argument that OP thyroidectomy is at least as safe as IP, in properly selected patients. The future of outpatient thyroid surgery lies in two factors. The first, establishing safety outcomes comparable with traditional inpatient thyroidectomy, is described in this article. The second is proper patient selection.21,31,43–46 NSQIP tracks a number of previously established criteria, including age, ASA class, body mass index (BMI), presence of disseminated cancer, and ownership of a telephone. Of these factors, all were matched between cohorts, and a bias in favor of outpatient thyroidectomy remained (with respect to decreased odds of readmission and AE). These findings suggest that other, unidentified factors are involved. These include intraoperative surgical findings for which no adequate proxy is tracked in NSQIP. A number of specific complications merit note. Surgical site infection (the major recorded surgical complication) would not be affected by outpatient versus inpatient thyroidectomy, because it would most likely occur 5–7 days postoperatively. Also, NSQIP outcomes data are generic, and do not routinely track thyroid-specific complications, such as hypocalcemia or hematoma.28,47 As a surrogate, NSQIP does track the incidence of bleeding requiring transfusion (which in our cohort occurred much more frequently in inpatients; Table 3). Alternatively, one could instead track unplanned reoperation, which also was increased significantly in the IP cohort. These factors may bear importance as bleeding and hypocalcemia/hypothyroidism were major reasons for readmission (2012 data available only in Table 4), although the majority of reasons for readmission were listed as ‘‘other’’ (56.8 and 48 %). Statistical matching for procedural complexity via the use of RVU total and operative time has a number of advantages. This process allowed us to either match patients with adjunctive procedures (e.g., tracheal surgery, esophageal resection, laryngectomy) between cohorts or eliminate them from analysis (if an appropriately matched counterpart could not be found), thus removing any differences in the CPT codes between cohorts. This process further eliminated (or matched) patients who intraoperatively (based on events during the primary surgery)
required additional procedures (e.g., radical neck dissection, recurrent laryngeal nerve injury and repair, tracheal injury and repair), provided that surgeons reported and coded events appropriately. Conversely, there is no way to reliably track variables (including intraoperative or postoperative PTH levels), which could reasonably convert a given patient from OP to IP status. The closest surrogate for this information is the ‘‘reason for readmission’’ variable (Table 4), which reports equal numbers of readmission for bleeding, hypocalcemia, nerve injury, and so on. Our study has a number of limitations. NSQIP is only designed to detect if a given patient stayed in the hospital less than 23 h, neither what was initially planned, nor if a given patient was discharge the same day. Similarly, NSQIP cannot distinguish between same-day versus overnight (i.e., 23 h) observation. In this regard, the best indicator for a significant AE may be unplanned reoperation (e.g., postoperative bleeding, airway compromise), which also was significantly increased in the IP cohort. While the authors acknowledge this important limitation of the NSQIP database, they emphasize the importance of the current study’s findings. Further refinement of this variable in future iterations of NSQIP may be warranted. In addition, individual surgeon experience, another variable excluded from NSQIP, has been linked to adverse events.48 Finally, NSQIP does not specifically track the most critical complications for which monitoring is recommended following outpatient thyroidectomy (as noted above). The authors emphasize that the main outcome of interest in this study, namely comparison of readmission and incidence of adverse events after inpatient versus outpatient thyroidectomy, is not altered by this relative insensitivity to thyroidectomy-specific complications, in that both cohorts would be equally affected. Outpatient thyroidectomy appears to be safe, with rates of adverse events comparing well with other outpatient procedures. After matching, complications remain lower in the outpatient versus inpatient cohort, arguing that OP thyroidectomy is at least as safe as IP, in properly selected patients. As procedures continue to transition into outpatient settings and financial penalties associated with readmission become a reality, these findings will serve to optimize outpatient surgery utilization. Future research should aim towards tracking thyroidectomy-specific complications in NSQIP, and identifying additional patient selection criteria for outpatient thyroidectomy. FINANCIAL SUPPORT This particular research received no internal or external grant funding. CONFLICT OF INTEREST The authors report no relevant financial disclosures related to this current work.
Outpatient Versus Inpatient Thyroidectomy Outcomes ETHICAL APPROVAL De-identified patient information is freely available to all institutional members who comply with the American College of Surgeons National Surgical Quality Improvement Program (NSQIP) Data Use Agreement. The Data Use Agreement implements the protections afforded by the Health Insurance Portability and Accountability Act of 1996. DISCLAIMER The NSQIP and the hospitals participating in the NSQIP are the source of the data used herein; they have not been verified and are not responsible for the statistical validity of the data analysis, or the conclusions derived by the authors of this study.
REFERENCES 1. Hockenberry JM, Burgess JF Jr, Glasgow J, et al. Cost of readmission: can the Veterans Health Administration (VHA) experience inform national payment policy?. Med Care. 2013; 51(1):13–9. 2. Jencks SF, Williams MV, Coleman EA. Rehospitalizations among patients in the Medicare Fee-for-service program. N Engl J Med. 2009;360:1418–28. 3. Orszag PR, Emanuel EJ. Health care reform and cost control. N Engl J Med. 2010;363:601–3. 4. Medicare Payment Advisory Commission. A path to bundled payment around a rehospitalization. Report to the Congress: reforming the delivery system. Washington, DC, 2005:83–103. 5. Application of incentives to reduce avoidable readmissions to hospitals. Fed Regist. 2008;73:23673–7. 6. Centers for Medicare and Medicaid Services. Readmission Reduction Program. Web http://www.cms.gov/Medicare/Medicare-Feefor-Service-Payment/AcuteInpatientPPS/Readmissions-ReductionProgram.html (2013). Accessed 30 Jan 2013. 7. Patient Protection and Affordable Care Act Pub. L. No, 111-148 § 3025(a), 124 Stat. 119,408 (2010). 8. Hospital Readmissions Reduction Programs, 76 Fed. Reg. 51663 (Aug 18, 2011). 9. Fontanaros PB, McNutt RA. Revisiting hospital readmissions. JAMA. 2013;309(4):398–400. 10. McCarthy D, Johnson MB, Auget AM. Recasting readmissions by placing the hospital role in community context. JAMA. 2013;309(4):351–2. 11. Vaduganathan M, Bonow RO, Gheorghiade M. Thirty-day readmissions: the clock is ticking. JAMA. 2013;309(4):345–6. 12. Joynt KE, Jha AK. Characteristics of hospitals receiving penalties under the Hospital Readmissions Reduction Program. JAMA. 2013;309(4):342–3. 13. Snyder SK, Roberson CR, Cummings CC, Rajab MH. Local anesthesia with monitored anesthesia care vs general anesthesia in thyroidectomy: a randomized study. Arch Surg. 2006; 141(2):167–73. 14. Spanknebel K, Chabot JA, DiGiorgi M, Cheung K, Curty J, Allendorf J, LoGerfo P. Thyroidectomy using monitored local or conventional general anesthesia: an analysis of outpatient surgery, outcome and cost in 1,194 consecutive cases.World J Surg. 2006;30(5):813–24. 15. Terris DJ, Moister B, Seybt MW, Gourin CG, Chin E. Outpatient thyroid surgery is safe and desirable.Otolaryngol Head Neck Surg. 2007;136(4):556–9. 16. Inabnet WB, Shifrin A, Ahmed L, Sinha P. Safety of same day discharge in patients undergoing sutureless thyroidectomy: a comparison of local and general anesthesia. Thyroid. 2008;18(1):57–61. 17. Champault A, Vons C, Zilberman S, Labaille T, Brosseau S, Franco D. How to perform a thyroidectomy in an outpatient setting.Langenbecks Arch Surg. 2009;394(5):897–902.
18. Trottier DC, Barron P, Moonje V, Tadros S. Outpatient thyroid surgery: should patients be discharged on the day of their procedures?.Can J Surg. 2009;52(3):182–6. 19. Seybt MW, Terris DJ. Outpatient thyroidectomy: experience in over 200 patients. Laryngoscope. 2010;120(5):959–63. 20. Snyder SK, Hamid KS, Roberson CR, Rai SS, Bossen AC, Luh JH, Scherer EP, Song J. Outpatient thyroidectomy is safe and reasonable: experience with more than 1,000 planned outpatient procedures. J Am Coll Surg. 2010;210(5):575–82, 582–4. 21. Hessman C, Fields J, Schuman E. Outpatient thyroidectomy: is it safe and reasonable option?. Am J Surg. 2011;201(5):565–8. 22. Houlton JJ, Pechter W, Steward DL. PACU PTH facilitates safe outpatient total thyroidectomy. Otolaryngol Head Neck Surg. 2011;144(1):43–7. 23. Sklar M, Ali MJ, Solomon P. Outpatient thyroid surgery in a Toronto community hospital. J Otolaryngol Head Neck Surg. 2011;40(6):458–61. 24. Tuggle CT, Roman S, Udelsman R, Sosa JA. Same-day thyroidectomy: a review of practice patterns and outcomes for 1,168 procedures in New York State. Ann Surg Oncol. 2011;18(4):1035– 40. 25. Mazeh H, Khan Q, Schneider DF, Schaefer S, Sippel RS, Chen H. Same-day thyroidectomy program: Eligibility and safety evaluation. Surgery. 2012;152(6):1133–41. 26. Sahmkow SI, Audet N, Nadeau S, Camire’ M, Beaudoin D. Outpatient thyroidectomy: safety and patient’s satisfaction. J Otolaryngol Head Neck Surg. 2012;41(Suppl 1):S1–12. 27. Rutledge J, Siegel E, Belcher R, et al. Barriers to same-day discharge of patients undergoing total and completion thyroidectomy. Otolaryngol Head Neck Surg. 2014;150(5):770–4. 28. Dixon JL, Snyder SK, Lairmore TC, et al. A novel method for the management of post-thyroidectomy or parathyroidectomy hematoma: A single-institution experience after over 4,000 central neck operations. World J Surg. 2014;38(6):1262–7. 29. Stack BC Jr, Moore E, Spencer H, et al. Outpatient thyroid surgery data from the University Health System (UHC) Consortium. Otolaryngol Head Neck Surg. 2013;148(5):740–5. 30. Lang BH, Yih PC, Lo CY. A review of risk factors and timing for postoperative hematoma after thyroidectomy: is outpatient thyroidectomy really safe?. World J Surg. 2012;36(10):2497–502. 31. Terris DJ, Snyder S, Carneiro-Pla D, et al. American thyroid association statement on outpatient thyroidectomy. Thyroid. 2013;23(10):1193–202. 32. Mlodinow A, Ver Halen J, Rambachan A, Gaido J, Kim JY. Anemia is not a predictor of free flap failure: a review of NSQIP data. Microsurgery. 2013;33(6):432–8. 33. Kim BD, Ver Halen JP, Lim S, Kim JYS. Predictors of 61 unplanned readmission cases in microvascular free tissue transfer patients: Multi-institutional analysis of 774 patients. Microsurgery. 2014. doi:10.1002/micr.22230. 34. Khanavin N, Mlodinow A, Kim JYS, et al. Predictors of 30-day readmission after outpatient thyroidectomy. An analysis of the 2011 NSQIP dataset. Am J Otolaryngol. 2014;35:332–9. 35. Available at: site.acsnsqip.org/participant-use-data-file/. Accessed 14 Mar 2014. 36. Shiloach M, Frencher SK Jr, Steeger JE, et al. Toward robust information: data quality and inter-rater reliability in the American College of Surgeons National Surgical Quality Improvement Program. J Am Coll Surg. 2010;210:6–16. 37. Sellers MM, Merkow RP, Halverson A, et al. Validation of new readmission data in the American College of Surgeons National Surgical Quality Improvement Program. J Am Coll Surg. 2013;216(3):420–7. 38. Henderson WG, Daley J. Design and statistical methodology of the National Surgical Quality Improvement Program: why is it what it is?. Am J Surg. 2009;198(5 Suppl):S19–27.
N. Khavanin et al. 39. Austin PC. Some methods of propensity-score matching had superior performance to others: results of an empirical investigation and Monte Carlo simulations. Biom J. 2009;51(1):171–84. 40. Mioton LM, Buck DW 2nd, Gart MS, et al. A multivariate regression analysis of panniculectomy outcomes: does plastic surgery training matter? Plast Reconstr Surg. 2013;131(4):604e–12e. 41. Austin PC. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm Stat. 2011;10(2):150–61. 42. Thoemmes F. Propensity score matching in SPSS 2012. Available at: http://arxiv.org/ftp/arxiv/papers/1201/1201.6385.pdf. Accessed 12 Dec 2013. 43. Sun GH, DeMonner S, Davis MM. Epidemiological and economic trends in inpatient and outpatient thyroidectomy in the United States, 1996-2006. Thyroid. 2013;23(6):727–33.
44. Doran HE, England J, Palazzo F. Questionable safety of thyroid surgery with same day discharge. Ann R Coll Surg Engl. 2012;94:543–7. 45. Mowschenson PM, Hodin RA. Outpatient thyroid and parathyroid surgery: a prospective study of feasibility, safety, and costs. Surgery. 1995;118:1051–3; discussion 1053–4. 46. Materazzia G, Dionigic G, Bertia P, et al. One-day thyroid surgery: retrospective analysis of safety and patient satisfaction on a consecutive series of 1,571 cases over a three-year period. Eur Surg Res. 2007;39:182–8. 47. Doran HE, Palazzo F. Ambulatory thyroid surgery: do the risks overcome the benefits? Presse Med. 2014;43(3):291–6. 48. Stavrakis AI, Ituarte PH, Ko CY, Yeh MW. Surgeon volume as a predictor of outcomes in inpatient and outpatient endocrine surgery. Surgery. 2007;142(6):887–99.
