|
 
 |
| ORIGINAL ARTICLE |
|
| Year : 2019 | Volume
: 7
| Issue : 3 | Page : 119-127 |
|
Predictors of postoperative complications in retroperitoneal sarcoma surgery
Mohammed O Nassif
Department of Surgery, King Abdulaziz University, Jeddah, Saudi Arabia
| Date of Web Publication | 4-Nov-2019 |
Correspondence Address:
Mohammed O Nassif
Department of Surgery, King Abdulaziz University, Jeddah
Saudi Arabia
 Source of Support: None, Conflict of Interest: None  | Check |
DOI: 10.4103/ssj.ssj_28_19
Introduction: Retroperitoneal sarcomas (RPSs) are large in size and often involve adjacent organs or vital structures. Completeness of resection is critical for long-term survival; however, this often involves extensive surgeries. This study aimed to identify predictors of early severe postoperative complications after RPS surgery.
Methodology: In patients who underwent surgery for RPS, intraoperative variables and patient characteristics were assessed to determine predictors for severe postoperative complications.
Results: Two hundred and thirty-three patients were included. In comparison to patients who had no comorbidity, those with one or more comorbidities were more likely to have postoperative complications (odds ratio [OR]: 2.38; confidence interval [CI]: 1.03–5.48). Patients who avoided admission to the intensive care unit (ICU) within 24 h of surgery had less complications postoperatively (OR: 0.08; CI: 0.02–0.30). Multiple organ resection during surgery and patients' age had no impact on the occurrence of severe complications.
Conclusion: This study showed that a high patient comorbidity index, male gender, and early admission to the ICU were independently associated with an increased risk of postoperative severe complications. However, the age of the patient and degree of surgical resection had no impact on this occurrence. These findings suggest that age and extent of resection should not be used as a sole determinant of patient's eligibility for curative surgery. Keywords: Elderly, postoperative morbidity and mortality, retroperitoneal sarcoma
How to cite this article:
Nassif MO. Predictors of postoperative complications in retroperitoneal sarcoma surgery. Saudi Surg J 2019;7:119-27 |
| Introduction | |  |
Sarcomas are rare neoplasms that account for <1% of all adult cancers, and retroperitoneal sarcomas (RPSs) represent only 9% of all soft-tissue malignancies.[1] Based on the data collected from 1973 to 2008 by the Surveillance, Epidemiology, and End Results program of the National Cancer Institute,[2] the annual incidence of RPS was approximately 2.7 cases/million of population. The mean age of diagnosis was 64 years with women and men having similar incidence.
Due to the rareness of this disease, there is a lack of established evidence-based treatment guidelines. That being said, there are many aspects that are extrapolated from large trials on extremity sarcomas and are universally agreed upon. The National Comprehensive Cancer Network summarizes these in its yearly guidelines.[3] At present, surgical resection with negative margins is the only potentially curative treatment for RPS.[4] Obtaining a negative margin at the time of surgery is the most important survival prognostic facto.[4],[5],[6],[7],[8],[9],[10],[11],[12],[13],[14] These surgeries are often extensive and require multiple organ resections.[15],[16],[17]
Given the lack of preoperative guidelines for patient selection criteria and studies concerning postoperative complications, it is important to study population-based RPS perioperative data as opposed to institution based. The Province of Quebec has a universal health insurance program (Regie de l'assurance-maladie du Quebec [RAMQ]) and maintains administrative database on every insured resident including data from all claims for medical services rendered. Of the population of Quebec (7.75 million), 97.8% have provincial health insurance.[18],[19] This supplies us with a convenient chance to study predictive factors associated with postoperative complications for rare diseases. The objective of this study was to identify risk factors that predict the occurrence of early postoperative severe complications after RPS surgery in the Province of Quebec over the period limited by the database available between 1993 and 2007. All-cause mortality and return to the operating room (OR) within 30 days were also examined.
| Methodology | | |
Settings and data sources
A random population-based sample of over 5 million insured persons has been developed by the Clinical and Health Informatics Research Group, McGill University, Montreal Canada, to assess various components of health-care quality and physician competency. This resource has already supported a number of high-impact studies.[20],[21],[22],[23],[24] It includes information collected on each patient from the following data sources maintained by the RAMQ.
- Registrant database – Includes date of birth, gender, first three digits of the postal code, insurance eligibility coverage periods, and if applicable, date of death
- Med-Echo database – Includes hospitalization dates, principal diagnosis and up to 10 secondary diagnoses, and intramural procedures and transfers.
Medical service claims database
It includes information on all services provided by physicians under the fee-for-service payment model. This includes information on diagnosis (International Classification of Diseases-9th edition [ICD]-9), procedure (service code), date, location, physician specialty, physician identification, and patient identification. This database was used as the main pool that the cohort in this study was selected from and has been validated in a previous study. It has been found that if a fee-for-service billing is the prevalent method of payment, data collected form billing claims by the physicians are reliable and valid. It could be reliably used to be study prevalence of an occurrence and its associated risk factors.[25] Additional demographic information was obtained from census data from Statistics Canada and linked using postal codes.[26] RPS diagnosis was ascertained using the ICD codes, ICD-9, and the RPS surgery service codes. The index date was the date of the RPS surgery. Appropriate ethical clearance was obtained.
Cohort selection
From the random population-based cohort described above, a subcohort was built, based on the following inclusion criteria:
- Adults – Patients over the age of 18 years were identified by their date of birth and the index date (the index date used was the first RPS excision surgery done)
- Incidental RPS – Newly diagnosed RPS, recognized by a claim with an ICD-9 codes for RPS and verified [Appendix A]
- Diagnosis date – All diagnoses between 1993 and 2007
- Underwent RPS surgery – Identified through claim codes that were verified by specialized surgeons who operate on RPS regularly
- With 1 year of medical history – Availability of administrative claims data for 365 days before the index date. Reason was to rule out any previous diagnosis with RPS
- With 1-month follow-up – Availability of administrative claims data for a minimum of 30 days following the index date. Reason is to find all complications within a month of the index date. Patients who die before the completion of 30 days are also included.
Two groups of patients were excluded from the cohort:
- Pregnant patients – Patients who had pregnancy ICD-9 codes in claims within 9 months before the index date
- Patients with nonrelated diseases – These include repeated ICD-9 codes of sarcomas elsewhere in the body, retroperitoneal cancers of solid organs, and other retroperitoneal benign diseases. Reason was to validate the cohort by having a high specificity cutoff.
A stepwise approach was taken when building the cohort [Figure 1]. Each step was established after extracting multiple random patient samples before the exclusion and manually searching their codes to verify that the path was correct.
Independent analytic variables
Patient related
- Age – Age at the index date calculated using the date of birth from the registrant database
- Sex – Sex of the patient, male or female, from the registrant database
- Place of residence – The municipalities of Quebec were divided into rural and nonrural according to Statistics Canada's definition of rural census as “the population outside settlements with 1,000 or more population with a population density of 400 or more inhabitants per square kilometer,” Statistics Canada, 2007[27]
- Income and education estimates – Census data were used to measure the mean family income and educational level percentage in the patients' residential area as described by Wilkins.[20],[28] Ten (4.3%) patients had missing data on one or more demographic variables (postal code, income, or education) at the time of their index date. By carrying forward or backward of the nonmissing values (i.e., using single imputation methods [29]) for adjacent years for each patient, data were imputed for five patients. For the remaining five patients, values for adjacent years were not available and thus the mean of the most frequently appearing value and the year and area was used to replace the missing data
- Distance to cancer center with sarcoma program – The mean latitude and longitude was calculated for each patient's forward sortation area (FSA) (postal code area) and used to define the center of the FSA. The data on the geographical coordinates of all postal codes were obtained from Statistics Canada's 2006 postal code conversion file.[30] These centers were then used to calculate the distance between it and the coordinates of cancer centers with a sarcoma program. These are centers that treat sarcoma patients, have multidisciplinary teams, and deliver chemo/radiation therapy
- Charlson Comorbidity Index (CCI) – The Charlson score was determined using the enhanced ICD-9-CM coding algorithms for Charlson comorbidities [31] with original Charlson weights [32] in the 365 days before the index date. Our database only contained four-digit diagnostic codes and all the 5th digits from the diagnostic codes used by the enhanced algorithm were ignored. However, similar to the Dartmouth–Manitoba algorithm, severity of comorbidities was adjusted to avoid double-counting one chronic condition that may be characterized using multiple diagnosis codes.[33] For example, a patient with metastatic solid tumor had the code for metastasis marked, while no code was inserted for the specific primary organ malignancy. Moderate-to-severe liver disease and complicated diabetes were treated the same way. Since every patient in the cohort had sarcoma, the malignancy points were given only to patients that had other previous malignancies. The patients were then divided according to their score into groups: 0 score (no comorbidity, healthy), 1–2 score (low CCI), and 3 or more score (high CCI).[34]
Perioperative
- Hospital setting – Defined as academic or nonacademic, based on assignations from the Ministry of Health
- Specialty of treating surgeon – Specialty codes were provided from billing claims data. In the case of more than one surgeon having billed for the index procedure, the principal role was assigned to the one who was remunerated the maximum amount as opposed to others who received differential pay, based on the assistant's roles. Recognized medical specialties were identified from service claims at the day of the surgery and they were grouped into general surgery or other
- Planned intensive care unit (ICU) admission – Patients who were admitted to the ICU were identified by billing codes with the ICU as a location for the billing. If these patients were admitted within the first 24 h window after surgery, they were considered an elective ICU admission
- Additional organs resected at surgery – Additional claims billed on the index date were reviewed. Those that corresponded to surgical procedures were classified according to the system/organ involved. Any surgery from one of the groups was counted as one additional organ resected
- Length of stay (LOS) – The Med-Echo hospitalization database was used to identify the number of days the patient was admitted to the hospital after the index date.
Outcome
- All-cause mortality – Date of death was captured from the registrant database for patient who dies within 30 days of the index date
- Return to the OR – Patients who returned to the OR were identified by the occurrence of a surgical billing claim in the operative room within 30 days of the index date
- Severe postoperative complications – Severe complications were defined as those that were grades III, IV, and V according to the Clavien morbidity and mortality classification.[35] Grade III included patients who had a complication requiring surgery, endoscopy, or interventional radiology with or without general anesthesia within 30 days of surgery. In addition to patients who returned to the OR “see above,” specific codes for endoscopic procedures and interventional radiology were identified within 30 days of the index date. Grade IV were patients who were admitted to the ICU after 24 h from the index date and within 30 days after. They were identified by claims that occurred in the ICU or coronary care unit during that period. Grade V were patients who died within 30 days of surgery “see above.”
Assessment of outcomes
Two main outcomes were identified from the databases: all-cause mortality and severe complications within 30 days of RPS. Severe complications were defined as those that were grades III, IV, and V according to the Clavien morbidity and mortality classification.[35]
Statistical analysis
Age, income, education, distance, Charlson index, distance, and intraoperative variables were categorized on the basis of their frequency distributions in the cohort. Univariate analysis was used to assess the unadjusted correlations between each predictor and the outcome (postoperative complications). Subsequently, a multivariate logistic regression was constructed. Because all patients had unique surgeons, adjusting for clustering of patients within physicians was not necessary. Collinearity between variables was assessed via a Pearson's correlation matrix, and subsequently, distance, income, and ICU admissions were left out of the model due to their correlation (coefficient of ≥0.25) with residence, secondary education, and Charlson index, respectively. Selected interaction terms were introduced in the model one at a time, but none was found significant. All P values were for two-tailed tests with statistical significance defined as P ≤ 0.05. SAS software (SAS version 9.2, Institute, Inc., Cary, North Carolina, USA) was used for all analyses.[36]
| Results | | |
From an initial 5,212,853 patients in the database, 233 patients met our inclusion/exclusion criteria [Figure 1]. These patients were over the age of 18 years and who have undergone surgery for verified newly diagnosed RPS between January 1993 and March 2007. One patient was excluded from the cohort due to pregnancy.
Descriptive statistics of this cohort are listed in [Table 1]. The mean age was 57.6 years (standard deviation, 14.7), of whom 52.8% were male. The majority of patients (60.5%) were living in an urban setting, mainly in the metropolitan areas of Montreal and Quebec city. Two-thirds of the patients lived in the areas where the average individual overall income was between 30,000 and 60,000 Canadian Dollars (CAD) per year, whereas only 9.4% lived in impoverished areas with an average household income of <30,000 CAD per year. The majority of patients (47.6%) lived in the areas that had a population consisting of 60%–80% who were at least high school graduates. Only 1.3% of patients lived in the areas that had an average of <40% high school graduates. Two-thirds of the patients lived within 50 km of a cancer center with a sarcoma program and only 23.2% had to travel more than 100 km to seek treatment. In terms of comorbidity, 33.1% of patients were found to have a CCI score of 0 at the time of surgery, whereas 27.9% had a score of 1–2 and 39.1% had a score of 3 or more.
As for perioperative findings, it was found that the university hospitals were the center of choice for the majority of patients (87.5%) and only 12.5% were treated in nonacademic hospitals [Table 2]. The primary surgeon was a general surgeon in 76.47% of cases, with urologist staffing 9.27% of cases. Two-thirds of the patients had only a RPS resection without any additional organs removed during surgery, 26.6% had 1–2 organs resected, and 6.9% had three or more. An elective ICU admission during the first 24 h postoperatively was arranged for 24 (10%) patients. The average LOS in the hospital postoperatively was 12.6 days, median 9 days, and the mode 7 days. LOS ranged between 1 day (death) and 135 days and only 17 patients had LOS above 30 days, and if excluded, the average of the remaining was 9.5 days.
Severe postoperative complications within 30 days [Table 3] occurred in a third of the patients, of these 7 (3%) patients died. However, 52 (22%) patients had a Grade IV complication and were required to have an emergency admission to the ICU. Only 19 (8%) patients had a Grade III complication without an ICU admission. The OR was revisited by 35 (15%) patients within 30 days postoperatively and was considered a complication.
Multivariate analysis using a generalized estimating equation [Table 4] showed that, in comparison to patients who had no comorbidity, those with a CCI score of ≥3 were more likely to have early severe postoperative complications (odds ratio [OR]: 2.58; confidence interval [CI]: 1.05–6.35). Those with a CCI score of 0 or a low score of 1–2 showed no significant relationship. In addition, patients who lived in the areas with an average higher income had a less change of having complications (OR: 0.82; CI: 0.67–0.99). Furthermore, the distance between patients address and the nearest cancer center with a sarcoma program had no effect on complication rate.
Patients who avoided admission to the ICU within 24 h of surgery had less complications postoperatively (OR: 0.07; CI: 0.02–0.25). Male patients had a higher risk of complications (OR: 2.4; CI: 1.05–5.48). Multiple organs resected during the RPS surgery had no significant effect on the occurrence of complications (OR: 2.09; CI: 0.97–4.5). Surprisingly, the increase in patients' age as well had no impact on the occurrence of early severe complications postoperatively (OR: 0.89; CI: 0.68–1.16).
| Discussion | | |
There is an increased demand for reliable and informative studies on postoperative outcome and prognosis, both by the health professionals and by the public. A better quality of health care delivered with a lower cost is a shared goal by everyone. This study addresses early severe postoperative complications and mortality after patients underwent curative RPS surgery. Severe postoperative complications occurred in a third of the patients and 7 (3%) patients died. This percentage is similar to previous studies of patients treated for RPS, which showed that 30%–35% had early postoperative severe complications and 5%–6% died within a month.[37]
Predictors of risk factors on early postoperative complications after RPS were also identified. Overall, it was found that high patient comorbidity index, admission to the ICU within 24 h of surgery, male sex, and low average house hold income predicted a higher rate of postoperative severe complications. However, the age of the patient, the distance from a cancer center with a sarcoma program, low CCI, and degree of surgical resection had no impact on this occurrence.
The relationship between preoperative comorbidity levels and postoperative complications is expected and has been well documented in previous studies with other surgeries.[38],[39],[40],[41] Although we have found no statistical significance between patients who had a CCI score of 1–2 and postoperative risk, the cutoff between this group and ≥3 CCI is not very big. Any comorbidity is believed to affect outcomes in surgery and CCI of 3 should not be taken as a magic number.
ICU requirement during the early postoperative period has been linked to early postoperative deaths.[42] Many major surgeries are initially observed in the ICU postoperatively as a precaution. A study of 148 patients admitted postliver resection in the ICU showed that 88.1% of patients only received closed monitoring during their stay and the majority of patients who were treated in the ICU had either bled significantly in the OR or underwent an extended resection of the liver.[43] Many patients are being admitted to the ICU unnecessarily causing an increased demand of beds and a large waste resources,[44],[45] not to mention the effects on the patients.[46] Many surgeries have selective criteria for admission to the ICU;[47],[48],[49] this can be done for RPS surgery as well.
In regard to male sex having a higher risk of severe postoperative complications, previous studies have shown associations between male sex and worse outcome in other surgeries (neurosurgery interventions, bile duct exploration, and colorectal surgery [50],[51],[52],[53]). The reasons are still unknown, although some speculate that the protective role of estrogen might contribute to faster recovery and less complications.[50],[54]
Another point to add is that, even in a health-care system that is universally available and free for all residents, there was a difference in complication rates between patients living in low and high income areas. While we consider the health system in Quebec tiered-free, this could be due to other factors not accounted for in this study (e.g., social well-being, family support, and general knowledge of the disease). That being said, many studies have shown an association between low socioeconomic status and worse outcomes in cancer patients' status.[55],[56],[57],[58]
What comes as a surprise is that the age of the patients and number of organs resected did not predict a higher postoperative severe complication rate. Assuming that many oncologists would put plenty of weight on the patients' age and the expected extent of resection when making their treatment decision,[59],[60] it is fair to conclude that some patients who were not candidates for surgery might have benefited from it and vice versa.
| Limitations | | |
To begin, this database was derived from a registration data obtained from patients' charts and designed for administrative purposes. In addition, the clerical personnel who are responsible for entering these data are usually not physicians themselves. As a result, medical conditions might have been overlooked, and consequently, the relative contribution of these factors to postoperative death might have been underestimated.[61] However, given the rarity of this disease, having such a large number of cases to review is beneficial.
Second, there were limitations to the amount of information on each patient collected and available in this dataset. There might have been additional (potential) confounding factors that were not available in the database, including the intent to cure versus debulking, the clinical condition of the patient before the procedure, and pharmacotherapy. Therefore, residual confounding of other factors not available to us may still exist.
Third, as per the design, analyses were restricted to patients who underwent surgery. No information was included from patients who were screened, but did not undergo surgery because their mortality risk was perceived as prohibitive. Obviously, exclusion of patients at risk of adverse outcomes might have diluted estimates of relative risk.
Fourth, the database from 1993 to 2007 would be considered somewhat old, but the conclusion still stands the same as the surgical approach and management of RPS has not differed since then.
| Conclusion | | |
Regardless of a surgeon's capabilities and experience with RPS, many will face postoperative complications routinely in their everyday management of these patients. These complications have a significant impact on patients and health systems, yet our knowledge on appropriate patient selection and incidence of early postoperative complications is still limited. This study shines the light and some of the risk factors of these complications. Better understanding of them can be an aid to decision-makers in their selection of candidates, preoperative preparation, meticulous surgeries, and postoperative readiness in anticipation of any postoperative complication.
Acknowledgments
The Canadian Cancer Society Research Institute, Canadian Institutes of Health Research, and Fonds de la recherche en Santé du Québec provided funding for the database used for this study. King Abdulaziz University provided funds for training at McGill University. My research was conducted during training at Dr. Ari Megurditchian's research lab and analysis was conducted by Mrs. Stanimira Krotneva.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
| References | | |
| 1. | Bethesda MD, Howlader N, Noone AM, Krapcho M, Neyman N, Aminou R, Waldron W, et al. SEER Cancer Statistics Review, 1975-2008. Bethesda, MD: National Cancer Institute;2011.  |
| 2. | Siegel R, Ward E, Brawley O, Jemal A. Cancer statistics, 2011: The impact of eliminating socioeconomic and racial disparities on premature cancer deaths. CA Cancer J Clin 2011;61:212-36. |
| 3. | |
| 4. | Stoeckle E, Coindre JM, Bonvalot S, Kantor G, Terrier P, Bonichon F, et al. Prognostic factors in retroperitoneal sarcoma: A multivariate analysis of a series of 165 patients of the French Cancer Center Federation Sarcoma Group. Cancer 2001;92:359-68. |
| 5. | Catton CN, O'Sullivan B, Kotwall C, Cummings B, Hao Y, Fornasier V, et al. Outcome and prognosis in retroperitoneal soft tissue sarcoma. Int J Radiat Oncol Biol Phys 1994;29:1005-10. |
| 6. | Lewis JJ, Leung D, Woodruff JM, Brennan MF. Retroperitoneal soft-tissue sarcoma: Analysis of 500 patients treated and followed at a single institution. Ann Surg 1998;228:355-65. |
| 7. | Hassan I, Park SZ, Donohue JH, Nagorney DM, Kay PA, Nasciemento AG, et al. Operative management of primary retroperitoneal sarcomas: A reappraisal of an institutional experience. Ann Surg 2004;239:244-50. |
| 8. | Pierie JP, Betensky RA, Choudry U, Willett CG, Souba WW, Ott MJ, et al. Outcomes in a series of 103 retroperitoneal sarcomas. Eur J Surg Oncol 2006;32:1235-41. |
| 9. | Anaya DA, Lev DC, Pollock RE. The role of surgical margin status in retroperitoneal sarcoma. J Surg Oncol 2008;98:607-10. |
| 10. | Lehnert T, Cardona S, Hinz U, Willeke F, Mechtersheimer G, Treiber M, et al. Primary and locally recurrent retroperitoneal soft-tissue sarcoma: Local control and survival. Eur J Surg Oncol 2009;35:986-93. |
| 11. | Puri A, Gulia A. Multimodality management of extremity soft tissue sarcomas – An Indian perspective. Indian J Surg Oncol 2011;2:291-7. |
| 12. | Bhangu AA, Beard JA, Grimer RJ. Should soft tissue sarcomas be treated at a specialist centre? Sarcoma 2004;8:1-6. |
| 13. | Gustafson P, Dreinhöfer KE, Rydholm A. Soft tissue sarcoma should be treated at a tumor center. A comparison of quality of surgery in 375 patients. Acta Orthop Scand 1994;65:47-50. |
| 14. | Rydholm A. Improving the management of soft tissue sarcoma. Diagnosis and treatment should be given in specialist centres. BMJ 1998;317:93-4. |
| 15. | Nathan H, Raut CP, Thornton K, Herman JM, Ahuja N, Schulick RD, et al. Predictors of survival after resection of retroperitoneal sarcoma: A population-based analysis and critical appraisal of the AJCC staging system. Ann Surg 2009;250:970-6. |
| 16. | Gronchi A, Casali PG, Fiore M, Mariani L, Lo Vullo S, Bertulli R, et al. Retroperitoneal soft tissue sarcomas: Patterns of recurrence in 167 patients treated at a single institution. Cancer 2004;100:2448-55. |
| 17. | Schwarzbach MH, Hormann Y, Hinz U, Leowardi C, Böckler D, Mechtersheimer G, et al. Clinical results of surgery for retroperitoneal sarcoma with major blood vessel involvement. J Vasc Surg 2006;44:46-55. |
| 18. | |
| 19. | |
| 20. | Tamblyn R, Abrahamowicz M, Dauphinee WD, Hanley JA, Norcini J, Girard N, et al. Association between licensure examination scores and practice in primary care. JAMA 2002;288:3019-26. |
| 21. | Tamblyn R, Abrahamowicz M, Brailovsky C, Grand'Maison P, Lescop J, Norcini J, et al. Association between licensing examination scores and resource use and quality of care in primary care practice. JAMA 1998;280:989-96. |
| 22. | Wenghofer E, Klass D, Abrahamowicz M, Dauphinee D, Jacques A, Smee S, et al. Doctor scores on national qualifying examinations predict quality of care in future practice. Med Educ 2009;43:1166-73. |
| 23. | Tamblyn R, Abrahamowicz M, Dauphinee D, Wenghofer E, Jacques A, Klass D, et al. Physician scores on a national clinical skills examination as predictors of complaints to medical regulatory authorities. JAMA 2007;298:993-1001. |
| 24. | Cadieux G, Tamblyn R, Dauphinee D, Libman M. Predictors of inappropriate antibiotic prescribing among primary care physicians. CMAJ 2007;177:877-83. |
| 25. | Tamblyn R, Reid T, Mayo N, McLeod P, Churchill-Smith M. Using medical services claims to assess injuries in the elderly: Sensitivity of diagnostic and procedure codes for injury ascertainment. J Clin Epidemiol 2000;53:183-94. |
| 26. | Pampalon R, Hamel D, Gamache P, Raymond G. A deprivation index for health planning in Canada. Chronic Dis Can 2009;29:178-91. |
| 27. | |
| 28. | Wilkins R. Use of postal codes and addresses in the analysis of health data. Health Rep 1993;5:157-77. |
| 29. | Zhang S, Liao J, Zhu Z. A SAS ® Macro for Single Imputation. In: Proceedings of the Annual Pharmaceutical Industry SAS Users Group. Atlanta, GA, USA; 2008. Available from: https://www.statcan.gc.ca/eng/start. [Last accessed on 2019 Oct 01]. |
| 30. | |
| 31. | Quan H, Sundararajan V, Halfon P, Fong A, Burnand B, Luthi JC, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care 2005;43:1130-9. |
| 32. | Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chronic Dis 1987;40:373-83. |
| 33. | Romano PS, Roos LL, Jollis JG. Adapting a clinical comorbidity index for use with ICD-9-CM administrative data: Differing perspectives. J Clin Epidemiol 1993;46:1075-9. |
| 34. | Radley DC, Gottlieb DJ, Fisher ES, Tosteson AN. Comorbidity risk-adjustment strategies are comparable among persons with hip fracture. J Clin Epidemiol 2008;61:580-7. |
| 35. | Dindo D, Demartines N, Clavien PA. Classification of surgical complications: A new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004;240:205-13. |
| 36. | |
| 37. | Koenig AM, Reeh M, Burdelski CM, Wengert C, Gawad KA, Izbicki JR, et al. Long-term results of primary and secondary resections in patients with retroperitoneal soft tissue sarcoma. Langenbecks Arch Surg 2012;397:1251-9. |
| 38. | Campbell PG, Yadla S, Nasser R, Malone J, Maltenfort MG, Ratliff JK, et al. Patient comorbidity score predicting the incidence of perioperative complications: Assessing the impact of comorbidities on complications in spine surgery. J Neurosurg Spine 2012;16:37-43. |
| 39. | Calvillo-King L, Xuan L, Zhang S, Tuhrim S, Halm EA. Predicting risk of perioperative death and stroke after carotid endarterectomy in asymptomatic patients: Derivation and validation of a clinical risk score. Stroke 2010;41:2786-94. |
| 40. | Wirth M, Fröhner M. The significance of comorbidity and age in radical prostatectomy. Urologe A 2004;43:935-41. |
| 41. | Froehner M, Koch R, Litz RJ, Oehlschlaeger S, Twelker L, Hakenberg OW, et al. Detailed analysis of Charlson comorbidity score as predictor of mortality after radical prostatectomy. Urology 2008;72:1252-7. |
| 42. | Azoulay E, Adrie C, De Lassence A, Pochard F, Moreau D, Thiery G, et al. Determinants of postintensive care unit mortality: A prospective multicenter study. Crit Care Med 2003;31:428-32. |
| 43. | Kim SH, Lee JG, Kwon SY, Lim JH, Kim WO, Kim KS, et al. Is close monitoring in the intensive care unit necessary after elective liver resection? J Korean Surg Soc 2012;83:155-61. |
| 44. | Angus DC, Shorr AF, White A, Dremsizov TT, Schmitz RJ, Kelley MA, et al. Critical care delivery in the United States: Distribution of services and compliance with leapfrog recommendations. Crit Care Med 2006;34:1016-24. |
| 45. | Miranda DR, Rivera-Fernández R, Nap RE. Critical care medicine in the hospital: Lessons from the EURICUS-studies. Med Intensiva 2007;31:194-203. |
| 46. | Reich M, Rohn R, Lefevre D. Surgical intensive care unit (ICU) delirium: A “psychosomatic” problem? Palliat Support Care 2010;8:221-5. |
| 47. | To EW, Tsang WM, Lai EC, Chu MC. Retrospective study on the need of intensive care unit admission after major head and neck surgery. ANZ J Surg 2002;72:11-4. |
| 48. | Bertges DJ, Rhee RY, Muluk SC, Trachtenberg JD, Steed DL, Webster MW, et al. Is routine use of the intensive care unit after elective infrarenal abdominal aortic aneurysm repair necessary? J Vasc Surg 2000;32:634-42. |
| 49. | Hobart DC, Nicholas GG, Reed JF 3 rd, Nastasee SA. Carotid endarterectomy outcomes research: Reduced resource utilization using a clinical protocol. Cardiovasc Surg 2000;8:446-51. |
| 50. | McArdle CS, McMillan DC, Hole DJ. Male gender adversely affects survival following surgery for colorectal cancer. Br J Surg 2003;90:711-5. |
| 51. | Hendifar A, Yang D, Lenz F, Lurje G, Pohl A, Lenz C, et al. Gender disparities in metastatic colorectal cancer survival. Clin Cancer Res 2009;15:6391-7. |
| 52. | Davis MC, El-Sayed AM, Ziewacz JE, Jayachandran P, Geisert WC, Zamora-Berridi GJ, et al. Sex disparities in postoperative outcomes after neurosurgical intervention: Findings from the UMEND project. Neurosurgery 2012;70:959-64. |
| 53. | Noble H, Whitley E, Norton S, Thompson M. A study of preoperative factors associated with a poor outcome following laparoscopic bile duct exploration. Surg Endosc 2011;25:130-9. |
| 54. | Buchanan FF, Myles PS, Cicuttini F. Effect of patient sex on general anaesthesia and recovery. Br J Anaesth 2011;106:832-9. |
| 55. | Harper S, Lynch J, Meersman SC, Breen N, Davis WW, Reichman MC, et al. Trends in area-socioeconomic and race-ethnic disparities in breast cancer incidence, stage at diagnosis, screening, mortality, and survival among women ages 50 years and over (1987-2005). Cancer Epidemiol Biomarkers Prev 2009;18:121-31. |
| 56. | Freeman HP. Patient navigation: A community based strategy to reduce cancer disparities. J Urban Health 2006;83:139-41. |
| 57. | Ward E, Halpern M, Schrag N, Cokkinides V, DeSantis C, Bandi P, et al. Association of insurance with cancer care utilization and outcomes. CA Cancer J Clin 2008;58:9-31. |
| 58. | Roetzheim RG, Gonzalez EC, Ferrante JM, Pal N, Van Durme DJ, Krischer JP, et al. Effects of health insurance and race on breast carcinoma treatments and outcomes. Cancer 2000;89:2202-13. |
| 59. | Turner NJ, Haward RA, Mulley GP, Selby PJ. Cancer in old age – Is it inadequately investigated and treated? BMJ 1999;319:309-12. |
| 60. | Leonard RC, Barrett-Lee PJ, Gosney MA, Willett AM, Reed MW, Hammond PJ, et al. Effect of patient age on management decisions in breast cancer: Consensus from a national consultation. Oncologist 2010;15:657-64. |
| 61. | Iezzoni LI. Using administrative diagnostic data to assess the quality of hospital care. Pitfalls and potential of ICD-9-CM. Int J Technol Assess Health Care 1990;6:272-81. |
[Figure 1]
[Table 1], [Table 2], [Table 3], [Table 4]
|
Search Pubmed for