Accepted Manuscript Are Procedures Codes in Claims Data a Reliable Indicator of Intraoperative Splenic Injury Compared to Clinical Registry Data? Anne M. Stey, MD, MSc Clifford Y. Ko, MD, MS, MSHS, FACS Bruce Lee Hall, MD PhD, MBA, FACS Rachel Louie, MS Elise H. Lawson, MD, MSHS Melinda M. Gibbons, MD, MSHS, FACS David S. Zingmond, MD, PhD Marcia M. Russell, MD, FACS PII:
S1072-7515(14)00322-6
DOI:
10.1016/j.jamcollsurg.2014.02.029
Reference:
ACS 7367
To appear in:
Journal of the American College of Surgeons
Received Date: 10 November 2013 Revised Date:
24 February 2014
Accepted Date: 25 February 2014
Please cite this article as: Stey AM, Ko CY, Lee Hall B, Louie R, Lawson EH, Gibbons MM, Zingmond DS, Russell MM, Are Procedures Codes in Claims Data a Reliable Indicator of Intraoperative Splenic Injury Compared to Clinical Registry Data?, Journal of the American College of Surgeons (2014), doi: 10.1016/j.jamcollsurg.2014.02.029. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT 1
Are Procedures Codes in Claims Data a Reliable Indicator of Intraoperative Splenic Injury Compared to Clinical Registry Data?
RI PT
Anne M Stey, MD, MSc (a, b), Clifford Y Ko, MD, MS, MSHS, FACS (b, c, d), Bruce Lee Hall, MD PhD, MBA, FACS (c, e), Rachel Louie, MS (b), Elise H Lawson MD, MSHS (b), Melinda M Gibbons. MD, MSHS, FACS (b), David S Zingmond MD, PhD (b), Marcia M Russell MD,
SC
FACS (b, d)
M AN U
(a) Icahn School of Medicine at Mount Sinai Medical Center, NY, NY;
(b) David Geffen School of Medicine, University of California Los Angeles, Los Angeles, CA; (c) American College of Surgeons, Chicago, IL;
(d) VA Greater Los Angeles Healthcare System, Los Angeles, CA;
TE D
(e) Washington University in St Louis, Department of Surgery, Olin Business School, and Center for Health Policy; St Louis VA Medical Center; BJC Healthcare St Louis.
EP
Disclosure Information: Nothing to disclose.
Presented at the American College of Surgeons 99th Annual Clinical Congress, Washington, DC,
AC C
October 2013.
Support: Dr Stey’s time was supported for this publication by The Robert Wood Johnson Foundation Clinical Scholars program and the U.S. Department of Veterans Affairs.
Correspondence address:
Anne M. Stey 10940 Wilshire Blvd, Suite 710 Los Angeles, CA 90024. Phone: 310-794-2507
ACCEPTED MANUSCRIPT 2
AC C
EP
TE D
M AN U
SC
RI PT
Fax: 310-794-3288 Email:
[email protected] Running Head: Procedure Codes Capture Iatrogenic Injury
ACCEPTED MANUSCRIPT 3
ABSTRACT Background: Identifying iatrogenic injuries using existing data sources is important for improved transparency regarding the occurrence of intraoperative events. There is evidence that
RI PT
procedure codes are reliably recorded in claims data. The objective of this study was to assess whether concurrent splenic procedure codes in patients undergoing colectomy procedures are reliably coded in claims data as compared to clinical registry data.
SC
Study Design: Patients who underwent colectomy procedures in the absence of neoplastic diagnosis codes were identified from American College of Surgeons National Surgical Quality
M AN U
Improvement Program (ACS-NSQIP) data linked with Medicare inpatient claims data file (20052008). A kappa statistic was used to assess coding concordance between ACS-NSQIP and Medicare inpatient claims, with ACS-NSQIP as the reference standard. Results: A total of 11,367 colectomy patients were identified from 212 hospitals. There were
TE D
114 patients (1%) who had a concurrent splenic procedure code recorded in either ACS-NSQIP or Medicare inpatient claims. There were 7 patients who had a splenic injury diagnosis code recorded in either data source. Agreement of splenic procedure codes between the data sources
EP
was substantial (kappa statistic 0.72, 95% CI 0.64- 0.79). Medicare inpatient claims identified 81% of the splenic procedure codes recorded in ACS-NSQIP, and 99% of the patients without a
AC C
splenic procedure code.
Conclusions: It is feasible to use Medicare claims data to identify splenic injuries occurring during colectomy procedures as claims data have moderate sensitivity and excellent specificity for capturing concurrent splenic procedure codes compared to ACS-NSQIP.
ACCEPTED MANUSCRIPT 4
INTRODUCTION The spleen is injured in 0.4-1% of colectomies, affecting a total of 6,000 patients annually in the United States.(1,2) Such iatrogenic injuries account for approximately 20% of all
RI PT
splenic procedures performed.(3-6) Patients subsequently have a 5% lifetime risk of
overwhelming post-splenectomy infection with an associated 38-70% mortality rate.(4,7)Despite these clinical implications, providers and payers have not focused on intraoperative splenic
SC
injury as a target for quality improvement. Yet, how tissue is handled and intraoperative injury to surrounding organs, such as the spleen, may be viewed as an indicator of poor technical
M AN U
quality.(8)
Iatrogenic splenic injury may be an identifiable adverse occurrence in administrative claims data because injury is often treated with an additional procedure to repair or remove the spleen. Administrative claims data have been used to measure quality for colectomy procedures
TE D
by identifying surgical complications based on the presence of diagnosis codes. However, this use of diagnosis codes for capturing surgical complications in claims data is limited by variability in coding.(9-13) Additionally, pay-for-performance programs that impose financial
EP
losses on providers based on the presence of surgical complication diagnosis codes may
AC C
encourage under-coding of these diagnoses, effectively decreasing transparency within the health system.(14-16) There is a need for more valid and reliable data elements than the current diagnosis codes if claims data are to be used for surgical quality measurement. One such data element that has shown promise in claims data is the procedure code. Previous studies using claims data have demonstrated that procedure codes may be more reliably coded than diagnosis codes. (17-19) Some of these authors have postulated that this difference in
ACCEPTED MANUSCRIPT 5
the reliability of coding may be due to the higher reimbursement potential associated with procedures.
RI PT
The overarching goal of this study was to analyze the reliability of using procedure codes recorded in Medicare inpatient claims data to identify intraoperative adverse events. We
hypothesized that concurrent splenic procedure codes recorded in Medicare inpatient claims data
SC
for patients undergoing colectomy reliably reflect the concurrent splenic procedures recorded in the American College of Surgeons National Surgical Quality Improvement Program (ACS-
M AN U
NSQIP). If procedure codes are as reliable for capturing intraoperative events as hypothesized, these codes could be a means of improving transparency surrounding intraoperative events. This increased transparency would allow for the health care system to measure and therefore focus on reducing the rates of iatrogenic intraoperative adverse events, such as splenic injury.
Data Source and Measures
TE D
METHODS
The American College of Surgeons’ National Surgical Quality Improvement Program
EP
(ACS-NSQIP) is a surgical clinical registry that collects high quality clinical data with dedicated trained clinical abstractors in participating hospitals. The Medicare inpatient claims data file is an
AC C
administrative dataset comprised of demographic data, as well as diagnoses and procedures billed to Medicare for all Medicare beneficiaries. Patient records accrued from the ACS-NSQIP clinical registry from 2005 to 2008 were linked to the Medicare inpatient claims data file from the same period using indirect patient identifiers and a deterministic linkage algorithm. The details of the linkage procedure are provided elsewhere.(20) As a result of the linked data source, the study sample was restricted to patients 65 years of age and older who underwent a colectomy at an ACS-NSQIP participating hospital from 2005 to 2008. Colectomy procedures were
ACCEPTED MANUSCRIPT 6
identified by Current Procedures Terminology (CPT) codes recorded in the ACS-NSQIP “CPT primary procedure” data field (Appendix 1, online only). The exclusion criterion was the presence of colon cancer diagnoses (due to possibility that the cancer could have invaded the
RI PT
spleen and therefore splenectomy may have been indicated) in any of the ten Medicare inpatient claims International Classification of Diseases, 9th edition (ICD-9) diagnosis data fields.
Additionally, colectomies in the setting of trauma admissions were not included in this dataset
SC
because ACS-NSQIP does not accrue any patients who undergo operations as a result of blunt or penetrating trauma.
M AN U
The primary occurrence queried was the presence of a concurrent splenic procedure code. This was determined in ACS-NSQIP based on the presence of designated CPT codes in any of ACS-NSQIP’s ten “concurrent procedure” or ten “other procedure” data fields. This was determined in Medicare inpatient claims based on the presence of designated ICD-9 procedure
TE D
codes in any of the six Medicare inpatient claims “procedure” data fields. CPT and ICD-9 codes used to identify splenic procedures are reported in Table 1. The secondary occurrence queried was the presence of a diagnosis code of splenic injury.
EP
This was determined based on the presence of ICD-9 diagnosis codes denoting splenic injury in the one “postoperative diagnosis” data field in ACS-NSQIP (Appendix 2, online only). In
AC C
Medicare inpatient claims, this was determined based on the presence of the same ICD-9 diagnosis codes in any of the ten “ICD-9 diagnosis” Medicare inpatient claims data fields. Statistical Analysis
Initial descriptive statistics were performed to compare the demographic and clinical characteristics of patients who underwent colectomy with concurrent splenic procedure (in ACSNSQIP and Medicare inpatient claims) to those who did not. The raw rates of ACS-NSQIP
ACCEPTED MANUSCRIPT 7
postoperative complications, including 30-day composite morbidity and 30-day mortality were calculated to understand the clinical implication of concurrent splenic procedures on postoperative outcomes. Thirty-day composite morbidity was defined as a dichotomous variable
RI PT
representing the occurrence of any one or more of the following 30-day postoperative
complications: surgical site infection (superficial, deep and/or organ-space), wound disruption, sepsis, pneumonia, unplanned reintubation, prolonged intubation, bleeding requiring transfusion,
SC
cardiac arrest, myocardial infarction, deep vein thrombosis, pulmonary embolus, coma, stroke,
M AN U
peripheral nerve injury, renal failure and urinary tract infection.
Next the clinical implication of a splenic injury requiring a splenic procedure was quantified. This was done by calculating the association between odds of 30-day morbidity and 30-day mortality and a concurrent splenic procedure while controlling for clinical variables. This association was calculated using a hierarchical multivariate logistic model including, the hospital
TE D
as a random intercept, and the following independent variables selected using clinical relevance and significance on bivariate analysis, p
