Open circles represent procedures for which prophylaxis was indicated; closed circles represent procedures for which prophylaxis was not indicated; and horizontal bars represent interquartile range by hospital. ORIF indicates open reduction internal fixation.
Horizontal lines inside the box represent the median appropriate utilization rate; the borders of the box are the interquartile range (IQR); the plus sign represents the mean appropriate utilization rate; the whiskers are drawn to the most extreme points in the group that lie within the fences (the upper fence is defined as the third quartile plus 1.5 times the IQR, and the lower fence is defined as the first quartile minus 1.5 times the IQR); and circles outside the fences represent extreme outliers.
Each circle represents 1 hospital with the size of the circle reflecting the volume of procedures; vertical and horizontal lines represent overall median rates.
eTable. Categorization of Surgical Antibiotic Prophylaxis Indications for the 45 Most Commonly Performed Operations.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Sandora TJ, Fung M, Melvin P, Graham DA, Rangel SJ. National Variability and Appropriateness of Surgical Antibiotic Prophylaxis in US Children’s Hospitals. JAMA Pediatr. 2016;170(6):570–576. doi:10.1001/jamapediatrics.2016.0019
Copyright 2016 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.
Appropriate use of surgical antibiotic prophylaxis (AP) reduces surgical site infection rates, but prior data suggest variability in use patterns.
To assess national variability and appropriateness of AP in pediatric surgical patients.
Design, Setting, and Participants
Retrospective cohort study of 31 freestanding children’s hospitals in the United States using administrative data from 2010-2013. The study included 603 734 children younger than 18 years who underwent one of the 45 most commonly performed operations.
Receipt of surgical AP.
Main Outcomes and Measures
Primary outcomes included procedure- and hospital-specific rates of AP use and appropriateness of use based on clinical guidelines and consensus statements. We also assessed rates of Clostridium difficile infection and potential allergic reactions (using epinephrine administration as a surrogate event) after AP receipt.
Of the 603 734 eligible patients, the mean (SD) patient age was 4.8 (4.4) years and 384 571 (63.7%) were boys. For the 671 255 operations evaluated, AP was administered for 348 119 (52%) of procedures. Intrahospital variation in AP use by procedure ranged from 11.5% to 100% (median, 78.1%). Overall, AP use was considered appropriate for 64.6% of cases. Appropriate use of AP by hospital varied from 47.3% to 84.4% with large variability by procedure within each hospital. For procedures for which AP was indicated, the median rate of appropriate use by hospital was 93.8%; however, for procedures for which AP was not indicated, the median rate of appropriate use by hospital was 52.0%. The odds of C difficile infection and epinephrine administration were significantly higher among children who received AP (odds ratio, 3.34; 95% CI, 1.66-6.73 and odds ratio 1.97; 95% CI, 1.92-2.02; respectively).
Conclusions and Relevance
There is substantial national variability in the overall and appropriate use of AP for the most commonly performed operations in children both at a procedure and hospital level. A high proportion of AP use is inappropriate, potentially exposing many children to avoidable adverse events. Urgent attention should be directed to efforts to standardize the use of surgical AP in pediatrics.
Surgical site infection (SSI) is one of the most common complications of adult and pediatric surgery and is associated with significant morbidity and mortality. Previous studies, including the initial phase of the American College of Surgeons’ National Surgical Quality Improvement Pediatric Program, suggest that SSI occurs in 1% to 4% of children after surgery.1-5 Appropriate use of perioperative surgical antibiotic prophylaxis (AP) to reduce the burden of microorganisms at the surgical site and thereby minimize intraoperative wound contamination reduces the incidence of SSI for procedures for which AP is indicated.6-9 At the same time, there is increasing evidence that inappropriate use of AP has potentially negative consequences, including antibiotic-related adverse drug events and Clostridium difficile infection (CDI) at the patient level, the emergence of resistant organisms, and an increase in health care costs at the population level.10-17
While data and guidelines for surgical AP use in adult patients are available, the trend in surgical AP use for pediatric patients is not well understood.1,4-6,18 We previously examined AP use among children undergoing general surgical and urologic operations and found that 40% of patients received antibiotics when there was no indication for prophylaxis.17 Given that more than 200 000 pediatric inpatient surgical procedures are performed in the United States each year,19 ensuring appropriate use of surgical AP in children should be a public health priority.
The objectives of this study were to characterize the variability in surgical AP use in the pediatric population across a wide variety of surgical subspecialties and to examine the appropriateness of surgical AP administration, with a focus on indications for antibiotic prophylaxis for specific operations. In addition, we sought to identify adverse events potentially attributable to inappropriate use of AP.
Question What is the variability and appropriateness of surgical antibiotic prophylaxis in US children’s hospitals?
Findings In this cohort study using administrative data for 671 255 operations, surgical antibiotic prophylaxis was appropriate for 64.6% of cases, with substantial variability at a procedure and hospital level.
Meaning Strategies to standardize use of surgical antibiotic prophylaxis among children are required to reduce unnecessary antibiotic exposure and minimize unintended consequences.
We conducted a retrospective cohort study using the Pediatric Health Information System (PHIS) database of the Children’s Hospital Association (Overland Park, Kansas). The PHIS is a deidentified administrative database containing data from 47 freestanding tertiary pediatric hospitals located in nearly all major metropolitan areas throughout the United States. Participating hospitals provide data from inpatient, outpatient, and emergency encounters including diagnostic and procedural International Classification of Diseases, Ninth Revision, Clinical Modification codes, demographic information, and billing data (including detailed pharmacologic billing data). These data undergo multiple reliability and validity checks prior to entry in the database. This study was approved by the Boston Children’s Hospital Committee on Clinical Investigation. A waiver of informed consent was granted because the study was determined to be minimal risk and because the PHIS dataset is deidentified.
Diagnostic and procedural International Classification of Diseases, Ninth Revision, Clinical Modification codes were used to identify a cohort of patients younger than 18 years who underwent one of the 45 most commonly performed operative procedures by volume between January 1, 2010, and December 31, 2013. To restrict our analysis to cases where AP was relevant, we excluded cases where patients were already receiving antibiotics before surgery (defined as receipt of any type of antibiotic more than 1 day prior to the procedure but during the same admission) and cases where cultures were obtained within 48 hours prior to surgery (indicating the possibility of a contaminated case). To ensure that AP use could be attributed to an individual procedure, cases where multiple procedures were performed at the same time were excluded. We excluded cases in which another procedure was performed within 30 days before or after the primary procedure to attribute subsequent CDI episodes to a single procedure. We also excluded patients who were hospitalized for more than 2 days prior to the surgical procedure to identify operations that were the primary reason for hospitalization and to increase the likelihood that antibiotics on the day of the procedure were given for surgical prophylaxis. We excluded ophthalmologic and dentistry cases and patients with a prior cardiac operation because a subset of these patients may have received antibiotic prophylaxis for an underlying condition. Cases where AP use may have been associated with underlying immunodeficiency were excluded by the presence of a diagnosis code for immunodeficiency disorder, cancer, history of transplant, diabetes mellitus, malnutrition, or receipt of oral or intravenous steroids (other than dexamethasone, which was assumed to be given as postoperative nausea and vomiting prophylaxis) on the day before or day of surgery. Finally, we excluded 10 hospitals for which review of data quality reports suggested concerns about data completeness and quality for that time.
Patients were defined as having received AP if there was a charge for parenteral administration of any antibiotic on the same calendar day as the surgical procedure. Procedure-specific AP utilization rates were calculated for all 45 procedures by dividing the number of patients undergoing a specific procedure who received AP by the total number of patients undergoing that procedure. Procedure-specific utilization rates were calculated for the entire cohort and for each individual hospital. To capture intrahospital variation among the surgical procedures, a median AP utilization rate and interquartile range was calculated for each procedure based on each of the 31 hospitals’ utilization rate for that procedure.
To define procedures for which AP was indicated, we conducted a comprehensive literature search for specialty-specific AP guidelines and consensus statements, which was then cross-referenced with Centers for Disease Control and Prevention recommendations and a multisociety clinical practice guideline for antimicrobial prophylaxis in surgery.20 Based on this literature review, we defined AP use as indicated for all clean-contaminated procedures (entrance of gastrointestinal, genitourinary, or respiratory tract with minimal soilage) and clean procedures with potential for excessive morbidity from an infectious complication (procedures involving the central nervous system, open fractures, and the insertion of central catheters and other prosthetic devices). See the eTable in the Supplement for our categorization scheme for the 45 procedures in this study.
Procedure-specific appropriate AP use was calculated by adding the number of cases receiving AP for a particular procedure where AP use was indicated and the number of cases in which AP was not given when not indicated, and dividing this sum by the total number of patients undergoing that procedure.
Hospital-specific appropriate AP use was calculated by adding the number of cases receiving AP where AP use was indicated and the number of cases where AP was not given when not indicated. This sum was then divided by the total number of cases for that hospital. Intraprocedural variation in appropriate AP utilization rates among the 31 hospitals was identified by calculating, for each hospital, a median appropriate AP utilization rate based on the 43 to 45 procedure-specific appropriate AP utilization rates for that hospital. A scatterplot was created to assess the relationship between giving AP when indicated and not giving AP when not indicated within a given hospital.
The Mantel-Haenszel test for trends was used to evaluate the change in the overall annual appropriate use of AP from the beginning to the end of the study.
We examined the occurrence of 2 specific adverse drug events potentially associated with AP: CDI and allergic reactions. C difficile infection associated with AP was captured by identifying patients with an International Classification of Diseases, Ninth Revision, Clinical Modification code for CDI (008.45) within 30 days after the index procedure. This diagnostic code has been shown to reliably and accurately identify children with CDI.21 Receipt of diphenhydramine or epinephrine by any route on the day of surgery was identified as a surrogate for the potential presence of an allergic reaction. Logistic regression analyses were used to determine whether AP use was associated with CDI within 30 days and perioperative allergic reactions.
The final cohort included 603 734 eligible patients from 31 hospitals who underwent 671 255 procedures during the study period (Figure 1). A median of 19 851 surgical cases (range, 4710-52 438) were performed per hospital. The mean (SD) patient age was 4.8 (4.4) years and 384 571 (63.7%) were male. Most procedures (n = 566 304; 84%) were performed in an ambulatory surgery setting, while 49 327 (7%) were inpatient and 49 177 (7.3%) were observation cases. Surgical antibiotic prophylaxis was administered for 348 119 of the 671 255 procedures (51.9%), with the proportion receiving AP ranging from 24.6% to 78.9% across centers. The median procedure-specific AP utilization rate was 78.1%, ranging from a low of 11.5% of cases receiving AP for laryngeal procedures to a high of 100% for ventriculostomy, ventriculoperitoneal shunt, and spinal cord procedures. The procedures with the greatest variability by hospital for AP use included chordee repair, uretheral meatoplasty, and circumcision (Figure 2).
For the entire cohort, AP was defined as indicated for 250 287 of 671 255 cases (37.3%), and AP was actually administered for 180 483 of these cases (72.1%). For the remaining 420 968 procedures for which AP was not indicated, AP was administered for 167 636 cases (39.8%). Overall, AP use was considered appropriate for 433 815 of 671 255 cases (64.6%; 95% CI, 64.5%-64.7%). Appropriate use of AP by hospital varied from 47.3% to 84.4%, with large variability by procedure within each hospital (Figure 3). Figure 4 displays the relationship by hospital between giving AP when indicated and not giving AP when not indicated. Over the 4-year study, there was a minimal but statistically significant increase in the annual appropriate use of AP (63.5% in 2010 compared with 64.8% in 2013, P < .001).
Clostridium difficile infection was documented in 46 of 671 255 cases (0.7 per 10 000 cases) within 30 days after surgery. Children who received AP had significantly higher odds of developing CDI (Table). Diphenhydramine was administered after surgery in 7292 of 671 255 cases (1.1%) and epinephrine was administered in 26 199 of 671 255 cases (3.9%); children who received AP had significantly higher odds of receiving these medications (Table).
We found that there is substantial national hospital- and procedure-level variation in the use of antibiotic prophylaxis in children undergoing surgery in the United States, and a substantial proportion of this AP use may be inappropriate. To our knowledge, this is the first large-scale examination of national surgical AP use in children with a scope from procedure-specific to hospital-specific AP practices. Our results are consistent with existing literature describing variation in AP use,22-29 although pediatric data about surgical AP use are limited. In a study of AP use at 4 children’s hospitals between 1999 and 2000, practice was not in accordance with American Academy of Pediatrics guidelines in almost half of the procedures; use of AP in clean procedures where it may not have been necessary was identified as an opportunity for improvement.29 More than 50% of surveyed pediatric surgeons in the United Kingdom and Ireland do not recommend AP for pyloromyotomy.30 In a survey of US pediatric surgeons, 99% reported using intravenous antibiotics for colorectal surgery, but there was variability in the use of oral antibiotics as part of preoperative bowel preparation.31 Although we did not attempt to examine postoperative duration of AP use in our study, this issue is also extremely important because data show that prolonged AP use after surgery does not reduce the risk of SSI but does increase the risk of antibiotic-resistant bacteria.32 Two additional aspects of appropriate antibiotic prophylaxis that we did not evaluate for this study include selection of an antibiotic with appropriate spectrum of activity for the type of operation and timing of administration of the perioperative dose.
We found that many children did not receive antibiotics when prophylaxis was indicated, and an even greater proportion received AP despite a lack of clear benefit. The importance of proper AP use cannot be overemphasized because appropriate use of AP is associated with decreased incidence of SSI after a variety of operations.18,20 However, antibiotics are one of the most common medication classes associated with potential adverse drug events in children,33 and studies have demonstrated that CDI can occur after exposure to perioperative AP.11,17,34,35 In this study, we found that across a wide spectrum of pediatric surgical procedures, administration of AP was associated with both CDI and proxy measures of potential allergic reaction after surgery. Further studies should assess adverse drug reactions related to AP use in more detail.
There are several possible explanations for the observed variation in AP use between and within hospitals and procedures. Perhaps the most likely contributing factor is the lack of pediatric-specific guidelines for AP use.6,20 In this regard, some disagreement exists between adult-derived consensus guidelines and pediatric-focused observational data. For example, while pyloromyotomy is considered a clean procedure for which AP is not indicated in adults, data suggest that AP may provide benefit for pediatric patients.36 In addition, SSI risk classification systems such as those from the Study of Efficacy of Nosocomial Infection Control and the Centers for Disease Control and Prevention’s National Healthcare Safety Network have largely been derived from data from adult surgical populations. Whereas underlying comorbidities heavily influence adult SSI risk, data indicate that procedure-specific factors and markers of acute physiologic status strongly predict SSI risk in children.2,3 Therefore, some pediatric surgeons may not consider adult AP guidelines to be relevant for their patients. Finally, factors such as physician training environment, patient demands, and medical reimbursement structures may contribute to physician decisions to prescribe antibiotics.37,38
This study has several limitations. As with any study using administrative data, the possibility of misclassification exists. In particular, miscoding of surgical procedures and lack of charges for antibiotic administration could affect our findings. However, PHIS data undergo several validation steps prior to inclusion in the database, which reduces the likelihood of these errors. In addition, we assumed that antibiotics administered on the day of surgery were being used for surgical prophylaxis. Although this is likely a valid assumption in most cases, the precise indication for antibiotic administration for a given patient cannot be determined using PHIS data. We did not evaluate operations that were not among the 45 most commonly performed procedures (notably, no cardiac surgeries were among this group), but variation in other procedures is equally important to understand. Finally, the hospitals contributing data to PHIS are metropolitan, tertiary-care, academic pediatric hospitals, which may limit the generalizability of our findings for community or nonacademic settings.
Surgical antibiotic prophylaxis is associated with both potential benefits and risks for individual patients, and it has important public health implications on a population level with respect to antibiotic resistance and health care costs. Additional research is urgently needed to document the procedure-specific risk of SSI among pediatric patients and to establish strategies to improve AP use for children to prevent SSI and minimize unintended consequences.
Corresponding Author: Thomas J. Sandora, MD, MPH, Division of Infectious Diseases, Department of Medicine, Boston Children’s Hospital, 300 Longwood Ave, Boston, MA 02115 (email@example.com).
Accepted for Publication: December 30, 2015.
Published Online: April 18, 2016. doi:10.1001/jamapediatrics.2016.0019.
Author Contributions: Dr Sandora and Ms Melvin had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Sandora, Fung, Graham, Rangel.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Sandora, Fung, Melvin, Rangel.
Critical revision of the manuscript for important intellectual content: Sandora, Graham, Rangel.
Statistical analysis: Sandora, Melvin, Graham.
Administrative, technical, or material support: Rangel.
Study supervision: Sandora, Graham, Rangel.
Conflict of Interest Disclosures: None reported.
Create a personal account or sign in to: