eAppendix 1. Predictors of hip fracture revision
eTable 1. Predictors of revision of hip fractures within 1 y of surgery
eAppendix 2. Adding type of assistant to the matched variables
eTable 2. Characteristics of hip fracture cohort after matching
eTable 3. Outcomes (hazards ratio) for postoperative complications in matched hip fracture patients
Customize your JAMA Network experience by selecting one or more topics from the list below.
Ravi B, Pincus D, Wasserstein D, et al. Association of Overlapping Surgery With Increased Risk for Complications Following Hip Surgery: A Population-Based, Matched Cohort Study. JAMA Intern Med. 2018;178(1):75–83. doi:10.1001/jamainternmed.2017.6835
What is the association of overlapping surgery with increased risks for complications following surgical treatment of hip fractures and end-stage arthritis?
In this population-based cohort study of patients with hip fracture and hip arthritis, there were 960 and 1560 overlapping procedures, respectively. For patients undergoing overlapping procedures, there was an approximately 90% increase in the risk for surgical complications at 1 year, although the association was weaker in elective hip replacements than in hip fractures.
Overlapping surgery is associated with an increased risk for complications in hip surgery, particularly for nonelective procedures.
Overlapping surgery, also known as double-booking, refers to a controversial practice in which a single attending surgeon supervises 2 or more operations, in different operating rooms, at the same time.
To determine if overlapping surgery is associated with greater risk for complications following surgical treatment for hip fracture and arthritis.
Design, Setting, and Participants
This was a retrospective population-based cohort study in Ontario, Canada (population, 13.6 million), for the years 2009 to 2014. There was 1 year of follow-up. This study encompassed 2 large cohorts. The “hip fracture” cohort captured all persons older than 60 years who underwent surgery for a hip fracture during the study period. The “total hip arthroplasty” (THA) cohort captured all primary elective THA recipients for arthritis during the study period. We matched overlapping and nonoverlapping hip fractures by patient age, patient sex, surgical procedure (for the hip fracture cohort), primary surgeon, and hospital.
Procedures were identified as overlapping if they overlapped with another surgical procedure performed by the same primary attending surgeon by more than 30 minutes.
Main Outcomes and Measures
Complication (infection, revision, dislocation) within 1 year.
There were 38 008 hip fractures, and of those, 960 (2.5%) were overlapping (mean age of patients, 66 years [interquartile range, 57-74 years]; 503 [52.4%] were female). There were 52 869 THAs and of those, 1560 (3.0%) overlapping (mean age, 84 years [interquartile range, 77-89 years]; 1293 [82.9%] were female). After matching, overlapping hip fracture procedures had a greater risk for a complication (hazard ratio [HR], 1.85; 95% CI, 1.27-2.71; P = .001), as did overlapping THA procedures (HR, 1.79; 95% CI, 1.02-3.14; P = .04). Among overlapping hip fracture operations, increasing duration of operative overlap was associated with increasing risk for complications (adjusted odds ratio, 1.07 per 10-minute increase in overlap; P = .009).
Conclusions and Relevance
Overlapping surgery was relatively rare but was associated with an increased risk for surgical complications. Furthermore, increasing duration of operative overlap was associated with an increasing risk for complications. These findings support the notion that overlapping provision of surgery should be part of the informed consent process.
Overlapping surgery refers to a situation in which a single attending surgeon supervises 2 or more operations, in different operating rooms, at the same time.1 Where overlapping surgery is performed, the attending surgeon is required to be present for critical parts of the procedure and readily available at other times, as needed.2 This longstanding practice is used to improve hospital resource utilization, provide opportunities to educate surgical trainees, and promote timely access to care, particularly for emergent cases.3,4
Recently, media reports have suggested a link with postoperative complications,5 resulting in considerable concern and discussion among the public, surgical leaders, and policymakers.6,7 Foremost is the concern that the attending surgeon cannot appropriately supervise the procedure if not physically present, resulting in a potential quality-of-care deficit.5 Transparency in the informed consent process is also a concern.7 The American College of Surgeons (ACS) released guidelines around concurrent and overlapping surgery.6 While 2 recently published peer-reviewed studies8,9 have suggested that overlapping surgery is a safe practice, they both have significant limitations that limit their generalizability.
The current study sought to address this gap by examining the impact of overlapping surgery for hip fractures and for primary elective total hip arthroplasty (THA) at more than 75 hospitals in Ontario, Canada. The specific objectives of this study were to (1) determine the extent of overlapping surgery in patients receiving hip surgery; (2) assess whether overlapping procedures are at increased risk for relevant surgical complications; and (3) to explore whether there is an association with the duration of overlap and risk for complications.
We conducted a population-based, matched-cohort study using data from several health administrative databases from Ontario, Canada (population of 13.5 million). Ontarians are insured under a single-payer system, which covers medically necessary procedures, including hip fracture surgery and elective hip arthroplasty. The study protocol was approved by the research ethics board at Sunnybrook Health Sciences Centre, Toronto, Ontario.
Data on every interaction a person has with our health care system are captured within administrative databases. These databases capture relevant demographic information on each patient (eg, age, sex, comorbidity) and physician (eg, years in practice, volume). The main data sources were hospital discharge abstracts from the Canadian Institute for Health Information Discharge Abstract Database (CIHI-DAD) and physician service claims from the Ontario Health Insurance Plan (OHIP) Claims History Database.
We defined 2 cohorts: the first comprised patients older than 60 years who had undergone an acute surgical procedure (fixation or replacement) for a hip fracture between April 1, 2009, and March 31, 2014. The second comprised patients older than 40 years who received a primary elective THA for arthritis between April 1, 2009, and March 31, 2015. During this study period, entry and exit times from the operating room for every surgical procedure in the province were recorded. We excluded procedures that were longer than 3 hours because this length of time is atypical and likely represented challenging cases that are not representative. See Table 1 for a full list of exclusions.
The start and end times captured in the database for each procedure reflect the entry and exit times into the operating room. As such, for each procedure, we chose the duration of overlap to define a procedure as overlapping that would indicate that the 2 procedures overlapped for a significant portion of time, after accounting for the time required to transfer the patient to the operating table, administer anesthesia, and properly position and drape the patient. We set the threshold of overlap to define a procedure as overlapping at 30 minutes, which is approximately 30% of the median duration (door-to-door) of both THAs and hip fracture surgery.
The primary outcomes were the occurrence of an infection, revision, or dislocation within a year.10 These were selected because they are the most likely to result from technical errors, which in turn may occur during unsupervised portions of the procedure. Other outcomes of interest were duration of surgery and death within 90 days.
We measured and adjusted for several patient and physician covariates that have been previously shown to affect the risk of occurrence of complications following hip surgery.10-14 Patient age and sex were obtained from the OHIP Registered Persons Databases (RPDB). Comorbidities were assessed via an adaptation of the Charlson Comorbidity Index with a look-back of 5 years.15 The Adjusted Clinical Groups (ACGs) indicators were used to identify “frail” patients.16,17 Diabetes mellitus, prior myocardial infarction (MI), congestive heart failure (CHF), chronic obstructive pulmonary disease (COPD), cerebrovascular disease, chronic renal failure, and dementia were identified using validated algorithms.18-24 Median neighborhood household income quintile was used as a surrogate for socioeconomic status.25-27
Hospital volume was defined as the number of hip fracture or THA procedures performed at the hospital in the 365 days prior to the index procedure. Surgeon volume was similarly defined for each hip fracture and THA patient. Presence of a billing surgical assistant was determined, as were his or her specialty (orthopedic surgeon, nonorthopedic surgeon, or family physician).
For each patient undergoing hip fracture repair or THA that overlapped with another procedure, we matched a patient who underwent the same surgical procedure by the same primary surgeon without an overlapping procedure; we used “greedy” matching.28 Patients were further matched according to age, sex, and hospital where the surgery was performed. We estimated standardized differences for all covariates after matching with more than 10 differences considered to be indicative of imbalance.29
Baseline cohort characteristics were described using proportions and medians as appropriate and were compared between groups using Wilcoxon rank sum tests for continuous variables and χ2 tests for categorical variables. Cox proportional hazard tests for the occurrence of a surgical complication were performed before and after matching, for the latter pair-matching into account by stratifying on matched pairs. All analyses were performed using SAS statistical software (version 9.3 and EG 6.1; SAS Institute). The type I error probability was set to α = .05 for all analyses.
We also performed a secondary analysis in the overlapping procedures. First, restricted cubic splines were used to visualize the relationship between duration of overlap (measured in minutes) and the probability as well as relative risk for a complication. Second, generalized estimating equations were used to estimate the odds of complications following hip fracture surgery based on the duration of overlap, after controlling for relevant confounders. Other variables in the model included income quintile, rurality, Charlson Comorbidity Index, frailty, type of hip fracture, type of fixation, teaching hospital status, and physician volume, as well as controlling for clustering by the primary surgeon.
Between April 1, 2009, and March 31, 2014, there were 38 008 eligible patients with hip fracture (mean age, 66 years [interquartile range, 57-74 years]; 52.4% were female) (Table 1 and Table 2) and 52 869 eligible THA recipients(mean age, 84 years [interquartile range, 77-89 years]; 82.9% were female) (Table 1 and Table 3). A total of 960 (approximately 2.5%) hip fracture procedures and 1560 (approximately 3%) THAs were performed overlapping with another procedure. More than 70% of overlapping procedures were performed in teaching hospitals. Prior to matching, overlapping patients with hip fracture were not significantly different from patients who did not overlap, in terms of age, sex, comorbidity, or fracture type (Table 2). Among THAs, patients in overlapping procedures were slightly younger and had a lower prevalence of comorbid conditions (prior MI, dementia, diabetes, COPD). (Also see Appendices 1 and 2 and eTables 1-3 in the Supplement.)
A total of 855 overlapping hip fracture procedures (89%) were successfully matched (by age, sex, procedure, attending surgeon, and hospital) to nonoverlapping procedures (Table 2). After matching, absolute standardized differences were 10.0 or less for all measured confounders (Table 2), except for prior MI, which was higher in the nonoverlapping group, and was controlled for in subsequent analyses.
A total of 1456 overlapping THAs (93%) were successfully matched (by age, sex, attending surgeon, and hospital) to nonoverlapping procedures (Table 3). After matching, absolute standardized differences were 10 or less for all measured confounders.
After matching, overlapping hip fractures were at a higher risk for surgical complications (hazard ratio [HR], 1.85; 95% CI, 1.27-2.71) (Table 4), particularly infection and early revision. Similarly, overlapping THA procedures were at a higher risk for complications (HR, 1.79; 95% CI, 1.02-3.14), but we could not determine the relative risk for individual complications as we were underpowered.
The restricted cubic splines relating duration of overlap to the risk for a surgical complication (Figure) suggest that these risks increase with increasing duration of overlap, but differently in hip fracture procedures than in THAs. Our regression analysis also indicated that increased duration of overlap is associated with an increased risk for complications in hip fracture procedures (adjusted odds ratio, 1.07 for every additional 10 minutes of overlap, P = .009).
We found that approximately 2.5% of patients with hip fracture and 3.0% of THA recipients were operated on by a surgeon who was also performing another operation. After matching, patients undergoing overlapping hip fracture repairs had a significantly increased risk for surgical complications (an increase from 6.4% to 10.4%). This risk worsened as the duration of overlap increased, with every additional 10 minutes of overlap raising the risk for a complication by approximately 7.0%. Overlapping THA procedures were also at increased risk; however, the relationship was not as pronounced (an increase from 1.4% to 2.3%).
The primary outcome of our study was the occurrence of surgical complications that directly related to the technical performance of the surgery itself. Overlapping procedures were at increased risk for the occurrence of a complication, particularly infection and early revision. The most likely mechanism for increased risk arises from having less experienced surgeons or surgical trainees perform portions of the case. We know from previous work that lack of experience and low surgical volumes are associated with an increased risk for surgical complications related to technical error.10 One of the tasks most frequently delegated to trainees and assistants is closure of the operative wound, as this is not typically considered “critical.”30 However, inadequate closure increases the risk for prolonged wound healing and infection,31 which in turn increases the likelihood of early revision. Interestingly the risk associated with overlapping surgery was much lower in elective hip arthroplasty than in nonelective hip fracture procedures. This suggests that the provision of overlapping surgery may be safer when the entire surgical team can prepare to deliver care to multiple patients, and perhaps when the primary surgeon can select assistants that he or she believes are suitable to have certain parts of the case delegated to them.
We selected hip fracture repair and elective THA as the study procedures for several reasons. The rates of these hip fractures are increasing in North America32 and there is a large body of evidence pointing toward improved outcomes with early surgical intervention.33 In addition, timely treatment of hip fractures is a quality indicator benchmark used by the provincial government, and in other jurisdictions, to assess hospital performance.34-36 As nonelective procedures, scheduling of hip fracture procedures must be balanced with the needs of patients who have been on waiting lists for elective procedures. All of these factors contribute to overlapping hip fracture surgery. However, THAs are one of the most common elective orthopedic procedures in North America,37 and many centers have developed standardized care pathways designed to safely maximize operating room throughput for these procedures. In addition, we wished to explore the safety of overlapping surgery in the setting of an elective procedure, and indeed found that it may be safer than for a nonelective procedure.
We defined overlapping in our cohort as an overlap that lasted for at least 30 minutes. This represents approximately 30% of the median time in the operating room for both hip fracture procedures and THAs and allows for time to position the patient and administer anesthesia. As such, procedures labeled as overlapping likely represent ones with a genuine surgical overlap, and that the overlap resulted in someone other than the attending surgeon operating unsupervised for at least a portion of the case. Our secondary analysis found that increased duration of overlap in hip fracture procedures was associated with an increased risk for complications, indicating that as the likelihood that the surgical procedures themselves overlapped (as opposed to positioning or induction of anesthesia), so did the risk for a complication.
Our findings may not generalize to hospitals where overlapping surgical procedures are more commonly performed.38 Indeed, overlapping hip fracture surgery seems to be an anomaly in our jurisdiction, rather than standard of practice, which may have had an impact on the quality of its performance. We also cannot assume our findings are generalizable to other surgical procedures, such as more acute surgical emergencies, where timely care may necessitate overlapping surgery.3 However, we were able to assess overlapping surgery in both elective and nonelective circumstances. While our findings in the arthroplasty cohort indicate that overlapping surgery might lead to slightly increased risk for complications, the absolute increase in the rates of complications is low. These findings further suggest that overlapping surgery may be safer when performed in the context of an elective procedure, when the surgeon and the rest of the health care team can adequately prepare to deliver care in an overlapping fashion.
While overlapping surgery has come under increasing scrutiny of late, this study is, to our knowledge, the first population-based study of this issue. We had the ability to capture complications that presented at centers other than where the original procedure was performed. We were also able to examine long-term outcomes (≤1 year), which allowed us to compare complications that may result from technical error (eg, infection, revision surgery).11,39 We controlled for various confounders, including patient (age, sex, comorbidity) and physician (surgeon volume, hospital volume, teaching hospital) factors. Recently, 2 studies have been published that indicated that overlapping surgery does not confer any increased risk for complications. Zhang et al9 performed a retrospective review of a single institution over a 3-year period, and compared the rates of complications at 30 days in multiple outpatient procedures.9 However, they did not perform any risk adjustment for procedure, or by patient and physician factors. Hyder et al8 looked at multiple surgical procedures at a single center, using an m:n matching strategy in which all overlapping procedures of a particular type of surgery were matched to all nonoverlapping procedures of the same type of surgery, with subsequent adjustment for the individual surgeon. Neither study was able to capture complications that presented at a different center.
This study used operating room entry and exit times on a population-wide scale (ie, across an entire geographic area as opposed to a single center or group of hospitals) to identify overlapping surgical procedures.40 The data on surgery times have high face validity; basic tests of their quality included that (1) start times always occurred after the hospital admission time and that (2) “sensible” surgical durations are predicted by surgery start and end times.41 We were able to measure and balance a wide variety of patient and physician factors between groups. Matching by procedure, surgeon, and hospital levels allowed us to control for unmeasured physician factors. Limitations to this study primarily relate to unmeasured factors that were not captured in our database(s), and the study design. We attempted to mitigate the impact of unmeasured confounders for physicians by matching by surgeon and hospital. However, there were other patient-level variables we were unable to assess or control for, such as body mass index and smoking history.
We selected 30 minutes of overlap as our threshold to identify a procedure as overlapping because it was approximately one-third of the median door-to-door times for both cohorts; this would provide sufficient time for induction of anesthesia, positioning, and draping. However, it is possible that in some procedures these steps would have taken longer, particularly in patients with hip fracture because they generally have a higher burden of medical comorbidities. As such, we performed a second analysis wherein we determined the impact of the duration of overlap on the risk for complications following hip fracture repair. We also generated splines for both cohorts to visualize the relationship between overlap duration and complication risk. These splines suggested that increasing overlap was associated with an increased risk for complications; this was more pronounced in hip fracture procedures than in elective hip arthroplasty.
Most important, we could not identify the actual location of the attending surgeon, and could not determine where the attending surgeon was during “critical portions” of the case.7,38 According to the ACS,2 “concurrent” procedures occur when “the critical or key components of the procedures for which the primary attending surgeon is responsible are occurring all or in part at the same time,” whereas in overlapping procedures, the surgeon is present for all critical portions while a “qualified practitioner” performs the noncritical portions in the primary surgeon’s absence. Although the distinction is potentially important, this concept is severely limited by the fact that there are no widely accepted definitions of what constitutes a critical portion of a procedure; this is determined by the primary surgeon. Every surgeon might have a unique definition for what might be a critical aspect of a specific procedure, which might vary depending on the assistant(s) available to them. Our contention is that without having knowledge of which portions of a procedure the attending surgeon would identify as critical, then potentially every step of the procedure from incision to closure may meet this threshold. This might explain our findings that overlapping (as opposed to concurrent) hip surgery is also associated with an increased risk for complications.
To our knowledge, this is the first population-based study examining the consequences of overlapping surgery at the level of a health care system. Our results indicate that while overlapping surgery is uncommon in our jurisdiction for hip surgery, it is a risk factor for complications, particularly for hip fractures. Furthermore, an increasing duration of overlap with another procedure is associated with progressively increasing risk for complications. Most important, our findings reinforce the notion that overlapping provision of surgery must be part of the informed consent process.
Corresponding Author: Bheeshma Ravi, MD, PhD, Department of Orthopaedic Surgery, Sunnybrook Health Sciences Centre, 43 Wellesley St E, Room 315, Toronto, ON M4Y 1H1, Canada (email@example.com).
Accepted for Publication: September 16, 2017.
Published Online: December 4, 2017. doi:10.1001/jamainternmed.2017.6835
Author Contributions: Dr Ravi had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Ravi, Pincus, Wasserstein, Jenkinson, Paterson, Kreder.
Acquisition, analysis, or interpretation of data: Ravi, Pincus, Govindarajan, Huang, Austin, Jenkinson, Henry, Paterson.
Drafting of the manuscript: Ravi, Pincus, Wasserstein, Huang.
Critical revision of the manuscript for important intellectual content: Ravi, Pincus, Wasserstein, Govindarajan, Austin, Jenkinson, Henry, Paterson, Kreder.
Statistical analysis: Ravi, Pincus, Wasserstein, Huang.
Administrative, technical, or material support: Ravi, Wasserstein, Jenkinson, Paterson.
Study supervision: Wasserstein, Jenkinson, Kreder.
Conflict of Interest Disclosures: None reported.
Funding/Support: This study was supported by the Institute for Clinical Evaluative Sciences (ICES), an independent research institute funded by the Ontario Ministry of Health and Long-Term Care. Parts of this material are based on data and information compiled and provided by the Canadian Institute for Health Information (CIHI). This study was supported by internal funding from the Marvin Tile Chair in Orthopaedic Surgery at Sunnybrook Health Sciences Centre, Toronto, Canada.
Role of the Funder/Sponsor: The supporting organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.
Disclaimer: The opinions, results, and conclusions reported in this study are those of the authors and are independent from the funding sources and CIHI. No endorsement by the Institute for Clinical Evaluative Sciences, the Ontario Ministry of Health and Long-Term Care and CIHI is intended or should be inferred.