2.1 Study design

This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols (PRISMA) guidelines.¹¹ The study protocol is registered within the PROSPERO database (ID#: CRD42023454868), and was exempt from Institutional Review Board approval.

A comprehensive literature search was conducted in PubMed and Embase via Elsevier for articles published without date restrictions through August 2023. Searches were developed by an experienced medical librarian (NA) using keywords in the title and abstract fields in combination with database-specific indexing terms. Search concepts included proximal femur fracture fixation and surgical training, incorporating variations of learning, education, and surgeon experience (e.g., resident, junior, novice, inexperienced, experienced), along with synonyms for femoral anatomy and fixation techniques. In Embase, results were limited to articles, articles in press, conference papers, errata, and reviews. Detailed search strategies are provided in Supplement 1.

Search results were exported to EndNote. Duplicates were removed using The Systematic Review Accelerator Deduplicator tool¹² before they were added to Covidence for screening. Screening included two stages: title/abstract screening and full text screening. All imported references were independently screened by two investigators (ES and EL). Disagreements were resolved by discussion.

2.2 Eligibility criteria

English language articles were included if they reported on learning curve metrics associated with surgical fixation of proximal femur or peritrochanteric femur fractures. Articles were excluded if they did not report learning curve metrics, if they were not published in English, or were review articles or case reports.

2.3 Risk of bias assessment

Study quality was assessed using the Newcastle-Ottawa Scale (NOS), evaluating observational studies across three domains: selection of study groups, comparability of cohorts, and ascertainment of outcomes. Each study can be awarded up to nine stars, reflecting the highest methodological quality. In accordance with established criteria, studies receiving 7–9 stars were classified as high quality, 4–6 stars moderate quality, and 0–3 stars low quality.¹³

2.4 Statistical analysis

The following information and variables were extracted using a standardized Excel form: study design, number and level of training of studied surgeons, patient characteristics including number of patients, specific injuries, specific surgical treatments, and complication rates, time spent in the operating room, estimated blood loss, fluoroscopy time, and reoperations. All outcome variables were described qualitatively.

3.1 Operating room time

Operating room time was the most consistently reported a measure of the learning curve in fifteen of the twenty-two included studies (Table 2). Of these, nine studies reported that operative durations were significantly longer in the presence of residents or when surgeries were performed by less experienced residents or attending surgeons. Four studies found no significant difference in operating room time across levels of surgical experience, while two studies observed shorter operative times among residents relative to more senior colleagues.

Faraj et al. and Prat et al. were the only studies that directly compared unsupervised residents with attending surgeons, showing that residents performing hemiarthroplasty took 29 and 34.5 min longer, respectively.^14,15 Solarino et al. found that supervised residents required an additional 24 min on average for hemiarthroplasty, with operative times decreasing by approximately 10 min after 5 cases.¹⁶ Similarly, Bjorgul et al.,¹⁷ Kagan et al.,¹⁸ and Neuwirth et al.¹⁹ reported significantly longer operative times among residents than attendings, though the extent of supervision was not specified.

Several studies quantified the effect of cumulative experience on operating time. Authen et al. reported that residents with fewer than three years of experience took an average of 7 min longer per procedure than those with greater experience.²⁰ Seo et al. demonstrated that new attending surgeons (<50 cases completed) required significantly more operating time than experienced surgeons (>500 cases).²¹ Altintas et al. observed that residents achieved a marked reduction in proximal femoral nailing time after approximately 15 cases, suggesting a measurable transition along the early learning curve.⁶

In contrast, four studies comparing supervised residents and attending surgeons across intramedullary nailing and total hip arthroplasty found no significant difference in operating times.^22–25 Two studies reported paradoxically longer OR times for more experienced surgeons. Biber et al. found that attendings took an average of 4 min longer than residents during proximal femoral nailing—a difference that, while statistically significant, was likely clinically negligible and attributed to greater case complexity.²⁶ Traven et al. reported that each additional year of training was associated with an increase of 4.5 min in operative time, although this may reflect increased resident autonomy rather than a true decline in efficiency.²⁷

3.2 Fluoroscopy use

Fluoroscopy use was analyzed in five studies to assess the impact of surgical experience (Table 3). Four studies demonstrated that less experienced surgeons, including both residents and attendings, were associated with significantly greater fluoroscopy usage times or increased total radiation dose. Only one study found no significant difference in intraoperative fluoroscopy times between residents and attending physicians.

Bruce et al. found junior residents had significantly higher radiation dose exposure (261 Gy/cm²) compared to senior residents (178 Gy/cm²) and attendings (203 Gy/cm²).⁹ The higher radiation exposure among attendings relative to senior residents was attributed to more complex case selection and more patient comorbidities.

Buxbaum et al. compared fluoroscopy time during intramedullary nailing across residents of different postgraduate years and attending surgeons. They found second-year residents had significantly longer fluoroscopy times (122 s) compared to attendings (95 s), while first-year residents’ longer times (130 s) did not reach statistical significance, likely reflecting small sample size and variability.²⁸ Similarly, Quah et al. demonstrated that residents with 3–8 years of training had significantly longer fluoroscopy times than attendings, but first- and second-year residents exhibited lower exposure, a finding attributed to their treatment of simpler fracture patterns.²⁹

Botchu et al. stratified attending surgeons by experience level, observing a trend of decreasing fluoroscopy time with greater experience. Attendings with fewer than three years averaged 1.19 min versus 0.49 min for those with over ten years, though these differences were not subjected to formal statistical testing.³⁰ Kelly et al. showed that residents were exposed to more radiation than their attendings (265 versus 233 Gy/cm²), however this difference was not statistically significant.³¹

3.3 Intraoperative blood loss

Intraoperative blood loss was investigated in four studies, either as estimated blood volume or postoperative hemoglobin drop (Table 4). Three studies identified no significant association between the amount of blood loss and surgeon experience, whereas one study documented greater blood loss during procedures involving residents. None of the included studies reported increased blood loss with procedures performed by more experienced surgeons.

Dilernia et al. and Faraj et al. found no significant differences between residents and attendings for total hip arthroplasties and hemiarthroplasties (approximately 1.0 L vs 1.0 L and 350 mL vs 300 mL, respectively).^14,23 Rose et al. measured postoperative hemoglobin changes across several procedures including total and hemiarthroplasty, dynamic hip screw fixation, and femoral nailing and did not find a significant difference between residents and attendings (3.1 g/dL for junior residents, 2.7 g/dL for senior residents, and 2.4 g/dL for attendings).³² Although these studies did not demonstrate significant differences, intraoperative blood loss inversely trended with surgeon experience. In contrast, Kagan et al. reported significantly greater estimated blood loss in resident-involved cases (235 mL vs 177 mL), potentially explained by a higher proportion of less invasive procedures in the attending cohort.¹⁸

3.4 Reoperation rate

Reoperation rates were reported in seven studies as a surrogate marker of technical failure and secondary outcome reflecting surgical proficiency (Table 5). The majority of these studies (six studies) found no significant difference in reoperation rates according to surgeon experience. One study identified a statistically significant but not clinically meaningful increase in reoperation rates for procedures performed by less experienced surgeons. No studies observed higher reoperation rates among surgeries conducted by more experienced surgeons.

For hemiarthroplasty, Prat et al. reported reoperation rates of 3.7% for residents and 5.0% for attending,¹⁵ while Spaans et al. observed an overall 4.5% reoperation rate with no significant differences between surgeon groups.³³ Among intramedullary nailing cases, four studies reported reoperation rates ranging from 5.5% to 12.4% for residents and 3.4% to 11.6% for attendings, none reaching statistical significance.^6,15,22,24 Seo et al. noted four reoperations in their resident cohort compared to two in attendings.²¹ Authen et al. reported a statistically higher reoperation rate in surgeons with fewer than three years of experience (5.3%) versus more experienced surgeons (4.2%), though this difference lacked clinical significance.²⁰

3.5 Complication rate

Twelve studies reported overall complication rates (Table 6). Nine studies concluded that complication rates did not differ significantly in relation to surgeon experience. Three studies documented significantly higher postoperative complication rates among less experienced residents or attending surgeons. No studies identified a higher complication rate associated with greater surgical experience.

Complications, depending on the type of fracture fixation, included hematoma formation, infection, nonunion, screw cut out, implant malpositioning, fracture malreduction, intractable postoperative pain, loose prostheses, prosthetic dislocation, peri-implant fracture, wound dehiscence, mortality, and overall complication rate. For intramedullary fixation, Biber and Buecking et al. reported overall complication rates of 5.3 – 9.3% for residents versus 5.8 – 6.9% for attendings.^22,26 Halonen et al. reported 10 peri-implant fractures in their resident cohort and 8 in their attending cohort.²⁴ Schutze and Neuwirth et al. reported no significant differences in mortality rates between residents (7.8 – 18%) versus attendings (6.3 – 23%) in attendings.^19,25 Altintas et al. showed no difference in overall complication rates between residents before and after completing 10 intramedullary nailing procedures.⁶

For hip total and hemiarthroplasty, Dilernia and Spaans et al. reported no significant difference in complication rates between residents and attendings.^23,33 This was in contrast to Solarino et al. who reported residents having a significantly higher overall complication rate (15.7%) compared to attendings (7.5%).¹⁶ Faraj et al. reported a significantly higher rate of intraoperative fracture, peri-implant fracture, postoperative thigh pain, and loose prostheses in residents compared to attending surgeons.¹⁴ In addition, Traven et al. reported a significantly higher rate of wound dehiscence in cases with residents compared to no residents, and an increasing rate of wound dehiscence as resident experience by year increased. However, this study included a variety of different fracture types including proximal femur fractures, femoral and tibial shaft fractures, and ankle fractures, and complication rates were not stratified accordingly.

Supplement 2 presents NOS-based quality assessments for the included cohort studies. All twenty-two studies were rated as high quality, with scores ranging from seven to nine stars, indicating a low risk of bias.

4.1 Operating room time

Operating room time remains a practical indicator of procedural efficiency and technical progression, offering a reproducible metric for assessing learning curves across surgeon experience levels. Cumulatively, evidence supports the presence of a measurable learning curve in operative efficiency, with the steepest improvements typically occurring within the first 15–30 procedures. Variability between studies likely reflects institutional workflow, supervision models, and case selection, underscoring the importance of contextual factors when interpreting time-based measures of surgical proficiency.

Although increased operating times were consistently observed among less experienced surgeons, the magnitude of these differences, typically ranging between 5 and 35 min, may hold limited clinical significance. While prolonged procedures may modestly increase anesthesia duration and resource utilization, such intervals are unlikely to affect patient outcomes in isolation. Instead, they primarily reflect the expected efficiency gap inherent to skill acquisition early in surgical training. Importantly, several studies demonstrated a rapid reduction in operative time after only a small number of cases, suggesting that proficiency in proximal femoral fixation is achieved relatively early in the learning curve.

4.2 Fluoroscopy use

Collectively, these studies consistently demonstrate an association between greater surgical experience and reduced fluoroscopy use. Intraoperative fluoroscopy time and radiation exposure thus serve as practical surrogate markers for surgical proficiency. However, variability in findings likely reflects differences in case complexity, surgical technique, and institutional protocols, underscoring the need to interpret time- and dose-based metrics within appropriate clinical contexts.

4.3 Intraoperative blood loss

Although intraoperative blood loss is important for estimating patient transfusion requirements, it is often inconsistent and prone to error. Rothermel and Lipman conducted a study where surgeons, anesthesiologists, nurses, and technicians estimated blood volume in three simulated operative scenarios. Ninety-five percent of participants had errors exceeding 25% in at least one scenario, and only 27% demonstrated consistent over- or underestimation across cases. Accuracy did not correlate with professional role or experience level. Adopting standardized categorical ranges such as “minimal” (<100 mL), “moderate” (100–500 mL), and “large” (>500 mL) blood loss may improve reliability and clinical utility in practice.³⁴

4.4 Reoperation rate

Collectively, these studies do not illustrate a significant association between surgeon experience and reoperation risk for proximal femur fracture fixation. Confounding factors such as case complexity, procedure type, supervision level, and patient comorbidities likely mediate learning curve effects. Future investigations should employ standardized definitions and adjust for case difficulty to clarify these relationships.

4.5 Complication rate

Although early-stage trainees appear more prone to certain intraoperative complications, particularly technical nuances during total and hemiarthroplasty and wound-related events, the overall complication rates appear variable. It is important to recognize these inconsistencies likely reflect differences in case mix, complication definitions, supervision structures, and institutional protocols, rather than experience alone.

4.6 The learning curve

Certain proximal femur fracture fixation procedures exhibit minimal to no learning curve effect, enabling trainees to perform at a high level early in their experience when foundational techniques are strictly followed. For instance, Bjorgul et al. found that residents achieved a significant reduction in operative time for internal fixation of femoral neck fractures with cannulated screws after only 15 cases.¹⁷ This contrasts with Lee et al., who found no significant learning curve for internal fixation of nondisplaced femoral neck fractures using cumulative sum analysis.¹⁰ Notably, Lee et al.’s cohort consisted of procedures performed by a single attending surgeon with a year of fellowship experience at a tertiary referral center, suggesting that prior advanced training may account for absent observable learning curve effects.

Other techniques demonstrate steeper learning curves requiring a higher case volume to achieve proficiency. Li et al. reported that 90% of proficiency with the Inter-Tan intramedullary nail for intertrochanteric fractures was reached only after 18 cases.³⁵ This finding aligns with Altintas et al., who observed a statistically significant decrease in operative time after 15 cases of proximal femoral nailing, as well as Bjorgul et al., who found significant reductions only after 20 cases for short intramedullary nails and 30 cases for long nails.^6,17 These findings reinforce that procedural complexity and implant type can significantly influence the slope and duration of the surgical learning curve.

For hemiarthroplasty, Bjorgul et al. demonstrated that residents required completion of 25 procedures before achieving substantial reductions in operative time, compared to their first five cases.¹⁷ Solarino et al. also reported a significant decrease in operating times after only five cases, from 95 min to 85 min. However, this difference was deemed clinically marginal.¹⁶ Bjorgul et al. additionally showed first attempts averaged 97 min, with times decreasing to 66 min after 25 cases. Together, these studies indicate that while initial improvements in operative efficiency can be observed after only a handful of procedures, attainment of true proficiency in hemiarthroplasty likely requires a higher threshold—potentially 25 cases or more.

4.7 Evaluating the impact of resident involvement on surgical outcomes: risk vs. benefit

The involvement of residents in the operating room has long been debated, particularly regarding whether their participation increases surgical risk. Resident work hour restrictions have also been scrutinized for their influence on surgical outcomes. One study evaluating Dynamic Hip Screw (DHS) fixation before and after the implementation of work hour reforms found that although the mean operative time decreased from 1.63 to 1.25 h, there were no significant differences in wound infection rates, hospital length of stay, postoperative ambulation, fixation adequacy, need for repeat surgery, or mortality,³⁶ suggesting no compromise of surgical quality. Broader analyses of surgical education and fatigue culture indicate that most surgeons and residents believe experience and technical proficiency mitigate the effects of fatigue, even amid concerns about performance decline after prolonged shifts.³⁷ Despite these restrictions, many surgical educators continue to emphasize the importance of developing strategies to manage fatigue effectively, as extended hours remain an inherent aspect of surgical practice.

4.8 Evaluating ACGME case minimums in orthopedic surgery residency: balancing quantity, quality, and supervision

The Accreditation Council for Graduate Medical Education (ACGME) has established case minimums to ensure orthopedic surgery residents gain sufficient experience in essential procedures. As of October 2024, residents must perform at least 60 operative managements of pertrochanteric, intertrochanteric, or femoral neck fractures.³⁸ While operative volume remains important for technical development, supervision quality and case complexity are equally critical. Enhanced supervision of junior residents has been associated with lower reoperation rates, underscoring the value of mentorship beyond simply meeting quantitative requirements.³⁹ Residency programs must therefore balance case volume with instructional quality and diverse operative exposure to prepare trainees for independent practice.