2.1 Study design

This was a retrospective review conducted at an academic acute care hospital in an urban area of the western United States. Institutional Review Board (IRB) approval was obtained. The electronic health record (EHR) was reviewed from the time period between January 1st^, 2015–December 31st, 2019 to identify patients managed operatively for a hip fracture using surgeon case logs. Two surgeon case logs were used, both of which are board certified fellowship trained orthopaedic trauma surgeons.

2.2 Setting

The study was performed at an academic acute care hospital in an urban area of the western United States.

2.3 Participants

Included patients were those who underwent operative fixation of a hip fracture, including OTA/AO classification 31-A, 31-B, and 32.1.¹⁸ Exclusion criteria included age <40 years old, death during admission for the hip fracture, hip fractures sustained from high energy trauma, implant failure requiring revision as the primary indication for surgery, non-operative management, patients who did not have a scheduled POV, patients who had a first POV scheduled outside of four weeks from discharge, patients who were still admitted or were re-admitted prior to their scheduled first scheduled post-operative visit, prophylactic fixation for oncological process or stress fracture, and peri-prosthetic fractures below a total hip arthroplasty. If a patient sustained a contralateral hip fracture within the analyzed time-period, only the first encounter was included in the analysis, but the contralateral hip fracture was noted as a complication if it occurred within the 90-day period.

2.4 Variables

Patient charts were reviewed for demographics, comorbidities, fracture characteristics, fixation method, time to operating room, pre-injury ambulatory status (independent, community ambulator with assistive device, household, or non-ambulatory), 90-day re-admission, 90-day complications, and mortality. Comorbidities evaluated included dementia, coronary artery disease, atrial fibrillation, chronic obstructive pulmonary disease requiring oxygen supplementation at home, hypertension, diabetes mellitus, end-stage renal disease, smoking, and congestive heart failure.

Attendance or absence at the scheduled first orthopaedic POV was collected. At our institution, all patients are provided a scheduled follow up visit within two to four weeks from date of discharge. Our EHR records presence or absence at all scheduled visits. If a patient misses their scheduled appointment, a medical assistant reaches out to the patient and their family to attempt a re-schedule. For this study, a patient was marked as “absent” for their first POV if they had a scheduled orthopaedic follow up within four weeks from discharge but failed to make it to a POV within that same timeframe.

90-day complications included deep surgical site infection, dislocation, implant failure, periprosthetic fracture, return to the operating room, anemia requiring transfusion, deep venous thrombosis, pulmonary embolism, contralateral hip fracture, and a systemic infection or illness requiring hospitalization. Mortality was determined by EHR review, review of the state death registry, and by internet searches with the online obituaries.^19,20 If these methods were inconclusive, a family member was contacted at the number provided in the chart.

Comorbidity effect across patient data was evaluated by tallying the presence of each comorbidity per subject within the group who made their first post-operative visit and the group who missed their first post-operative visit. Each comorbidity tally count listed within patients who made and patients who missed their first post-operative visit were normalized by the total number of subjects within each group.

2.5 Statistical method

Data analysis was conducted via statistical analysis software JMP and SAS (Cary, NC). Two-tailed T-test for continuous numerical data and chi-square test for nominal categorical data were used for comparisons. Bivariate analysis was performed through chi-square and Fisher's Exact test. Multivariate analysis was performed through logistic regression model. For categories with more than two parameters, Cramer's V analysis was performed to validate associations between groups. For all tests, a p value of 0.05 was considered significant.

3.1 Participants

A total of 285 hip fractures were managed operatively and met inclusion criteria from January 1st^, 2015–December 31st, 2019. There were 150 (52.6%) femoral neck, 115 (40.4%) intertrochanteric, and 20 (7.0%) subtrochanteric fractures. Of these, 279 had sufficient data for final analysis. Average age was 80 years old (range 42–103), and the majority were aged >70 years old (82.3%) and female (68%).

3.2 Descriptive data

Of the 279 patients, 213 made their first post-operative visit for a follow-up success rate of 76.3%. Male patients, ASA class 3 and 4, and discharge to skilled nursing facility (SNF) were significantly more likely to miss their first POV (Table 1). Age, BMI, and comorbidity index did not differ significantly between the groups. ASA classes 1&2 and 3&4 were combined during statistical analysis, due to a low distribution of samples.

3.3 Outcome data

Of the 279 patients, 129 (46.2%) were found to have died since their surgery. The average time from date of surgery to date of death was 598 days, with a one-year mortality rate of 21.9% combined between the two groups. Missing the first POV was significantly associated with 90-day and 1 year mortality (Table 2).

Bivariate analysis showed that patients with ASA class 3 or 4 (OR 5.71, 95% CI 2.36–13.85, p < 0.005), COPD on home oxygen (OR 4.35, 95% CI 1.27–14.84, p < 0.005), CAD (OR 2.66, 95% CI 1.43–4.96, p = 0.0321), pre-injury status below independence without assist (p < 0.005), and who missed their first post-operative visit (OR 7.78, 95% CI 4.18–14.49, p < 0.005) were significantly associated with 90-day re-admission. Multivariate analysis revealed that missing the first POV, COPD on home oxygen, and Pre-injury status were independent predictors of 90-day re-admission (Table 3). Missing the first POV had the highest OR at 8.04 (95% CI 4.06–15.93, p < 0.005).

Bivariate analysis demonstrated that patients who missed their first POV (OR 10.41, 95% CI 5.52–19.60, p < 0.005), ASA class 3 or 4 (OR 5.71, 95% CI 2.36–13.85, p < 0.005), pre-injury status being below independent (p < 0.005), COPD on home oxygen (OR 4.35, 95% CI 1.27–14.83, p = 0.0112), CAD (OR 2.66, 95% CI 1.42–4.96, p < 0.005), HTN (OR 1.72, 95% CI 1.01–2.96, p = 0.0446), and CHF (OR 2.05, 95% CI 1.04–4.04, p = 0.0352) were all significantly associated with 90-day complications. Multivariate analysis revealed that missing the first POV, CAD, and pre-injury status were independent predictors of 90-day complication (Table 4). Missing the first POV had the highest OR for 90-day complications at 10.65 (95% CI 5.24–21.64, p < 0.005).

4.1 Key results

Our study is the first to evaluate attendance at the first POV following a hip fracture surgery as a risk factor for complications, re-admission, and mortality. The novel finding of this study was that missing the first post-operative visit was the strongest independent factor associated with both 90-day readmission and 90-day complication in patients post-operative from hip fracture surgery, thus confirming our hypothesis. The odds ratios were profound, at 8.0 and 10.6, respectively. Patients who missed their first POV were also significantly more likely to have 90-day and one-year mortality. To our knowledge, this is the first study highlighting this specific risk factor.

Prior studies have questioned the utility of the post-operative visit following a hip fracture. Since patients are most commonly weight bearing as tolerated following surgery, routine radiographs without a traumatic event or change in symptoms might not be necessary.²¹ In a study of 423 patients status post hip fracture fixation, Kuorikoski et al. found that the first post-operative visit led to a change in treatment course in just one patient.²² A more recent study found a 0.9% change in treatment in the care of 995 hip fracture cases.²³ Reich et al. found a higher yield of 9.4%, but noted that 80% of these cases were problems noticed by the patient or care provider, who then called for the specific visit.²⁴

Missing the first post-operative visit was associated with the highest odds ratio of predicting 90-day complication and re-admission when compared across multiple different comorbidities and patient factors. We argue that while few post-operative visits will lead to a change in management, for example a return to the operating room for a revision surgery, the results of this study demonstrate that there may be a potential benefit of the first POV that affects patient outcome with or without a change in surgical or medical management. There are potential benefits of the first POV that may be overlooked because they do not lead to a change in management. The first POV allows the patient and the patient's family to see the surgeon and treatment team and feel that there is support and investment in the patient's recovery. The first POV allows the surgeon to address concerns about returning to life at home, safety at home, potential need for additional care at home and to encourage ambulation and nutrition. While addressing these social factors may not lead to a change in management it is possible that it may affect patient outcome.

The findings of this study can be used by surgeons to give reassurance to patients present at their first POV that they are likely to have a better outcome with evidence of lower complication and re-admission rates, as well as lower early mortality. A question for future research will be whether the presence at the first POV is a modifiable risk factor. It remains to be seen if complications, re-admission, and mortality could be decreased by increasing attendance at the first POV. Lack of transportation is a potential cause of missing the first POV. If further study revealed a causal relationship between missing the first POV and the associated increased complication rate, then simply getting patients to the first POV may be able to improve outcomes.

A secondary goal of our study was to evaluate other risk factors for poor outcomes following hip fracture surgery for comparison. Multiple studies have shown that the post-operative period following a hip fracture is rife with complications, and often involves a rapid decline in health that can lead to early mortality.^25–27 Decreased pre-injury ambulation status has consistently been shown to predict higher complications and mortality in patients with hip fractures.^28,29 Our study reinforces this by showing an independent association of pre-injury ambulation with 90-day complication and re-admission. Odds ratio for these outcomes rose as functional status declined. This finding, in tandem with prior research, can be helpful for surgeons when discussing recovery expectations with patients and their family.

The presence of comorbidities is another risk factor that the orthopaedic surgeon should be aware of. CAD and chronic respiratory diseases have been shown to be significantly associated with post-operative complications following hip fracture fixation.^8,30 Our study reflected this trend. CAD was an independent risk factor for 90-day complications, while COPD requiring home oxygen was an independent risk factor for 90-day re-admission.

4.2 Limitations & strength

Our study is not without limitations, the first of which is its retrospective design, which can introduce bias. The timing of the first post-operative visit was not conducted at a specific time point but rather within two to four weeks from discharge. We do not consider this a major issue given that this follow up should be consistent with other orthopaedic practices, and variances among different providers are to be expected. It is possible that an earlier first POV at two weeks could be better than a first POV closer to four weeks but that was not looked at in this study. Future studies could narrow the timing of the first POV or separate early versus late first POVs to determine if there is any significant diference with timing less than 4 weeks. This study does not include functional analyses for patients who did or did not make their first POV. For statistical analysis we had to combine ASA classes 1 with classes 2 and ASA classes 3 with 4 due to a low sample number of ASA class 1 patients. However, ASA classes 3 and above have been shown to be associated with higher rates of post-operative complications when compared to ASA classes 1 and 2²⁷.