Complications and its impact in patients with closed and open tibial shaft fractures requiring open reduction and internal fixation

The primary objective was to estimate the real-world short-term and mid-term rates of infection and nonunion in patients with open and closed tibial shaft fractures after ORIF. Secondary objectives were to:

The study included 19,383 commercially insured, 2654 Medicare and 2299 Medicaid patients (Figure 1). There was a greater proportion of women in the Medicare (67%) and Medicaid groups (57%) compared with the commercially insured group (47%; Table 1). The majority (64%) of commercially insured patients were under PPO (Preferred Provider Organization) plans. The Medicare population, overall, had the highest mean comorbidity score based on the Charlson Comorbidity Index (1.03 [SD: 1.46], followed by the Medicaid population (0.83 [1.43], and commercial (0.29 [0.78]).

Across all payer groups, there was a greater proportion of closed tibial shaft fractures compared with open (commercial: 91.9 vs 8.1%; Medicare: 91.9 vs 8.1%; Medicaid: 91.6 vs 8.4%).

Table 2 shows the rates of infection and nonunion, by payer and time of follow-up, in which rates for both outcomes increased over study follow-up. Within the first 30 days after index, almost 3% of patients developed infection, with the lowest infection rate (1.8%) observed in patients under commercial insurance, followed by 2.7% in Medicaid, and 2.9% under Medicare. At day 180, infection rates rose almost threefold, to 4.8% (commercially insured), 9.3% (Medicare) and 7.3% (Medicaid), whereas at 365 days, these rates rose to 6.2, 10.1 and 9.2%, respectively.

Nonunion rates were lower than infection rates across payer types within the first 30 days post-index (range 0.3–0.5). At day 180, the rates of nonunion exceeded infection rates specifically in the commercial insurance population (6%), 4.5% under Medicare and 4.7% under Medicaid. At the 365-day time point, these rates were as high as 8.2% for commercially insured patients, 5.8% for the Medicare population and 7% for the Medicaid patients.

By study end (24 months), the rate of infection was 7.4% in 8159 commercially insured patients, 12.3% in 1274 Medicare patients and 11.3% in 1002 Medicaid patients. Nonunion rates were 8.8% in commercially insured, 6.0% in Medicare and 8.0% in Medicaid.

Patients with a diagnosed infection post-ORIF had significantly higher (p < 0.0001 for all) rates of hospital readmissions as well as emergency room visits across every time period during study follow-up, compared with patients with no diagnosed infection (Figure 2). At the 90-day time point, 55% of patient (commercially insured) with diagnosed infection had readmissions compared with only 4% of those without the infection diagnosis. At 365 days, 60% of patients with infection were rehospitalized compared with those without infection (9%). Within the Medicare population, 26% of patients with an infection diagnosis were re-hospitalized within 3 months compared with 6% of those without infections; and 44% in the Medicaid subsample with infection compared with 8% without infection.

The total healthcare cost to the payer incurred was significantly higher for the group developing infections during follow-up (Table 3). At day 30, the average healthcare expenditure for commercially insured patients with infection was US$25,942 (SD: 38,278) which is 11-times greater than the group without infections (US$2360 [SD:8053]; p < 0.0001). At the end of day 365, the average cost incurred by patients with infections rose to as much as US$67,494 (SD: 94,657), more than five-times higher than the group without infections (US$13,434 [SD: 27,436]; p < 0.0001). At 1 year follow-up, younger commercially insured patients were less likely to develop infections, compared with those aged 46–64 years (vs 18–24 years, odds ratio (OR): 0.63; 95% CI: 0.48–0.83; vs 25–45 years, OR: 0.74; 95% CI: 0.63–0.88). Patients with open fractures were 69% more likely to develop infections compared with closed fractures (OR: 1.69; 95% CI: 1.35–2.12). Similarly, patients with open fractures were significantly more likely to develop nonunions compared with those with closed fractures (OR: 1.67; 95% CI: 1.36–2.04) Commercially insured patients diagnosed with diabetes in the pre-index period had a 45% increased likelihood of nonunion (OR: 1.45; 95% CI: 1.08–2.96), whereas Medicare patients with diabetes had an 85% increased likelihood of nonunion (OR: 1.85; 95% CI: 1.03–3.33; Supplementary Material 2)

Tibial shaft fractures, apart from directly adding to the healthcare burden, can be complicated by nonunion and infections. These complications significantly increase the rates of emergency room visits and readmissions with significantly increased total healthcare costs. Small, single setting, studies allow hypothesis generation related to nonunion and infection risk factors, while large real-world data (data relating to patient health status and/or the delivery of healthcare routinely collected from a variety of sources, for example, registries, collections of electronic health records and administrative healthcare claims databases) [14] studies expand our understanding of risks post-ORIF and test the hypothesis. To our knowledge this study is the first to assess nonunion and infection rates in the same population in a large-scale, nationally-representative claims database. This is also the first study to measure these outcomes at different follow-up periods and assess the risk within each timeframe thus, can assist in developing the surgical and post-surgical treatment plan. The study further validates the need to address this patient population to help control costs and improve outcomes.

Recent studies reported post-surgical infection-specific rates of 2–3% at 30 days [9] and about 8% at 7 years [8]. Shepard et al. conducted a retrospective analysis at four Johns Hopkins Health System acute care hospitals and found the surgical site infection rate was 2.76 per 100 surgical procedures; the length-of-stay for patients with infection was almost twice that for patients without infection (10.56 vs 5.64 days in hospital, respectively) with statistically higher 30-day readmission rate (51.94 vs 8.19%, p < 0.001, respectively) [9]. Our nationwide claims-based study showed similar results, with an average 3% infection rate in the 30-day post-ORIF; 30-day readmission rates were 55% for patients with infection and 3% for those without infections. Another center-specific retrospective study of patients treated at a Level 1 trauma center reported a 7-year infection rate of 7.8% (20 out of 256 cases) [8], very similar to our 2-year infection rate of 7.4%, although their 4-week infection rate of 4.7% was higher than our 3% 30-day rate.

Antonova and colleagues [1] conducted a retrospective US managed care claims study of 853 patients with tibial shaft fractures, and reported an overall nonunion rate of 12% over the 2-year follow-up, with a higher rate observed among patients with open fractures (23%) compared with closed (10%). For a similar follow-up duration, we observed a 9% rate of nonunion among patients with both closed and open tibial shaft fractures.

Our study found an 11-fold difference between the mean 30-day total healthcare costs for tibial shaft fractures post-ORIF for those experiencing an infection (US$25,942) compared with those without infection (US$2360); there was a fivefold difference at 1 year. In contrast, the Johns Hopkins Health System study [9] of multiple inpatient surgical procedures (not including tibial shaft fracture [TSF]), reported average post-surgery hospital charges (as opposed to costs) to be 1.8-fold higher for those experiencing infection compared with those who do not (US$79,134/10.56 day median length of stay [LOS] vs US$44,727/5.64 day median LOS), and the average duration of follow-up is unclear. In another cost analysis of nonunion after TSF, Antonova et al. [1] found median 2-year total healthcare costs to be more than twice as high compared with those without nonunion (median US$25,556 vs US$11,686).

Despite study design differences, mainly in case-finding criteria, these closely aligned results offer good external validity to our infection, nonunion and readmission rates. For example, Lin et al. [8] limited study inclusion to patients who had treatment for an infection within 6 months post-surgery, while our study assessed infection rates up to 24 months. Other study design factors provide a rationale for differences in costs between our research and others. Shepard et al. [9] evaluated multiple inpatient surgical procedures, not including TSF, in a single center analysis involving multiple surgical procedure hospital charges over a maximum period of 4 years. Antonova et al. [1] specifically assessed nonunion rates and costs, in an analysis based on treatment approaches from 2006 to 2008, while our cost analysis was specific to surgical infections. Further, our findings are consistent with other literature suggesting greater hospitalization and higher healthcare costs in individuals experiencing surgical complications, including infection [15,16]. This research also provides insight into patient characteristics that may be associated with more intense healthcare utilization immediately post-surgery. Within the acute timeframe, it was observed that males and the elderly patient population were more likely to develop infections and therefore probably require more individualized postoperative/postprocedure care. This care should extend beyond acute management as similar result trends are observed even at a longer follow-up duration (results not reported). Patients with open fractures also need alternative therapy management because their risk of developing infection and nonunion is consistently much higher, across all time frames.

The findings of our study should be considered notwithstanding a few limitations. The inclusion criteria and database used for analysis is limited to individuals with medical and prescription benefit coverage, thereby limiting generalizability to those without health insurance or coverage. The diagnosis codes for infection may lack specificity regarding the exact anatomy of the surgical site and there could be under-reporting or missed diagnoses, based on patient’s choice (not to seek care) or access challenges. In this study, postoperative infection may not have been at the site of the ORIF, however the clinicians on the team reviewed and included codes to limit this potential bias. The presence of postoperative infection was identified based on ICD-9-CM, ICD-10-CM diagnosis and CPT codes, without the availability of laboratory confirmation. Results have been mixed regarding the accuracy of ICD-9-CM diagnosis codes for infection after various procedures [17]. However, administrative data using ICD-10-CM codes have shown reasonable test performance characteristics for identification of complex infections after surgery [18]. Future research using newer data with ICD-10 codes and laboratory confirmation is warranted to validate study findings [19]. Patients with adherence problems may be less likely to seek or comply with care, causing under-reporting of outcomes, and contribute to a selection bias among the sampled population. Additionally, the identification of patients and outcomes relied upon on clinician’s use of diagnosis codes for billing purposes and therefore cannot completely exclude the likelihood of reporting errors or biases. This study is the first, to our knowledge, to investigate differences in rates of infection, nonunion and readmission, as well as healthcare costs by insurance type (commercial, Medicaid, Medicare); however, the association between fracture-related variables (e.g., compound, displaced, comminuted, among others) or patient characteristics (e.g., smoking, gender, age) on the outcome measures deserves further investigation. Productivity costs and indirect costs were not assessed; other studies have suggested that these costs may be substantial, with indirect costs exceeding direct medical healthcare costs by a factor of two in patients with tibia fractures [20]. This finding may be exacerbated by the fact that many trauma patients are young otherwise productive adults that may see substantial drop in productivity as a result of trauma. Future studies should evaluate this as well as explore other patient-reported outcomes related to health-related quality of life. Despite these limitations, this study provides essential and needed perspective on post-surgical tibial shaft fracture complication rates and specific medical costs associated with infection over specific time periods and by type of insured population.

In this nationwide retrospective real-world analysis of patients with tibial shaft fractures undergoing ORIF, an increasing rate of infection up to 2 years was determined. These post-surgical complications are accompanied by added healthcare resource use and expenditure.

To view the supplementary data that accompany this paper please visit the journal website at: Supplementary Material

This study was funded by Johnson and Johnson. ASC, MV, CS, JM and CEH are employees and stock holders of Johnson & Johnson. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Writing assistance for the initial draft was provided by eMAX Health, White Plains, NY, which received funding from Johnson and Johnson.

IBM MarketScan database is in compliance with the Health Information Portability and Accountability Act of 1996 (HIPAA). Only de-identified patient records from IBM MarketScan database were used in this study; hence Institutional Review Board approval was not required.

This work is licensed under the Attribution-NonCommercial-NoDerivatives 4.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-nd/4.0/