Safety and Outcomes of Inpatient Compared with Outpatient Surgical Procedures for Ankle Fractures

Qin, Charles BA; Dekker, Robert G. MD; Blough, Jordan T. BS; Kadakia, Anish R. MD

Journal of Bone & Joint Surgery - American Volume:
doi: 10.2106/JBJS.15.01465
Scientific Articles
Abstract

Background: As the cost of health-care delivery rises in the era of bundled payments for care, there is an impetus toward minimizing hospitalization. Evidence to support the safety of open reduction and internal fixation (ORIF) of ankle fractures in the outpatient setting is largely anecdotal.

Methods: Patients who underwent ORIF from 2005 to 2013 were identified via postoperative diagnoses of ankle fracture and Current Procedural Terminology codes; patients with open fractures and patients who were emergency cases were excluded. Patients undergoing inpatient and outpatient surgical procedures were propensity score-matched to reduce differences in the baseline characteristics. Primary tracked outcomes included medical and surgical complications, readmission, and reoperation within 30 days of the procedure. Binary logistic regression models were created that determined the risk-adjusted relationship between admission status and primary outcomes.

Results: Outpatient surgical procedures were associated with lower rates of urinary tract infection (0.4% compared with 0.9%; p = 0.041), pneumonia (0.0% compared with 0.5%; p = 0.002), venous thromboembolic events (0.3% compared with 0.8%; p = 0.049), and bleeding requiring transfusion (0.1% compared with 0.6%; p = 0.012). Outpatient status was independently associated with reduced 30-day medical morbidity (odds ratio, 0.344 [95% confidence interval, 0.201 to 0.589]). No significant differences were uncovered with respect to surgical complications (p = 0.076), unplanned reoperations (p = 0.301), and unplanned readmissions (p = 0.358).

Conclusions: In patients with closed fractures and minimal comorbidities, outpatient ORIF was associated with reduced risk of select 30-day medical morbidity and no difference in surgical morbidity, reoperations, and readmissions relative to inpatient. Factors unaccounted for when creating matched cohorts may impact our results. Our findings lend reassurance to surgeons who defer admission for low-risk patients.

Level of Evidence: Therapeutic Level III. See Instructions for Authors for a complete description of levels of evidence.

Author Information

1Department of Orthopaedic Surgery, Northwestern University Feinberg School of Medicine, Chicago, Illinois

E-mail address for C. Qin: c-qin@northwestern.edu

E-mail address for R.G. Dekker: robbertgdekker@gmail.com

E-mail address for J.T. Blough: joblough@northwestern.edu

E-mail address for A.R. Kadakia: Kadak259@gmail.com

Article Outline

Ankle fractures are a common injury in adults, representing more than half of all fractures of the foot and ankle seen at major trauma centers in the United States1. Over the last 50 years, there has been a steady increase in the incidence as well as severity of ankle fractures, most notably in the elderly and those with osteoporosis2. A substantial proportion of these injuries will require open reduction and internal fixation (ORIF), a procedure not without risk3-6. Although perioperative admission is often elected to accommodate the operating room schedule and for patient convenience given various social factors including the pre-injury home environment, this practice has important financial implications. Recent studies have documented the substantial cost associated with inpatient ORIF in both the United States and Europe7-9.

As the price of health-care delivery and the popularity of bundled payments for care rise, there is an impetus toward minimizing unnecessary hospitalization. As a result, greater attention has been paid to assessing the safety of outpatient surgical procedures in various fields, including orthopaedic spine procedures and total joint arthroplasty10-13. However, there is a paucity of similar literature in the field of foot and ankle and orthopaedic trauma, with the only existing study limited by its single-institution sample size and inability to effectively minimize selection bias13. In short, evidence to support ORIF of ankle fractures in the outpatient setting is largely anecdotal.

Although under certain circumstances inpatient admission may be unavoidable, increased scrutiny of the patient safety implications of alternative approaches to care is warranted. In the absence of retrospective or prospective studies to address this question, we used the American College of Surgeons-National Surgical Quality Improvement Project (ACS-NSQIP) database to compare perioperative outcomes between properly selected cohorts of patients undergoing inpatient and outpatient ORIF, and we reduced selection bias with propensity score matching. We tested the null hypothesis that ORIF performed in the outpatient setting is as safe as ORIF performed in the inpatient setting.

Back to Top | Article Outline

Materials and Methods

Population

The details of the ACS-NSQIP data collection methods have previously been described and validated14,15. Patients undergoing ORIF from 2005 to 2013 were first identified by the postoperative diagnosis of ankle fracture (International Classification of Diseases, Ninth Revision [ICD-9] codes 824.0-824.9). Patients with primary Current Procedural Terminology (CPT) codes 27766 (ORIF of medial malleolus fracture), 27769 (ORIF of posterior malleolus fracture), 27792 (ORIF of lateral malleolus fracture), 27814 (ORIF of bimalleolar fracture), and 27822/27823 (ORIF of trimalleolar fracture) were included in the study. Emergency cases, cases with preoperative sepsis, or cases with open fractures were excluded as these injuries were more likely to warrant inpatient admission and, as such, did not qualify as standard closed ankle fractures that the study aimed to capture. The remaining patients were stratified by whether they underwent an outpatient or inpatient surgical procedure as defined by the facility at which the patient underwent the procedure. The stratification and all analyses regarding this stratification were determined by the initial designation of the patient prior to the surgical procedure.

Back to Top | Article Outline
Variables

Demographic variables (e.g., age, sex, race, body mass index), lifestyle variables (e.g., smoking, alcohol use), and medical comorbidities (e.g., American Society of Anesthesiologists [ASA] classification, ischemic heart disease, diabetes, chronic obstructive pulmonary disease, hypertension, peripheral vascular disease, corticosteroid use) were queried. Procedural factors included fracture type, anesthesia type, and total operative time. Primary tracked outcomes included surgical and medical complications, unplanned readmissions, and unplanned reoperations. Surgical complications included superficial and deep surgical site infections and wound dehiscence. Medical complications included deep venous thrombosis, pulmonary embolism, unplanned reintubation, ventilator dependence for >48 hours, progressive renal insufficiency, acute renal failure, coma, stroke, cardiac arrest, myocardial infarction, peripheral nerve injury, pneumonia, urinary tract infection, blood transfusions, and sepsis or septic shock. All readmissions and reoperations tracked were related to the index procedure. Finally, admission to the hospital of an outpatient was queried on the basis of a length of hospital stay of ≥24 hours16. All outcomes occurred within 30 days of the index procedure. All variables were used as defined in the NSQIP user guide.

Back to Top | Article Outline
Statistical Analysis

Patients undergoing inpatient and outpatient ORIF were propensity score-matched to reduce differences in the 15 patient and procedural factors mentioned above. Via logistic regression modeling, propensity scores were derived for each case and were used to establish pairs in a 1:1 fashion17. In short, propensity score matching minimizes selection bias, allowing a more accurate assessment of treatment effect. Chi-square tests for categorical variables and Student t tests for continuous variables were used to identify differences in perioperative variables between inpatient and outpatient cohorts. In addition, a subgroup analysis of the elderly population, defined as those who were ≥65 years of age, was performed given the relevance of age to inpatient admission. Finally, to determine whether the subset of outpatients who were admitted impacted our findings, we performed a second subanalysis comparing outpatients (excluding those who had a hospital stay of ≥24 hours) and inpatients.

The problem of multiple comparisons was addressed using the Bonferroni adjustment, with significance established at 0.0026 for 19 comparisons. Binary logistic regression models were created that determined the risk-adjusted relationship between admission status and primary outcomes. In the logistic regression, significance was set at p < 0.05. The Hosmer-Lemeshow test was calculated to assess model calibration and c-statistics were calculated to assess discriminatory capability18,19. All analysis was performed using SPSS version 22 (IBM).

Back to Top | Article Outline

Results

After excluding patients undergoing an emergency surgical procedure, those with preoperative sepsis, and those with open fractures, a total of 5,256 patients met inclusion criteria, of whom 2,344 were designated as inpatients and 2,912 were designated as outpatients. Upon derivation of the propensity score, a c-statistic of 0.804 was calculated. Upon matching on the propensity score, cases without a match were eliminated, producing a final cohort of 3,732 patients, of whom 1,866 were designated as inpatients and 1,866 were designated as outpatients. Table I summarizes the comparisons between the 2 subgroups based on the matched variables. With the exception of the difference in age (51.0 years in the inpatient group compared with 49.0 years in the outpatient group; p = 0.043), the differences in all matched variables were successfully reduced.

Table II summarizes univariate comparisons of postoperative morbidity, reoperations, and readmissions. Prior to the Bonferroni correction, outpatient surgical procedures were associated with lower rates of urinary tract infection (0.4% compared with 0.9%; p = 0.041), pneumonia (0.0% compared with 0.5%; p = 0.002), venous thromboembolic events (0.3% compared with 0.8%; p = 0.049), and bleeding requiring transfusion (0.1% compared with 0.6%; p = 0.012). Upon adjusting for multiple comparisons, only the difference in pneumonia remained significant (p = 0.002).

Upon subanalysis of only elderly patients (those ≥65 years of age), outpatient surgical procedures were similarly associated with lower rates of the aforementioned complications, but the difference was no longer significant (p > 0.0026, Bonferroni adjusted) (Table III). We uncovered a significantly higher rate of non-home-bound discharge among inpatients compared with outpatients in both the original analysis (44.5% compared with 26.7%) and the subanalysis (56.9% compared with 28.0%). When the 5.1% of outpatients who had to be admitted to the hospital for a stay of ≥24 hours (Table II) were eliminated from the analysis, rates of pneumonia remained significantly lower in the outpatient cohort (p = 0.002) (Table IV).

Age, the only factor for which the difference was not effectively reduced upon matching, was included in a multivariate analysis along with other variables meeting inclusion criteria to determine the risk-adjusted relationship between outpatient status and outcomes (Table V). Outpatient status was independently associated with reduced 30-day medical morbidity (odds ratio [OR], 0.344 [95% confidence interval (95% CI), 0.201 to 0.589]) and decreased likelihood of non-home-bound discharge (OR, 0.457 [95% CI, 0.399 to 0.525]). Relative to inpatients, outpatient management was not associated with significantly different risk (p < 0.0026) of surgical complications, unplanned reoperations, and unplanned readmissions.

Back to Top | Article Outline

Discussion

The safety of outpatient surgical procedures has become an active area of investigation involving both the orthopaedic spine and total joint arthroplasty10,11,20,21. To our knowledge, previous studies in the area of foot and ankle trauma, specifically examining the treatment of ankle fractures, have been rare. In 476 patients with an ankle fracture, Weckbach et al. found fewer postoperative complications and unanticipated surgical revisions in patients discharged home after ORIF, but they failed to balance differences in procedural and patient factors between cohorts, which resulted in selection bias13. Recent literature has suggested that common comorbidities such as obesity and diabetes have a dramatic effect on health-care utilization and costs among patients surgically managed for ankle fracture9. In the current study, we used propensity score matching to create cohorts for whom the baseline patient and procedural characteristics were balanced to minimize the potential for selection bias. With the exception of a 2-year difference in age (51.0 years compared with 49.0 years; p = 0.043), all factors were not significantly different. We also performed a subanalysis of patients ≥65 years of age and included age in our regression models to elaborate what effect, if any, the difference in age played in our findings.

The most important finding of our study was that, in a properly selected cohort of patients, outpatient ORIF was associated with reduced 30-day medical morbidity and there were no differences in surgical morbidity, reoperations, and readmissions relative to inpatient ORIF. With respect to medical morbidity, outpatients experienced reduced rates of venous thromboembolic events, urinary tract infection, pneumonia, and bleeding requiring transfusion, but only the difference in pneumonia was significant upon Bonferroni correction. Upon subanalysis, similar differences were detected, but they were not significant. Our findings suggest that outpatient ORIF of a fractured ankle heralds no additional medical or surgical risk to the patient and is as safe as an inpatient surgical procedure.

Intubation, poor oral care, supine positioning, and immobility all may contribute to the increased risk of postoperative pneumonia that our study uncovered22. Unfortunately, many of these factors are present during an inpatient stay. Given that hospital-acquired pneumonia is often multidrug-resistant and associated with higher rates of mortality relative to that which is community-acquired, our findings support outpatient surgical procedures as an acceptable alternative23.

Although it seems plausible that increasing age may also be driving the greater complication rates seen in inpatient ORIF in our study, our subanalysis of only patients ≥65 years of age demonstrates no significant difference in the rate of postoperative events except for non-home-bound discharge. Currently, there is conflicting literature as to whether age itself or comorbidity burden is the major risk factor for postoperative complications5,24. Our findings lend reassurance to surgeons when deferring postoperative admission of patients who have either no comorbidities or comorbidities that are largely stable.

It is our experience that elderly patients are often admitted for safety concerns, namely involving fear of home injury, ability to walk with assistance, and stability of medical conditions. Our analysis found that an outpatient surgical procedure was associated with a similar risk of reoperations and readmissions. It would not be unreasonable to strongly consider home care nursing as a viable alternative to postoperative admission for patients who undergo ORIF and whose health is largely unaffected by a hospital stay. In light of recent evidence of the impressive economic burden of inpatient admission for ankle fracture, operative intervention without admission to the hospital becomes an attractive option7.

Our study was not without limitations. First, although propensity score matching is an accepted method to minimize selection bias, our findings may have reflected intrinsic differences of the surgical setting or perioperative details unaccounted for in the NSQIP; therefore, there may have been residual confounding factors at play. In a similar vein, the NSQIP data set does not specify whether the admission occurred before or after the surgical procedure or both and the reason for admitting a patient initially designated as an outpatient. Our inability to account for these factors precludes us from creating truly identical cohorts with only a difference in surgical setting. We found that 92.8% of inpatients stayed in the hospital after an ankle surgical procedure for a mean time of 2.24 days. Lastly, the NSQIP only captures events within 30 days of the index procedure, which may have biased our results. We recognize that inpatients may benefit from heightened surveillance perioperatively prior to discharge, and therefore there may be a bias toward detecting major complications such as unplanned reintubation, myocardial infarction, or pulmonary embolism. In our study, the rates of these complications were not significantly different between the two cohorts. The rates of complications typically detected in the outpatient setting were also similar in our study and were reinforced by a similar readmission rate. It is our experience that patients with ankle fracture who are treated as outpatients will follow up in the clinic setting at least once in the 30-day period, reducing the potential bias that our follow-up period may generate. Further investigation through a prospective randomized approach will help to fully elucidate differences in safety between inpatient and outpatient ORIF.

In conclusion, in patients with closed ankle fractures and few comorbidities, outpatient ORIF was associated with reduced 30-day pneumonia and no difference in surgical morbidity, reoperations, and readmissions relative to inpatient ORIF. Our findings may increase the acceptability of ORIF in the outpatient setting in both cost-savings and patient safety models.

Investigation performed at the Department of Orthopaedic Surgery, Northwestern University Feinberg School of Medicine, Chicago, Illinois

Disclosure: There was no external funding source for this study. The Disclosure of Potential Conflicts of Interest forms are provided with the online version of this article.

Back to Top | Article Outline

References

1. Shibuya N, Davis ML, Jupiter DC. Epidemiology of foot and ankle fractures in the United States: an analysis of the National Trauma Data Bank (2007 to 2011). J Foot Ankle Surg. 2014 ;53(5):606–8. Epub 2014 Apr 29.
2. Kannus P, Palvanen M, Niemi S, Parkkari J, Järvinen M. Increasing number and incidence of low-trauma ankle fractures in elderly people: Finnish statistics during 1970-2000 and projections for the future. Bone. 2002 ;31(3):430–3.
3. SooHoo NF, Krenek L, Eagan MJ, Gurbani B, Ko CY, Zingmond DS. Complication rates following open reduction and internal fixation of ankle fractures. J Bone Joint Surg Am. 2009 ;91(5):1042–9. Epub 2009 May 5.
4. Miller AG, Margules A, Raikin SM. Risk factors for wound complications after ankle fracture surgery. J Bone Joint Surg Am. 2012 ;94(22):2047–52.
5. Belmont PJ Jr, Davey S, Rensing N, Bader JO, Waterman BR, Orr JD. Patient-based and surgical risk factors for thirty-day postoperative complications and mortality following ankle fracture fixation in hospitalized patients. J Orthop Trauma. 2015 ;29(12):e476–82.
6. Basques BA, Miller CP, Golinvaux NS, Bohl DD, Grauer JN. Morbidity and readmission after open reduction and internal fixation of ankle fractures are associated with preoperative patient characteristics. Clin Orthop Relat Res. 2015 ;473(3):1133–9. Epub 2014 Oct 22.
7. Murray AM, McDonald SE, Archbold P, Crealey GE. Cost description of inpatient treatment for ankle fracture. Injury. 2011 ;42(11):1226–9. Epub 2010 Sep 23.
8. Regan DK, Manoli A 3rd, Hutzler L, Konda SR, Egol KA. Impact of diabetes mellitus on surgical quality measures after ankle fracture surgery: implications for “value-based” compensation and “pay for performance”. J Orthop Trauma. 2015 ;29(12):e483–6. Epub 2015 Nov 26.
9. Cavo MJ, Fox JP, Markert R, Laughlin RT. Association between diabetes, obesity, and short-term outcomes among patients surgically treated for ankle fracture. J Bone Joint Surg Am. 2015 ;97(12):987–94.
10. Pugely AJ, Martin CT, Gao Y, Mendoza-Lattes SA. Outpatient surgery reduces short-term complications in lumbar discectomy: an analysis of 4310 patients from the ACS-NSQIP database. Spine (Phila Pa 1976). 2013 ;38(3):264–71. Epub 2012 Jul 21.
11. Adamson T, Godil SS, Mehrlich M, Mendenhall S, Asher AL, McGirt MJ. Anterior cervical discectomy and fusion in the outpatient ambulatory surgery setting compared with the inpatient hospital setting: analysis of 1000 consecutive cases. J Neurosurg Spine. 2016 ;24(6):878–84. Epub 2016 Feb 5.
12. Qin C, Antony AK, Aggarwal A, Jordan S, Gutowski KA, Kim JY. Assessing outcomes and safety of inpatient versus outpatient tissue expander immediate breast reconstruction. Ann Surg Oncol. 2015 ;22(11):3724–9. Epub 2015 Feb 5.
13. Weckbach S, Flierl MA, Huber-Lang M, Gebhard F, Stahel PF. [Surgical treatment of ankle fractures as an outpatient procedure. A safe and resource-efficient concept?]. Unfallchirurg. 2011 ;114(10):938–42. Epub 2011 Jul 23. German.
14. Cohen ME, Ko CY, Bilimoria KY, Zhou L, Huffman K, Wang X, Liu Y, Kraemer K, Meng X, Merkow R, Chow W, Matel B, Richards K, Hart AJ, Dimick JB, Hall BL. Optimizing ACS NSQIP modeling for evaluation of surgical quality and risk: patient risk adjustment, procedure mix adjustment, shrinkage adjustment, and surgical focus. J Am Coll Surg. 2013 ;217(2):336–46.e1. Epub 2013 Apr 28.
15. Birkmeyer JD, Shahian DM, Dimick JB, Finlayson SR, Flum DR, Ko CY, Hall BL. Blueprint for a new American College of Surgeons: National Surgical Quality Improvement Program. J Am Coll Surg. 2008 ;207(5):777–82. Epub 2008 Sep 19.
16. Centers for Medicare & Medicaid Services. Medicare benefit policy manual, chapter 1 – inpatient hospital services covered under Part A. 2012 Apr 5. https://www.cms.gov/Regulations-and-Guidance/Guidance/Manuals/downloads/bp102c01.pdf. Accessed 2016 Jun 29.
17. Austin PC. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm Stat. 2011 ;10(2):150–61. Epub 2010 Oct 7.
18. Merkow RP, Hall BL, Cohen ME, Dimick JB, Wang E, Chow WB, Ko CY, Bilimoria KY. Relevance of the c-statistic when evaluating risk-adjustment models in surgery. J Am Coll Surg. 2012 ;214(5):822–30. Epub 2012 Mar 21.
19. Paul P, Pennell ML, Lemeshow S. Standardizing the power of the Hosmer-Lemeshow goodness of fit test in large data sets. Stat Med. 2013 ;32(1):67–80. Epub 2012 Jul 26.
20. Lied B, Rønning PA, Halvorsen CM, Ekseth K, Helseth E. Outpatient anterior cervical discectomy and fusion for cervical disk disease: a prospective consecutive series of 96 patients. Acta Neurol Scand. 2013 ;127(1):31–7. Epub 2012 May 10.
21. Helseth Ø Lied B, Halvorsen CM, Ekseth K, Helseth E. Outpatient cervical and lumbar spine surgery is feasible and safe: a consecutive single center series of 1449 patients. Neurosurgery. 2015 ;76(6):728–37, discussion: 737-8. Epub 2015 Apr 9.
22. Myrianthefs PM, Kalafati M, Samara I, Baltopoulos GJ. Nosocomial pneumonia. Crit Care Nurs Q. 2004 ;27(3):241–57.
23. Hoo Lee J, Hyung Kim Y. Comparison of clinical characteristics between healthcare-associated pneumonia and community-acquired pneumonia in patients admitted to secondary hospitals. Braz J Infect Dis. 2012 ;16(4):321–8.
24. Lynde MJ, Sautter T, Hamilton GA, Schuberth JM. Complications after open reduction and internal fixation of ankle fractures in the elderly. Foot Ankle Surg. 2012 ;18(2):103–7. Epub 2011 Apr 30.
Copyright 2016 by The Journal of Bone and Joint Surgery, Incorporated