Background: The demand for total hip arthroplasty (THA), total knee arthroplasty (TKA), and unicompartmental knee arthroplasty (UKA) is growing rapidly because of the proven success of these procedures and an increase in the aging population. However, resources are limited and health-care budgets are finite. Recently, a number of care providers have started performing these procedures on an outpatient basis, with the patients being discharged from the hospital on the day of surgery. The primary objective of this systematic review was to examine the evidence regarding the safety and feasibility of performing THA, TKA, or UKA on an outpatient basis.
Methods: An electronic search of 3 online databases (Embase, MEDLINE, and HealthSTAR) was conducted to identify eligible studies. The reference lists of identified articles were then screened. All studies evaluating outcomes following outpatient THA, TKA, or UKA were included. Eligible articles that included a comparative group were assessed for methodological quality with use of the Cochrane Risk of Bias Assessment Tool for Non-Randomized Studies of Interventions (ACROBAT-NRSI). Non-comparative studies were assessed for quality with use of the Effective Public Health Practice Project (EPHPP) Quality Assessment Instrument.
Results: The electronic literature search yielded 805 articles. Following a review of the titles, abstracts and reference lists, 26 articles remained and were assessed for eligibility. Of those, 17 articles (≈79,500 patients) met the inclusion criteria and were included in the review. Although both quality-assessment tools showed that the majority of studies included in the review were of poor quality, there was no increase in readmission rates or perioperative complications among patients who underwent outpatient procedures. Studies assessing satisfaction illustrated a high level of satisfaction for the majority of patients. The average age of the patients in the THA studies ranged from 53.5 to 63 years. The TKA and UKA studies included an older population, with mean ages ranging from 55 to 68 years. The majority of the included studies included a larger percentage of males as compared with females. Of the 17 included studies, 4 were cohort studies with a control group and 13 were case series. All 4 cohort studies indicated that the complication rates and clinical outcomes were similar between the inpatient and outpatient groups. Furthermore, the 3 studies that involved an economic analysis indicated that outpatient arthroplasty is financially advantageous.
Conclusions: In selected patients, outpatient THA, TKA, and UKA can be performed safely and effectively. The included studies lacked sufficient internal validity, sample size, methodological consistency, and standardization of protocols and outcomes. There is a need for high-quality prospective cohort and randomized trials to definitively assess the safety and effectiveness of outpatient THA, TKA, and UKA.
Level of Evidence: Therapeutic Level IV. See Instructions for Authors for a complete description of levels of evidence.
Total hip arthroplasty (THA), total knee arthroplasty (TKA), and unicompartmental knee arthroplasty (UKA) are established and proven procedures for the treatment of advanced arthritis. Patient outcomes and complication rates are continually improving because of advanced surgical techniques and better prosthetic designs1,2. The annual demand for these procedures is expected to increase substantially in the near future because of the rapidly growing elderly population and the increased public awareness of the ability of these procedures to improve the quality of life3. As the demand for joint replacements grows, it is imperative to improve patient safety and satisfaction3 while minimizing costs and optimizing the use of health-care resources. To accomplish these goals, surgical teams, nursing staff, and physiotherapists work together to discharge patients from the hospital as soon as is safely possible. Recent studies involving THA, TKA, and UKA have analyzed outpatient care pathways, in which patients are discharged from the hospital on the day of surgery1-20.
To successfully achieve outpatient arthroplasty, clinicians have developed accelerated clinical pathways featuring a multidisciplinary approach involving a range of health-care professionals. The pathway includes extensive preoperative patient education, early mobilization, and intensive physical therapy1,2. These procedures can be safely performed in an outpatient setting as a result of a multifactorial approach, including improved analgesia, early physiotherapy, and advanced surgical technique. Changes to surgical techniques include careful hemostasis and decreased muscle damage. Preservation of neuromuscular units, with resultant minimized soft-tissue damage, leads to less pain, quicker recovery, and rapid mobilization following surgery21,22.
All patients undergoing an outpatient joint replacement enroll in a comprehensive clinical pathway that includes preoperative, intraoperative, and postoperative care1,2,4-6,13. For successful outpatient surgery, patients must be given appropriate goals and expectations through preoperative education1,3,12. Perioperatively, the important goals are to prevent pain and nausea and to provide appropriate volume resuscitation1,19. Physical therapy is initiated when there are no symptoms of nausea or hypotension7,19. When patients appropriately complete standard physical therapy discharge criteria and are medically stable, they are cleared for discharge1,3,12. With outpatient clinical pathways becoming more popular and with the impending financial constraints on the health-care system, it is crucial to review the current literature to ensure that medical centers adopting this practice are well informed.
The primary objective of this systematic review was to determine the safety and feasibility of outpatient THA, TKA, and UKA. The secondary objective was to compare the complication rates, clinical outcomes, and costs associated with the same procedure conducted with an inpatient care pathway. We hypothesized that outpatient arthroplasty procedures would be safe and feasible. We also hypothesized that outpatient arthroplasty procedures would have (1) similar complication, readmission, and revision rates, (2) similar clinical outcomes, and (3) decreased costs in studies comparing outpatient with inpatient cohorts.
Materials and Methods
The findings are reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement23.
Literature Search and Study Selection
We conducted a systematic search of the online bibliographic databases Embase, MEDLINE, and HealthSTAR from inception through November 2014 to identify eligible studies on outpatient THA, TKA, and UKA. We used database-appropriate search terms, including “hip arthroplasty” and “knee arthroplasty.” We also used terms relating to outpatient surgery, such as “length of stay,” “outpatient surgery,” “ambulatory surgical procedure,” and “same day discharge” (see Appendix). The search was limited to articles written in the English language. Additionally, we performed a review of the reference lists of potentially relevant papers to identify articles that may have been missed through the literature search.
Two independent reviewers (M.P. and A.F.) assessed titles and abstracts for eligibility. The inclusion criteria were (1) a study population undergoing THA, TKA, or UKA, (2) discharge on the day of surgery, and (3) inclusion of at least 1 reported outcome. Articles were restricted to those that were published in full and written in English. If the eligibility of an article was in question following a review of the title and abstract, the full text was obtained and was independently reviewed with use of the same eligibility criteria. Any disagreement between authors was resolved through third-party consensus (B.L. or L.S.). Agreement was calculated by means of the kappa statistic.
Two reviewers independently extracted data from eligible studies into an electronic screening form. For each article, the author names, study title, journal name, issue, and page numbers were collected. Sample characteristics were collected, including demographic characteristics (mean patient age, mean length of stay, etc.) as well as intraoperative and perioperative characteristics. Two groups of 2 reviewers (M.P. and A.F.; B.L. and L.S.) independently extracted outcome data on complications, readmissions, revisions, cost, satisfaction, physical milestones in rehabilitation, and quality-of-life scores. The quality-of-life scores comprised pain scores, the Harris hip score (HHS), the 12-Item Short Form Health Survey (SF-12) score, the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), the Knee Society Clinical Rating System (KSCRS) score, and the Knee injury and Osteoarthritis Outcome Score (KOOS) (Table I).
Methodological Quality Assessment
Three independent reviewers assessed the methodological quality of eligible studies with use of the Cochrane Risk of Bias Assessment Tool for Non-Randomized Studies of Interventions (ACROBAT-NRSI)24 for comparison studies and the Effective Public Health Practice Project (EPHPP) Quality Assessment Instrument for non-comparison studies25,26. Quality was evaluated according to bias in the following domains: (1) confounding factors, (2) selection of patients, (3) measurement of interventions, (4) sufficient follow-up, (5) study design, (6) measurement of outcomes, (7) selection of the reported results, and (8) blinding. Any disagreements were resolved through consensus (see Appendix).
Studies involved the use of a variety of instruments and questionnaires to measure patient outcomes. Commonly reported outcomes included complications, readmission rates, and revision rates. Complications were categorized as either acute (occurring intraoperatively or in the immediate perioperative period) or post-discharge. Frequently reported acute complications included pain, fracture, deep-vein thrombosis (DVT), pulmonary embolism, nausea or dizziness, hypotension, oversedation, cardiac issues, inability to complete physiotherapy, infection, transfusions, and gastrointestinal ulcers. Post-discharge complications included dislocation, joint stiffness necessitating manipulation, anemia, wound drainage or infection, the need for a visit to the emergency department, revision or readmission, gastrointestinal or bowel-related issues, death, benign swelling, hematoma, mechanical complications, and accidental falls.
Additional clinical outcomes that were reported in the studies included validated measures such as the HHS27, SF-1228, WOMAC29, KSCRS30, and KOOS31. When applicable, means, standard deviations, ranges, and proportions were collected. For comparative studies, the mean differences and relative risks of outcomes were reported.
Summary of Evidence
The initial search strategy generated a total of 807 potentially relevant studies. Review of titles, abstracts and reference lists generated 26 eligible articles for full-text assessment. Agreement between reviewers was good (kappa = 0.775). Following full-text review, 17 articles remained eligible for inclusion in the systematic review (kappa = 0.897). Of the included studies, 9 assessed outpatient procedures in a THA cohort; 4, in a TKA cohort; 3, in a UKA cohort; and 1, in both TKA and UKA cohorts. One study was excluded because only an abstract was published; 3 were excluded because the mean length of stay was >24 hours; 2 were excluded because the population studied did not consist of patients undergoing THA, TKA, or UKA; 1 was excluded because the intervention description did not include length of stay and therefore could not be assessed; and 2 were excluded because they were preliminary reports of already included studies. All 17 studies were nonrandomized, and only 4 included a control group. There were a total of 79,243 patients in the included studies. Figure 1 summarizes the literature search.
The Cochrane Collaboration Tool showed moderate to high risk of bias for the majority of comparative studies (see Appendix), whereas the EPHPP showed moderate to weak quality for the majority of non-comparative studies (see Appendix). As none of the studies were randomized controlled trials, there was an inherent selection bias associated with the majority of included studies. Multiple studies only selected healthy patients on the basis of age, body mass index (BMI), and medical history. Although certain studies used broad selection criteria, selection bias is hard to avoid without patient randomization.
Variability Across Studies
Surgical technique varied across studies. The techniques used for THA included the direct anterior Smith-Peterson approach (1 study3), a 2-incision technique (4 studies7,16-18), a posterolateral approach (1 study8), an anterolateral Watson-Jones approach (1 study9), and a posterior minimally invasive surgical technique (1 study12). The techniques used for TKA included a minimally invasive quadriceps-sparing approach (2 studies4,6) and a medial parapatellar approach (1 study13). The techniques used for UKA included a mini-midvastus approach (1 study10), a medial arthrotomy approach (1 study11), and incision without patellar eversion (1 study20) (Table II).
Selected Versus Unselected Populations
There is a common distinction made in the literature between selected and unselected populations. The majority of studies employed a selected population, with restrictions on age (<65 years old), BMI, and the severity of comorbidities. These studies strictly enrolled healthy patients with characteristics that are presumed to be associated with positive utcomes. However, 6 studies4,8,10,14,15,19 involved an unselected population, with the inclusion of any patient undergoing a joint replacement who was discharged on the same day as the surgery. These unselected populations provide a more useful assessment of effectiveness because the study populations are more generalizable. The discrepancy in patient populations made comparing study results difficult.
The majority of studies assessed complication rates, revisions, and readmissions. In the literature, there is a clear distinction between acute complications (either intraoperative or immediately postoperative) (Tables III and IV) and long-term complications that occur outside the hospital (Tables V and VI). In general, the case series investigating the safety and feasibility of outpatient joint replacements demonstrated low complication rates, ranging from 0% (0 of 160)20 to 25.3% (38 of 150)7 in the acute phase and from 0% (0 of 100)17 to 9.3% (8 of 86)4 over the long term. Lovald et al.15 reported that the majority of adverse events that occur following outpatient joint replacements are joint pain, joint stiffness, deep-vein thrombosis, mechanical complications, revision, readmission, and wound complications.
More importantly, the comparative studies did not show higher complication rates following outpatient procedures as compared with inpatient procedures. As readmission rates are increasingly used as a performance indicator, there is concern that outpatient surgery may increase readmission rates. However, Lovald et al.15 illustrated that, in a Medicare sample from 1997 to 2009, readmission rates were similar between outpatient TKA (n = 454) and inpatient TKA (involving a standard 3 to 4-day stay) (n = 71,341) at 90 days (p = 0.1305), 1 year (p = 0.0968), and 2 years (p = 0.7534) postoperatively. Similarly, in both comparative studies investigating outpatient THA3,8, no difference in complication rates was found between the inpatient and outpatient groups.
The studies evaluating outpatient THA assessed HHS scores, SF-12 scores, cost, and patient satisfaction (see Appendix). Two studies by Berger et al.7,17 demonstrated improvements in HHS scores from preoperatively to 3 months postoperatively. Furthermore, Berger et al.17 demonstrated that, in a sample of 100 patients, both the physical (p < 0.001) and mental (p < 0.05) scores on the SF-12 improved from preoperatively to 12 weeks postoperatively.
Two economic analyses demonstrated a marked decrease in costs in association with outpatient THA as compared with inpatient THA. Aynardi et al.3 illustrated that the overall cost was significantly lower in the outpatient setting than in the inpatient setting ($24,529 compared with $31,327; p = 0.0001). Bertin8 found that the total average charge (including preoperative, intraoperative, and postoperative charges) was approximately $2,500 less for the outpatient group than for the inpatient group.
Two studies evaluated satisfaction after outpatient THA7,12. Berger et al.7 assessed satisfaction at 2 weeks postoperatively by asking the question, “Would you be discharged home the same day following the same clinical pathway again?”7 Of the 150 patients who were included, 144 were satisfied, 5 were unsatisfied because of postoperative nausea, and 1 was unsatisfied because of postoperative pain7. Dorr et al.12 administered a 6-item questionnaire at the 6-week follow-up visit. Of the 52 patients who completed the questionnaire, 96% reported that they would have same-day surgery again, 19% reported that postoperative pain was a problem, 87% claimed that outpatient THA gave them more confidence and accelerated their recovery, and 94% reported that they would recommend same-day surgery to others.
The studies investigating outpatient TKA and UKA assessed clinical outcomes such as cost and various outcome measures (WOMAC, KOOS, and KSCRS) (see Appendix). Lovald et al.15 found that, at 2 years following surgery, the costs associated with outpatient TKA were $1,967 lower than those for their standard-of-care group, who stayed in the hospital for 3 to 4 days following surgery. Dervin et al.11 studied outpatient UKA involving a multimodal analgesia regimen and concluded that mean WOMAC total scores (48.02 preoperatively, 14.64 at 1 year, and 12.12 at 2 years), KOOS sports/recreation scores (25.15 preoperatively, 55.56 at 1 year, and 46.88 at 2 years), and KOOS quality-of-life scores (21.42 preoperatively, 66.67 at 1 year, and 66.80 at 2 years) improved postoperatively (p < 0.001). Gondusky et al.20, in a study of patients who were discharged on the day of surgery following UKA, reported significant improvement in the mean KSCRS score between the preoperative and latest follow-up evaluations (from 99.51 to 182.66; p < 0.001). Finally, Kolisek et al.13 reported no difference between outpatient and inpatient TKA in terms of the KSCRS knee score (p = 0.26) or the KSCRS function score (p = 0.966) after a mean duration of follow-up of 24 months.
The aims of this systematic review were to summarize the literature on outpatient joint replacement and to assess the safety and feasibility of the outpatient model of care. The goals of THA, TKA, and UKA are to relieve pain and to restore function with minimal complications and adverse events. The desire to increase patient autonomy and satisfaction and to respond to economic pressures, in conjunction with analgesic and technical advancements, has increased the popularity of outpatient hip and knee arthroplasty. Despite a lack of high-level evidence, our review of the literature demonstrated that outpatient procedures result in similar complication rates and clinical outcomes, while substantially minimizing the cost, in comparison with inpatient procedures3,8,13,15.
A number of studies have indicated that careful patient selection (e.g., restrictions on age, BMI, and severity of comorbidities) is required for successful outpatient surgery. This careful selection is reflected by the demographic characteristics of the patients involved in the various studies included in the present review (Table II). In this review, the mean age for patients undergoing hip replacement ranged from 53.5 to 63 years and the mean age for those undergoing knee replacement ranged from 55 to 68 years. However, as of 2014 in the United States, the average age was 67.7 years for patients undergoing hip arthroplasty and 66.1 years for those undergoing knee arthroplasty32. This discrepancy highlights the gap between current outpatient candidates and the average THA candidate. Currently, safe outpatient joint replacement procedures require careful patient selection. Appropriate patient characteristics are essential to ensure safe and successful joint replacements in an outpatient setting.
Six studies4,8,10,14,15,19 in this review involved an unselected population, with all patients undergoing a joint replacement being included. Those were largely feasibility studies4,8,10,19 or studies retrospectively examining the Medicare database14,15. The purpose of those studies was not to suggest that outpatient joint replacement is appropriate for all patients; rather, they were done to identify factors that are associated with a longer hospital stay. Both Mears et al.19 and Bozic et al.33 identified 4 factors: (1) female sex (p < 0.001), (2) increasing age (p < 0.001), (3) increasing blood loss (p < 0.001), and (4) an American Association of Anesthesiologists (ASA) classification of 3 or 4 (p < 0.01). An interesting factor to note is the sex inequality present in the current literature. The majority of studies in the present review involved more males than females in the outpatient cohorts (Table II). Notably, as of 2014 in the United States, patients were more frequently female than male for both hip arthroplasty (56.8%) and knee arthroplasty (61.6%)32. The under-representation of females in outpatient studies could be due to a number of factors, such as sex differences in societal roles of caregiving or an underlying sex bias toward males in the preliminary outpatient trials. Regardless, this is a fascinating trend that will need to be further examined with higher-level evidence.
Of the 4 studies that included a comparison inpatient group, 2 examined THA3,8 and 2 focused on TKA13,15. Those studies demonstrated that outpatient procedures are proving to be safe in terms of perioperative and postoperative complications (Tables III through VI)3,8,13,15. Kolisek et al.13 found no significant difference between the outpatient and inpatient groups in terms of the mean Knee Society knee score (p = 0.26), Knee Society function score (p = 0.966), range of motion (p = 0.289), or patient satisfaction (p = 0.930). The studies comparing inpatient and outpatient groups3,8,13,15, although informative, did not control for differences between the outpatient and inpatient cohorts in terms of patient demographics. Certain patient characteristics, such as BMI and age, are different between groups, preventing results from being generalizable.
One of the benefits of the movement toward outpatient joint replacements is the potential for considerable cost savings. Savings associated outpatient procedures were reported by Lovald et al.15, who found a marked reduction in cost of $8,527 at 2 years of follow-up, and Aynardi et al.3, who reported that the overall cost of for the outpatient group was $6,978 less than that for the inpatient group (p = 0.0001). Reducing the cost of health care improves access to services for both patients and health-care centers. Bertin8 determined that the total average charge (including preoperative, intraoperative, and postoperative charges) was approximately $2,500 less for the outpatient group than for the inpatient group. In addition to the potential cost savings, in certain physician-owned centers in the United States, an increased payment is gained by the physician when THA, TKA, or UKA is performed on an outpatient basis. These centers are motivated to perform these outpatient procedures, so it is imperative to develop a pathway that ensures patient safety.
It is important to note that there are variations in surgical techniques used throughout the literature (Table II). Differences in operative time, blood loss, and type of anesthesia exist depending on the technique used (see Appendix). These differences could account for differing perioperative care and could cause performance bias. It is encouraging, however, that outpatient joint replacement can be performed with use of multiple techniques. Surgeons can be reassured that there are numerous techniques to safely and effectively perform outpatient joint replacement procedures without having to drastically change their practice.
In contrast to the findings of previous studies13,34, Lovald et al.15 found that, in certain instances, there were less-favorable outcomes for the groups that had a shorter stay following TKA. Lovald et al.15 had performed a review of the Medicare 5% sample (1997 to 2009) and attributed this finding to sites that implemented outpatient protocols that were beyond their capabilities. Patient care should not be disregarded in the name of outpatient surgery; each center performing these procedures needs to develop and implement operative and postoperative care protocols that work for them and their patient populations. Currently, it is recommended to gradually reduce the length of stay until the clinical staff are comfortable performing joint replacements in an outpatient setting4,15.
The major limitation of the present systematic review is that only 4 of the included studies compared an outpatient group with an inpatient control group. The lack of control groups makes it difficult to establish definitive conclusions. However, the purpose of the majority of the single-cohort studies was to investigate whether outpatient joint replacements are safe and feasible. The overarching conclusion from the present review confirms that safety is achieved in selected patients, as illustrated by low complication rates (Tables III through VI). An additional limitation of the present review was the heterogeneity of outcome scales reported. This discrepancy made it difficult to pool the data and to make any definitive conclusions.
The quality of the studies that were included in the present systematic review was poor, as evidenced by the quality-assessment scales that we utilized (see Appendix). Joint replacements in an outpatient setting are a relatively new pathway that first required feasibility and safety studies to ensure that these procedures do not lead to an increase in complication rates. However, in the absence of high-quality comparative evidence, the success in same-day discharge may be limited because previous research was largely driven by observational studies of highly selected populations. To ensure that these procedures are more generalizable, it is imperative to complete prospective cohort and randomized controlled trials to diminish any bias associated with these findings.
In selected patients, outpatient THA, TKA, and UKA can be performed safely and effectively. However, studies lacked sufficient internal validity, sample size, methodological consistency, and standardization of protocols and outcomes. There is a need for more rigorous and adequately powered randomized trials to definitively conclude the safety and feasibility of outpatient THA, TKA, or UKA.
The search strategy and tables showing a risk-of-bias assessment of comparative studies, a quality assessment of non-comparative observational studies, functional outcomes and questionnaire findings, surgical data, and discharge data are available with the online version of this article as a data supplement at reviews.jbjs.org.
Investigation performed at London Health Sciences Centre, University Hospital, London, Ontario, Canada
Disclosure: No external funding source played a role in the investigation of this systematic review of the literature. On the Disclosure of Potential Conflicts of Interest forms, which are provided with the online version of the article, one or more of the authors checked “yes” to indicate that the author had a relevant financial relationship in the biomedical arena outside the submitted work.
- Copyright © 2016 by The Journal of Bone and Joint Surgery, Incorporated