Introduction

With the advancement of medicine, life expectancy of the United States (US) population continues to rise. Research suggests between the year 2000 and 2050 the number of individuals living over the age of 65 will increase by 135% [1]. By the year 2050, there will be more than 69 million Americans over the age of 65, with 19 million older than 85 years of age [2]. As the population continues to age, one of the major concerns in society will be the costs associated with treating chronic musculoskeletal conditions [3]. An analysis based on a national representative survey finds that 25% of physical therapy episodes are among Medicare beneficiaries [4]. The World Health Organization has identified four chronic musculoskeletal conditions that will continue to become more prevalent as the population ages. Two of these conditions include osteoarthritis (OA) and rheumatoid arthritis (RA), both of which affect millions of individuals worldwide [5–7].

OA affects approximately 27 million adults in the US alone [8]. It is characterized by progressive articular cartilage breakdown that often leads to pain and loss of function [9]. OA costs an average of $2,600 a year per patient, with combined direct and indirect annual costs averaging $5,700 for each individual [8, 10, 11]. In addition, over 80% of those with OA present with some degree of movement limitation, and 40% of these individuals rate their quality of life as fair or poor [12]. RA is a systemic, autoimmune disease that affects approximately 1.5 million adults in the US [8]. It is characterized by destruction of the cartilage and synovial lining of multiple joints in the musculoskeletal system, which often leads to pain and functional loss [13]. RA costs an average of $2,085 a year per patient, with combined direct and indirect costs averaging $3,200 per patient each year in the US [14–17]. In addition, individuals with RA are 40% more likely to report their health as fair or poor, and 30% more likely to require assistance with daily personal care [18].

When conservative management of RA and OA has failed and the overall quality of life continues to decline for an individual, total hip and knee arthroplasty are the surgical treatments of choice to alleviate joint destruction, decrease pain, and improve quality of life [19–22]. It is predicted that by the year 2030, there will be approximately 3.5 million primary total knee arthroplasties (TKA) and almost 600,000 primary total hip arthroplasties (THA) performed each year in the US [9]. The estimated US hospital expenditures of THA and TKA surgery in 2009 were 13.7 billion and 28.5 billion dollars, respectively [23]. Unless improvements with conservative care dramatically change, it is reasonable to assume that these trends in costs will continue to rise as the population ages and a greater number of individuals elect to have these surgical procedures.

Physical therapy is an integral component in the management of musculoskeletal conditions. One study reported a direct connection between duration of physical therapy and the improvement in functional status of patients hospitalized for lower extremity orthopedic issues [24]. In addition, early initiation of physical therapy has been found to improve health outcomes among patients with cerebrovascular accident and low back pain [25–27]. A recent systematic review published by Ojha et al [28] favorably reported the cost-effectiveness of early physical therapy for a variety of musculoskeletal conditions without compromising patient outcomes.

Physical therapy has been shown to reduce swelling, increase range of motion, improve strength, and return individuals to a higher level of function following THA and TKA [19–22]. Inpatient physical therapy services are commonly utilized following THA or TKA and are often considered an important part of post-operative management. Over the past decade, research has focused on early post-operative physical therapy with these patients and its impact on functional outcomes, costs, length of stay (LOS), and adverse reactions in the inpatient setting [29–40].

A recent systematic review [41], which included five randomized clinical trials (RCT) assessed the role of early mobilization after hip or knee joint arthroplasty on LOS in the acute care setting. Clinical homogenous data were analyzed using meta-analysis and the results demonstrated a reduced LOS of approximately 1.8 days without increasing the risk of negative outcomes. While the Guerra et al [41] systematic review reported a positive result on LOS in the acute care setting, several limitations exist. First, only RCTs were included, which led to an overall small sample size (622 participants). Secondly, it only examined the impact of early mobilization on LOS and did not discuss associated cost savings.

Further research is needed to investigate the overall costs of care, in addition to the clinical effectiveness demonstrated. Considering that there will be a significant increase in THA and TKA over the next 15 years, a more comprehensive systematic review, which includes various study designs that analyze both LOS and cost effectiveness is warranted to synthesize evidence regarding the safety and effectiveness, as well as economic impact of early physical therapy initiation after joint replacement procedures.

Therefore, the purpose of this manuscript was to conduct a systematic review of the literature analyzing the role that early initiation of physical therapy has on LOS and cost following THA, TKA, or uni-compartmental knee arthroplasty (UKA). The review will provide a summary of existing literature on acute management strategies following hip or knee arthroplasty, and subsequently identify potential gaps in the literature that may lead to future inquires in this line of research.

Methods

This article is a systematic review with meta-analysis that followed the PRISMA [42] guidelines (S1 Table. PRISMA 2009 Checklist.). This manuscript reviewed management of post-operative joint replacement in the acute care setting and the influence that early physical therapy initiation had on LOS and costs.

Identification and selection of trials

An electronic search was conducted in July 2016 using the databases PubMed, CINAHL (EBOSCO Host), Pedro, and AMED (Ovid), Embase, and the Cochrane Library for all pertinent articles relevant to this systematic review. Key words were used independently and in combination including early rehabilitation, costs, early mobilization, early rehabilitation, hip arthroplasty, immediate physical therapy, joint arthroplasty, knee arthroplasty, length of stay, and outcomes. Specific search strategies are outlined in Table 1. The goal behind the search strategy was to identify all potential articles that discussed THA, TKA, or UKA and the role that early rehabilitation had on LOS and/or costs following surgery. After the computerized search was completed, reference lists of all selected articles were searched by hand to identify any other related articles pertaining to this research study. One author, (MM), examined all titles and abstracts to determine initial study eligibility. Full text articles were then revaluated for specific inclusion criterion. A second author, (KK), independently reviewed all full text articles for eligibility. A third author, (WJH), determined final eligibility should a discrepancy exist between the initial authors.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Database search strategy.

https://doi.org/10.1371/journal.pone.0178295.t001

Inclusion criteria.

To be included in this systematic review, an article needed to meet the following inclusion criteria: (1) published in a peer reviewed journal; (2) published in the English language; (3) at least 2 groups in which the definition of timing of rehabilitation was clearly stated; (4) at least one measure of LOS or cost; (5) patients undergoing THA, TKA, or UKA.

Exclusion criteria.

Case studies and case series were excluded. Also, any potential disagreements were resolved through consensus between two authors, (MM and KK), and if consensus could not be agreed upon, a third author, (WJH), was consulted for a final opinion.

Outcomes.

The primary outcomes for this systematic review were LOS and costs following joint arthroplasty surgery of the hip and knee. For the purpose of this review LOS was defined as the number of days spent in the hospital from surgery to discharge from the acute care setting. Costs, on the other hand, were defined differently among the included studies. Chen et al [30] defined total medical expenses as rehabilitation expenses related to rehabilitation services in the index visit, as well as outpatient and inpatient expenses related to prosthetic infection and deep vein thrombosis within one year after discharge. Larsen et al [34] defined total medical expenses as costs associated with the information day, hospital stay, care in the hospital, rehabilitation in the hospital, patient needs, primary care in the follow-up period, and hospital re-admission in the follow-up period of one year after discharge. Reilly et al [43] defined total average costs as the combination of fixed costs (surgery, anesthesia, prosthesis, etc), average hospital stay cost, additional outpatient appointments, and cost of specialist registrar time over a six month period. Lastly, Pua and Ong [37] solely assessed hospitalization costs at discharge, which included room and ward charges, professional fees, laboratory investigations, pharmaceutical supplies, implant, and rehabilitation services.

Results

Study selection

The search strategy identified 1029 studies, and five additional studies were located through manual searching. After duplicates were removed, 974 remained to be screened based on title and abstract. Thirty-one full-text articles were assessed and 17 were included for review after screening to determine eligibility (Fig 1). The 17 studies included patients who had undergone THA, TKA, and UKA and implemented a variety of inpatient rehabilitation paradigms.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. PRISMA flow diagram.

https://doi.org/10.1371/journal.pone.0178295.g001

Characteristics of included trials

A total of 26,614 patients undergoing either THA, TKA, or UKA were included across all 17 studies and received either early rehabilitation or standard rehabilitation. For the purposes of this systematic review, early and standard rehabilitation are relatively defined for each individual study. Therefore, across studies, these operational definitions may overlap in regards to timeframe.

Quality.

Seventeen studies were individually scored for their methodological quality, which are presented in Table 2. The average score across 17 studies was 21/28 on the modified Downs and Black checklist, with a range of 16–26. Overall, reduced quality across studies can be attributed to inadequate reporting of blinding, loss to follow up, randomization, and adjustment for confounding variables in the analyses. Of the 17 studies included, 4 were randomized clinical trials, [32, 36, 43, 45] 1 was a quasi-experimental study, [33] 1 was a cost-effectiveness study, [34] based on the Larsen et al 2008 randomized clinical trial, [36] 6 were prospective cohort studies, [29, 35, 39, 40, 46, 47] and 5 were retrospective cohort studies [30, 31, 37, 38, 50].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Downs and Black methodological quality.

https://doi.org/10.1371/journal.pone.0178295.t002

Participants.

Five studies included participants only with TKA [30, 32, 37, 45, 47], 5 studies included participants with only THA [31, 33, 38, 40, 46], 1 study included participants with only UKA [43], and 6 studies included participants with either THA, TKA, or UKA [29, 34–36, 39, 50] (Table 3). Details of all included studies including demographic data, outcome measures, interventions in each group, and results are presented in Table 3. A total of 26,614 participants were included across all 17 studies with a mean age range of 50.3 to 72.3 years. The experimental group had a mean age range of 50.3 to 72.3 years, compared to 52.3 to 71.3 years in the control group. Among the studies that reported the distribution of male and female participants (15 studies), 27.1% were male and 72.9% were female (Table 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Data extracted from the included studies.

https://doi.org/10.1371/journal.pone.0178295.t003

Experimental group.

Patients in the experimental groups received early rehabilitation following surgery. Early rehabilitation was defined as rehabilitation that commenced on either the day of surgery, or post-operative day one. This operational definition of early rehabilitation applied to 16 out of 17 studies. One study, however, had a slightly different methodological design in which three separate groups were analyzed: [30] an experimental group, a comparison group, and a control group. For this particular study, early rehabilitation is defined relative to the two other treatment groups and was initiated within the first two weeks following surgery. The authors would suggest that this is not the ideal timeframe for early initiation of rehabilitation, nevertheless, this article demonstrated differences in outcomes between groups.

Comparison / Control group.

Patients in the comparison groups received standard or delayed rehabilitation following surgery. Standard rehabilitation was defined as rehabilitation that commenced on either post-operative day one, or post-operative day two. Delayed rehabilitation was defined as rehabilitation that commenced after two weeks of the surgery, which was only implemented in the Chen et al [30] study. This study also implemented a control group in which no formal rehabilitation was performed.

Outcomes

The outcome measures reviewed in this manuscript included LOS (15 studies) [29, 31–33, 35–40, 43, 45–47, 50] and cost (4 studies) associated with treatment [30, 34, 37, 43] (Table 3). Data from the included articles, including data in the experimental group, data in the comparison/control group, as well as between group differences are presented in Table 4.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Results of included studies.

https://doi.org/10.1371/journal.pone.0178295.t004

Length of stay.

All 15 studies that measured LOS as an outcome variable demonstrated a shorter hospital stay in the group receiving early physical therapy, with 12 of the studies demonstrating a statistically significant reduction in LOS following THA, TKA, or UKA (Table 4) [29, 31, 35–40, 45–47, 50]. While the Larsen et al [33], Reilly et al [43], and Wellman et al [40] studies did not calculate inferential statistics, a shorter mean LOS favoring the experimental group was reported (Table 4). The results of the meta-analysis that included the four RCTs (n = 548) demonstrated a large effect size favoring the early rehabilitation group (standardized mean difference = -1.90; 95% CI -2.76 to -1.05; I² = 93%) (Fig 2). The results of the meta-analysis that included 6 studies (5 prospective studies and 1 quasi-experimental study, n = 1,321) demonstrated a large effect size favoring the early rehabilitation group (standardized mean difference = -1.47; 95% CI -1.85 to -1.10; I² = 88%) (Fig 3). Both meta-analyses however, demonstrated a large I² value indicating substantial heterogeneity among studies.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Length of stay for RCTs meta-analysis.

Abbreviations: df, degrees of freedom; IV, inverse variance; RCT, randomized clinical trial; Std, standardized.

https://doi.org/10.1371/journal.pone.0178295.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Length of stay for prospective and quasi-experimental studies meta-analysis.

Abbreviations: df, degrees of freedom; IV, inverse variance; Std, standardized.

https://doi.org/10.1371/journal.pone.0178295.g003

Costs.

Of the four studies that measured cost as an outcome variable [30, 34, 37, 43], two, Larsen et al [34] and Pua and Ong [37], demonstrated a statistically significant between group mean cost savings favoring the experimental group that received early post-operative physical therapy. The third study by Chen et al [30], was unique to this systematic review because it implemented a third group, in which participants did not receive physical therapy. The findings of this study were similar to those of Larsen et al [34] and Pua and Ong [37] that demonstrated a statistically significant between group mean cost savings favoring the experimental group (Table 4). It was not appropriate to conduct a meta-analysis on costs due to differing methodological designs in the studies assessing costs. Overall, three studies [30, 34, 43] assessed total medical expenses following joint arthroplasty, while one study, Pua and Ong [37], only assessed hospitalization costs at discharge.

Discussion

The results of this systematic review demonstrated an overall reduction in LOS (15 studies) [29, 31–33, 35–40, 43, 45–47, 50] and cost (4 studies) [30, 34, 37, 43] in the group that received early rehabilitation following joint arthroplasty, without any increase in adverse reactions compared to participants in the standard rehabilitation group. This is consistent with other research in which early physical therapy has resulted in a shorter LOS following proximal hip fractures and hemiarthroplasty [51–53]. In addition, research has reported that early physical therapy can decrease adverse reactions such as venous thromboembolism following THA and TKA [54–56]. While the meta-analyses demonstrated an overall decrease in LOS favoring the experimental group with large effect sizes, the heterogeneity of the included studies were also high (Figs 2 and 3). In addition, lack of detail provided about the interventions in the different groups makes it difficult to assess and replicate the treatment factors that led to early discharge from the acute care setting.

Findings from the current systematic review exhibit both similarities and differences with the previous systematic review on early mobilization following hip or knee joint arthroplasty. Guerra et al [41] conducted a systematic review and meta-analysis, that focused only on RCTs and demonstrated a moderate to large effect size favoring early mobilization. In contrast, the authors of the current review chose to allow a variety of study designs, including prospective and retrospective cohort studies to achieve a broader overview of the literature. Thus, the current review includes several cohort studies that have large sample sizes, and two RCTs considered by Guerra et al [41] (Dowsey et al [57] and Whitney et al [58]) were excluded because they did not meet current inclusion criteria. While prospective and retrospective cohort studies represent a lower quality of evidence, they have the potential to bridge the gap between research and clinical practice, thereby increasing generalizability of findings. Moreover, the strict internal rigors necessary for RCTs may not be logistically possible in many clinical settings and thus not reflect the manner in which the intervention is likely to be applied in practice. Despite the lesser quality of evidence considered, the authors believe that the included cohort studies add important clinical information that must be combined with RCTs to provide a comprehensive summary of optimal treatment interventions.

Clinicians may question the safety of these early mobilization algorithms and the potential for an increase in adverse reactions. Of the 17 studies identified in this systematic review, 11 reported adverse reactions [30, 35–38, 40, 43, 45–47, 50], examples of which can be found in Table 5. Three studies, Larsen et al [35], Pua and Ong [37], and Gulotta et al [46] reported no statistically significant differences in the number of adverse reactions between groups. Raphael et al [50] reported no statistically significant differences in number of emergency department visits or re-admissions to the hospital within 30 days following surgery between groups. Chen et al [30] reported a statistically significant difference (p < .0001) in the number of adverse reactions between groups, however, the percentage of participants who developed prosthetic infection was lower in the group that received rehabilitation in the first 2 weeks. Moreover, the group that received rehabilitation within the first two weeks had a lower rate of deep vein thrombosis [30]. The remaining six studies [36, 38, 40, 43, 45] did not calculate between group differences for the number of adverse reactions.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Adverse effects associated with included studies.

https://doi.org/10.1371/journal.pone.0178295.t005

This review suggests that early initiation of physical therapy can be carried out without increased risks to patients when properly designed treatment paradigms are implemented. In practice, quality assurance measures should be in place to ensure that earlier discharge from the inpatient hospital setting does not lead to a higher re-admission rate following THA and TKA. Before widespread implementation of treatment paradigms that promote earlier initiation of physical therapy and quicker discharge from the hospital setting, the authors would recommend that future studies be conducted with less heterogeneity of treatment interventions and stronger methodological quality focusing on principles of both internal and external validity.

With the exception of the previous discussion on adverse reactions, no other long-term outcomes or quality of life measures were considered across the included studies, preventing the authors of this systematic review from making any further conclusions. In order to assess the long-term outcomes of early initiation of rehabilitation following joint arthroplasty, larger prospective cohort studies focusing on functional outcomes and quality of life measures with a one to two year follow-up are required.

With the costs of healthcare continuing to rise, it is necessary that all treatment interventions be assessed for cost-effectiveness. In this systematic review, 4 studies assessed cost-effectiveness as a primary outcome variable [30, 34, 37, 43]. Two studies, Larsen et al [34] and Pua and Ong [37] showed a statistically significant reduction in mean cost favoring the experimental group, which received earlier physical therapy. These results are similar to a recent systematic review published by Ojha et al [28].

One included study conducted by Chen et al [30] had a third group, in which no physical therapy was provided (Table 3). This study reported that the group that did not receive rehabilitation demonstrated the lowest total medical expenses compared with the group that received physical therapy within 2 weeks and the group that received physical therapy after 2 weeks. However, Chen et al [30] cautioned that while the group which did not receive physical therapy treatment demonstrated lower costs overall, patients in this group were more likely to experience a higher incidence of prosthetic infection (odds ratio 1.29, p = .0409) and deep vein thrombosis (odds ratio 1.51, p = .0099) compared to those in the group that initiated physical therapy within a 2 week time frame [30]. While it is impossible to identify a definitive reason for the increase in adverse reactions, one plausible theory is that increased blood pooling in the lower extremities associated with longer periods of inactivity after surgery may increase the risk of deep vein thrombosis. The authors would encourage clinicians and health policy makers to view these results in context and avoid making premature decisions that no physical therapy is the best option following joint replacement surgery. Since the studies included in this review focused on LOS and costs, there were no validated outcome measures implemented to assess treatment effectiveness. However, the authors suggest that the clinical effectiveness of these early initiation programs can be manifested in both decreased costs to the healthcare system, as well as an improved quality of life, less overall complications, and faster discharge from the hospital setting.

Although the authors’ original intention was to perform meta-analysis on costs as well as length of stay, differences in sample size, methodological design, reported data, and forms of currency hindered further statistical analysis. In accordance with the Center for Medicare and Medicaid Services’ [59] recently implemented initiative, total medical costs are bundled together over a 90-day period, including skilled nursing facility (SNF), rehabilitation, visiting nurse services, emergency department visits, and office visits. While this serves as the standardized definition of total medical costs in the US, other countries may have different reimbursement systems. Since the four included studies that discussed cost were from countries other than the US, it is necessary to consider various operational definitions of costs within this systematic review. Nonetheless, early initiation of physical therapy demonstrated a consistent trend in cost savings across studies.

To our knowledge, this systematic review provided the most comprehensive summary to date assessing the role of early initiation of physical therapy on LOS and costs following joint arthroplasty. The results of this systematic review suggest that patients in the group that received early physical therapy intervention had shorter LOS, lower medical cost, and experienced no greater risk of adverse events. This review is a timely contribution to the literature because recent healthcare reform has renewed providers’ interest in shortening the LOS after hip and knee replacement procedures. A policy established on April 1, 2016 mandates hospitals located in 67 geographic areas, defined by metropolitan statistical areas, to participate in the Comprehensive Care for Joint Replacement (CCJR) model [59]. In this model, participant hospitals are financially accountable for the quality and cost of care of patients receiving joint replacement procedures in an episode between hospital admission and 90 days post-discharge. Participant hospitals have the financial incentives to discharge patients as early as possible when clinically appropriate. Traditionally, Medicare patients receiving joint replacement procedures had to stay in the hospital for at least 3 days to qualify for Medicare covered SNF stay. As part of the CCJR payment model, patients could have the 3-day qualifying hospital stay waived as long as they are discharged to a SNF that has an overall 3-star rating or better under the Centers for Medicare and Medicaid Services 5-Star Quality Rating System.

While timing of physical therapy initiation may be one critical factor, another important consideration may be the overall plan of care. Among the included studies, treatment protocols varied in the pre-surgical, rehabilitation, and post-surgical management strategies. Prior to surgery, the following conditions varied among studies: type of education provided before surgery, timing of hospital admission, location of residence within the hospital, and healthcare practitioners involved in the care. Rehabilitation and post-surgical follow-up varied in the number of hours and treatment sessions per day, types of interventions performed, and whether or not home care services were provided (Table 3). While variability is inevitable between treatment protocols following joint arthroplasty, all components of rehabilitation should be addressed within a biopsychosocial management approach that considers the multi-faceted aspects of patient care.

Conclusions

This review suggests that early initiation of physical therapy following THA, TKA, or UKA is associated with a shorter LOS and lower overall cost. Furthermore, the results presented in this review show no evidence of increased number of adverse effects when physical therapy is initiated early following joint replacement surgery.

Supporting information

Acknowledgments

We would like to thank Dr. Rebecca States, PhD, MA for her advisement on the search strategy and data analysis section.