Inpatient Compared with Home-Based Rehabilitation Following Primary Unilateral Total Hip or Knee Replacement: A Randomized Controlled Trial

Inpatient Compared with Home-Based Rehabilitation Following Primary Unilateral Total Hip or Knee Replacement: A Randomized Controlled Trial
of 8
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Inpatient Compared with Home-Based RehabilitationFollowing Primary Unilateral Total Hip or KneeReplacement: A Randomized Controlled Trial By Nizar N.Mahomed, MD, ScD, FRCSC, Aileen M.Davis, PhD, Gillian Hawker, MD, FRCPC, ElizabethBadley, PhD,J. RodDavey, MD, FRCSC, Khalid A.Syed, MD, FRCSC, Peter C.Coyte, PhD, RajivGandhi, MD, FRCSC,and James G.Wright, MD, MPH, FRCSC Investigation performed at Toronto Western Hospital, University of Toronto, and North York General Hospital, Toronto, Ontario, Canada Background:  Home-based rehabilitation is increasingly utilized to reduce health-care costs; however, with a shorter hospital stay, the possibility arises for an increase in adverse clinical outcomes. We evaluated the effectiveness andcost of care of home-based compared with inpatient rehabilitation following primary total hip or knee joint replacement. Methods:  We randomized 234 patients, using block randomization techniques, to either home-based or inpatientrehabilitation following total joint replacement. All patients followed standardized care pathways and were evaluated,with use of validated outcome measures (Western Ontario and McMaster Universities Osteoarthritis Index [WOMAC],Short Form-36, and patient satisfaction), prior to surgery and at three and twelve months following surgery. The primary outcome was the WOMAC function score at three months after surgery. Results:  The mean length of stay (and standard deviation) in the acute care hospital was 6.3  ±  2.5 days for the groupdesignated for inpatient rehabilitation prior to transfer to that facility compared with 7.0  ±  3.0 days for the home-basedrehabilitation group prior to discharge home (p  =  0.06). The mean length of stay in inpatient rehabilitation was 17.7  ± 8.6 days. The mean number of postoperative home-based rehabilitation visits was eight. The prevalence of postoperativecomplications up to twelve months postoperatively was similar in both groups, which each had a 2% rate of dislocationand a 3% rate of clinically important deep venous thrombosis. The prevalence of infection was 0% in the home-basedgroup and 2% in the inpatient group. None of these differences was clinically important. Both groups showed substantialimprovements at three and twelve months, with no significant differences between the groups with respect to WOMAC,Short Form-36, or patient satisfaction scores (p > 0.05). The total episode-of-care costs (in Canadian dollars) for theinpatient rehabilitation and home-based rehabilitation arms were $14,532 and $11,082, respectively (p < 0.01). Conclusions:  Despite concerns about early hospital discharge, there was no difference in pain, functional outcomes,or patient satisfaction between the group that received home-based rehabilitation and the group that had inpatientrehabilitation. On the basis of our findings, we recommend the use of a home-based rehabilitation protocol followingelective primary total hip or knee replacement as it is the more cost-effective strategy. Level of Evidence:  Therapeutic Level I. See Instructions to Authors for a complete description of levels of evidence. L imited health-care resources have created a demand for amore efficient delivery of care for patients undergoing total joint replacement. In Canada, the most notablerecent change has been a substantial reduction in the lengths of stay in acute care hospitals following both total hip and kneereplacements 1 . The average length of stay in a Canadian hos- Disclosure:  In support of their research for or preparation of this work, one or more of the authors received, in any one year, outside funding or grants inexcess of $10,000 from Physicians’ Services Incorporated. Neither they nor a member of their immediate families received payments or other benefitsor a commitment or agreement to provide such benefits from a commercial entity. No commercial entity paid or directed, or agreed to pay or direct, any benefits to any research fund, foundation, division, center, clinical practice, or other charitable or nonprofit organization with which the authors, or amember of their immediate families, are affiliated or associated.A commentary is available with the electronic versions of this article, on our web site ( and on our quarterly CD-ROM (call our subscription department, at 781-449-9780, to order the CD-ROM). 1673 C OPYRIGHT    2008  BY  T HE  J OURNAL OF  B ONE AND  J OINT  S URGERY , I NCORPORATED  J Bone Joint Surg Am.  2008;90:1673-80  d doi:10.2106/JBJS.G.01108  pital for total hip replacement has declined by 36%, fromfourteen days in 1994 to 1995 to nine days in 2004 to 2005. Fortotal knee replacement, the hospital length of stay over thesame period declined by 42%, from twelve days to seven days 1 .This reduced length of stay has resulted in an increased de-mand for both inpatient and home-based rehabilitation ser-vices 2 . In 2005, in the United States, the average length of stay was 3.8 days for total knee replacement and 4.9 days for totalhip replacement, which was less than that in Canada 3 . Lengthof stay is declining even more with the advent of minimally invasive total hip arthroplasty  4 ; however, early clinical findingshave shown no significant differences in functional outcomescores between the shorter length of stay associated with theminimally invasive technique and the longer length of stay associated with the conventional technique 4,5 .Traditionally in Canada after total joint replacement,patients were routinely sent to inpatient facilities for rehabil-itation 6-10 . More recently, there has been an increased trend tosend patients directly home after total joint replacement withsupporting home-care services 11 . However, large variationsexist in current practice patterns with regard to where patientsreceive postoperative rehabilitation 12,13 . For example, in theprovince of Ontario, the rate of utilization for inpatient re-habilitation following total joint replacement has been re-ported to range from 3% to 79% 14 .Unfortunately, to date, there is limited informationabout the effectiveness and costs of home-based rehabilitationcompared with inpatient rehabilitation for total joint re-placement 9,10,15 . Tribe et al. 16 compared the functional out-comes for patients who were managed with either home-basedor inpatient rehabilitation after undergoing total hip and kneereplacements for primary osteoarthritis. In that nonran-domized Australian trial, there were no differences in func-tional outcomes between the two groups at the one-yearfollow-up evaluation. The authors of that study emphasized aneed for randomized trials to explore which patients are mostsuitable for the two types of rehabilitation.Clinical trials to date have suggested that patient pref-erence and institutional factors rather than clinical parametersprimarily influence the decision regarding the type of reha-bilitation 6,17 . Following total joint replacement, it is unclearwhich method of rehabilitation is superior in terms of patient-related outcomes and cost of care. We evaluated the efficacy and cost of home-based compared with inpatient rehabilita-tion in a randomized controlled trial following primary totalhip and knee replacement. To our knowledge, this study pres-ents the longest follow-up to date (twelve months) and in-cludes a per diem cost analysis. Materials and Methods T his randomized controlled trial enrolled patients under-going primary total hip or knee replacement from twoinstitutions: a tertiary-care referral center and a community hospital in the same city. Patients were deemed eligible for thestudy if they were over the age of eighteen years; were un-dergoing unilateral hip or knee replacement for osteoarthritis,inflammatory arthritis, or osteonecrosis; were residents of thecity where the two institutions were located; were fluent inEnglish; and provided informed consent to participate. Pa-tients requiring total joint replacement for a fracture, malig-nant tumor, revision of a total joint replacement, or bilateraltotal joint replacement were excluded. Eligible patients wereidentified by participating surgeons and were then approachedto participate in the trial by a study coordinator. Patientsproviding informed consent were then enrolled in the trial.The study protocol and patient consent forms were reviewedand approved by the Human Subject Review Committee.The primary outcome of the trial was to evaluate theefficacy of inpatient compared with home-based rehabilitationat three months after surgery (the time at which both inter-ventions were discontinued), with use of the function subscaleof the Western Ontario and McMaster Universities Osteoar-thritis Index (WOMAC) 18 . Secondary outcomes included themeasurement of health status with use of the Short Form-36(SF-36) 19 and patient satisfaction with use of the Hip and KneeSatisfaction Scale 20 .At baseline and at the three and twelve-month follow-upvisits, the study subjects completed the WOMAC, SF-36, andHip and Knee Satisfaction Scale questionnaires. The WOMACis a self-administered, disease-specific questionnaire with a5-point Likert scale used to assess pain, function, and stiffnessof the hip and knee 18 . The descriptors range from no difficulty with respect to pain, functional activities, and stiffness (a scoreof 1 point) to extreme pain, difficulty, and stiffness (5 points).The SF-36 is a thirty-six-item questionnaire that includes eightscales measuring physical function, role limitations related tophysical health problems, bodily pain, perceptions of generalhealth, vitality, social functioning, role limitations related toemotional problems, and mental health. Mental and physicalcomponent summary scores are also calculated with use of these items 19 . In this study, the physical function, physicalcomponent score, and mental component score measures of the SF-36 were assessed. The Hip and Knee Satisfaction Scaleis a simple, reliable, and validated instrument that evaluatespatient satisfaction with the outcome of total joint arthroplasty in terms of improvement in pain and function 20 . It consists of four questions with scores ranging between 0 and 100, wherehigher scores represent greater satisfaction.All subjects were evaluated approximately two weeksprior to surgery during a preadmission hospital visit. Thisbaseline evaluation included data on demographics, height,weight, primary diagnosis, medical comorbidities, and socialsupports as well as completion of the selected outcome mea-sures, WOMAC, SF-36, and Hip and Knee Satisfaction Scale.One week prior to surgery, the subjects were random-ized, with use of block randomization, and were informed of their treatment allocation to either home-based or inpatientrehabilitation. The inpatient rehabilitation facility is a separate,independent facility wherein patient care is supervised by arehabilitation doctor. The subjects were informed of theirrandomization prior to surgery in order to allow sufficienttime to prepare their home settings if they were allocated to 1674 T HE  J OURNAL OF  B ONE  & J OINT  S URGERY  d JBJS . ORG V OLUME  90-A   d N UMBER   8  d A  UGUST  2008I NPATIENT  C OMPARED WITH  H OME -B ASED  R  EHABILITATION A  FTER   U NILATERAL  H IP OR   K NEE  R  EPLACEMENT  home-based rehabilitation. All subjects were admitted to theacute care hospital on the day of surgery and were managedpostoperatively according to previously established and stan-dardized care guidelines for total joint replacement. All pa-tients received prophylactic antibiotics and prophylaxis againstdeep-vein thrombosis. Patients were excluded from the trial if they had development of serious postoperative complicationsthat precluded participation in the rehabilitation protocol.These included myocardial infarction, cerebrovascular acci-dent, intraoperative fracture, neurovascular injury, deep in-fection, or a return to the operating room for any reason.The target length of stay at the acute care hospital forboth treatment groups was five days. All patients received thesame physiotherapy protocol during their stay (see Appendix).This included deep breathing and coughing, active and active-assisted bed and chair exercises, and gait training beginning with assisted walking. On postoperative day 1, the physio-therapy goals were to transfer from lying to sitting and fromsitting to standing with minimal assistance, to stand with awalker, and to sit in a chair for one hour. On postoperative day 2, physiotherapy goals were independent transfers and walking from bed to bathroom with assistance. From postoperativeday 3 until discharge, the goals were to achieve independentwalking, including climbing stairs. For the home-based reha-bilitation group, subjects were deemed safe for discharge fromthe acute hospital when they had achieved four key functionalmilestones, as described by Zavadak et al. 21 , including theability to independently transfer from the supine position tositting and from sitting to standing, walking a distance of 30 m,and climbing stairs if necessary.The subjects who were randomized to inpatient reha-bilitation were transferred to one of two independent inpatientrehabilitation institutions depending on the availability of re-habilitation beds. There they were managed along previously established care pathways with a target of a fourteen-day lengthof stay (see Appendix). The subjects who were randomized tohome-based rehabilitation were referred to their respectiveCommunity Care Access Centre (a provincially funded home-care agency that includes nursing, rehabilitation, and homesupport as necessary for the client) (see Appendix). The Com-munity Care Access Centre provided an early interventionprogram that ensured that each subject was seen at home by aphysiotherapist within forty-eight hours of discharge. Eachsubject was managed along a multidisciplinaryclinical pathway for home-based rehabilitation that was developed for this trial.The subjects were discharged from the home-based programwhen the physiotherapist thought that they had achieved suf-ficient functional improvement to either attend an outpatientphysiotherapy clinic or maintain a self-directed program. Allsubjects returned to their respective acute care hospital at threeand twelve months postoperatively for a follow-up evaluationby their treating surgeon.Cost analysis was conducted from a health system per-spective; therefore, only direct health-care costs were evaluatedfor acute care hospitals, inpatient rehabilitation hospitals, andhome-based rehabilitation services. These data did not includephysician fees, medications, or indirect costs to the patients orsecondarycaregivers.Weusedperdiemcostsfromtherespectiveinstitutionsandcalculatedthecostsbymultiplyingperdiemcostbytheactuallengthofstayforthepatient.Wewereabletoobtainpatient-level costs for services provided by home care using thecentralized data system. All costs were measured in Canadiandollars and were adjusted for inflation over the study period(with use of the Consumer Price Index) to the base year 2006 22 . Statistical Methods The primary analysis was on an intention-to-treat basis. Wecompared the two treatment groups at baseline (prior to sur-gery) to ensure they were equivalent on all measured variables.This was done to ensure that our randomization strategy wassuccessful and to identify any potential variables we needed toadjust for in our final analysis. Univariate analyses for differ-ences between groups were conducted with use of the Student ttest, chi-square test, or Fisher exact test as appropriate, withthe level of significance set at 0.05. Since there were no sig-nificant differences in the baseline parameters, we used ananalysis of variance to evaluate differences between groups inthe two treatment arms, adjusting for baseline severity of therespective dependent variable and the number of medical co-morbidities. Similar strategies were used for analysis of ourprimary outcome (the WOMAC function score at threemonths) as well as all secondary outcomes. For ease of pre-sentation, the WOMAC scores were rescaled from 0 to 100,with high scores representing better function and pain relief.Differences between groups in satisfaction scores were evalu-ated with use of Wilcoxon rank-sum tests. Twenty subjectsrequested a crossover from their assigned treatment groupof home rehabilitation to inpatient rehabilitation (primarily because of patient preference). Therefore, we replicated theanalysis on the basis of the actual discharge destination. Theresults were identical to those based on an intention-to-treatanalysis; therefore, we elected to present the results of theintention-to-treat analysis for clarity of presentation. Power Calculation The sample size was calculated to detect a minimal clinically important difference in the WOMAC function score at twelveweeks after total joint arthroplasty between the home-based andinpatient rehabilitation groups. We defined the minimal clini-cally important change accordingtocriteriasuggestedbyBellamy et al. 23 : a standard deviation of 14 forcalculating samplesize withuse of this scale, and a change of 7.4 points following a weak intervention, such as a low dose ofVoltaren (diclofenac sodium).The effect size (the mean difference divided by the standarddeviation) associated with such a change was approximately 0.5(7.4/14). Therefore, we calculated the sample size to detect aneffect size of 0.5, with a type-I error of 5% and 95% power. Results F rom 2000 to 2002, 234 subjects were entered into the trial;119 of them were randomized to the inpatient rehabilita-tion arm and 115 to the home-based rehabilitation arm (Fig. 1675 T HE  J OURNAL OF  B ONE  & J OINT  S URGERY  d JBJS . ORG V OLUME  90-A   d N UMBER   8  d A  UGUST  2008I NPATIENT  C OMPARED WITH  H OME -B ASED  R  EHABILITATION A  FTER   U NILATERAL  H IP OR   K NEE  R  EPLACEMENT  1). Our study ended in the middle of a randomization block,and thus there was some imbalance between the numbersassigned to the home-based rehabilitation and inpatient re-habilitation groups. The mean age of the subjects was sixty-eight years, and approximately two-thirds were women (seeAppendix). None of the patients randomized to the trial waslost to follow-up. The mean body mass index was 28 kg/m 2 .Osteoarthritis was the primary diagnosis in the vast majority of patients, and most had two or more medical comorbidities.Nearly two-thirds were white, and about 50% had postsec-ondary education. Nearly 25% lived alone, and 20% wereworking at the time of surgery. There were nearly equal pro-portions of hip and knee replacements. There was no signif-icant difference between the treatment groups in any of themeasured baseline variables (see Appendix).Preoperative baseline functional outcomes are outlinedin Table I. At baseline, both treatment arms had substantialpain and functional disability on the basis of the WOMAC painand physical function scores; however, no preoperative dif-ferences were noted between the two groups. Fig. 1 Schematic diagram showing the number of study participants in each group who participated,crossed over, or were lost to follow-up. The statistical analysis for this study was conducted withuse of the intention-to-treat group of 115 patients who received home-based rehabilitation and119 patients who had inpatient rehabilitation. 1676 T HE  J OURNAL OF  B ONE  & J OINT  S URGERY  d JBJS . ORG V OLUME  90-A   d N UMBER   8  d A  UGUST  2008I NPATIENT  C OMPARED WITH  H OME -B ASED  R  EHABILITATION A  FTER   U NILATERAL  H IP OR   K NEE  R  EPLACEMENT  Since we did not have control over the availability of either inpatient rehabilitation beds or home-based rehabilita-tion services, the actual hospital length of stays exceeded thetarget of five days for both groups (the mean length of stay [and standard deviation] was 6.3  ±  2.5 days for the inpatientrehabilitation group and 7.0  ±  3.0 days for the home-basedrehabilitation group; p  =  0.06). The average length of stay in inpatient rehabilitation was 17.7  ±  8.6 days. The averagenumber of postoperative home-based rehabilitation visits waseight (range, four to sixteen visits). All visits were to performphysiotherapy and were approximately one hour in duration.The prevalence of postoperative complications up totwelve months after surgery was similar in both groups, with a2% rate of hip dislocation and a 3% rate of clinically importantdeep venous thrombosis in each group. There were no reportedinfections in the home-based group and a 2% prevalence of infection inthe inpatientgroup. No differenceswere observedinhospital readmissions for the management of infections be-tween the treatment groups for twelve months after surgery.The self-reported WOMAC, SF-36, and satisfactionscores at three and twelve months after surgery are detailed inTable I. Overall, both treatment groups had dramatic im-provements in the WOMAC pain, physical function, andstiffness scores at three months compared with the preopera-tive scores. There was slight further improvement in theWOMAC scores at twelve months. There were no significantdifferences between the groups in any of the WOMAC scores ateither three or twelve months after surgery. There was a similarpattern for the SF-36 physical component scores, with bothgroups reporting improvements in the physical function andphysical component summary scores at three months andfurther improvement at twelve months after surgery comparedwith baseline. However, there were no significant differencesbetween the groups in any of the SF-36 scores at either three ortwelve months. With regard to satisfaction, both treatment armsreported equally high scores on the Hip and Knee SatisfactionScale at three and twelve months after surgery (Table I).We also conducted an analysis of the data on the basis of the site of the joint replacement (i.e., knee or hip) and foundno differences in functional outcomes or patient satisfactionon the basis of whether patients had undergone a hip or a kneereplacement.The mean costs for the stay in the acute care hospital(excluding the day of surgery) for the inpatient rehabilitationand home-based rehabilitation groups were $9411 and$10,191, respectively (p = 0.18) (Table II). This difference may have reflected the slightly longer length of stay in the acute carehospital for the home-based rehabilitation group comparedwith the inpatient rehabilitation group (7.0 compared with 6.3days). The greatest difference in costs, however, resulted fromthe postdischarge rehabilitation portion of the total episode of care. There was a nearly sixfold difference in the mean cost forinpatient rehabilitation compared with that for home-basedrehabilitation ($5120  ±  $7552 and $891  ±  $1316, respectively;p < 0.001). The resulting total episode-of-care cost for theinpatient rehabilitation and home-based rehabilitation armsof the trial was a mean of $14,531  ±  $11,555 and $11,082  ± $7747, respectively (p < 0.01). Discussion W  hen the results of home-based rehabilitation and inpa-tient rehabilitation following primary total hip or kneereplacement were compared with use of validated outcomemeasures, we found no differences in clinical outcomes atthree and twelve months after surgery. Both treatment groups TABLE I Outcomes of Patient-Reported Measures (Based on anIntention-to-Treat Analysis) Parameter*Home-BasedRehabilitation † (N  =  115)InpatientRehabilitation † (N  =  119) P ValuePreoperativeWOMACPain 47  ±  21 47  ±  17 0.90Stiffness 47  ±  21 47  ±  21 0.83Physical function 43  ±  18 44  ±  18 0.86Short Form-36Physical function 26  ±  20 26  ±  21 0.93Physical componentsummary 29  ±  7 27  ±  7 0.13Mental componentsummary 43  ±  11 45  ±  10 0.15Three months after total joint replacementWOMACPain 80  ±  19 80  ±  17 0.85Stiffness 70  ±  20 68  ±  20 0.51Physical function 72  ±  19 71  ±  19 0.92Short Form-36Physical function 47  ±  25 49  ±  24 0.25Physical componentsummary 34  ±  9 36  ±  10 0.11Mental componentsummary 44  ±  10 45  ±  11 0.83Satisfaction score 87  ±  15 89  ±  14 0.37Twelve months after total joint replacementWOMACPain 87  ±  16 83  ±  20 0.08Stiffness 79  ±  19 75  ±  22 0.12Physical function 80  ±  19 76  ±  19 0.07Short Form-36Physical function 57  ±  28 50  ±  27 0.11Physical componentsummary 39  ±  12 38  ±  11 0.99Mental componentsummary 45  ±  9 44  ±  10 0.80Satisfaction score 90  ±  14 90  ±  15 0.94 *WOMAC  =  Western Ontario and McMaster Universities Osteo-arthritis Index.  † The values are given as the mean score and thestandard deviation. 1677 T HE  J OURNAL OF  B ONE  & J OINT  S URGERY  d JBJS . ORG V OLUME  90-A   d N UMBER   8  d A  UGUST  2008I NPATIENT  C OMPARED WITH  H OME -B ASED  R  EHABILITATION A  FTER   U NILATERAL  H IP OR   K NEE  R  EPLACEMENT
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks