Ninety-three patients, each undergoing simultaneous bilateral total knee replacement, were enrolled in a prospective, randomized, multicenter clinical investigation in which the primary objective was to compare the knee motion metrics between a high-flexion prosthesis and a standard knee prosthesis in patients who received bilateral total knee arthroplasty. A secondary objective of the study was to compare the preoperative goals and postoperative satisfaction of each patient through the interpretation of data collected with use of Palm technology.
Approval by the institutional review board at each of the participating clinical sites was obtained. The study is registered as study number NCT00380861 at http://clinicaltrials.gov/. Informed consent was obtained. Each surgeon's office had a dedicated study coordinator who enrolled the patient into the electronic database and then set up the patient's Palm device (Palm m500; Palm, Santa Clara, California). Preoperatively, the patient received training on how to use the device and entered practice data into the Palm device. Patients completed the KOOS survey, which is a patient questionnaire that inquires about pain, knee symptoms, quality of life, difficulty with sports and recreation activities, and difficulties with activities of daily living. Patients also completed a questionnaire in which they assigned a degree of importance to their ability to perform the activities of daily living, sports, and recreation activities, as based on elements in the KOOS questionnaire. American Knee Society scores were also collected preoperatively. Patients were then sent home with the Palm device.
Patient-reported outcomes were collected in two ways: (1) with use of traditional paper-based case report forms that were completed preoperatively and then at two weeks, six weeks, six months, and twelve months postoperatively; and (2) with use of preprogrammed Palm devices to electronically record patient-recorded outcomes.
After undergoing simultaneous bilateral total knee replacement with the randomized knee implants, patients were discharged. The study coordinator entered the patient's discharge date, pain medication, and physical therapy facility. With use of the Palm device, patients then answered questions (based on elements of the KOOS questionnaire) about the degree of difficulty that was involved in the performance of activities of daily living. Patients were asked specific questions about the severity and location of pain; methods of managing pain with prescription medications, nonprescription medications, or other methods; and the degree of difficulty they had in descending and ascending stairs, rising from sitting, standing, bending to pick up an object, walking on a flat surface, getting in and out of a car, shopping and doing chores, putting socks or stockings on or off, rising from bed, lying in bed, getting in and out of the bath, sitting, getting on and off the toilet, performing light domestic duties, driving a car, and returning to his or her usual job function. Patients also answered questions about quality of life. Specifically, questions were asked about the level of awareness that the patient had of his or her knee problem, the degree to which the patient's lifestyle was modified because of the knee, the patient's confidence in the knee, and how much difficulty in general that the patient was having with the knee. Patients uploaded the Palm data via the eCaseLink system (DSG, Malvern, Pennsylvania) on a daily basis for six weeks. Daily data were a subset of additional data that the patients provided on a weekly basis.
In addition, patient outcomes data were entered postoperatively from week three to week six by the treating physical therapists to track knee range of motion, lower-extremity strength, and the reported pain and functional status. Data recorded on the electronic patient-recorded outcome technology included milestones regarding knee motion and strength, functional milestones, home-environment assessment, and pain-management goals.
Finally, patients also completed a paper-based questionnaire regarding their satisfaction with performing the same twenty-two activities, which were based on elements of the KOOS questionnaire, during clinical follow-up appointments at six and twelve months postoperatively.
This study had multiple goals: (1) to assess utilization and compliance issues related to electronic patient-reported outcomes; (2) to examine patient subjective and objective findings recorded on a daily basis for the first six weeks following simultaneous bilateral total knee arthroplasty; and (3) to compare preoperative expectations and postoperative satisfaction in patients undergoing total knee replacement.
Statistical Analysis
Data were gathered for each patient with use of electronic patient-recorded outcome technology, from the day of hospital discharge through six weeks after discharge. Data were transmitted on completion of the daily questionnaire; patients were not allowed to skip any questions, so there were no missing responses in a completed questionnaire. The Palm device had an internal clock, which would only allow data to be entered in the twenty-four-hour period of that day. If the window of time was missed, the patient was not able to enter data for that missed day and the next day's data were captured. This enabled the collection of real-time responses. Daily compliance was calculated as the proportion of the ninety-three patients who completed the questionnaire on each respective day postoperatively.
Responses to questions over time were matched by patient so that trends could be identified. Many questions solicited a response from among five choices (i.e., not difficult, slightly difficult, somewhat difficult, difficult, or extremely difficult) on an ordinal scale. Responses to these questions were graphically represented on an interval scale of 1 through 5 so that averages over time could be observed. One particular question on pain or discomfort solicited a response on a visual analog scale of 0 to 100, with 0 denoting no pain at all and 100 denoting pain as bad as it can be. Responses on an interval scale, whether 1 through 5 or 0 through 100, were analyzed with use of the PROC GPLOT procedure of the SAS statistical software package, version 9.2 (SAS Institute, Cary, North Carolina) by fitting a smoothed line through the data under a spline routine with the option INTERPOL = SM50S. These smoothed plots were overlaid onto plots of individual responses; each individual response at each time interval was indicated with a dot. Minimal random noise was added to individual responses so that plots would distinguish the frequency of individual responses.
Daily, patients reported qualities of pain (e.g., aching, dull, burning, throbbing, sharp, or stabbing), pain locations (e.g., medial, lateral, center, left, right, above, below, hip area, or lower back) on both anterior and posterior diagrams of their legs, ambulatory aids being used, and pain medications (both prescription and nonprescription); more than one response was allowed for each of these questions. The percentages of patients responding at least once during the week were computed for each respective question.
Patients were asked preoperatively how important each of twenty-two daily-living and sports activities were to them (from the KOOS questionnaire; e.g., kneeling, getting in and out of a car, rising from sitting, and descending stairs). Responses were recorded on a 1 to 4 scale, with 1 meaning not at all; 2, a little; 3, somewhat; and 4, very. At both six and twelve months postoperatively, patients were asked about their postoperative satisfaction regarding each of these same twenty-two activities, on the same scale of 1 to 4. The proportion of patients whose postoperative satisfaction was equal to or greater than their preoperative importance was calculated for each of these twenty-two categories, at both six and twelve months after surgery, and this was defined as the proportion of patients whose expectations were met or had been exceeded.
Source of Funding
There was no external funding source for the present study; three authors of this paper (K.A.D., J.L., and T.O'D.) are employees of DePuy Orthopaedics and thus receive a salary from DePuy.
Patient demographic data are found in Table I. Patient ages ranged from forty years to seventy-three years, with a mean age (and standard deviation) of 60.8 ± 8.5 years. Osteoarthritis was the preoperative diagnosis in 99% of the patients. Body mass index was calculated for all patients and ranged from 19 to 47 kg/m2 with a mean body mass index of 31.6 ± 5.3 kg/m2.
Recording of Data
Real-time data were collected electronically every twenty-four hours. All of the required questions were answered, with no missing responses identified. Patient compliance in recording outcome data was 82% for days 4 through 38 after hospital discharge. Compliance was lower for the first few days following discharge (average, 62% for days 0 through 3) and the last few days of the study (average, 50% for days 39 through 42). In the patient satisfaction survey at the conclusion of the study, electronic documentation was rated as user friendly across all ages in the study.
Pain and Stiffness
As expected, pain after total knee arthroplasty varied among the patients, but pain levels also varied markedly within patients on a day-to-day basis. Pain intensity for all patients was greatest immediately after hospital discharge and decreased by approximately 50% over the course of the first forty-four days following discharge (Fig. 1). The patients’ answers regarding the nature of the pain varied at different points in the recovery process as well. Sharp, stabbing pain decreased during the first six weeks after discharge, but postoperative dull and aching pain remained virtually unchanged from week one through week six (Fig. 2).
The present study focused on simultaneous bilateral knee replacement, so the patients provided pain data for both knees. There were, however, no significant differences in the pain location for the left or right knee. Figure 3 depicts the location of pain for both lower extremities, although depicted as only on the left side. There was a higher prevalence of anterior knee pain than posterior knee pain. Anterior knee pain, especially central knee pain, persisted the longest, and with the least amount of change, while thigh pain and calf pain decreased. Pain in the posterior aspect of the thigh, knee pain, and calf pain decreased significantly, with thigh pain showing the biggest decrease. Low back pain and hip pain did not change appreciably throughout the six weeks.
Confidence and Awareness
Although awareness is a broad concept, and awareness to one patient may equate with a perception of pain to another, patients were asked to report on how aware they were of knee problems after discharge. Patients reported gradual improvements in the degree to which they were aware of their knee over the six-week interval, with the data demonstrating that the patients were between "somewhat aware" and "aware" at six weeks after total knee arthroplasty. Likewise, patients reported gradual improvements in the degree of confidence that they had in their knees. Data demonstrated that patients "mildly" to "moderately" lacked confidence in their knees. Perhaps reflecting confidence and awareness, over the six-week period, patients became less inclined to modify their lifestyle to avoid "damaging" the knee. At six weeks after discharge, patients were still modifying their lifestyle moderately.
Function
Mobility
Patients became more independent with gait over the six-week period of time after discharge from the hospital. Figure 4 depicts the percentage of assistive device utilization for the six-week period. The use of a walker decreased steadily during the six weeks, with a concomitant increase in the use of crutches or a cane. At six weeks, 14% of the patients still utilized a walker for their primary means of assistance.
As can be seen in Figure 5, the steeper slope between weeks one and two after discharge as compared with the slope from weeks three to six indicates that recovery with respect to stair descent improved rapidly during the first two weeks and that the rate of improvement was slower after two weeks. Difficulty in ascending stairs followed a similar recovery pattern to that of stair descent, with a persistent difficulty with this activity at six weeks. It appears that at the six-week point in time after discharge, stair ascent was less problematic than stair descent (Fig. 6).
Preoperative Expectations Versus Postoperative Function
Six months after total knee arthroplasty, more than 60% of the patients reported that their satisfaction did not match their preoperative expectations in their ability to kneel, squat, get in and out of a car, descend stairs, rise from a chair, get on and off the toilet, ascend stairs, or bend to retrieve an object from the floor.
All other activities at six months met the preoperative expectations of at least 50% of the patients. Two activities (light domestic activities and jumping) met preoperative expectations in at least 70% of the patients. Over the time period from six to twelve months, more patients had their preoperative expectations met than they did at the six-month point in time, except for the activity of running. At twelve months, kneeling was the only activity in which 50% of the patients’ preoperative expectations were not met. At twelve months, more than 90% of the patients’ expectations were met or exceeded with regard to the patient's ability to perform light domestic activities or walk on a flat surface (Fig. 7).
The use of an electronic format to record patients’ outcomes provided a dynamic, real-time glimpse into the recovery process in patients following bilateral total knee arthroplasty. This format facilitated patient compliance in recording data during the first six weeks after discharge. Historically, electronic patient-recorded outcomes were initially utilized to assess patient response to the effect of pharmaceuticals with regard to pain management. The patient's recording of pain parameters and pain management comprised an appreciable amount of data analyzed in the present study as well. Data gathered via electronic patient-recorded outcomes can be used to better educate patients about postoperative pain and pain management following total knee arthroplasty. Electronic data capture of specific pain medication dosage and administration coupled with a more detailed analysis of daily fluctuations tied to specific interventions in the formal postoperative physical therapy program may allow for greater success in achieving goals of knee motion and function.
The results of this study also demonstrate clear and consistent variability in pain between individual patients as well as widespread variability in day-to-day pain intensity within individual patients. With consistent patient feedback in the form of daily electronic data entry, the surgeon and therapists are provided insight into daily recovery between office visits. Timely and accurate data via electronic documentation, shared in an effective manner with the health-care professional, allow for closer monitoring of patient progress. The ability to be informed and to react to changes reported on a daily basis may lead to increased patient satisfaction.
Electronic reporting of patient outcomes in this study provided meaningful data regarding patient function following bilateral total knee arthroplasty. Data revealed a clear trend in progressing from a walker to a cane or crutches as the primary assistive device at four weeks. It is unknown whether this progression was dictated by the surgeon's postoperative protocol or a directive from the therapist or if it was a function of the patient's self-perceived ability and independence. Negotiating stairs was difficult for this study group. Only 56% of patients’ expectations in ascending stairs were met at six months, and only 68% were met at one year. Preoperative expectations to descend stairs were met in only 53% of patients at six months and in 66% of patients at one year after surgery. According to body mass index criteria, 92% of the patients were overweight or obese. A closer correlation between body mass index and difficulty on stairs, quadriceps strength, and/or exploration of improved component design to alter loads on the patellofemoral joint may be warranted. Transferring in and out of the bathtub did not meet expectations in 62% of patients at six months and in 68% at one year after surgery. Altering the survey to provide a choice between the ability to transfer into a shower or the ability to transfer into a bathtub may provide greater insight to patients’ ability to perform this vital function. At the higher end of functional activities, there was only a one-point difference between the values at six months and at one year in the percentage of patients whose ability to jump (77% compared with 78%, respectively) and to perform heavy domestic duties (66% compared with 67%, respectively) met or exceeded preoperative expectations. At six months, 68% of the patients’ expectations to run were met or exceeded, compared with only 64% at one year. This finding may prove beneficial for patients who expect to run following bilateral total knee arthroplasty.
While this study provides insight into the postoperative recovery process, it must be kept in mind that findings in patients undergoing simultaneous bilateral total knee arthroplasty cannot be generalized to patients undergoing unilateral total knee arthroplasty or staged bilateral total knee arthroplasty. Although there were no major differences in pain location between right and left knees, pain locations may change in patients undergoing unilateral total knee replacement. Preoperative variables, including articular cartilage involvement, lower-extremity and knee alignment, leg preference, knee motion, contractures, and other factors can differ from one knee to the other. These side-to-side differences may impact postoperative recovery when a unilateral total knee arthroplasty is performed.
Patient compliance in maintaining electronic reported outcomes in this study was far superior to reported compliance for paper recording. Compliance may have been impacted by the amount of help received from others assisting in the care of the patient following surgery. This study did not attempt to correlate compliance with the location to which the patient was discharged following surgery, nor did it assess the amount of assistance provided in tracking outcomes with the Palm device. To the authors’ knowledge, this is the first application of electronic patient-recorded outcomes in the orthopaedic population and it proved successful in this mostly elderly population even in the early recovery period. While the present study demonstrates the effectiveness of electronic means to record patient data following total knee arthroplasty, the specific technology utilized in the study is quickly becoming outdated with the rapid advances in digital technology. Data were entered by the patients into the Palm and transferred from their device into the collective data pool of the study with use of an analog signal. Future studies utilizing electronic data capture will need to move to the digital format.
The present study provided detailed patient-reported outcomes regarding the location, intensity, and nature of pain along with pain-management modalities. In addition, the present study provided patient-reported data relative to the patients’ confidence and trust in their knees following bilateral total joint arthroplasty along with a general awareness of the knee following surgery. This study, however, did not explore the relationship between function, confidence, or awareness and the patient's pain. Exploring the relationship between pain, function, and subjective indicators of confidence in the knee may prove insightful.
Finally, although a primary goal of the study was not to assess pain in the lower back and hips, patients were asked to rate postoperative pain in these areas at given time points following surgery. Any referred pain that is reported postoperatively by the patient could very well be a vestige of preoperative dysfunction, but it could also be the result of changes in the patient's postoperative manner of movement. Future studies should look closer at preoperative pain in the lower back and hips and track changes closely in these areas following total knee arthroplasty.
The current study provided keen insight into pain intensity, the nature of pain, and modalities of pain management through the use of electronic patient-recorded outcomes. Electronic reported outcomes also provided meaningful data regarding the patient's confidence in the knee following total knee arthroplasty as well as the patient's overall awareness of the knee following surgery. Most importantly, the current study gives the health-care provider a detailed look at the recovery of function from the patient's perspective, which may prove useful in the preoperative education of patients who are considering having bilateral total knee arthroplasty.