| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Research Reports |
Lieutenant Colonel TG Sutlive, PT, PhD, OCS, is Assistant Professor, US Army-Baylor University Graduate Program in Physical Therapy, Fort Sam Houston, Tex.
Lieutenant SD Mitchell, PT, MPT, is Staff Physical Therapist, Federal Medical Center Devens, Devens, Mass
Lieutenant SN Maxfield, PT, MPT, is Staff Physical Therapist, Federal Medical Center Carswell, Fort Worth, Tex
Captain CL McLean, PT, MPT, is Chief Physical Therapist, 67th Combat Support Hospital, Wuerzburg, Germany
Captain JC Neumann, PT, MPT, is Staff Physical Therapist, Ireland Army Community Hospital, Fort Knox, Ky
Captain CR Swiecki, PT, MPT, is Staff Physical Therapist, Walter Reed Army Medical Center, Washington, DC
Major RC Hall, PT, MS, ATC, SCS, is Chief, Physical Therapy Element, 86th MDOS, Ramstein Air Force Base, Germany
Captain AC Bare, PT, MPT, ATC, is Director of Sports Medicine, US Army World Class Athlete Program, Fort Carson, Colo
TW Flynn, PT, PhD, OCS, FAAOMPT, is Associate Professor, Department of Physical Therapy, Regis University, Denver, Colo
Address all correspondence to Dr Sutlive at Academy of Health Sciences (Attn: MCCS-HMT), Physical Therapy Branch, 3151 Scott Rd, Suite 1303, Fort Sam Houston, TX 78234-6138 (USA) (thomas.sutlive{at}cen.amedd.army.mil)
Submitted December 6, 2002;
Accepted July 8, 2003
 | Abstract |
|---|
2 degrees (+LR=4.0, 95% confidence interval [CI]=0.7–21.9), great toe extension of 
78 degrees (+LR=4.0, 95% CI=0.7–21.9), and navicular drop of 3 mm (+LR=2.4, 95% CI=1.3–4.3). Discussion and Conclusion. The results suggest that patients with PFPS who have forefoot valgus alignment of 2 degrees, passive great toe extension of 78 degrees, or navicular drop of 3 mm are most likely to respond favorably to initial intervention with an off-the-shelf foot orthosis and instruction in activity modification.
Key Words: Knee pain • Orthotics • Patellofemoral • Physical examination • Rehabilitation
| Introduction |
|---|
 TopAbstract Introduction MethodResultsDiscussionConclusionAppendix 1Appendix 2.References |
|---|
Despite its prevalence, the etiology of PFPS is not clearly understood. Factors that are hypothesized to contribute to PFPS include vastus medialis muscle weakness and reduced control,8–11 shortened lower-extremity muscles, and abnormal foot and ankle biomechanics. Because of the multifactorial nature of PFPS, numerous intervention strategies have been proposed. Intervention options for PFPS include strengthening and stretching exercises,9,11–14 use of orthotic devices for the knee,15–17 taping of the patella,8,10 use of foot orthoses,18–20 and acupuncture.21
The patient characteristics that might predict success for any of these interventions are based largely on opinion or on arguments of biomechanical or biological plausibility. There is currently limited evidence available to the clinician matching patients to specific interventions.22 We conducted our study in an attempt to identify characteristics that could predict which patients are likely to respond to a specific intervention. The intervention strategy we investigated was a combined program of the use of foot orthoses and modified activity. Several studies18–20 have demonstrated the benefits of foot orthoses for managing the pain associated with PFPS. We opted to use a combined program of interventions because we believe it is representative of what is typically done in clinical practice. That is, we contend that most physical therapists would initially prescribe a physical intervention and would then advise the patient to modify his or her physical activity level.
Although we believe some patients will improve with the use of foot orthoses and a modified physical training regimen, we suspect that not all patients respond to this combination of interventions. For instance, many patients with PFPS are thought to respond favorably to programs of strengthening or flexibility exercises.9,11–14 Orthotic devices, however, can be relatively expensive, and therefore determining which patients are most likely to benefit from the use of orthoses is, in our opinion, an important consideration for the therapist.
No one, to our knowledge, has identified those examination findings or factors in a patient's history that could be used to predict which patients with PFPS will respond to a specific intervention. The purpose of our study was to identify the patient characteristics that can be used to determine which individuals with PFPS might respond favorably to intervention consisting of use of an off-the-shelf orthotic insert combined with instruction in activity modification. We believed, therefore, that all participants needed to receive identical treatment and needed to be examined by use of the same tests compared against the same reference standard.23 In this type of study, the reference standard was the participants' response to intervention.22,23 The diagnostic value of each finding and observation related to a patient's history and physical examination was calculated to determine the characteristics of the participants who responded best (reduction in pain) to the intervention. Thus, the research design we used was one traditionally used in diagnostic accuracy studies.24
| Method |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionAppendix 1Appendix 2.References |
|---|
=28.1, SD=6.2), participated. In order for people to be included as participants, they needed to have symptoms of PFPS. The predominant symptom of PFPS is retropatellar pain that increases during weight-bearing activities such as running, squatting, and stair climbing.1,10,25 Therefore, the diagnosis of PFPS was determined by clinicians based on the complaint of retropatellar pain that was provoked by either a partial squat or stair descent.10 People were not allowed to participate in the study if they had a recent history of trauma to the knee, ligamentous laxity of the painful knee(s), history of surgery for the affected knee(s), or history of systemic or neurologic disease. People who reported having lower-extremity problems such as stress fractures or shinsplints and those already receiving intervention for their knee pain also were excluded. Participants were required to be fluent in the English language. Each participant signed an informed consent form prior to enrollment in the study.
Procedure and Examination
Participants met with the same investigators twice. All questions asked during the history-taking session are shown in Appendix 1. During that session, the participants also completed a visual analog scale (VAS) to characterize their current level of knee pain. The VAS is believed to provide a valid measurement of pain intensity and clinical change in a sample of patients diagnosed with PFPS.26 Because the pain of PFPS is typically induced by activity,3,10,27 participants were asked to record the maximum knee pain they experienced during the most recent activity that brought on their pain. Participants were then asked to place a vertical mark on a 10-cm horizontal line to indicate the amount of pain they experienced. One end of the line was labeled "No pain at all," and the other end of the line was labeled "Worst possible pain." The distance from the left extreme point of the line ("No pain at all") to the participant's mark was measured and recorded (in millimeters).
The physical examination we used consisted of measurements that we believe are routinely obtained from patients with knee pain. This examination included palpation of bony landmarks and measurement of range of motion to reflect muscle length. Additionally, the following lower-extremity measurements were taken: rear-foot alignment in subtalar joint neutral (STJN), forefoot–to–rear-foot alignment, navicular drop, relaxed calcaneal stance, and Q angle. To our knowledge, no one has reported the reliability of data obtained with these tests and measures in a population of individuals with PFPS. We included an extensive list of lower-extremity measures in order to avoid missing any possible predictors of success with the intervention used in our study. The measurements were obtained for all patients, and data were recorded on a data collection form. Two different testers (SDM and CLM) examined the first 30 participants sequentially to determine the interrater reliability of the measurements. One rater (CLM) tested the remaining 20 participants, and only that rater's measurements were entered into the data analysis for the predictive portion of the study. Two researchers (JCN and CRS) served as recorders for the values measured by the raters.
Initial visit.
Measurements were taken bilaterally while participants were in the following positions: prone, standing, seated, and supine. All goniometric measurements were taken using a 17.8-cm (7-in) plastic goniometer marked in single-degree increments. After the informed consent and history were obtained, participants were asked to lie prone with one leg extended and the ipsilateral foot off the end of the plinth. The contralateral knee was flexed to 90 degrees, with the hip laterally (externally) rotated. This position was identical to the "figure-four position" for subtalar joint measurement as described by Donatelli.28 In this position, 2 lines were drawn with a felt-tip pen by the examiner. One line bisected the distal one third of each participant's leg, and the other line bisected the calcaneus. Although the reliability of STJN measurements is questionable, the foot was positioned in what we believed was STJN, and the rear-foot measurement was recorded. The forefoot–to–rear-foot alignment also was measured and recorded in this neutral position.29 Next, ankle dorsiflexion was measured both with the knee fully extended and with the knee flexed to 90 degrees. Tibial torsion was measured as described by Gross.29 For the final measurement obtained in the prone position, the Craig test was used to determine if femoral retroversion or anteversion was present.30
Measurements of tibial varum and tibial valgum, the rear foot in a relaxed calcaneal stance position, and navicular drop were obtained with the participants standing on a step stool with their feet shoulder-width apart. To our knowledge, there is no standardized method of measuring the degree of tibial varum or tibial valgum. We measured the angle formed between the line previously drawn on each participant's leg and the horizontal plane formed by the surface of the step stool. We defined tibial varum as being present when the proximal end of the line drawn on the participant's leg being more lateral than the distal end. The rear-foot position in a relaxed calcaneal stance position was measured as described by Jonson and Gross.31 For the navicular drop test, the examiner palpated the participant's navicular tuberosity and made a horizontal mark on the tuberosity. The examiner then positioned the participant's foot in the subtalar neutral position.32 An index card then was placed vertically next to the navicular bone and flush with the surface of the stool. The examiner made a mark on the index card at the height of the navicular tuberosity. The participant then was asked to relax his or her stance on the extremity, and the new position of the navicular was marked on the index card. The difference between the 2 marks was measured (in millimeters) and reflected what we called the navicular drop (or rise). The final procedure performed with the participants in a standing position was the measurement of Q angle.33
Medial, neutral, or lateral patellar orientation was determined visually by the testers (SDM, CLM) with the participant in a seated position with both knees flexed to 90 degrees.30 Passive great toe extension was measured with the participant in a long sitting position. The stationary arm of the goniometer was positioned parallel to the first metatarsal. The axis was at the first metatarsophalangeal (MTP) joint, and the movable arm of the goniometer followed the proximal phalanx of the great toe.34 The examiners put the participant in the Thomas test position to measure hip flexor tightness and in the position for the 90/90 straight-leg-raising test to measure hamstring muscle length. The leg-length difference as perceived by the examiners were measured with the participant positioned supine. The Thomas test and the 90/90 straight-leg-raising test were conducted as described by Magee.30 If the participant was unable to extend the knee to within 20 degrees of full extension during the 90/90 straight-leg-raising test, the test was considered positive for hamstring muscle tightness. Leg length was determined by measuring from the anterior superior iliac spine to the distal medial malleolus (in centimeters) with a tape measure.29,30,35 Finally, with the participant in a side-lying position, the Ober test was conducted to assess the presence or absence of iliotibial band tightness.30
Intervention.
After all measurements had been taken, the participants were given orthotic inserts with instructions for their use. The orthotic inserts used were First Step premolded full-length insoles* with a firm arch support and heel cushion (Fig. 1). No adjustments or postings were made to the orthosis. The participants were instructed to use the orthoses at all times for 21 days. Participants were instructed to take the orthotic splints out of their work boots and use them in other shoes they wore. They were told to wear only footwear that allowed for orthosis use during the 21-day intervention period.
|
Second visit.
Participants returned between days 20 and 23 for a follow-up visit. Each participant completed a second VAS to determine if there was a change in his or her pain level after 3 weeks of orthosis use and modified activity. Each participant also completed a Global Rating of Change (GRC) questionnaire to assess the overall change in the status of his or her PFPS (Appendix 2).36 All participants were told they could seek further intervention if their knee pain persisted. No data were collected to assess the degree of adherence for either the orthoses or the activity modification.
Data Analysis
Data were collected and formatted into a spreadsheet usable in SPSS for Windows, version 10.1.
To determine the interrater reliability for continuous variables, intraclass correlation coefficients (ICC [2,1]) were calculated. Interrater reliability for categorical measurements (Thomas test, hamstring muscle 90/90 straight-leg-raising test, Ober test, patellar orientation) was determined using the kappa coefficient.
Prescriptive validity studies are designed to identify predictor variables that serve as the basis for choosing a specific intervention.37 For the prescriptive validity portion of the study, the results for each participant were classified as being either successful or not successful. If 50% or greater improvement was noted on the VAS, the intervention was considered a success. If improvement was less than 50%, the intervention was not considered to be successful. The classification of results as successful or unsuccessful allowed us to examine how data from the history and the physical examination could be used as a predictor of success.
Univariate analyses (using chi-square tests and individual t tests) were conducted to determine which variables had a significant relationship with the reference standard (intervention success) and served as a first-pass screening procedure to determine which variables would be entered into a binary logistic regression model. Chi-square analysis was done to determine which of the discrete variables were predictive of intervention success.
Discrete variables included questions from the history (Appendix 1), the participant's sex, the Thomas test, hamstring muscle 90/90 straight-leg-raising test, Ober test, and patellar orientation. Continuous variables were analyzed for their relationship with intervention outcome using independent t tests. Continuous variables included age, duration of symptoms, rear foot in SJTN position, forefoot–to–rear-foot alignment, ankle dorsiflexion with knee extended, ankle dorsiflexion with knee flexed, tibial torsion, Craig test, relaxed calcaneal stance position, tibial varus and tibial valgus, navicular drop, Q angle, great toe extension, and leg-length difference. The alpha level for all univariate analyses was set at 
=.15. This liberal alpha level was chosen to avoid discarding any potentially useful predictors from being entered into the logistic regression model due to a Type II error. Given the exploratory nature of this study and influence of the subsequent regression procedure, the inflated Type I error rate was determined to be acceptable. Successful and unsuccessful results were analyzed to ensure homogeneity of variance, normal distribution, and random assignment using the Levene test.
Sensitivity, specificity, and likelihood ratios (LRs) were calculated for those elements of the history and physical examination that were found to be predictive of intervention success (P<.15). Sensitivity reflects the ability of a diagnostic test to determine a true positive for a given disease or condition, and specificity reflects the ability of a test to determine a true negative for a disease or condition.38 Intervention success was what we were attempting to predict, and our goal was to identify the information from the history and measurements obtained during the physical examination that best predicted the likelihood of intervention success. We also were interested to see if any combination of the diagnostic variables generated higher LRs than the individual variables alone. Stepwise logistical regression analysis was conducted to determine which, if any, combination of the diagnostic variables was most predictive of intervention success. Only variables that demonstrated statistical significance from the univariate analysis were entered into the regression model.
Although sensitivity and specificity are commonly cited in the research literature, some authors24,39 contend that they are of limited clinical use. An LR is another diagnostic accuracy measure that is used to determine the probability of whether a person is more or less likely to have a condition (target condition). Likelihood ratios combine the best features of sensitivity and specificity, and they are convenient summary measures of diagnostic test performance.40 Likelihood ratios provide clinicians with a measure that can be used to determine which elements of the clinical examination are most predictive of the target condition being present.24 A positive LR increases the likelihood of a target condition given a positive test result, and a negative LR decreases the likelihood of the condition given a negative test result.24 Positive and negative LRs are calculated as follows:
|
|
|
|
An LR greater than 1 means the probability that a given condition exists is high, and an LR less than 1 means the probability that the condition exists is low. When an LR approaches 1, the odds of showing whether a condition is present diminishes, and the test results are indeterminate.41 In our study, the condition of interest was intervention success. We were most interested in identifying measurements with a large +LR; that is, those measurements that were most helpful in ruling in the condition of intervention success.
For the diagnostic predictors that consisted of discrete variables, we used a 2x2 contingency table to calculate sensitivity and specificity. For each of the continuous variables, a receiver operating characteristic (ROC) curve was generated. The ROC curve plots the sensitivity along the Y-axis and 1–specificity along the X-axis for multiple values of each variable, and it displays the coordinates for both axes in tabular format. The data from the table were transferred into a spreadsheet, and the sensitivity, specificity, and LR for each point on the curve were calculated. The cutoff score for each of the continuous variables was the value of the variable that generated the highest +LR.
| Results |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionAppendix 1Appendix 2.References |
|---|
=28.1, SD=6.2), participated. Data from 5 participants were excluded from the data analysis for the predictive portion of the study. Four participants were unable to attend their final appointment, and their data were not used. One participant did not complete the questionnaire. Thirty-three of the remaining 45 participants had bilateral PFPS. Calculations for the predictive portion of the study, therefore, were based on 78 knees from 45 subjects (34 men, 11 women). Descriptive statistics for the continuous variables measured on each of the 50 subjects are shown in Table 1.
|
=.07; hamstring muscle 90/90 straight-leg-raising test, =.19; Ober test, =.08; and patellar orientation, =.08.38
|
|
|
Findings related to 8 variables appeared to be related to intervention success based on the chi-square and independent t-test analyses (Tab. 3). Of those variables, 6 characteristics emerged as predictors of intervention success based on their LRs. The sensitivity, specificity, positive LRs, and cutoff scores for the predictors are shown in Table 4. No diagnostic test cluster emerged from the logistical regression analysis.
|
|
| Discussion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionAppendix 1Appendix 2.References |
|---|
Fagan42 developed a nomogram (Fig. 4) to facilitate the use of LRs and to provide clinicians with a tool for determining the probability that a condition is present given the results of a diagnostic test. In our study, 60% of all participants were considered to have initial intervention success after 3 weeks of the combined interventions. Therefore, the pretest probability of success with the intervention used for the participants in our study was 60%, as shown in the first column of Figure 4. The second column of the nomogram represents the LR for a given diagnostic test. The range of positive LRs for the forefoot–to–rear-foot alignment, great toe extension, and navicular drop measurements are plotted in the second column of the nomogram. The third column of the nomogram shows the change in the probability of intervention success after applying the LRs to the pretest probability.
|
Brody43 described the navicular drop test as a convenient clinical method for estimating the amount of foot pronation. He considered 10 mm to be a normal amount of navicular drop, whereas he considered values greater than 15 mm as indicating excessive motion and reason to consider the use of foot orthoses in runners. In studies of people without known impairments or pathology, mean navicular drop test values ranged from 6.2 to 9.0 mm.32,44 The mean navicular drop test value for the participants in our study was 5.3 mm (Tab. 1). The mean navicular drop test value for the successful group was 3.6 mm, whereas the mean value for the unsuccessful group was 6.3 mm (P<.01). Therefore, the participants in our study with minimal motion on the navicular drop test responded best to intervention. Furthermore, based on the ROC analysis, the cutoff score for the navicular drop test was 3 mm. That is, patients with a navicular drop of 3 mm or less were most likely to respond favorably to the combined interventions used in this study (Tab. 4).
The amount of great toe extension required during the propulsion phase of gait ranges from 50 to 90 degrees.34,44,45 Several authors34,44,45 have described static non-weight-bearing (NWB) techniques for assessing first MTP joint extension, such as the one we used in our study. Other investigators examining subjects without knee pain have reported passive great toe extension values of 82 to 96 degrees when measured in a NWB position.34,44,45 Our finding of mean great toe extension of 94 degrees in our patients with PFPS is consistent with the earlier reports. There was a difference in great toe extension between participants in the successful group (=92°) and participants in the unsuccessful group (=99°) (P<.05). Based on the ROC analysis, participants who had reduced great toe extension (78° or less) were more likely to benefit, presumably from intervention (Tab. 4).
Forefoot–to–rear-foot alignment is a measurement of the frontal-plane relationship between the 2 foot segments and is usually measured in a NWB position.28,29,46 McPoil et al47 studied the incidence of forefoot types in 58 female subjects without knee pain and found that forefoot valgus was present in 44.8%, while forefoot varus was present in 8.6%. In contrast, Garbalosa et al46 reported that forefoot varus malalignments (86.7%) were far more common than valgus types (8.8%) in 120 subjects without known knee pathology. In our study, 83% of the feet were classified as being in forefoot varus, and the remaining 17% were considered to be in forefoot valgus. The mean forefoot alignment measurement for all participants in our study was 7.3 degrees of varus. Participants in the successful group had less varus (=4.9°) than those in the unsuccessful group (=10.2° of varus) (P<.005). The ROC analysis showed that patients with PFPS who appeared to respond best to the intervention used in this study had forefoot valgus (Tab. 4).
Participants' Response to Intervention
Because our study was not a randomized trial, we were unable to determine whether the participants' response to intervention was due to the effects of the orthotic device, due to the modified physical training regimen, or reflective of improvement due to the passage of time. We used our study design as a first step in determining which participants might respond best to a specific intervention. In order to do so, all participants had to undergo an identical battery of historical questions and tests and measures, receive identical intervention, and be compared against the same reference standard.23
Given that the predictors of intervention success were foot and ankle measurements, it is our opinion that the orthotic device was primarily responsible for the relief of symptoms experienced by the participants. Our research design, however, does not allow us to make any stronger claim than one that is opinion-based. To our knowledge, there are no reports on the effects of modified activity or rest on patellofemoral pain, and therefore no such data are available for comparison with our results. Additionally, it would be helpful to know the short-term natural history of PFPS. We found, however, only 2 studies in which there were descriptions of the natural history of PFPS.5,48 Unfortunately, the mean length of follow-up in the 2 studies was 3.8 and 6 years, respectively, which prevents us from making comparisons with our study.
Customized foot orthoses are often prescribed for patients with knee pain due to PFPS who have excessive foot and ankle motion.28,43 Brody43 originally described the navicular drop test as a convenient clinical method for estimating the amount of foot pronation, and he considered values greater than 15 mm as indicating excessive motion and reason to consider the use of foot orthoses in injured runners. Presumably, patients with excessive motion would respond best to a customized orthosis that is supposed to control motion. However, the participants in our study who appeared to respond best to initial intervention had minimal motion on the navicular drop test, a lesser amount of passive extension of the first MTP joint, or generally a lesser degree of forefoot varus alignment compared with those in the unsuccessful group. The combination of these findings suggests to us that the patients with PFPS who appeared to respond best to our intervention had relatively inflexible feet. Such feet have a limited ability to become mobile and a relative inability to attenuate the forces imparted to the feet during the stance phase of gait.49 Forces introduced at the foot will then most likely be transmitted up the lower extremity to be absorbed by joints such as the knee and patellofemoral joints, and this may lead to dysfunction in these regions.49–52 We did not measure ground forces or any biomechanical effects of the orthoses, and therefore we do not know how well the insert used in our study absorbed shock. Our findings suggest to us, however, that if there is a therapeutic effect of the intervention, it may have been achieved by shock attenuation rather than by motion control. Evidence suggests that shock-absorbing inserts can reduce the incidence of lower-extremity overuse injuries.20 Recently, based on reviews of the literature, some authors53,54 have suggested that the shock-absorbing effects of orthoses, and not their ability to correct alignment and control motion, may be their most useful asset.
Outcome Measures
A 50% improvement on the VAS was set as the criterion used to identify the group of participants who had a sufficiently high and, in our view, clinically meaningful reduction in their pain. Eng and Pierrynowski18 reported a 37.5% improvement on the VAS for subjects with PFPS who used a customized insole over the course of 8 weeks. Furthermore, it has been argued that a change of 30% on any numerical pain rating scale represents a clinically meaningful reduction in pain in subjects with a variety of disorders,55 although evidence showing how this change is beneficial for all situations is not clear. Therefore, we felt that a 50% threshold was sufficiently high to identify individuals who responded to the intervention.
Yet, the results of the GRC questionnaire suggest that the participants improved by more than just their pain levels. The GRC is a self-report instrument that measures how much the participant's condition changed since starting the intervention.36 Juniper et al36 proposed that a score of –1, 0, or 1 on the GRC questionnaire indicates that there is no real change in a person's condition. They also contended that people who score ±2 to 3 on the GRC questionnaire experience minimal change, those who score ±4 to 5 experience moderate changes, and those who score ±6 to 7 experience large changes in their condition. Positive scores indicate an improvement in the person's condition, whereas a negative score indicates that the person's condition has worsened.
The mean GRC questionnaire score of the successful group in our study was 3.6, which suggested to us that those participants experienced minimum to moderate improvement in their overall condition. In contrast, there was no apparent global self-perceived change in the participants in the unsuccessful group (mean score=1.3). There was a difference between the successful and unsuccessful groups (P<.0001).
Research has shown that the use of foot orthoses can be an effective intervention for people with PFPS.18,20,56 After a retrospective study of clinical records, Clement56 reported that most patients with PFPS who were treated with an off-the-shelf or customized orthosis responded favorably within 2 to 6 weeks and then were able to resume running without further injury. Again, Eng and Pierrynowski18 reported a 37.5% improvement in VAS scores for subjects who used a customized insole combined with an exercise program over the course of 8 weeks. In contrast, the successful group in our study showed a mean improvement in VAS scores of 72.2% (SD=14.5%) after 3 weeks of orthosis use and activity modification. Given that the mean symptom duration for participants in our study was more than 3 years, we are encouraged by these results, and believe that our data indicate the need for a randomized clinical trial comparing the interventions we studied with other commonly used approaches.
Limitations
Interrater reliability.
In addition to identifying the characteristics of those participants who responded best to off-the-shelf orthoses and modified activity, we also examined the interrater reliability of the measurements obtained in our study. No data were previously available on the quality of these measurements in a sample of patients with PFPS. The generally low interrater reliability values for our measurements pose a threat to the internal validity of our investigation and may limit the interpretation and application of the clinical prediction rule.
Our ICC of .51 for the navicular drop measurement is similar to the interrater reliability of .57 for the same measurement reported by Picciano et al,32 but compares less favorably with the value of .73 reported by Sell and colleagues.57 Additionally, our ICC of .25 for the measurement of forefoot–to–rear-foot alignment is much lower than the values of .58 to .92 reported in previous studies.46,58 The higher reliability values reported in the earlier studies46,58 may be due to the fact that each examiner performed multiple measurements. That is, taking a mean value of multiple measurements may provide a more reliable measurement than a single value.35,38 We chose to obtain each measurement just once because we believe this is most representative of what is done in clinical practice. We were unable to find articles describing investigations of interrater reliability for measurements of great toe extension.
We believe that a more clinically meaningful way to examine the reliability data for our study is to interpret the data within the context of its intended use.59 We believe it is more useful to determine how often the raters agreed for each predictor with regard to the cutoff score. For instance, how often did the raters agree that a participant had a navicular drop test score greater than or less than the cutoff score of 3 mm? The percentages of agreement between the raters with regard to the cutoff score were 75% for great toe extension, 73% for forefoot alignment, and 70% for the navicular drop test. Cohen kappa coefficients () were calculated to determine the chance-corrected agreement between raters for each predictor: great toe extension, =.17; forefoot alignment, =.04; and navicular drop test, =.43.
Lack of independence of data.
A second limitation of our study was the fact that a majority of our data were taken from people with bilateral knee pain. Because prescriptive validity coefficients were computed using 78 knees from 45 subjects, our results were potentially biased due to a lack of independence of data. In order to examine this issue, we compared VAS pain scores between subjects with unilateral PFPS and those with bilateral pain. Descriptive statistics were calculated for the VAS change scores of the 12 subjects with unilateral knee pain and compared with the VAS change scores of 12 randomly selected subjects with bilateral pain. The mean VAS change score in the subjects with unilateral pain was 52.6% (SD=58.9%), and the mean VAS change score in the subjects with bilateral pain was 52.0% (SD=36.2%). The correlation between the 2 groups was r=.65 (P<.05).38 These results suggest to us that the subjects with bilateral knee pain responded similarly to those with unilateral knee pain. The internal validity of data from future studies of PFPS would be strengthened by including only subjects with unilateral knee pain.
Future Research
Clinical prediction rules are clinical tools that quantify the contribution that elements of the history and physical examination make toward a diagnosis, prognosis, or likely response to intervention.23 A 3-step process for the development and testing of a clinical prediction rule has been recommended.23 In our line of inquiry, the first step was to develop a clinical prediction rule, which could be used to identify individuals with PFPS who respond favorably to an initial episode of care using an off-the-shelf orthosis and modified activity. In our study, we identified variables that could be used in a clinical prediction rule for identifying people with PFPS who are likely to respond favorably to an orthosis and instruction in activity modification. The second step is validation. Validation of the clinical prediction rule should be a goal in a future RCT in which the effectiveness of off-the-shelf orthoses can be compared with that of customized orthoses (or a placebo) in individuals with PFPS who have been identified using the predictors we found. The use of these predictors as inclusion criteria could strengthen the power of the RCT by identifying those patients who are most likely to benefit from the use of orthoses. The third step is to assess the impact of the rule on clinical success. Ultimately, any clinical prediction rule must be shown to improve outcomes and clinical decision making before it can be advocated for widespread use.23,60
| Conclusion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionAppendix 1Appendix 2.References |
|---|
2° of valgus), limited passive extension of the first MTP joint (78°), or minimal motion on the navicular drop test (3 mm) will respond more favorably to intervention with an unmodified, off-the-shelf orthotic insert and modified activity. Based on the results of this exploratory study, we suggest that clinicians consider prescribing off-the-shelf foot orthoses combined with a modified physical training regimen as an intervention strategy for patients with long-standing PFPS who have one or more of the characteristics identified in this study. However, validation of the proposed clinical prediction rule is still needed and should be the goal of a future RCT in which the effectiveness of off-the-shelf foot orthoses is compared with that of customized orthoses (or a placebo) in patients who have been identified using the predictors we found. | Appendix 1 |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionAppendix 1Appendix 2.References |
|---|
|
| Appendix 2. |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionAppendix 1Appendix 2.References |
|---|
|
| Footnotes |
|---|
This study was approved by the Institutional Review Board, Department of Clinical Investigation, Brooke Army Medical Center, Fort Sam Houston, Tex.
Opinions or assertions herein are the private views of the authors and are not to be construed as official or as reflecting the views of the US Army or the Department of Defense.
* Wrymark Inc, 11833 Westline Industrial Dr, St Louis, MO 63146. 
SPSS Inc, 233 S Wacker Dr, Chicago, IL 60606.
| References |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionAppendix 1Appendix 2.References |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  C N. van Dijk and W. M van der Tempel Patellofemoral pain syndrome BMJ, October 24, 2008; 337(oct24_1): a1948 - a1948. [Full Text]
|
|
|
|
 |
|
 L. L Currier, P. J Froehlich, S. D Carow, R. K McAndrew, A. V Cliborne, R. E Boyles, L. T Mansfield, and R. S Wainner Development of a Clinical Prediction Rule to Identify Patients With Knee Pain and Clinical Evidence of Knee Osteoarthritis Who Demonstrate a Favorable Short-Term Response to Hip Mobilization Physical Therapy, September 1, 2007; 87(9): 1106 - 1119. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Evidence based journal watch Br. J. Sports Med., June 1, 2004; 38(3): 369 - 371. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
