Abstract
Purpose:
The Knee injury and Osteoarthritis Outcome Survey (KOOS) profile of outcome measures are among the most commonly used outcome measures in knee arthroplasty (KA). The purpose was to develop and externally validate “score maps” (one-page figural depictions of most likely scores) for KOOS Pain and Function subscales to facilitate a variety of clinical decisions related to shared decision making prior to KA.
Methods:
Presurgical KA data collected within one year of surgery and obtained in two independent studies were used in this cross sectional study. Score maps were designed to be easily understandable, single-page graphical depictions of predicted KOOS Pain, and KOOS Function, daily activity subscales. To create the score maps, individual item scores from one dataset were used to determine the most probable responses for each item for the entire range of possible scores. Predicted KOOS score maps were derived from Osteoarthritis Initiative (OAI) data and externally validated using an independent single site KA cohort study. Score map predicted scores from OAI were compared to actual presurgical KOOS subscale scores using Weighted Kappa (Kw) agreement coefficients and actual versus predicted differences in scores.
Results:
The score maps derived from OAI and applied to actual scores in the validation sample demonstrated moderate to substantial chance-corrected agreement for both KOOS Pain and KOOS Function, daily activity subscale items. For example, KOOS Pain score map scores applied to the external validation dataset showed chance-corrected agreement with Kw ranging from 0.43 to 0.73. Score maps predicted actual item scores within ± 1 point at least 94% of the time. Findings for the KOOS Function, daily activity subscale items were similar.
Conclusions:
Score maps derived from OAI data agreed with actual KOOS scores obtained on an independent dataset at an acceptable degree of precision. Easy-to-use KOOS Pain and Function, daily activity score maps have potential to facilitate a variety of important clinical decisions during discussions between patients and surgeons prior to KA.
Keywords: knee, arthroplasty, clinical decision making, outcome, pain, function, PROMS, patient related outcome scoring
INTRODUCTION
Patient-reported outcome measures (PROMs) are the cornerstone of patient-centered data collection in knee arthroplasty (KA) and provide a quantifiable way of measuring patients’ functional status and knee pain during daily activity. There is strong emphasis to routinely use these measures but the challenge has always been to efficiently determine the meaning and value of scores obtained from PROMs.
PROM data has been shown to be used in less than 1% of clinical encounters between orthopaedic surgeons and patients considering arthroplasty [32]. Studies examining PROM usage by patient and clinician stakeholders have identified several deterrents to routine uptake including: a) limited understanding of how to incorporate PROMs into daily practice [1, 15, 31], b) limited time during patient encounters [1, 15, 31], c) perceived lack of actionable data [15, 31], d) implementation challenges [15, 31], e) no clear graphical depiction of scores[31], f) limited understanding of value by patients [31]. In summary, these studies suggest that while PROMs are considered very important, they create challenges in identifying clear, clinically impactful information in a time-efficient manner. Clinical uptake and clinical utility of PROMs could be enhanced if a simple graphical method (i.e., score maps) could be developed that clearly conveyed PROM score meaning that was understood by both patients and clinicians.
Score maps could assist patients and surgeons in quantifying expectations and determining when expectations are unrealistic. Shared decision making discussions could be enhanced if patients could better conceptualize typical outcome scores at baseline and follow-up and score maps may facilitate this process. Score maps could essentially provide an easy-to-use graphical display of complex latent variables represented by Knee Injury and Osteoarthritis Score (KOOS) subscales. One of the most commonly used PROMs worldwide for persons undergoing KA is the (KOOS) [25], The scale has strong psychometric support, captures a variety of constructs and is freely available [5, 15].
Rothrock and colleagues recently described a method for reporting and interpreting PROM scores from the PROMIS (Patient-Reported Outcomes Measurement Information System) family of measures in an easy-to-use “score map” graphical format [26]. Prior reports of similarly designed score maps for other conditions also have been reported [10, 29, 30] but none were found for patients considering KA. The purpose of this study was to apply similar methods to those described by Rothrock and colleagues [26] and others [10, 29, 30] to develop two KOOS subscale score maps (i.e., KOOS Pain, and KOOS Function, daily activities subscales) for use during patient/clinician discussions regarding KA candidacy. It was hypothesized that predicted scores derived from KOOS score maps obtained on one dataset would demonstrate moderate to substantial chance-corrected agreement [12] with actual preoperative KOOS scores obtained on an independent dataset [24, 25]. To provide context to score map interpretation, a second purpose was to derive evidence-based average KOOS Pain and KOOS Function, daily activity subscale scores from a recently published systematic review [27] for the measures of interest and obtained prior to and one-year following KA surgery.
METHODS
Score Map Derivation Dataset
Preoperative data from all participants in the Osteoarthritis Initiative (OAI) who had a study visit within one year of KA were used. The OAI is a National Institutes of Health (NIH) and privately funded prospective longitudinal cohort study with mostly yearly visits and a 9-year follow-up [13]. A primary objective of the OAI study was to develop diverse cohorts of persons for the study of the natural history, risk factors, onset and progression of knee OA. The four centers in OAI required all participants to read and signed IRB approved consent forms from the participating sites prior to participation. The OAI, a publicly available dataset, was chosen to provide a broad spectrum of OA symptom severity from mild to severe leading up to KA. Persons between the ages of 45 and 79 years with or at high risk for developing knee OA were recruited from communities at the following sites: 1) the University of Maryland School of Medicine in Baltimore, Maryland, 2) the Ohio State University in Columbus, Ohio, 3) the University of Pittsburgh in Pittsburgh, Pennsylvania, and 4) Memorial Hospital of Rhode Island, in Pawtucket, Rhode Island. Participants provided Institutional Review Board informed consent forms prior to study participation.
Participants recruited for the OAI study had to have KA for osteoarthritis after baseline data collection and before the final follow-up data collection. All participants completed KOOS Pain and Function, daily activity subscales at each yearly visit. The OAI is a time-varying study meaning that the amount of time between a baseline visit and KA surgery varied for each participant. Participants in the current study were seen by OAI investigators between two days and 365 days prior to KA. Systematic review data indicate that pain scores for patients awaiting arthroplasty remain unchanged if measured within a year of surgery [17]. A total of 407 participants from the OAI had at least one KA after recruitment and at least one completed KOOS subscale score at the visit prior to KA. More detail on study design and data collection is available on the OAI website (https://nda.nih.gov/oai).
Score Map Validation Dataset
Data from a single site prospective longitudinal cohort study were used to assess the validity of score maps derived from the OAI dataset. Inclusion criteria included men and women aged 45–90 years with a diagnosis of advanced symptomatic knee osteoarthritis (OA) and who had primary KA. Exclusion criteria were KA revision, simultaneous bilateral KA, unicondylar KA, or KA for inflammatory arthropathy or cancer. All participants meeting these criteria were recruited at the time of their preoperative KA educational class. After providing written informed consent, approved by the Institutional Review Board at Virginia Commonwealth University and the participating hospital, participants completed a set of baseline questionnaires including KOOS Pain and KOOS Function, daily activity subscales.
Outcome Measures
Knee Injury and Osteoarthritis Outcome Score (KOOS) [5, 24] subscales (i.e., KOOS Pain, KOOS Function, daily activity) obtained at the pre-operative visit were the primary outcome PROMs of interest. All KOOS subscale scores range from 0 to 100 with higher scores equating to better scores. Raw scores for all KOOS items range from 0 (none – a better score) to 4 (extreme – a worse score). A substantial literature supports the reliability, validity and unidimensionality of KOOS subscales [4–6, 11].
Participants in OAI were allowed to indicate if a KOOS item was not applicable or they could refuse to answer an item. As a result, approximately 4% of items were left blank. KOOS subscale scoring followed recommendations by the developer (http://www.koos.nu/).
KOOS Score maps are one-page figures (either KOOS Pain OR KOOS Function) that display the most likely responses for each item in the scale. The figures allow for a quick and simple determination of the meaning of PROM scores obtained from patients at a pre-surgery examination.
Statistical Analysis
Score maps were developed using OAI data. Preoperative KOOS Pain, and KOOS Function, daily activity subscale total and individual item scores were totaled and the proportion for each item response relative to the total for all options was calculated. Item responses for each subscale item with the highest proportion for a given score was chosen to generate the score map. The KOOS Pain score map is illustrated in Figure 1 and the KOOS Function, daily activity score map appears in Figure 2 [26]. For example, for the 24 participants’ scores between 51 and 55 on KOOS Pain in OAI, 16 participants scored “severe (3)” and 8 participants scored “extreme (4)”. In this case, the score map would include a score of 3 for this item because a score of 3 was the most likely score for that item. The score map was then used to determine predicted item scores by totaling the most probable score for each KOOS sub-score item. Score maps are reported in 5-point increments. Score map scores were collapsed into 5 point increments to make the score map easier to apply clinically (versus 101 separate scores). For both sub-scores, very severe (e.g. KOOS Pain ≤ 30 and KOOS, Function, daily activity ≤ 35 and very mild scores (e.g., KOOS Pain ≥ 76 and KOOS Function, daily activity ≥ 80) were collapsed because these scores were relatively uncommon in OAI and the validation dataset. Observed KOOS Pain scores of 30 or less, for example, occurred for 29 participants in OAI.
Figure 1.
Score map for the KOOS Pain Scale with KOOS Pain scores on the upper x-axis and individual item scores for each item in the horizontal bars. Vertical lines represent evidence-based weighted evidence-based averages of pre-operative and one-year postoperative scores.
Figure 2.
Score map for the KOOS Function, daily living scale with scores on the upper x-axis and individual item scores for each item in the horizontal bars. Vertical lines represent evidence-based weighted averages of pre-operative and one-year postoperative scores.
To determine the validity of OAI-derived predicted scores reflected in the score maps, each interval on the score map was compared to actual scores obtained from the validation dataset using weighted Kappa. The extent of chance-corrected agreement between actual and score-map predicted scores was determined using weighted Kappa (Kw) [12, 28] with quadratic weighting for KOOS Pain and KOOS Function, daily activity individual items. Quadratic weighting places greater consequences on more serious disagreements such as when larger differences between actual and predicted item scores occur. Kw with 95% confidence intervals were reported for each item comprising each KOOS subscale. The scale recommended by Landis and Koch for interpreting the magnitude of Kw was used (slight agreement = Kw of 0 to 0.2) to almost perfect agreement = Kw of 0.81 to 1.0) [12]. Moderate (Kw = 0.4 to 0.6) or higher chance-corrected agreement was used to judge acceptability of the coefficients. Additionally, the difference between actual score and predicted score for each item were calculated. For example, the predicted score for the “Pain with knee straightening” KOOS Pain item from the score map is “moderate” when the predicted score is 36–40. If the actual KOOS Pain score for a participant in the validation dataset was 38 and “moderate” was chosen for this item, the difference score for this participant would be 0. SPSS, version 28.0.1 was used for all analyses.
Because no evidence was found for estimating sample size using weighted Kappa, sample size estimation was determined for Kappa [3]. For a 5×5 kappa table (scores of 0 to 4 on individual KOOS items), a sample size of 89 participants would provide 90% power and an alpha of 0.05 to distinguish between a Kappa of 0.7 in one dataset and a Kappa of 0.5 in the other dataset. Sample sizes were 407 in the derivation dataset and 121 in the validation dataset.
To facilitate score map application for clinical practice, published weighted average estimates of preoperative and one year postoperative KOOS Pain and KOOS Function, daily activity scores were derived from a recently published systematic review of KA cohort studies. [27]. The purpose was to obtain evidence-based estimates of weighted average (based on study sample size) KOOS subscale scores across multiple studies to use as anchors to facilitate patient/surgeon discussions of the score maps. Pooled KOOS Function, daily activity scores, which are equivalent to WOMAC Function items, were estimated to be 42 preoperatively and 78 at one-year following surgery, based on 20 studies of 4,393 participants. Pooled KOOS Pain scores were estimated to be 45 at baseline and 80, one-year post-surgery based on 6 studies of 930 participants. These estimates were used to inform score map interpretation.
RESULTS
Sample characteristics from OAI and the validation dataset appear in Table 1. Data comparing the predicted scores for the OAI dataset (from the score map) and the actual scores for KOOS Pain, KOOS Function, daily activity subscales from the validation dataset appear in Tables 2 and 3, respectively. Comparisons between OAI predicted and OAI actual scores for KOOS Pain and KOOS Function, daily activity subscales for the OAI dataset appear in Additional File 1.
Table 1.
Characteristics of the samples with a preoperative visit <= 12 months prior to surgery
| Variable | Derivation Dataset OAI1 (n = 407) | Validation Dataset (n = 121) |
|---|---|---|
|
| ||
| Age in years, mean (SD) | 68.3 (8.5) | 67.7 (7.7) |
| Female, n (%) | 247 (60.7) | 64 (52.9) |
| Body mass index, mean (SD) | 30.2 (5.0) | Not collected |
| KOOS2 Pain, mean (SD) | 56.4 (18.2) | 44.2 (16.9) |
| KOOS Function, daily living, mean (SD) | 63.7 (17.8) | 46.0 (19.1) |
OAI = Osteoarthritis Initiative Study
KOOS = Knee injury and Osteoarthritis Outcome Score
Table 2.
KOOS Pain items (observed – predicted) for validation dataset and Weighted Kappa coefficients (Kw) with 95% Confidence Intervals between observed and predicted scores.
| Response Differences |
||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
|
| ||||||||||
| Observed – predicted response | Pain frequency Kw = 0.43 (0.20, 0.66) |
Pain with knee twisting Kw = 0.44 (0.24, 0.65) |
Pain with knee straightening Kw = 0.62 (0.41, 0.84) |
Pain with knee bending Kw = 0.62 (0.36, 0.88) |
Knee pain with walking Kw = 0.62 (0.36, 0.88) |
|||||
| N | % | N | % | N | % | N | % | N | % | |
| −4 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| −3 | 1 | 0.8 | 0 | 0 | 0 | 0 | 0 | 0 | 2 | 0.5 |
| −2 | 0 | 0 | 1 | 0.8 | 0 | 0 | 3 | 2.5 | 8 | 2.0 |
| −1 | 35 | 28.9 | 36 | 29.8 | 21 | 17.4 | 22 | 18.2 | 71 | 18.0 |
| 0 | 79 | 65.3 | 63 | 52.1 | 61 | 50.4 | 67 | 55.4 | 240 | 60.8 |
| +1 | 6 | 5.0 | 18 | 14.9 | 33 | 27.3 | 27 | 22.3 | 70 | 17.7 |
| +2 | 0 | 0.3 | 3 | 2.5 | 5 | 4.1 | 2 | 1.7 | 4 | 1.0 |
| +3 | 0 | 0 | 0 | 0 | 1 | 0.8 | 0 | 0 | 0 | 0 |
| +4 | 0 | 0 | 0 | 0 | 0 | 0.3 | 0 | 0 | 0 | 0 |
| Total sample | 121 | 121 | 121 | 121 | 121 | |||||
| Knee pain with stairs Kw = 0.48 (0.25, 0.70) |
Knee pain at night in bed Kw = 0.73 (0.58, 0.89) |
Knee pain sitting or lying Kw = 0.64 (0.39, 0.88) |
Knee pain standing Kw = 0.69 (0.51, 0.86) |
|||||||
| N | % | N | % | N | % | N | % | |||
| −4 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | ||
| −3 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | ||
| −2 | 0 | 0 | 2 | 1.7 | 3 | 2.5 | 2 | 1.7 | ||
| −1 | 16 | 13.3 | 23 | 19.0 | 18 | 15.0 | 27 | 23.1 | ||
| 0 | 65 | 54.2 | 70 | 57.9 | 69 | 57.5 | 60 | 51.3 | ||
| +1 | 37 | 30.8 | 23 | 19.0 | 27 | 22.5 | 28 | 23.9 | ||
| +2 | 2 | 1.7 | 3 | 2.5 | 3 | 2.5 | 0 | 0 | ||
| +3 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | ||
| +4 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | ||
| Total Sample | 120 | 121 | 120 | 117 | ||||||
Table 3.
KOOS Function, daily living items (observed – predicted) and Weighted Kappa coefficients (Kw) with 95% Confidence Intervals between observed and predicted scores.
| Response Differences |
||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
| ||||||||||||||
| Observed – predicted response | Descending stairs Kw = 0.54 (0.35, 0.73) |
Ascending Stairs Kw = 0.59 (0.32, 0.87) |
Rising from sitting Kw = 0.63 (0.31, 0.95) |
Standing Kw = 0.51 (0.26, 0.77) |
Picking up object off floor Kw = 0.66 (0.50, 0.82) |
Walking even surfaces Kw = 0.72 (0.50, 0.93) |
Getting in/out of car Kw = 0.68 (0.41, 0.94) |
|||||||
| N | % | N | % | N | % | N | % | N | % | N | % | N | % | |
| −3 | 1 | 0.9 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.8 | 0 | 0 | 0 | 0 |
| −2 | 0 | 0 | 0 | 0 | 1 | 0.8 | 0 | 0 | 5 | 4.2 | 2 | 1.7 | 0 | 4.5 |
| −1 | 23 | 19.7 | 12 | 10.3 | 21 | 17.9 | 26 | 22.2 | 43 | 36.4 | 34 | 28.8 | 24 | 23.0 |
| 0 | 60 | 51.3 | 60 | 51.3 | 70 | 59.8 | 59 | 50.4 | 54 | 45.8 | 69 | 58.5 | 71 | 48.3 |
| +1 | 33 | 28.2 | 41 | 35.0 | 23 | 19.7 | 22 | 18.8 | 14 | 11.9 | 13 | 11.0 | 54 | 20.1 |
| +2 | 0 | 0 | 4 | 3.4 | 2 | 1.7 | 10 | 8.5 | 1 | 0.8 | 0 | 0 | 10 | 3.7 |
| +3 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.4 |
| Total Sample | 117 | 117 | 117 | 117 | 118 | 118 | 118 | |||||||
| Going shopping Kw = 0.75 (0.56, 0.94) |
Putting on socks/ Stockings Kw = 0.71 (0.53, 0.89) |
Rising from bed Kw = 0.70 (0.51, 0.89) |
Taking off socks/ Stockings Kw = 0.76 (0.62, 0.91) |
Lying in bed Kw = 0.48 (0.30, 0.66) |
Getting in/out of bath Kw = 0.64 (0.43, 0.85) |
Sitting Kw = 0.59 (0.37, 0.81) |
||||||||
| N | % | N | % | N | % | N | % | N | % | N | % | N | % | |
| −3 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.9 | 0 | 0 |
| −2 | 1 | 0.8 | 2 | 1.7 | 1 | 0.8 | 1 | 0.9 | 1 | 0.8 | 1 | 0.9 | 1 | 0.9 |
| −1 | 12 | 10.2 | 23 | 19.7 | 34 | 28.8 | 21 | 17.9 | 13 | 11.0 | 30 | 26.5 | 17 | 14.7 |
| 0 | 75 | 63.6 | 66 | 56.4 | 63 | 53.4 | 73 | 62.4 | 51 | 43.2 | 54 | 47.8 | 54 | 46.6 |
| +1 | 30 | 25.4 | 26 | 22.2 | 19 | 16.1 | 22 | 18.8 | 37 | 31.4 | 25 | 22.1 | 39 | 33.6 |
| +2 | 0 | 0 | 0 | 0 | 1 | 0.8 | 0 | 0 | 16 | 13.6 | 1 | 0.9 | 5 | 4.3 |
| +3 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0.9 | 0 | 0 |
| Total Sample | 118 | 117 | 118 | 117 | 118 | 113 | 116 | |||||||
| Getting on/off toilet Kw = 0.72 (0.47, 0.98) |
Heavy domestic duties Kw = 0.49 (0.30, 0.68) |
Light domestic duties Kw = 0.67 (0.42, 0.93) |
||||||||||||
| N | % | N | % | N | % | |||||||||
| −3 | 0 | 0 | 0 | 0 | 0 | 0 | ||||||||
| −2 | 0 | 0 | 2 | 1.8 | 0 | 0 | ||||||||
| −1 | 26 | 22.4 | 6 | 5.4 | 25 | 21.7 | ||||||||
| 0 | 74 | 63.8 | 42 | 37.8 | 66 | 57.4 | ||||||||
| +1 | 16 | 13.8 | 57 | 51.4 | 24 | 20.9 | ||||||||
| +2 | 0 | 0 | 3 | 2.7 | 0 | 0 | ||||||||
| +3 | 0 | 0 | 1 | 0.9 | 0 | 0 | ||||||||
| Total Sample | 116 | 111 | 115 | |||||||||||
The KOOS Pain score map appears in Figure 1 while the KOOS Function, daily activity score map appears in Figure 2. The OAI derived score maps perfectly agreed with the actual scores from the validation dataset between 50.4% and 65.3% of the time for KOOS Pain items and between 37.8% and 63.8% of the time for the KOOS Function, daily activity subscale items. The score map predicted the actual score within ± one point 95% of the time, on average, for KOOS Pain and 94.8% for KOOS Function, daily activity. The Kw for KOOS Pain items ranged from = 0.43 (95%CI = (0.20, 0.66)) to 0.69 (95%CI = 0.51, 0.86) and for KOOS Function, daily activity, the Kw ranged from 0.48 (95%CI = 0.30, 0.66) to 0.76 (95%CI = 0.62, 0.91). The extent of perfect agreement and Kw for the OAI dataset were similar (see Additional File 1)
DISCUSSION
The most important clinical finding of this study, as hypothesized, was that KOOS Pain and Function, daily activity score maps predicted actual KOOS scores as demonstrated by moderate to substantial agreement across all individual items. This is the first study to demonstrate potential value and accuracy of a two simple and time-efficient KOOS score maps, applicable to several clinical practice applications in KA. The score maps in this study can either be applied separately, or as a more comprehensive set of key patient-important latent constructs. Clinicians using either the KOOS Pain or KOOS Function, daily activity measures could potentially benefit from incorporation of these score maps into daily practice decisions when examining patients considering KA. Pain and functional loss are the two most common complaints of persons seeking KA[8, 18].
Rothrock and colleagues [26] built on the work of others [10, 29, 30] and proposed this method of graphically mapping PROM scores (i.e., a score map) to facilitate uptake of PROMIS measures [26]. The investigators found strong associations between score maps (predicted) and actual scores on a validation dataset with 69.5% to 85.3% perfect agreement. Perfect agreement estimates were higher than those found in the current study. PROMIS measures and subsequent score maps were developed using state-of-the-art item response theory methods [19] and most likely response from probability curves for each item [26] while the KOOS family of measures were developed with more traditional methods though KOOS subscales have shown unidimensional characteristics [2]. The somewhat lower associations between actual and predicted scores in the current study compared to the work of Rothrock et al is likely attributable to a less rigorous psychometric method of scale development compared to the PROMIS family of measures.
Despite greater variation and lower perfect agreement between predicted and actual scores for KOOS compared to PROMIS, KOOS subscale score maps are likely still accurate enough (with approximately 95% accuracy within ±1 item point) to provide time-efficient and clinically useful information when discussing KA candidacy with patients. Discussions with patients that could benefit by incorporation of score maps include: interpreting the meaning of preoperative KOOS scores, assessment of patient expectations of outcome following future KA recovery, and during shared decision making discussions related to KA candidacy.
Clinical Applicability and Clinical Relevance
Patients undergoing evaluation for KA candidacy commonly complete a PROM, potentially the KOOS, prior to their visit. Once the score is obtained, a score map could be used during the discussion to illustrate for the patient how their score compares to the average patient who actually undergoes a KA. Using the score map for the KOOS Pain scale for example, if a hypothetical patient scored a 45, the patient could see that this score indicates that knee pain is severe during stairclimbing and knee twisting, moderate at night and with standing and knee straightening. The average patient who underwent KA scored very similarly prior to surgery. From the standpoint of knee pain with daily activity, this patient would appear to be a good candidate for KA. Patients who fall substantially above or below the average score may be at risk for minimal benefit [7, 21, 22]. For this hypothetical patient, the score map could facilitate clear discussions of KOOS Pain score meaning and how the score relates to other patients actually undergoing KA. This information could be incorporated with medical/social history and examination data gathered during the work-up to enhance KA candidacy decisions.
External validity of the findings is likely to be strong because OAI was a multicenter study and the external validation dataset was unrelated to OAI. It is possible that the findings are not generalizable to non-US sites but no evidence was found to support this assumption. Given that this study included multiple sites and an external validation dataset, the study suggests that surgeons should strongly consider use of score maps to enhance discussions with patients prior to arthroplasty surgical decisions to optimize patient selection for surgery. Optimization includes but is not limited to judgements of preoperative expectations and shared decision making discussions.
Judging Pre-operative Expectations
Unrealistic expectations are well-established risk factors for unsatisfactory outcomes following hip or knee arthroplasty [9, 33]. KOOS score maps could assist in determinations of patient expectations prior to surgery. Using the KOOS Pain score map and a hypothetical patient as an example, assume the patient scored a 30 on KOOS Pain and the patient’s one-year expectation was to be completely pain free. The score map could illustrate for the patient that average preoperative scores for persons prior to KA is 45, substantially less pain with activity as compared to our hypothetical patient’s score. The average one-year KOOS Pain outcome is 80 and is equivalent to having daily knee pain, moderate pain with stairclimbing and mild pain with walking. A powerful predictor of pain outcome is pre-operative pain severity with worse baseline pain associating with worse final outcome pain [14, 16]. Our hypothetical patient expects no knee pain at one year following KA, which occurs in approximately a third of patients [20]. Given the low (poor) preoperative KOOS Pain score, being pain free a year after surgery is unlikely for our hypothetical patient. The score map would likely assist in helping the patient to understand that their expectations are likely not realistic particularly given their low (poor) pre-operative KOOS Pain score.
Shared Decision Making Discussions and Knee Arthroplasty Candidacy
KOOS score maps could facilitate a shared decision making discussion [23]. For example, after our hypothetical patient completes the KOOS subscales, the clinician and patient could discuss score meaning, how it compares to average scores of others receiving KA [27] and the patient’s likely KOOS score outcome one year following surgery. Score maps can, in essence, convey meaning of these unitless scores in a way that may facilitate understanding. This discussion, part of a larger discussion of risks and benefits, could facilitate a discussion about the potential benefits with KA. For example, if our hypothetical patient scored a 40 on the KOOS Function, daily activity scale, this could be compared to a score of approximately 78, the average KOOS Function, daily activity score one year following surgery and the surgeon and patient could discuss this benefit relative to other benefits, risks and costs, as part of a shared decision making discussion. The score map would likely clarify and streamline discussions of magnitude of likely daily activity benefits related to surgery.
Limitations
The study has several strengths including external validation on an independent dataset but also some important limitations, most importantly a lack of evidence demonstrating score maps actually enhance patient/clinician interactions. Additionally, a Kw cutpoint of 0.4 or higher was used to indicate acceptable reliability and it may be that a 0.4 cutpoint is too liberal. Missing data was minimal with less than 1% missing data for all KOOS subscale scores in the validation dataset and less than 3% for KOOS subscale items. The OAI collected time-varying preoperative data and this variation may have influenced KOOS scores.
CONCLUSION
In conclusion, KOOS score maps appear to have potential for informing a variety of important KA candidacy and shared decision making discussions between patients and surgeons. Given that PROM scores show little uptake in routine clinical practice, these findings support the incorporation of KOOS score maps into daily practice. Future work should examine the value of score maps and their role in potentially enhancing shared decision making as applied to KA.
Supplementary Material
Acknowledgement:
The author wishes to thank Nancy Henderson for providing the validation dataset.
Funding:
The Osteoarthritis Initiative (OAI) is a public-private partnership comprising 5 contracts (N01-AR-2-2258, N01-AR-2-2259, N01-AR-2-2260, N01-AR-2-2261, N01-AR-2-2262) funded by the National Institutes of Health (NIH), a branch of the Department of Health and Human Services, and conducted by the OAI study investigators. Private funding partners include Merck Research Laboratories, Novartis Pharmaceuticals Corporation, GlaxoSmithKline, and Pfizer Inc. Private sector funding for the OAI is managed by the Foundation for the National Institutes of Health. This manuscript was prepared using an OAI public use dataset and does not necessarily reflect the opinions or views of the OAI investigators, the NIH, or the private funding partners.
List of Abbreviations:
- KA
Knee arthroplasty
- KOOS
Knee injury and Osteoarthritis Outcome Survey
- Kw
Weighted Kappa
- NIH
National Institutes of Health
- OA
Osteoarthritis
- OAI
Osteoarthritis Initiative
- PROMIS
Patient-Reported Outcomes Measurement Information System
- PROMs
Patient reported outcome measures
- SPSS
Statistical Package for the Social Sciences
Footnotes
The author reports no financial conflicts of interest.
Level of evidence: Level III prognostic study
Ethical approval: University IRB approval was obtained from all sites in both studies and all participants provided informed consent.
REFERENCES
- 1.Alshehri F, Alarabi A, Alharthi M, Alanazi T, Alohali A, Alsaleem M (2020) Use of patient-reported outcome measures (PROMs) by orthopedic surgeons in Saudi Arabia. J Orthop Surg Res 15:598. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Bellamy N, Buchanan WW (1986) A preliminary evaluation of the dimensionality and clinical importance of pain and disability in osteoarthritis of the hip and knee. Clin Rheumatol 5:231–241 [DOI] [PubMed] [Google Scholar]
- 3.Bujang MA, Baharum N (2017) Guidelines of the minimum sample size requirements for Cohen’s Kappa. Epidemiol Biostat Public Health 14: e12267–1 - 10 [Google Scholar]
- 4.Collins NJ, Misra D, Felson DT, Crossley KM, Roos EM (2011) Measures of knee function: International Knee Documentation Committee (IKDC) Subjective Knee Evaluation Form, Knee Injury and Osteoarthritis Outcome Score (KOOS), Knee Injury and Osteoarthritis Outcome Score Physical Function Short Form (KOOS-PS), Knee Outcome Survey Activities of Daily Living Scale (KOS-ADL) Lysholm Knee Scoring Scale, Oxford Knee Score (OKS) Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC, Activity Rating Scale (ARS) and Tegner Activity Score (TAS). Arthritis Care Res(Hoboken) 63 Suppl 1:S208–S228 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Collins NJ, Prinsen CA, Christensen R, Bartels EM, Terwee CB, Roos EM (2016) Knee Injury and Osteoarthritis Outcome Score (KOOS): systematic review and meta-analysis of measurement properties. Osteoarthritis Cartilage 24:1317–29 [DOI] [PubMed] [Google Scholar]
- 6.Collins NJ, Roos EM (2012) Patient-reported outcomes for total hip and knee arthroplasty: commonly used instruments and attributes of a “good” measure. Clin Geriatr Med 28:367–394 [DOI] [PubMed] [Google Scholar]
- 7.Dumenci L, Perera R, Keefe F, Ang D, Slover J, Jensen M, Riddle D (2019) Model-based pain and function outcome trajectory types for patients undergoing knee arthroplasty: A secondary analysis from a randomized clinical trial. Osteoarthritis Cartilage 27:878–884 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Frankel L, Sanmartin C, Hawker G, De CC, Dunbar M, Bohm E, Noseworthy T (2016) Perspectives of orthopaedic surgeons on patients’ appropriateness for total joint arthroplasty: a qualitative study. J Eval Clin Pract 22:164–170 [DOI] [PubMed] [Google Scholar]
- 9.Hafkamp FJ, Gosens T, de Vries J, den Oudsten BL (2020) Do dissatisfied patients have unrealistic expectations? A systematic review and best-evidence synthesis in knee and hip arthroplasty patients. EFORT Open Rev 5:226–240 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Jette AM, Tao W, Norweg A, Haley S (2007) Interpreting rehabilitation outcome measurements. J Rehabil Med 39:585–590 [DOI] [PubMed] [Google Scholar]
- 11.Juhl C, Lund H, Roos EM, Zhang W, Christensen R (2012) A Hierarchy of Patient-Reported Outcomes for Meta-Analysis of Knee Osteoarthritis Trials: Empirical Evidence from a Survey of High Impact Journals. Arthritis 136245. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Landis JR, Koch GG (1977) The measurement of observer agreement for categorical data. Biometrics 33:159–174 [PubMed] [Google Scholar]
- 13.Lester G (2012) The Osteoarthritis Initiative: A NIH Public-Private Partnership. HSS Journal 8:62–63 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Lingard EA, Katz JN, Wright EA, Sledge CB (2004) Predicting the outcome of total knee arthroplasty. J Bone Joint Surg Am 86-A:2179–2186 [DOI] [PubMed] [Google Scholar]
- 15.Mou D, Sisodia RC, Castillo-Angeles M, Ladin K, Bergmark RW, Pusic AL, del Carmen MG, Heng M (2022) The Surgeon’s Perceived Value of Patient-reported Outcome Measures (PROMs). Ann Surg 275:500–505 [DOI] [PubMed] [Google Scholar]
- 16.Nilsdotter AK, Toksvig-Larsen S, Roos EM (2009) A 5 year prospective study of patient-relevant outcomes after total knee replacement. Osteoarthritis Cartilage 17:601–606 [DOI] [PubMed] [Google Scholar]
- 17.Patten RK, Tacey A, Bourke M, Lane R, Woessner MN, Levinger I (2022) The impact of waiting time for orthopaedic consultation on pain levels in individuals with osteoarthritis: a systematic review and meta-analysis. Osteoarthritis Cartilage In Press [DOI] [PubMed] [Google Scholar]
- 18.Rastogi R, Davis AM, Chesworth BM (2007) A cross-sectional look at patient concerns in the first six weeks following primary total knee arthroplasty. Health Qual Life Outcomes 5:48. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Reeve BB, Hays RD, Bjorner JB, Cook KF, Crane PK, Teresi JA, Thissen D, Revicki DA, Weiss DJ, Hambleton RK, Liu H, Gershon R, Reise SP, Lai JS, Cella D (2007) Psychometric evaluation and calibration of health-related quality of life item banks: Plans for the Patient-Reported Outcomes Measurement Information System (PROMIS). Med Care 45:522–531 [DOI] [PubMed] [Google Scholar]
- 20.Riddle DL (2018) Prevalence and Predictors of Symptom Resolution and Functional Restoration in the Index Knee After Knee Arthroplasty: A Longitudinal Study. Arch Phys Med Rehabil 99(5):887–892 [DOI] [PubMed] [Google Scholar]
- 21.Riddle DL, Macfarlane GJ, Hamilton DF, Beasley M, Dumenci L (2022) Cross-Validation of Good versus Poor Self-reported Outcome Trajectory Types Following Knee Arthroplasty. Osteoarthritis Cartilage 30:61–68 [DOI] [PubMed] [Google Scholar]
- 22.Riddle DL, Perera RA, Jiranek WA, Dumenci L (2015) Using surgical appropriateness criteria to examine outcomes of total knee arthroplasty in a United States sample. Arthritis Care Res (Hoboken) 67:349–357 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Riddle DL, Sando T, Tarver T, Slover J, Sierra RJ, Brito JP, Montori VM (2021) Shared Decision-Making Applied to Knee Arthroplasty: A Systematic Review of Randomized Trials. Arthritis Care Res (Hoboken) 73:1125–1133 [DOI] [PubMed] [Google Scholar]
- 24.Roos EM, Lohmander LS (2003) The Knee injury and Osteoarthritis Outcome Score (KOOS): from joint injury to osteoarthritis. Health Qual Life Outcomes 1:64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Roos EM, Toksvig-Larsen S (2003) Knee injury and Osteoarthritis Outcome Score (KOOS) - validation and comparison to the WOMAC in total knee replacement. Health Qual Life Outcomes 1:17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Rothrock NE, Amtmann D, Cook KF (2020) Development and validation of an interpretive guide for PROMIS scores. J Patient Rep Outcomes 4:16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Sayah SM, Karunaratne S, Beckenkamp PR, Horsley M, Hancock MJ, Hunter DJ, Herbert RD, de Campos TF, Steffens D (2021) Clinical Course of Pain and Function Following Total Knee Arthroplasty: A Systematic Review and Meta-Regression. Journal of Arthroplasty 36:3993–4002 [DOI] [PubMed] [Google Scholar]
- 28.Sim J, Wright CC (2005) The kappa statistic in reliability studies: use, interpretation, and sample size requirements. PhysTher 85:257–268 [PubMed] [Google Scholar]
- 29.Wang YC, Hart DL, Stratford PW, Mioduski JE (2009) Clinical Interpretation of Computerized Adaptive Test-Generated Outcome Measures in Patients With Knee Impairments. Arch Phys Med Rehabil 90:1340–1348 [DOI] [PubMed] [Google Scholar]
- 30.Wang YC, Hart DL, Werneke M, Stratford PW, Mioduski JE (2010) Clinical interpretation of outcome measures generated from a lumbar computerized adaptive test. Phys Ther 90:1323–1335 [DOI] [PubMed] [Google Scholar]
- 31.Whitebird RR, Solberg LI, Ziegenfuss JY, Norton CK, Chrenka EA, Swiontkowski M, Reams M, Grossman ES (2022) What Do Orthopaedists Believe is Needed for Incorporating Patient-reported Outcome Measures into Clinical Care? A Qualitative Study. Clin Orthop Relat Res 480:680–687 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Ziegenfuss JY, Grossman ES, Solberg LI, Chrenka EA, Werner A, Asche SE, Norton CK, Nelson A, Reams M, Whitebird RR (2022) Is the Promise of PROMs Being Realized? Implementation Experience in a Large Orthopedic Practice. American Journal of Medical Quality 37:489–494 [DOI] [PubMed] [Google Scholar]
- 33.Zywiel MG, Mahomed A, Gandhi R, Perruccio A v, Mahomed NN (2013) Measuring expectations in orthopaedic surgery: a systematic review. Clin Orthop Relat Res 471:3446–56 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.