Remember me
ORIGINAL ARTICLE
Year : 2023 | Volume
: 14
| Issue : 2 | Page : 77-81
Development and preliminary evaluation of the CHARUSAT patellofemoral arthritis questionnaire: A psychometric study
Heta Pavan Patel, Balaganapathy Muruganantham
Department of Musculoskeletal Sciences, Ashok and Rita Patel Institute of Physiotherapy, Faculty of Medical Sciences, CHARUSAT, Anand, Gujarat, India
Correspondence Address:
Dr. Heta Pavan Patel
Department of Musculoskeletal Sciences, Ashok and.78 Rita Patel Institute of Physiotherapy, Faculty of Medical Sciences, CHARUSAT, Changa, Anand, Gujarat
India
Source of Support: None, Conflict of Interest: None
DOI: 10.4103/injms.injms_7_23
Background: Knee osteoarthritis is one of the leading causes of the increasing global burden of musculoskeletal disorders. The prevalence rate of patellofemoral arthritis is 39% in South Asia in >40 years of age. The census on early osteoarthritis recommends focusing on the prediagnostic stage. However, none of the tools is available to screen the early symptoms of patellofemoral arthritis. Purpose of the Study: The purpose of this study was to develop and evaluate the psychometric properties of the CHARUSAT Patellofemoral Arthritis Questionnaire. Methods: The questionnaire was developed using a mixed method stepwise approach. The participant with early patellofemoral arthritis (n = 14) was interviewed. We conducted scoping review and in-depth interviews to develop a conceptual framework. Initial content validation was completed through community consultation and expert opinion. Final psychometric validation was conducted on n = 310 individuals with early patellofemoral arthritis. Participants were recruited from tertiary hospitals' outpatient units and regional communities in India. Results: Conceptual framework was developed based on the current literature review, and qualitative data obtained from participants finalized the 4 domains and 12 content versions: crepitus (2 items), stiffness (2 items), pain (2 items), and difficulty in functional activity (6 items). The psychometric measurement demonstrated good internal consistency, Cronbach's alpha α was 0.90 and an acceptable level of inter-item redundancy. The intraclass correlation coefficient (ICC) value for the test and retest assessment was 0.93 (ICC:2,1; 95% confidence interval = 0.880–0.936; P < 0.000). Principal component analysis demonstrates a four-factor solution with 56.83% total variance. Conclusions: It is an acceptable, reliable, and valid self-administered questionnaire that can identify the early symptoms of patellofemoral arthritis. The questionnaire provides an excellent opportunity for self-analysis of the symptoms at the community level.
Keywords: Development, patellofemoral arthritis, psychometric properties, questionnaire
Patellofemoral osteoarthritis is a degenerative joint disease.[1],[2] Identifying early symptom detection is complicated because the symptoms can be manageable with daily routine. Indian population has high tolerance capacity compared to the Western population. They tolerate the initial symptom but when the symptom causes unable to do routine activities, then only they seek expert help. On the other hand, most of the time the solution from the expert opinion is towards arthroplasty which is a costly treatment approach. They try to live with existing symptoms which might be the reason for the increasing burden of arthritis among Indians. Preventive strategies are the best option to improve the quality of life and reduce the burden of such degenerative diseases. Patient-reported outcome measures are the instrument which can predict the symptoms and are easily and freely available self-assessment tools. Knee Injury and Osteoarthritis Outcome Score Patellofemoral (KOOS-PF) is a tool for patellofemoral disorders in general.[3] It deficits in establishing the validity for early cases of patellofemoral arthritis.[4] The prediagnostic stage of arthritis is consequently a 10th global burden in musculoskeletal disorders.[5] The knee joint starts from this sesamoid bone surface and progresses to the large condylar surface.[6] A systematic review concerning the prevalence of patellofemoral arthritis demonstrates a higher rate in Asian countries, 39% in the symptom-based cohorts.[7]
However, knee arthritis is being treated as general osteoarthritis in clinical settings for investigation and conservative treatment. Then, the clinical settings acquire poor outcomes and increase the treatment cost.[8] Here, identification of the early symptoms plays a vital role, and it is also demanded in the recent census.[9],[10] Patellofemoral pain consensus 2016 statements were favored for developing the tool to identify the patient-reported symptoms to compare and pool data for the patellofemoral arthritis population.[9] International patellofemoral osteoarthritis consortium 2018 also asked development of a patient-specific tool and the studies on psychometric properties measurement for that tool.[10] The participants are essential key informants in developing a self-administered tool, sharing their experiences through qualitative methods, and validating the shared content through experts. These are the systemic steps for the development of screening tools.[11]
The primary purpose of the study was to develop a self-administered questionnaire that could be used for self-assessment to identify the early symptoms of patellofemoral arthritis, which could guide the patient on the need for conservative management and consult with orthopedic professionals. A secondary objective was to establish the initial psychometric properties of the questionnaire. We hypothesized that the self-administered questionnaire demonstrates strong internal consistency and an acceptable level of factor loading as a structural validity analysis.
MethodsStudy participants
Participants were recruited from community camps, and experts from the orthopedic field confirmed the diagnosis. The duration of data collection was December 2020 to May 2022. Participants were recruited from community camps organized to identify knee problems. The recruited participants' characteristics had episodes of anterior knee pain in the past year, crepitus, and difficulty in weight-bearing activity. In addition, the individual should have received education for at least 5 years, and have no cognitive impairment. Informed verbal consent was obtained from all the volunteer participants.
Questionnaire development
It included three systemic steps: (1) item generation through qualitative data, (2) content validation, and (3) psychometric analysis for the final version of the questionnaire.[11]
Item generation
The open-ended questionnaire was prepared from the literature review of existing tools and census recommendations on patellofemoral arthritis.[12] The focused group discussions and face-to-face interview surveys were conducted in hospital-based settings; n = 14 participants with early patellofemoral arthritis shared their early experiences in vernacular language. Each session was recorded and audio transcribed into text. The contents were generated and framed into vernacular language.
Content validation
The content validation was performed in two sequential procedures 1) Expert review (n = 10)[13] Pretesting on individuals with patellofemoral arthritis. (n = 12). The responses were elicited through a five-point Likert scale of agreement. The pilot testing was conducted for initial validation.
Evaluation of psychometric properties and sample size estimation
Confirmation of psychometric measures for the questionnaire is a prerequisite for assuring the integrity of the development process.[14]
CHARUSAT Patellofemoral Arthritis Questionnaire (CPAQ)
It is a self-administered tool. It consists of four domains: Crepitus: two contents, Stiffness: two contents, Pain: two contents, Difficulty in functional activities: six contents. The responses were elicited using the five-point Likert scale of frequency and an agreement of symptoms. It is also chosen based on domain expectations. Symptom frequency was noted with never “1,” rarely “2,” sometimes “3,” often “4,” and always “5,” while the agreement was described by strongly disagree “1,” disagree “2,” undecided “3,” agree “4,” and strongly agree “5.” It also includes the instructions, duration of symptoms, scaling response, and interpretation. It was a scale from 0 to 100; the higher the score, the higher the severity.
The descriptive data collection included age, gender, affected side, and duration of symptoms. The following properties were the quality measures for scale: internal consistency, repeatability measures, and construct validity.[14]
Internal consistency
The extent to which scale contents were intercorrelated to measure the same construct. Value 0.7–0.9 was considered a good internal consistency of the subscale.[15] It was performed on n = 310 participants.
Repeatability measures
It can be established by measuring test–retest reliability and agreement (absolute measurement error).[15] We used the 2-week time interval between the first and second administration of the questionnaire for measuring test–retest reliability. A total of 40 samples were used to evaluate this measure.[16] Longitudinal precisions for the test–retest reliability coefficient calculated using standard error of the measurement difference (SEMdiff = SD × √1 − ICC), the minimal detectable change 95% confidence level (MDC95 = SEM diff × 1.96× √2), and floor and ceiling effects.[14]
Construct validity
It is the degree to which the scored self-administered questionnaire adequately reflects the dimensionality of the construct to be measured.[17] It was performed on n = 310 samples. There are two components for construct validity: (1) convergent validity and (2) discriminant validity. These both can be calculated by conducting factor analysis.
Data analysis
Data analysis was performed using IBM SPSS Statistics for Windows, Version 26.0. IBM Corp: Armonk, NY, USA. Demographic details were evaluated using descriptive statistics. Responses were calculated for missing data, and participants who returned the questionnaires with <20% of contents not filled were excluded.
Demographic characteristics
Descriptive statistics were performed for ordinal variables and frequency analysis for nominal variables.
Internal consistency
Cronbach's alpha (average correlation) was performed to measure the internal consistency of all the items on the scale. Value 0.7–0.9 was considered a good internal consistency of the subscale.
Repeatability Measures
Test–retest reliability
The intraclass correlation coefficient (ICC) assessed test–retest reliability at the 95% confidence interval (CI). An ICC value ≥0.70 indicates an acceptable level of reliability. The two-way mixed model effect and an absolute agreement were utilized to measure ICC. ICC value should be >0.80 for good test–retest reliability.[15]
Measurement error
Longitudinal precision for the test–retest reliability coefficient was analyzed with the (SEMdiff = SD×√1 − ICC) and the MDC 95% confidence level (MDC95 = SEM diff × 1.96× √2).[1]
Item statistics
Inter-item correlation and correlated item-item correlation were performed to assess the data distribution at the individual item level. It estimates the degree to which a single score correlates with the total scale score. It reflects the strength of the relationship among the items to score them together on one scale.
Construct validity
Exploratory factor analysis was conducted for structural validity using Principle component analysis with the varimax rotation method. The acceptable level of commonalities and factor loadings for items was 0.5, and an eigenvalue of more than one was considered for component factors. For the data reduction, the following norms were considered: principal component analysis, varimax rotation, commonalities >0.4, factor loading >0.5 >0.45, anti-image correlation matrix >0.45, correlation matrix >30%, and eigenvalue >1.[16],[17]
ResultsDemographic characteristics
A total of n = 310 participants with early patellofemoral arthritis data were included in the evaluation phase. The mean age was 51.86 ± 3.135 (range: 45–64). The frequency table for other demographic details is shown in [Table 1].
Internal consistency (n = 309)
The internal consistency for the self-administered questionnaire, Cronbach's alpha α, was 0.90, which is interpreted as good internal consistency.
Item statistics
The inter-item correlations were <0.80 and the item-total correlation was >0.20, which showed an acceptable level of item redundancy [Table 2].
Reliability
Test–retest reliability (n = 40)
The ICC value based on the total scores of the first (test) and second (retest) assessment was 0.93 (ICC 2,1; 95% CI = 0.880–0.936; P < 0.000); the test mean was 91.08 ± 8.50; the retest score of mean was 91.08 ± 7.32.
The Bland–Altman plot shows the difference in total scores against the mean total scores [Figure 1].
Figure 1: Bland–Altman plot for the agreement between test–retest measurementsMeasurement error
The test–retest reliability was 0.93, and the standard error of measurement was 1.95. The calculated MDC was 5.39, and the MCID was 3.29. The skewness of the questionnaire (n = 310) was − 0.92 (between + 1 and − 1), and there was no floor and ceiling effect.
Construct validity-exploratory factor analysis
Initially, the factorability of the 12 contents was examined in the field testing. A total of 12 contents correlated between 0.3 and 0.5 with at least one other item, and most of the content correlations were significant at the 0.000 level suggesting reasonable factorability. In addition, the Kaiser–Meyer–Olkin measure of sampling adequacy was 0.840, and Bartlett's test of sphericity was substantial (χ2 (66) =410.23, P < 0.000).
Principal component analysis was used because the primary purpose was to identify and compute composite coping scores for the factors underlying the scale [Table 3]. The initial analysis considering factors with more than one eigenvalue produced a four-factor solution with 56.83% total variance. The initial eigenvalues showed that the first factor explained 53.30% of the variance, the second factor was 14.78% of the variance, the third factor was 9.71% of the variance, and the fourth factor was 9.08% of the variance.
DiscussionOur aim of the study was to develop and validate the CHARUSAT Patellofemoral Arthritis Questionnaire. It must be suitable for the school-level literacy rate, so individuals in the field can self-analysis the symptoms. It is the first questionnaire that identifies the symptoms of early patellofemoral arthritis. We conducted validation testing on a sample size of n = 310. The mean age for the recruited participants was 51.9±3.1; female participants were more (57.4%) than males. Maximum participants were affected with unilateral involvement (88.4%). The total duration for occurrences of symptoms was <6 months (51%) and 1–2 years (48.1%). Similar demographic details were observed in Kim and Joo, 2012; Farrokhi et al., 2013; and Haj-Mirzaian et al., 2021, studies.[18],[19],[20]
It is the only questionnaire for early cases of patellofemoral arthritis. The questionnaire demonstrated an excellent level of reliability and validity. The reliability measures, such as good internal consistency; Cronbach's alpha (α) 0.90 and test–retest reliability (ICC); 0.93. The standard error of measurement, MDC, and MCID were 1.95, 3.29, and 5.39, respectively. It interpreted the good reliability measures for the self-administered questionnaire.[21] KOOS-PF was developed as a subscale of KOOS, but there was a deficit in the evaluation of psychometric measures, so no gold standard tool was available for this condition. Structural validation was performed using the principal component analysis to explore the initial factor structure of the scale. The sampling adequacy was evaluated using the Kaiser–Meyer–Olkin measure.[17] The value was adequate, KMO = 0.840, and Bartlett's test of sphericity was significant (χ2 (66) =410.23, P < 0.000). The factorability of 12 contents was examined in the field testing. A total of 12 contents correlated between 0.3 and 0.5 with at least one other item, and most of the content correlations were significant at P+ =0.000 level suggesting reasonable factorability.[17] Principal component analysis was used because the primary purpose was to identify and compute composite coping scores for the factors underlying the scale. The initial analysis considering factors with more than one eigenvalue produced a four-factor solution with 56.83% total variance. The initial eigenvalues showed that the first factor explained 53.30% of the variance, the second factor was 14.78% of the variance, the third factor was 9.71% of the variance, and the fourth factor was 9.08% of the variance. The construct validity of the questionnaire demonstrated a good level of measures.
The confirmatory factor analysis should be performed, which can be the limitation of the study. Therefore, we recommend future scope to conduct a large-scale survey in middle-aged population to identify the symptom and spread awareness of early arthritis.
ConclusionsIt is an acceptable, reliable, and valid self-administered questionnaire that can identify the early symptoms of patellofemoral arthritis. In addition, the questionnaire provides a good opportunity at the community level to self-analyze symptoms.
Financial support and sponsorship
None.
Conflicts of interest
There are no conflicts of interest.
References
Comments (0)