An online survey of individuals with and without chronic ankle pain and/or stiffness was used to address the following questions: (i) what are the self-reported differences in QoL, function, ankle instability and physical activity between individuals with ankle symptoms and asymptomatic controls, and (ii) which of these outcome measures are associated with QoL.
Recruitment
Between July 2015 and February 2017, Australian volunteers aged 30 to 75 years with and without a history of ankle pain and/or stiffness (present on most days for >3 months duration) participated in this cross-sectional survey. Participants were recruited via community advertisements placed in a local university staff and community newsletters, communications from National and State arthritis organisations, and social media. Participants were asked to indicate if they “experienced any of the following ankle symptoms for more than 3 months on most days”: 1) Pain or ache in/or around the ankle, 2) Ankle joint stiffness or reduced movement in the morning. Participants who answered “yes” to either of those questions were included in the symptomatic group. Participants who indicated they did not experience any ankle pain or stiffness in the last 3 months were included in the asymptomatic control group. Exclusion criteria for control participants were a history of ankle pain or injury. The study was approved by the institutional human research ethics committee and all participants provided informed consent.
Outcome measures
Participants provided information about their age, sex, body mass, height and history of ankle injuries and ankle related health-care consultations. They also completed the questionnaires and scales described below.
Severity of Pain and Stiffness
Participants rated their ankle pain at rest, average ankle pain over the past 24-hours, and worst pain over the past 7 days using an 11-point scale Numerical Rating Scale anchored at 0 with “no pain” and at 10 with “worst pain imaginable”. Participants also rated their usual level of ankle stiffness over the past week on an 11-point NRS anchored at 0 with “no stiffness” and at 10 with “worst stiffness imaginable”.
Quality of Life
The Assessment of Quality of Life questionnaire (AQoL-6D) is an Australian multi-attribute utility instrument used to evaluate QoL with age- and gender-based population norms (7). It comprises 20 questions in 6 separate dimensions (independent living, mental health, coping, relationships, pain, and senses). The unweighted responses of all questions are summed to create an overall profile score (0-100) and individual scores for each of the six dimensions. Higher scores indicate better QoL. This instrument has strong construct (8) and discriminative validity for use in OA populations (9).
Function
The Foot and Ankle Ability Measure (FAAM) was used to assess function (10). It consists of a 21-item Activities of Daily Living subscale (FAAM-ADL) and an 8-item Sports subscale (FAAM-sport). Each item is scored on a 5-point Likert scale (0-4) ranging from “no difficulty” (4) to “unable to do” (0). A “not applicable (NA)” option is available to indicate activities limited by factors other than foot or ankle problems. These items are excluded from scoring. Responses for rated items are summed, and the total scores for the FAAM-ADL and FAAM-Sport are presented as percentages, with a higher percentage indicating a higher level of function. The FAAM-ADL and Sport have excellent test-retest reliability and internal consistency (10). At the end of the FAAM, participants rated the current level of function as normal, nearly normal, abnormal or severly abnormal.
Pain and Disability
The Ankle Osteoarthritis Scale (AOS) is a disease-specific instrument used to evaluate pain and disability related to ankle OA. It consists of pain and disability subscales, with nine questions in each subscale. Participants indicate how much pain or difficulty they experience when performing certain activities over the past week. The original scoring of the two subscales is measured along a 100-mm visual analoge scale (VAS) anchored with “No pain” or “No difficulty” at 0mm and “Worst pain imaginable” or “So difficult, unable” at 100mm). To enable this questionnaire to be used in an online format, an 11-point (0-10) NRS was used rather than a 100 mm VAS, with the same anchors as the original scale (paper version). To assess if the online NRS version of the questionnaire captured the same measure as the paper VAS version, paper based and online versions were administered in random order to 10 participants with ankle pain approximately 3 days apart.
Ankle Instability
The Cumberland Ankle Instability Tool (CAIT) is a valid, and reliable tool used to measure perceived ankle instability (11). The tool contains 9-items with scores assigned based on the rank of the chosen response. Responses are summed separately for each limb. The maximum score is 30 with a higher score indicating less instability.
Physical Activity
The International Physical Activity Questionnaire- short form (IPAQ) was used to capture data on self-reported physical activity. The IPAQ measures the total amount of time spent performing moderate activity, vigorous activity, walking or sitting in bouts of 10 minutes or greater over the last 7 days (12). The time (in minutes) spent for each activity is multiplied by the defined metabolic equivalent (MET) of each task category and scores are combined and presented as total MET-minutes per week. The IPAQ categories physical activity into “low”, “moderate” or “high". Published guidelines for data processing and analysis of IPAQ were used (available from: http://www.ipaq.ki.se). The IPAQ has high reliability (Spearman's rho ranging from 0.66 to 0.88) (12).
Statistical analysis
Statistical analysis was performed using IBM SPSS Statistics for Windows (Version 25.0. Armonk, NY: IBM Corp). Kappa statistics were used to compare the online and paper based versions of the AOS, and agreement was categorized as poor (<0.00), slight (0.00–0.2), fair (0.21–0.4), moderate (0.41–0.6), substantial (0.61–0.8) or almost perfect (0.81– 1.0) (13).
A univariate analysis of covariance with age, sex and BMI entered as covariates and group as a fixed factor was used to compare between group differences for all outcomes. To ensure our asymptomatic participants reflected the Australian population, AQoL-6D data was compared between controls and published norms. Data representing point estimates of effect are presented as mean differences (MDs) and their 95% confidence intervals (CI) in tabular format and as standardized mean differences (SMDs) and (CI) in forest plots. The SMD was calculated as the difference between the two groups means divided by the pooled SDs. Differences in outcomes were calculated such that negative differences indicated a deficit in the measure for the symptomatic group compared to controls, with positive differences indicating the opposite. Effect sizes were interpreted as trivial: 0.0-0.2, small: 0.2-0.6, medium: 0.6-1.2, large: 1.2-2.0, very large: 2.0-4.0 and distinct: >4.0 (14). Chi-square tests were conducted to compare categorical variables (sex and categories of physical activity) between groups. Odds ratio (OR) and risk difference (RD) were reported for categorical and binary data.
As bivariate normality was not assumed, the relationship between variables (AQoL-6D, group, sex, BMI, age, ankle stiffness, CAIT, AOS-Pain, FAAM-Sport, FAAM-ADL and AOS-Disability) was investigated using nonparametric Spearman's Rank-Order Correlation. The correlation was interpreted as low (0.1 to 0.3), moderate (0.3 to 0.5), high (0.5-0.7) and very high(0.7-0.9) (14). A stepwise backward elimination regression was conducted to establish the most influential independent variables associated with the dependent variable of AQoL-6D. The independent variables included in the model were group, sex, BMI, age, ankle stiffness, CAIT, AOS-pain, AOS-disability, FAAM-ADL and FAAM-sport. Those with a higher correlation to AQoL-6D were entered first. The multiple regression model was tested for multicollinearity. If multicollinearity was present, we retained in the model the variable with the higher b value and that has been more commonly used in research of individuals with ankle problems (15). Statistical significance was set at p<0.05.