This cross-sectional study was conducted in accordance with resolution 466/2012 of the Brazilian National Health Council and was approved by the institutional review board (#1.757.451) of Universidade Estadual de Londrina (UEL), Londrina, PR, Brazil. All participants provided an informed consent form prior to participation in the study.

Female participants, between 18 and 40 years of age were screened for eligibility in a non-probabilistic design with a convenience sample. To be included in the study, they could not present any of the following conditions: history of traumatic knee injury; previous lower extremity surgery; patellar dislocation; clinical evidence of meniscal injury; ligament instability or patellar tendinopathy; neurological, cardiovascular, or rheumatic diseases; diabetes; sensory change on the plantar surface of the foot; using medication and/or performing physical therapy sessions in the previous six months; being pregnant at the time of the study.1 Additionally, participants in the PFP group needed to present the following features: (1) a clinical diagnosis of patellofemoral dysfunction by a physician; (2) anterior knee pain of four or more points on a 0–10 VAS for at least eight weeks before evaluation; (3) insidious onset of symptoms unrelated to trauma; (4) presence of knee pain in at least three of the following tasks: climbing/descending stairs, crouching, running, kneeling, or sitting for prolonged periods.11 Participants in the control group (i.e. without PFP) were included if they did not present any complaints of pain or disability, in any joints of the body. Participants were excluded if they did not complete all tests or attend the retest session.

During the initial visit, all participants were informed about the research and data collection procedures. In addition, on the same day, they completed a general questionnaire for sample characterization, the VAS for pain, and the AKPS (both translated and validated for Brazilian Portuguese).12 Participants, were then asked to perform five functional tests in a random sequence: sitting–rising test (SRT), sit-to-stand in 30s (STS30), stair-climbing test (SCT), stair descent test (SDT), and six-minute step test (6MST).

To assess inter-rater reliability, the set of 5 tests was performed twice and assessed by separate investigators. The sequence of tests and the order of the two investigators were chosen randomly on Day 1 (while one assessor performed the evaluation the other remained in a separate room to blind the assessment and result). Participants returned to the laboratory one week later to repeat the entire procedure (Day 2), with the same randomized sequence as the first visit to avoid differences due to fatigue. A total of three experienced investigators performed the data collection. To guarantee standardized operational procedures, all investigators were trained for the study protocol and participated in a pilot study of this protocol.

Sitting–rising test: The SRT assesses the ability to sit down and stand up from the floor with partial scores for each of the two actions. The participant is required to get up from the floor, and after standing up, return to the floor, trying to use minimal support.13 This test evaluates the quality of movements to sit and rise from the floor. The quality of movement is analysed and scored from zero to 11. For each external support necessary, for example if the participant need to put their hand on the floor to perform sitting, or if imbalance occurs during execution, the total score is reduced by a point and a half point, respectively. Higher values refer to better functional performance. Due to the lack of evidence describing the number of repetitions necessary to assure reliability of the test, on each visit the test was repeated three times with 1-min rest between each execution. The mean of all attempts was used as the test result.

Sit-to-stand in 30s: The STS30 evaluates the number of repetitions of sitting down and getting up from a 42cm-chair in 30s. The complete movement of sitting and getting up was counted as one repetition, where higher number of repetitions correspond to a better result. If participants did not fully extend their knees or sit on the chair, the repetition was not counted. On each visit, two executions of the test were performed with 2min rest between them, as proposed in the literature.14 The mean of the attempts was used as the test result.

Stair-climbing test and stair descent test: These tests evaluate the time taken to go up and down a set of stairs, respectively. In the SCT, the volunteer climbs eight steps in the shortest time possible without external support, with lower values corresponding to greater agility and better performance. The same protocol was used for stair descent. On each visit, participants performed two trials of each test with a 2min rest between attempts.15 The mean of the attempts for both SCT and SDT was used as the test result.

Six-minute step test: The 6MST measures the number of times the participant climbed and descended from a 20cm high step in 6min. The aim is to go up and down the step as fast as possible, with higher values corresponding to better performance. The number of repetitions was counted, considering one repetition when the participant went up and down the step. The test is similar to the six-minute walk test in terms of rhythm, timing, and encouragement. Participants performed the test twice on each day, one time with each assessor,16 and the best attempt for each assessor (done on different days) was used as the test result.

For safety reasons, heart rate (using a cardio-frequency metre) and the subjective sensation of dyspnoea and exertion (using the Borg scale) were monitored during all tests. The presence of syncope, vertigo, major tachycardia (100% of maximal heart rate), or cyanosis, were indications to discontinue testing; however, none of these events were observed in this study. To control fatigue and pain, the Borg scale and VAS were completed before and after each test, if there was increase in pain or fatigue participant waited before performing another trial or test.

Difference in the STS30 between the two groups was defined a priori as the primary outcome. Based on a pilot study, forty participants (2 groups of twenty) were needed to obtain 90% power, based on a difference of 5.16 repetitions in the STS30, assuming a standard deviation of 4.27 repetitions, and an alpha lower than 0.01.

Statistical analyses were performed using Statistical Package for the Social Sciences (SPSS®) 20.0. Data are described as mean±SD or median [interquartile range] according to data distribution. Data normality was assessed by the Shapiro–Wilk test. Characteristics of the participants and performance in the tests were compared between patients with PFP and the control group using unpaired t tests or Mann–Whitney tests, according to the data distribution. Descriptive data of the tests and the comparison between the two groups were done using the results of the first patients’ assessment during the first day, with the exception of 6MST which was the only one performed by two raters on both days, so the best attempt of the first patients’ assessment on both days was used.

Intra-rater analysis was done using the two-way mixed effects, absolute agreement and multiple raters (i.e. all test–retest data were used). Interrater analysis, was done using the two-way random effects, absolute agreement and multiple raters (i.e. all the three raters were used in the analysis).17,18 Reliability was deemed moderate if values were between 0.50 and 0.75, good if between 0.76 and 0.90, and excellent if above 0.90.19 To observe inter-evaluator and execution differences, two-way ANOVAs were performed, and data with non-parametric distribution were normalized with log10. The Standard Error of Measurement (SEM) was calculated to establish the absolute reliability, by means of the following equation: SEM=SD√(1−ICC), in which SD is the standard deviation and ICC is the Intraclass correlation coefficient.20

Although there is no gold standard to evaluate function in individuals with PFP, validity of the tests was assessed by determining the extent of association between scores on the tests and score on the AKPS, using the Spearman correlation coefficient.

In addition, Spearman correlation was used to determine the association between pain intensity (VAS) and performance for each test. The Mann–Whitney test was used to compare differences between pain scores between groups for each test. To compare pain between groups, the average of VAS in the end of all tests of the SRT, 30STS, SCT, and SDT and the reported VAS during the best attempt of the 6MST were used. Statistical significance was set at p<0.05 for all analyses.