Design and Procedure. This study was approved by the Ethics Committee of the Medical Association Westphalia-Lippe and the University of Muenster. Approval Code: 2021-684-f-S. Approval Date: 9 March 2022. Informed consent was obtained from all subjects and/or their legal guardian(s). All methods were carried out in accordance with relevant guidelines and regulations. All patients with suspected lipedema visiting our specialized clinic for an initial consultation were asked to fill in the paper-pencil questionnaires assessing general baseline information. Therefore, age, height, weight, age of symptom onset/diagnosis, family history and stage of lipedema, location and symptoms of pain, previous therapies, comorbidities, QoL, symptoms of depression, and physical activity behavior were assessed. Only patients later diagnosed with lipedema were included. Patients who already had received a liposuction or any bariatric surgery prior to consultation were excluded from the study.
Outcome Assessment. Patients were asked to rate their pain in the specific areas (i.e., legs and limbs) that were affected by lipedema. Therefore, a single-item questionnaire was used (“Do you have pain in the affected areas?”). The response was rated on a ten-point Likert scale ranging from one (“none”) to ten (“very strong”).
Physical activity was assessed via a questionnaire consisting of seven items that was specifically designed for the purpose of this study. Each item represents one dimension (i.e., gym, Nordic walking, swimming, biking, running, others, and none). The patients were asked to define the quantity of sporting activities per week.
Mental Health. Mental health (i.e., symptoms of depression and dimensions of QoL) was assessed via the World Health Organization Quality of Life-BREF (WHOQOL-BREF) and the PHQ-9 [29], [37] questionnaire. The PHQ-9 is a well-established assessment tool based on the diagnostics criteria set in the DSM-IV to screen for the presence and severity of depression [37], [38]. It entails nine items that are rated on a four-point Likert scale ranging from zero (“none”) to three (“nearly every day”). The sum score is interpreted as follows: Zero to four points represent none to minimal depression; five to nine points represent mild depression, ten to fourteen represent moderate depression, fifteen to nineteen represent moderately severe depression, and twenty to twenty-seven represent severe depression. Thus, higher levels of the PHQ-9 can be interpreted lower levels of health, respectively.
The WHOQOL-BREF is a proven self-report tool with high validity to assess general QoL. It comprises twenty-six items that are rated on a six-point Likert scale ranging from one to five [29]. The WHOQOL-BREF consists of four subscales including physical health, psychological health, social relationships, and environment. Physical health embodies aspects of mobility, pain, and energy. Psychological health (i.e., mental health) describes negative feelings, self-esteem, and memory/concentration. The social relationships scale represents social support and personal relationships. Environment describes home and physical environment, transport, and financial resources. Each subscale contains six items. Two additional items referring to the total score of QoL are used to assess aspects of general health and overall QoL. The items were recoded, and scores were calculated according to the current guidelines [29]. Higher scores represent higher self-reported QoL whereas lower scores represent lower self-reported QoL. The WHOQOL-BREF is a widely used tool to assess the QoL and has been demonstrated high validity [29], [39]. Thus, in total, higher level of the WHOQOL-BREF questionnaire can be interpreted as higher levels of health, respectively.
Data Analysis. Data analyses were performed via the programming language R [40] with the interface RStudio [41]. To investigate the hypotheses, RSA was conducted. The full polynomial model is defined as
$$\text{Z} = {\text{b}}_{0}+ {\text{b}}_{1}\text{X}+ {\text{b}}_{2}\text{Y}+ {\text{b}}_{3}{\text{X}}^{2}+ {\text{b}}_{4}\text{X}\text{Y}+ {\text{b}}_{5}{\text{Y}}^{2}+ {\epsilon }$$
in which the centered outcome variable (aspects of mental health and QoL) is regressed on pain in the referring areas (X), physical activity (Y), the squared terms of pain (X2) and physical activity (Y2), and the cross-product of pain and physical activity (XY). b0 to b5 represent the unstandardized regression weights. Within the full model described above, more parsimonious (i.e., statistically simpler) models can be identified that fit the data better and that are nested within the full model [36]. The rising ridge model allowing main effects of the predictor variables, forming a tilted ridge (RR model), a shifted and tilted ridge (SRR model), or a shifted and tilted ridge with an additional rotation (SRRR model). The flat ridge model does not allow a main effect, but a mismatch of the predictor variables, again without a tilted but with a shifted ridge (SSQD model), or with a shifted and rotated ridge (SRSQD model). Furthermore, additive model, interaction model (IA model), and models with single effects of the X-variable (onlyX or onlyX2 models) or Y-variable (onlyY or onlyY2 models) can be identified.
To test and select the best fitting model, model fit indices were compared and evaluated [36]. Therefore, the corrected Akaike Information Criterion (\(\text{A}\text{I}\text{C}\text{c}\)), \({\Delta }\text{A}\text{I}\text{C}\text{c}\), and model weight were used to inspect relative fit. The Comparative Fit Index (CFI; with values > .95 indicating sufficient model fit), and adjusted R2 were assessed to evaluate absolute model fit. According to Cohen [42], values were interpreted as substantial (R2adj = 0.26), moderate (R2adj = 0.13), or weak (R2adj = 0.02). Practically equivalent models were defined by \({\Delta }\text{A}\text{I}\text{C}\text{c}\) < 2, Implausible models were defined by a cumulative weight > .95. The models showing the best fit indices were selected for analysis.
Within the three-dimensional visualization of the RSA, the shape of the surface is guided by the lines of congruence and incongruence. Effects of congruence (both predictors are high, or both are low) on the outcome variable are made visible via a line of congruence (LOC). Similarly, effects of incongruence between the two predictor variables on the outcome variable are visualized via the line of incongruence (LOIC). Regarding the congruence effects described by the LOC, the coefficients a1 and a2 are interpreted. The slope of the LOC is defined as \({\text{a}}_{1} = {\text{b}}_{1}+ {\text{b}}_{2}\). A significant coefficient indicates a linear additive effect of the predictor variables. The coefficient \({\text{a}}_{2} = {\text{b}}_{3}+ {\text{b}}_{4}+ {\text{b}}_{5}\) describes whether there is a curvature on the LOC, i.e., if the effect is linear or curvilinear. Regarding the incongruence effects described by the LOIC, the coefficients a3 and a4 are interpreted. The slope of the LOIC is defined as \({\text{a}}_{3} = {\text{b}}_{1}- {\text{b}}_{2}\). A significant coefficient suggests that the ridge is shifted away from the LOC, i.e., indicating the direction of the predictor variables’ mismatch on the outcome variable. The coefficient \({\text{a}}_{4} = {\text{b}}_{3}- {\text{b}}_{4}+ {\text{b}}_{5}\) describes whether there is a curvature on the LOIC, i.e., indicating the degree of the predictor variables’ effect on the outcome variable.
Prior to the main analysis, the stage of lipedema was entered as a moderator variable into linear regression analyses regressing the outcome variables on pain and physical activity. Therefore, we could examine whether the severity of lipedema (i.e., stage I to III) is influencing the relationship between the predictor variables and the outcome variable. As high values of PHQ-9 represent lower levels of health, and high levels of WHOQOL-BREF represent higher levels of health, the colors in the RSA visualization were selected as follows. In all analyses, green color represents higher levels of health, whereas red color represents lower levels of health.