This study included 222 women from the Region Västra Götaland, Sweden. Assessments were conducted at three rehabilitation centers. Based on the guidelines of Ko & Li, we aimed to assess at least 30 participants at each center [27]. The women were invited to participate at antenatal and childcare centers, and via social media. Inclusion criteria were age of ≥ 18 years, vaginal delivery or caesarean section within the past 3 months, and ability to understand and respond in Swedish. Exclusion criteria were chronic pelvic girdle pain and/or low back pain (defined as pelvic or low back pain for over 3 months, not related to pregnancy) and/or pelvic floor tear grade III/IV.
The participants were contacted and booked for assessment at one of the three rehabilitation centers in the Västra Götaland region within 3 months after giving birth. Prior to the assessments, the participants completed a questionnaire about their age, BMI, mode of delivery, number of delivered children, self-reported pelvic floor tears, most recent baby’s birth weight, and the birth weights of previous children (if applicable).
Assessments were performed by six physiotherapists—two at each rehabilitation center. These physiotherapists had each completed a four-day (or longer) course in PFM assessment and treatment methods, and all had between one and nine years of experience in assessing PFM. During the design phase of this study, four hours of training in DRA measurement was planned. All included physiotherapists were novices at measuring the DRA by caliper, and on using the rating scales for depth and bulging. Two months after the start of the study, we conducted a preliminary data analysis because the physiotherapists expressed strong uncertainty regarding the right technique for using the caliper. This preliminary analysis showed low-to-negative ICC values and large differences between the measurements. Thus, the 61 measurements acquired between September and November of 2018 were excluded from the final analysis. The physiotherapists at all centers underwent additional training. At this time, rehabilitation center 3 had not yet started their assessments.
Clinical assessment of PFM
The PFM was assessed with the patient in the supine position, with the legs flexed and slightly abducted on a plinth, and a pillow under the head. Participants were assessed by observation and digital palpation.
During observation, the physiotherapist stood beside the plinth, holding the participant’s legs and observing the movement of the perineum. To observe involuntary contraction, the participant was asked to cough forcefully, and the physiotherapist rated the movement as moving downwards, perineal in-drawing, or no movement. To observe voluntary contraction, the participant was given the verbal cue “contract your pelvic floor muscles like you want to prevent the escape of gas/urine”. The physiotherapist then observed the movement of the perineum, and rated it as moving downwards, perineal in-drawing, or no movement.
Digital palpation of the PFM was performed by physiotherapists using their index and middle finger, with examinations gloves and water-based lubricant. These fingers were inserted 2–3 cm into the vagina, with the palmar side directed to the caudal part of the vagina. To assess involuntary contraction, the participant was asked to forcefully cough three times. The physiotherapist noted the absence or presence of a correct contraction, defined as a squeeze around the pelvic openings and an inward lift [17].
To assess maximal voluntary contraction (MVC), the participant was asked to contract the PFM. In the event of a downward movement, the participant was again given the verbal cue “contract your pelvic floor muscles like you want to prevent the escape of gas/urine”. If the physiotherapist felt a correct contraction, the participant was encouraged to activate their PFM “as strong and as long you can”. Of three MVCs, the strongest was rated on a 6-point modified oxford scale (Appendix 1). The participants rested 15 seconds between the contractions. If a participant, despite several attempts and verbal cues, failed to squeeze and lift and was instead straining, their PFM function was rated as “−1” and the participant was excluded from the statistical analysis of MVC and PFM endurance.
To assess PFM endurance by digital palpation, after 15 seconds of rest, the participant was asked to contract the PFM for as long as possible at approximately 50% of the previous contraction strength. The physiotherapist rated PFM endurance as positive if the participant was able to hold this contraction for longer than 30 seconds. Finally, to assess voluntary relaxation, the participant was given the verbal cue “try to relax your pelvic floor, let the vagina get larger and go downwards”. This function was rated as absent, partial, or complete.
Clinical assessment of DRA
DRA assessment was conducted in the same position as described above for PFM. The physiotherapists assessed DRA width using an electronic digital caliper (150 mm, carbon fiber, accuracy ± 0.2 mm, 24 se Sverige AB, Kalmar, Sweden). Caliper application is explained in Appendix 2. At the start, the physiotherapist used a water-soluble marker to mark the three measurement points: at the umbilicus, and at 4.5 cm above and 4.5 cm below the umbilicus [28, 29]. For accurate assessment, the participant had to lift her head 2–3 cm from the plinth, with no pillow. Before the assessment began, the physiotherapist assured that the participant correctly lifted her head 2–3 cm, which was trained by several repetitions.
To assess DRA width, the participant was asked to lift her head and then lower it slowly. During this movement, the physiotherapist palpated the outer edges of the linea alba with their index and middle finger, without examination gloves. Next, the participant was instructed to relax her muscles and perform the trained head lift of 2–3 cm. During this movement, the physiotherapist identified the distance between the two parts of the rectus abdominis with her fingers, and measured this felt distance using the caliper (Fig. 2a). The same procedure was conducted at all three measurement points.
To measure the linea alba depth, the participant was asked to repeat the exact same head lift of 2–3 cm. At all three measurement points, the physiotherapist palpated the resistance, and rated it as “good resistance at all points”, “resistance in the depth at measurement point x”, or “bottomless resistance at measurement point x”. To assess linea alba bulging under load, the participant performed a 3-step sit-up test [30]. During this test, the physiotherapist observed whether the linea alba bulged during the movement (Fig. 2b).
Upon completion of the assessment, the participants rested for 30 minutes in a sitting or lying position. After the 30-minute rest, the second physiotherapist conducted the same assessment as described above. The two investigating physiotherapists were blinded to each other’s findings, and were not allowed to talk about their assessments.
Statistical analysis
Statistical analyses were performed using IBM SPSS statistical package version 25 (SPSS Inc., Chicago, IL) and the Svensson Excel template from http://avdic.se/svenssonsmetod.html. Descriptive statistics are presented as mean, standard deviation (SD), and range for ratio data, and as number and percentage for nominal and ordinal data. To calculate statistically significant differences between the three rehabilitation centers, we used the one-way ANOVA test for interval and ratio data, and the Kruskal-Wallis test for ordinal data. A p value of ≤ 0.05 was regarded as statistically significant.
All PFM functional measures were rated on ordinal scales, except for PFM endurance and involuntary contraction by palpation. DRA depth was also rated on an ordinal scale. Ratings on ordinal scales were evaluated by Cohen’s weighted kappa values. PFM endurance, involuntary contraction by palpation, and linea alba bulging were rated on nominal scales, and these ratings were evaluated by Cohen’s kappa values. For interpretation of kappa values, we used the categories of Landis and Koch: <0.2, slight; 0.21–0.40, fair; 0.41–0.60, moderate; 0.61–0.80, substantial; and 0.81–1.0, almost perfect agreement [31]. Percentage agreement was calculated and presented for all nominal and ordinal data, and < 60% agreement was defined as faulty agreement [32].
We used the Svensson method, developed by Elisabeth Svensson [33], to distinguish the position and concentration variance between physiotherapists. Position variance was defined by the fact that one physiotherapist is systematically using a higher or lower value on a rating scale than the other physiotherapist. Concentration variance was defined by the fact that one physiotherapist was systematically using a smaller part of the scale. Relative rank variance explains an individual variation that cannot be explained by a systematic bias.
For assessment of DRA width, a continuous scale (in mm) was used. To evaluate the inter-rater reliability of the assessments on a continuous scale, we calculated the intraclass correlation coefficient (ICC) and 95% confidence interval (CI). ICC values were calculated in SPSS based on absolute agreement and a 2-way mixed effects model. ICC values of < 0.50 indicate poor reliability, 0.50–0.75 indicate moderate reliability, 0.75–0.90 good reliability, and values of > 0.90 indicate excellent reliability [27].
To further evaluate reliability, we calculated the standard error of measurements (), which represent the typical error in a single measurement and the minimal detectable change (). For calculation of the standard error of measurements (SEM), we used the standard deviation (SD) from the scores of all subjects. SEM and minimal detectable change values are presented in mm.