Clinical inter‐rater reliability of postural control techniques

Abstract Effectiveness of postural control techniques to compensate for oropharyngeal dysphagia have been recommended and used by several clinicians. However, the inter‐rater reliability of these techniques is not well understood. The purpose of this study was to clarify the ambiguity of postural control techniques using statistical analyses. A total of 50 clinicians involved in dysphagia treatment participated in this study, where a healthy male served as the simulated patient. The following clinically used postures were measured by two investigators on two separate days: chin down, right/left incline, and right/left rotation. Postural angles were measured twice by two investigators on each day. Data obtained for the angle of each posture were visually displayed. Data from both investigators were assessed for each posture using the Youden plot, which analyzes data variability for systematic errors and accidental errors separately. The correlation coefficient for examining the measurement error between investigators was calculated. The results showed considerable variation between clinicians regarding the postures used, and significant differences were noted each day. The correlation coefficient for a total of four measurements was more often lower on Day 2 than that on Day 1. The details of the instructions provided by clinicians were not fixed, and the same specified posture was not reproduced even when instructions were provided to the same subject. These findings suggest poor inter‐rater reliability because of the variability of selected postures when using statistical analyses. Therefore, standardized postures need to be developed that can be easily measured and reproduced.

tiple swallows, dietary modifications, and postural control techniques.
Postural control techniques are designed to reduce aspiration and penetration by changing the angle and position of the head and body (Ertekin et al., 2001;Logemann, Kahrilas, Kobara, & Vakil et al., 1989;Shanahan et al., 1993). Depending on the specific swallowing deficits found in a patient with dysphagia, a single posture or combination of postures is chosen to facilitate efficient and safe swallowing (Ota, Saitoh, & Matsuo et al., 2002). Effectiveness of postural control techniques has been studied by many clinicians and investigators (Logemann, Rademaker, & Pauloski et al., 1994) since Larsen (Larsen, 1973) recommended the flexed neck posture in 1973. The clinical benefit of applying a single postural control technique or a combination of several techniques has been reported to be effective in 80% to 90% of patients (Logemann et al., 1994;Fujishima et al., 2010). Logemann described these techniques based on videofluorographic data in a significant number of patients with dysphagia (Logemann, 1983). Postural control techniques improve dysphagia by altering the configuration of the oral cavity or pharynx, in an attempt to redirect the bolus or to change the speed of bolus flow (Logemann et al., 1989). Clinicians involved in dysphagia treatment should thoroughly understand how each posture impacts swallowing physiology before selecting a specific posture.
The use of certain postural control techniques is a challenge, as these are not clearly defined. Therefore, postural control techniques currently in use may not be standardized across clinicians. Additionally, there are no studies with statistical data reporting the reproducibility of these techniques. The criteria used to assume a correct posture (such as the angle of inclination of the body) remain ambiguous. The purpose of this study was to provide objective data about inter-rater This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. 1 | PARTICIPANTS AND METHODS

| Participants
This study was approved by the Ethics Committee of Showa University School of Dentistry (Approval No. 2013-004).
Before the initiation of the study, written informed consent was obtained from all participants. A total of 50 clinicians (48 dentists and two speech-language therapists) involved in dysphagia rehabilitation participated in this study. The years of the experiences were 1 through 17. It was verified that each clinician used postural control techniques in their practice.
A healthy male volunteer (26 years of age; BMI 19.0) served as the simulated patient, and it was confirmed that he had no history of orthopedic disease, abnormalities of cervical alignment, or abnormal findings such as muscular spasms, which may affect adjustment of the instructed posture. The simulated patient wore the same clothes throughout each measurement. Additionally, he wore a swimming cap to reduce systematic error due to hair and a singlet to allow easy assessment to the upper body and cervical region.
After the simulated patient was placed in the targeted posture, the range of motion angles was independently measured by two investigators (one male and one female dentist).

| Data analysis
We obtained the correlation coefficient to examine measurement error between the two investigators. Data on the angle obtained for each posture were visually displayed, and data from both investigators were assessed for each posture, using the Youden plot, which separately displays data variability for systematic errors and accidental errors (Skendzel & Youden, 1970;Skendzel & Youden, 1969;Youden, 1977). Data from the investigators were assessed for the first and second measurements for each day of the study.

| RESULTS
The results for the angle of the chin-down posture are shown in The average angle on Day 1 was 10.86°and that on Day 2 was 11.78°( p = .78, Table 1).
The results for the right cervical rotation posture are shown in  The average angle on Day 1 was 46.12°and that on Day 2 was 45.33°( p = .58, Table 1).
The results for the left cervical rotation posture are shown in  Table 1).
On the Youden plot, the 95% confidence interval was elliptical, and systematic error was noted for all measurements plotted. This result represents poor reproducibility because of the variability of selected postures. Nevertheless, no significant difference was noted between the angle of each posture on Days 1 and 2.

| DISCUSSION
The settings for patients with dysphagia include hospitals (Cherney, 1994), nursing homes (Siebens et al., 1986), or their own homes (Lindgren & Janzon, 1991), and theoretically, the precise compensation used for swallow safety must be determined for each patient. To accurately reproduce the specified posture, various media, such as verbal communication, written documents, illustrations, photos (still images), and videos (moving images), may be used in clinical practice. In such settings, reproducibility of postural control techniques might be lower than that in a physician's office. In this study, the Youden plot showed considerable variation in the measurement data. It is assumed that considerable variation due to bias would occur when an angle gauge is used to measure the range of motion in the clinical setting. However, there are no studies using statistical analysis that report the potential variance in assuming the correct posture. In contrast, in this study, findings are shown using Youden plots and correlation coefficients.
In addition, we used Spearman's rank correlation coefficient. The presence of bias (postural variation) could be related to the investigator's  Note. Spearman's rank correlation coefficient. expertise in postural control, the vagueness of evaluation terminology, and/or variation in the comprehension of instructions provided by the clinicians. It was expected that the correlation coefficient on Day 2 would be higher than that of Day 1 because the experience was replicated. However, the correlation coefficient for a total of four measurements was more often lower on Day 2 than that on Day 1.  To assess the reproducibility of posture, several devices such as specialized training chairs and head/neck fixing apparatus have been developed (Logemann, 2008). However, no method is available to evaluate the patients' postures while using these devices. It is also important to evaluate postural controls during actual swallowing as the conditions may change during the patient's attempt to swallow. In this study, we measured the reproducibility of postural control techniques conducted by clinicians. In the future, we need to study the potential differences between a posture set by the clinician and the actual posture the patient is in at the moment of swallowing. Establishing the most reliable measuring tool is also of interest for future study.
A limitation of this study was that only two-dimensional (x and y) measurements were taken, where there is a need to examine threedimensional changes including the trunk of the body (x, y, and z) as a large number of patients with dysphagia consume their meals while being seated on a chair. In this case, the posture is affected in three dimensions. This was a pilot study to determine the inter-rater reliability of postural control techniques using statistical analyses, and we used only one simulated patient. Further study involving patients with different physiques, physical characteristics, and age is required.

ACKNOWLEDGEMENT
This work was supported by JSPS KAKENHI Grant number 26861840.