The survey was designed using the QuestionPro survey software (Seattle, WA, USA). As it is easily accessible, saves time, and is cost-effective, an online survey format was selected in our study. The survey was designed to be taken anonymously—without personal identification data requested or stored—and was completed in approximately 7 minutes.
A cover letter describing the purpose of the study, informing participants of the voluntary nature of their participation, and assuring their anonymity was provided along with the survey questionnaire. Participants were encouraged to contact the research investigator for any queries pertaining to the study, using the provided contact information.
This cross-sectional study was designed to assess work-related musculoskeletal symptoms among clinical radiologists, including residents, specialists (junior staff), and consultants (staff) practicing across all hospitals (academic, public and private) in the major cities of the Eastern Province of Saudi Arabia, which included a total of 12 institutions.
Recruitment of participants
We sent a personalized message with a link to the online survey to all members (n = 110) of a WhatsApp (Facebook, Menlo Park, CA, USA) group of radiology residents practicing in the Eastern Province. The link to the survey was also sent to radiology specialists and consultants whose contact information was available to the investigators. A reminder message was sent three days later. The survey began on April 28, 2019, and was open to respondents for 14 days.
Each invited radiologist received a unique link for the online survey so that the survey could not be filled more than once from the same link. This ensured that the survey would not be compromised by duplicate responses or responses from individuals not included in the target population. Anonymity of the respondents’ identity was maintained using the QuestionPro respondent anonymity assurance feature.
Additionally, a paper-based survey questionnaire was distributed to the radiology departments of hospitals in the surveyed region to reach radiologists whose contact information was unavailable with the investigators. Investigators visited those departments a week later to collect the completed surveys. Overall, the survey was distributed to a total of 263 radiologists using both modes (online and paper).
Content of the questionnaire
The survey was comprised of 25 multiple-choice questions covering the following areas: (1) background demographic information, (2) work-related data, (3) workstation evaluation, and (4) identification of work-related musculoskeletal symptoms. There were questions asking about the physical activity and the methods used by radiologists to generate radiology reports; the results of these questions are outside the scope of the current study and will be reported separately. We conducted a pilot study with a group of 30 radiologists to assess the clarity of the questions and the time needed to complete the survey. After the pilot study, no major changes were made to the questions.
The proposed risk factors were determined based on the literature focusing on individual demographics and work-related information. Background demographic information included data pertaining to age group (<30 years, 30–39 years, 40–49 years, 50–59 years, and ≥60 years), sex, handedness, years of practice (<1 year, 1–5 years, 5–9 years, and ≥10 years), the current institution of practice, and the type of practice (part-time, full-time, or both). Work-related characteristics included the time spent at a computer workstation reviewing medical images (<4 hours, 4–7 hours, 7–9 hours, and >9 hours), the amount and type of imaging studies typically reviewed (plain radiography, ultrasound, computed tomography (CT), magnetic resonance imaging (MRI), fluoroscopy, and nuclear medicine), as well as the duration (<5 min, 5–10 min, 11–15 min, >15 min) and frequency of breaks taken (once, twice, every 2 hours, and at least every hour). Information about the average percentage of time spent on each imaging modality was grouped based on the time distribution (0%, 1–25%, 26–50%, 51–75%, and 76–100%). For example, if a participant spent more than 75% of his or her working time on a particular imaging modality, this modality was considered as the predominantly reviewed modality for the participant. A Likert-like scale (always, often, sometimes, rarely, and never) was used to record data about the performance of the workstation stretching exercises by participants. In addition, the workstation was evaluated in terms of the number of monitors and the adjustability of the height of the workstation and the viewing distance (adjustable vs. nonadjustable).
The standard Nordic Musculoskeletal Questionnaire (NMQ), a valid and reliable screening and surveillance tool , was used to determine which body regions were affected by musculoskeletal symptoms resulting from working as a radiologist. It included the following questions about nine body regions (neck, shoulder, elbow, wrist/hand, upper back, lower back, hip/thigh/buttock, knee, and ankle):
- Have you had trouble (ache, pain, or discomfort) in the last 12 months?
- Have you had trouble in the last 7 days?
- Have you been prevented from carrying out normal activities (e.g., job, housework, or hobbies) due to this trouble in the last 12 months?
In this study, the outcome was the presence of musculoskeletal symptoms in any of the nine body regions, which restricted the performance of normal activities in the last 12 months. The responses of the outcome variables were dichotomized: responses of “left”, “right” or “bilateral” in any body region were coded as a “yes”, whereas a respondent who indicated “no” for all body regions was coded as a “no”.
The obtained data were compiled using the QuestionPro platform and analyzed using IBM SPSS for Windows, version 25 (IBM Corp., Armonk, NY, USA). Two questionnaires were excluded from the analysis because of missing data. All variables used in the study were categorial. Descriptive statistics, such as percentages and frequency distribution of different characteristics, were used as appropriate. In questions based on a Likert-type scale, the responses were combined into two responses. The Chi-square test was used to confirm the bivariate relationship between the explanatory and outcome variables. A multivariable logistic regression analysis was performed to identify independent factors associated with musculoskeletal symptoms. Candidate variables were selected based on biologic plausibility, risk factors that have been identified in the literature, and bivariate analysis results. Manual backward stepwise regression was used to eliminate covariates whose inclusion did not significantly improve the Akaike Information Criterion for the model. In addition, the regression model was tested for multicollinearity using the variance inflation factor statistic. The result of the modeling process was a single multivariate model that encompassed all potential factors relating to the outcome variable. The covariates included age, sex, current institution of practice, and time spent at a computer workstation reviewing medical images. The unadjusted and adjusted odds ratios (OR) with their 95% confidence interval (95% CI) were reported in comparison to the designated referent. When estimating the OR for participants who predominantly review and interpret a particular imaging modality, those who spent their time on different modalities were used as the reference group. The statements of statistical significance were based on a significance level of α = 0.05.