Participation involved workbook preparation in advance of a data collection session that included a focus group discussion and an intervention idea generation collaborative activity. There were 20 participants (16 women, 4 men; with 6 months to 32 years of experience in sonography); each participated in one session. Two sessions were devoted to DMS (9 participants in total), one session to CS (5 participants), and one to VT (6 participants).

Workbooks were designed to elicit information on the participant’s background, organization of their work tasks, and descriptions of their work environment, equipment, and patients. Preparing and submitting workbooks in advance of the session provided participants an opportunity for reflection and priming; reviewing workbook contents in advance allowed researchers to include examples of each participant’s concerns in the discussion.

Each session consisted of four elements: introductions, review of some workbook entries and discussion of selected workbook photographs submitted by participants, (issues) card sorting and voting to identify participants’ most important challenges based on issues they had just discussed, and intervention idea generation. The latter element was conducted from the viewpoint of design as co-creation, in which researchers provide tools to non-designers to use to express their needs and solution ideas creatively. Sanders (2002) described this participatory methodology as a manifestation of ‘a belief that all people have something to contribute to the design process and that they (non-designers) can be both articulate and creative when given appropriate tools with which to express themselves’.

Phase I data were analyzed separately for each type of sonographer. Workbook entries were entered in spreadsheets to facilitate review, researchers made notes from audio recordings of the discussions and generative element presentations, and photos and video were analyzed for thematic and specific content. Through detailed review of these materials (grounded theory analysis), a document was created for each type of sonographer, which was organized by a number of major categories (such as Patient room design, Patient handling, Obesity and positioning, etc.), that emerged from the data. Within each category, every expressed need was listed, along with its source (audio recording, workbook entry, etc.), the interpretation of that expressed need, initial intervention ideas to address the need, an initial rating of the feasibility of prototyping and/or testing the intervention idea, and the researcher’s comments. Two researchers assessed data from each group of sonographers. Each researcher independently completed his/her analysis and then each pair of researchers worked together to develop an analysis for their sonographer group, based on consensus, which they presented to the research team for further analysis, discussion, and, finally, group consensus.

The next step was to develop intervention categories (for example, Reduce strains associated with portable exams – administrative controls; Reduce strains associated with portable exams – engineering controls; Address lack of arm support and awkward postures when scanning; etc.) along with a list of intervention concepts for each category and sonography group. Various types of brainstorming techniques, internet searches, and literature searches were employed to generate intervention concepts. The next step was to refine and reduce the number of concepts to those that were most promising, viable, and within the scope of the study. These concepts were then discussed with the sonographers in a concept review session (Phase 2).

Six DMS, 12 VT, and 6 CS participated in this phase of the study (23 female, 1 male; 1 to 32 years of experience). Eleven participants were new to the study.

Concept posters and some physical prototypes were presented to the participants who used evaluation forms and discussion to convey their opinions of the usability, usefulness, and desirability of each concept, potential barriers they foresaw to using or implementing them, and improvement suggestions. Time was allocated for each concept, which included presentation of the concept to participants, question time, and time for completing the evaluation form. CS and VT groups each reviewed 15 concepts and DMS reviewed 14. Hands-on interaction and role-playing were encouraged where prototypes were present. Once participants completed their evaluation form, the next concept was presented.

Usability items on the evaluation form addressed ease of use and ease of learning to use; usefulness items addressed predicted effects on sonographer’s physical effort, fatigue, and efficiency, and patient comfort; desirability items addressed level of interest in using the intervention by the participant and co-workers. Participants used a 5 point fixed response scale with a verbal anchor at each end ranging from Strongly Agree to Strongly Disagree to respond to each statement. Participants also provided an overall rating of each category, with a verbal anchor at each end, using a 1 (very poor) to 7 (very good) range. Example: ‘My overall rating of its potential usability is ____.’ At the end of the session, participants were asked to indicate which concepts should continue to be developed or pursued. They were each given 6 green voting dots to allocate to whichever concept(s) they chose.

Phase 2 data were analyzed separately for each type of sonographer. Evaluation form entries were entered in spreadsheets to facilitate their review. Median values and spreads of scores were calculated for each statement. The number of green dots and barriers were counted and were normalized to provide counts per participant for each concept.

The first pilot study was a proof of concept, with two male members of the research team serving as patients; neither would be considered nor was classified as overweight (BMI<25). The professional experience of the participating CS ranged from 1 to 20 years (cumulative experience = 82 years). Two were male and four female; two of the women had participated in a previous phase of the study. The CS were asked to obtain four particular scans, first as they normally would and then repeat all four scans using the articulating support arm system (ASAS). Before using the ASAS, each CS was shown how to use the locking mechanism of the arm and given a few minutes to become familiar with the device before using it to obtain images.

The sonographers were asked to think aloud and make comments while scanning in order to provide insight into how such a device might fit into the work system and issues or concerns they had. In addition to recording their verbal comments, data collected in the first pilot study included video recordings of the sessions, scanned sonographic images for quality assessment, the sonographer’s verbal expression of level of exertion, and written assessment of usability, usefulness, and desirability of the articulating arm system using a form similar to the form used during the concept review session, but including several detailed items about each element of the ASAS (transducer holder, articulating arm, locking mechanism).

Image quality was assessed by two highly experienced cardiovascular credentialed sonographers who were blinded to the imaging method. The sonographers provided individual assessments that took into account anatomical structures within the heart, display of pathologic function, proper display of color Doppler, and M-mode display. Their independent results were then compared to determine and, if necessary, resolve any intra-observer differences.

In general, CS were observed performing scanning tasks within a reasonable time frame when using the ASAS, and did not need much time to familiarize themselves with the operation of the device. In addition, while performing the scanning procedures, the CS verbally expressed that they were able to obtain comparable image quality when scanning with and without the device. They manipulated the ASAS with ease and did not appear to have any difficulties moving the arm. In fact, the flexible arm and light weight of the device pleasantly surprised some of the participants. Their verbal comments included ‘It turns really easy’ and ‘it’s easier to use than I thought it would be’. This is important, as limiting the range of motion would result in loss of fine tuning manipulation, which would compromise image quality. Overall, the design of the device was intuitive, even at this initial prototype stage.

Based on observation, scanning with the device appeared to reduce the duration of pinch gripping, forceful exertion, awkward postures, and static postures. A typical work posture when performing a left-handed scan is depicted in Figure 4 (left). In addition to prolonged pinching, the CS has to push the transducer against the patient’s chest. Greater pinch force is required when the ultrasound gel migrates to the gripping surface of the transducer. Additional force is also required if the patient is obese. By utilizing the locking mechanism, the ASAS maintained the location of the transducer and the requisite pushing force against the patient’s chest. As anticipated, this provided an opportunity for temporary breaks from pinching and physical exertion during the measurement phase of the exam (Figure 4, right). Ratings of perceived exertion (on a 0 to 10 scale, presented verbally, where 0 = no exertion at all and 10 = maximum possible exertion) ranged from 4 – 6 without the device and 2 – 3.5 with the device. The participants’ responses to several detailed items on the evaluation form suggest the reasons for these reduced exertion ratings. Median scores for items about their perceptions that the ASAS would “reduce the total amount of time I ________ during a scanning procedure” were as follows: grip transducer, 5.0, am pushing with transducer, 5.0, work in awkward postures, 4.5, work in a fixed posture, 5.0 (where 1= strongly disagree and 5= strongly agree).

Regarding image quality, for three scans (Apical 2 and 3 chamber 2-d, and Apical 4 chamber Doppler), 17 of 18 images collected while using the ASAS were judged to be as good or better than the images acquired without the device. For the fourth scan (Parasternal short axis (PSAX) M-mode) 3 of the 6 images collected using the device were judged to be of lesser quality than images made without the device.

In comparison to average overall ratings (1 – 7 scale) of usability, usefulness, and desirability from the concept review session (4.5, 4.2, 3.9, respectively), ratings were markedly higher from the CS using the functional prototype in the pilot study (6.2, 6.2, 6.1, respectively). Ratings for specific items provide useful information concerning design improvements. For example, some participants did not agree that it was easy to determine the correct orientation of the transducer in the prototype probe holder. The responses to some items revealed some concerns about potential effects of the device on patients including the possibility of the device making the exam more physically uncomfortable for some patients and the possibility of patients being intimidated by the device.

Having confirmed that high quality images could be obtained with the articulating support arm system, the next step was to determine if the system could be used to acquire high quality images in a more representative sample of patients.

One male and five female CS participated in the study. Their professional experience ranged from 1 to 25 years (cumulative experience = 62 years). Five healthy male adults each weighing over 90 kg (200 lbs) and with no affiliation to the study volunteered to act as patients. Based on BMI, three were classified as obese and two as overweight. Actor-patients ranged from under 30 to over 45 years of age. Before participating in the study, all participants reviewed and signed IRB-approved consent forms.

The CS were asked to obtain four particular scans and measurements, first as they normally would and then repeat all four scans using the ASAS. In advance of the scanning session, each CS was given time to become familiar with the ASAS, including how to use the locking mechanism. Two CS scanned the same actor-patient and all the others scanned a unique actor-patient. After scanning the actor-patients, four CS confirmed that their pilot study patient would be considered average or typical in size relative to patients they see in practice. Actor-patients did not report experiencing less comfort or more discomfort when the ASAS was used in comparison to when it was not used; this result addresses an important potential concern raised by some CS in the first pilot study.

Objective data collected during the study included electromyographic (EMG) data from the extrinsic finger flexor group, middle deltoid, and trapezius muscles of the left arm and image capture data to measure trunk lateral flexion and rotation and left arm shoulder abduction and external rotation. Before the actor-patient entered the exam room, Delsys^™ (Natick, MA) single differential electrodes were positioned on the CS’s shaved and alcohol-cleaned skin. A Motion Monitor^™ data acquisition system (Innsport, Chicago, IL) was used to collect the EMG data, which were sampled at 1000 Hz, bandpass filtered between 20 and 500 Hz, and notch filtered at multiples of 60 Hz. EMG data from a reference exertion that simulated pushing a transducer against a patient’s chest with 50 N of force was used for normalizing the EMG data collected during the scanning tasks. Tenth and 50^th (median) percentile values of the normalized EMG data were used to assess static and average levels of activity in the muscles. At the conclusion of the session the CS completed the same usefulness, usability, and desirability evaluation form that was used in the first pilot study.

Images were evaluated by a physician who is board certified in Cardiovascular Medicine and Echocardiography. PSAX M-mode and Apical 3 Chamber Doppler images were graded as good, fair, or poor. Apical 2 and 4 chamber 2-D images were graded on a 1=poor, 3=good, 5=excellent scale using standardized myocardial segmentation (Cerqueira et al., 2002). Ejection fraction was measured from the Apical 4 chamber view. The physician was blinded to condition when performing the image evaluation.

There was no statistically significant difference between grades for scans made with and without the ASAS, based on results from the sign test. The sign test with ties included (Marshall, 2014) was used because, for each CS, the majority of image segment scores were identical between the two conditions (from 63 to 84%). Only one CS did not show improvement in any image segment score with the ASAS, however, the cardiologist noted that the patient had ‘Overall technically difficult echo windows. Would have used contrast clinically.’ There was no pattern in the difference (larger v. smaller) between ejection fraction values with and without the articulating arm (average absolute difference was 5.7%).

Performing each scan was divided into three stages (tasks): orienting the transducer to the desired view (‘window Searching’), Optimizing the image, and making Measurements. During the window Searching stage for the apical views, 10^th and 50^th percentile trapezius m. activity tended to be lower when using the ASAS (Sign test, p<0.002 and p<0.001, respectively), while flexor m. group activity tended to be higher with the ASAS (Sign test, p<0.049 and p<0.002, respectively). During the Optimize phase, 10^th and 50^th percentile activity in all three muscles was reduced after the locking mechanism was engaged during that phase (p values ranged from 0.039 to 0.0005). While making Measurements, 10^th and 50^th percentile activity for all three sampled muscle areas tended to be lower when using the ASAS (p<0.013 for deltoid and flexor m. and p<0.0001 for trapezius m.). Figure 5 illustrates the difference in muscle activity level before and after the ASAS locking mechanism was activated during a scan.

By visual inspection, similar effects on muscle activity were seen for the parasternal view scans, but the small sample size precluded statistical analysis. Parasternal EMG data were not analyzed with apical EMG data because of differences in hand and arm posture when acquiring images from those two locations.

Postures were most notably affected across the group of sonographers during the Measure phase of scanning. For the apical view scans, spine rotation, external shoulder rotation, and shoulder abduction were reduced 6–8 deg (p<0.03), and for the parasternal view shoulder abduction was reduced by 11 deg (p<0.0022) when using the ASAS. These differences may appear small, but Palmerud et al. (2000) demonstrated the sensitivity of intramuscular pressure (IMP; which can affect blood flow in muscle and tendon, as well as nerve function and physiology) in the supraspinatus and infraspinatus muscles to small changes (15 deg. increments) in arm elevation. Reduced force (muscle) exertion combined with a small reduction in shoulder abduction could result in a decrease in IMP below levels that have been shown to be problematic regarding development of muscle fatigue. Lateral spinal flexion was statistically (p<0.028), but not practically greater (2 deg.) when using the ASAS (main effect, across all three scan phases).

Perceptions of usefulness, usability, and desirability of the ASAS were, again, positive overall, with the exception of one individual. Average scores (1 – 7 scale; very poor - very good) for five of the participants were 5.6, 5.6, and 5.8, respectively, but were all 2’s for the one exception. Interestingly, on the evaluation form this CS responded ‘disagree’ to the item ‘I am able to get quality images with this device’, while the cardiologist’s evaluation showed no difference between quality of the images acquired by this CS with or without the ASAS; both received an average score of 2.8, which equated to ‘good’ in the rating system used by the cardiologist. The CS responses on the prototype’s evaluation form provide direction for improvement, including designing the end effector of the ASAS to be comfortable to hold, providing more flexibility in the articulating arm, and providing a way to make fine orientation adjustments at the transducer (distal) end of the ASAS.