ORIGINAL RESEARCH published: 05 June 2019 doi: 10.3389/fpsyg.2019.01247 Support System for the Assessment and Intervention During the Manual Material Handling Training at the Workplace: Contributions From the Systematic Observation Mariona Portell 1* , Anna M. Sene-Mir 2 , M. Teresa Anguera 3 , Gudberg K. Jonsson 4 and José L. Losada 5 1 Department of Psychobiology and Methodology of Health Sciences, Universitat Autònoma de Barcelona, Cerdanyola del Vallès, Spain, 2 Physical Activity and Sports Studies Centre, University of Vic – Central University of Catalonia, Vic, Spain, 3 Faculty of Psychology, Institute of Neurosciences, University of Barcelona, Barcelona, Spain, 4 Human Behavior Laboratory, University of Iceland, Reykjavik, Iceland, 5 Faculty of Psychology, University of Barcelona, Barcelona, Spain Edited by: Efficacy of classical manual material handling (MMH) training interventions on back pain María Angeles Peláez-Fernández, prevention at the workplace has been called into question. The way that observation University of Málaga, Spain (self-observation or hetero-observation) is used in other areas to create feedback Reviewed by: addressed to modify motor activities can justify innovative components for these Antonio Hernández-Mendo, University of Málaga, Spain interventions. However, their implementation and evaluation cannot be done without África Borges, tackling the methodological challenge of developing a reliable observational instrument Universidad de La Laguna, Spain to measure manual handling practice during the training process. The aims of this *Correspondence: Mariona Portell study were: (1) justify and develop an hetero-observation (H-O) instrument to assess mariona.portell@uab.cat changes in the worker behavioral patterns with a level of analysis convenient to derive a parallel version for the systematic self-observation (S-O) during training on MMH; (2) Specialty section: This article was submitted to provide evidence on the inter-rater reliability of the H-O instrument; (3) provide evidence Quantitative Psychology on the usability of the S-O instrument and its perceived usefulness; and (4) provide and Measurement, evidence on the benefits that can be derived with the use of the H-O instrument to a section of the journal Frontiers in Psychology create feedback based on T-pattern and polar coordinate analysis. A mixed method Received: 23 February 2019 approach mainly grounded on systematic observation was used. A convenience sample Accepted: 13 May 2019 composed by blue-collar workers participated in the study. Based on literature review Published: 05 June 2019 and expert opinion, the H-O instrument proposed was composed by six dimensions Citation: Portell M, Sene-Mir AM, (feet, knee joints, back, elbow joints, load position, and interaction between back Anguera MT, Jonsson GK and tilt and displacement) plus a structural dimension which defined MMH phases. The Losada JL (2019) Support System inter-rater reliability of this instrument was almost perfect for all dimensions using a for the Assessment and Intervention During the Manual Material Handling tolerance level of 2 s (the range of time-unit kappa was from 0.93 to 0.97 and the Training at the Workplace: range of event-based kappa was from 0.82 to 0.9). The usability and usefulness of Contributions From the Systematic Observation. Front. Psychol. 10:1247. the S-O instrument was highly valued by workers. Regarding the way to use hetero- doi: 10.3389/fpsyg.2019.01247 observations to create feedback, the paper shows the great potential of T-pattern Frontiers in Psychology | www.frontiersin.org 1 June 2019 | Volume 10 | Article 1247 Portell et al. Systematic Observation at the Workplace and polar coordinate analysis. The observational instruments developed combined with these techniques make it possible to characterize the body positions adopted during manual handling performance, and this is crucial to create feedback on performance instead of only feedback on results. Keywords: systematic observation, health promotion in the workplace, T-pattern analysis, polar coordinate analysis, mixed method approach, manual material handling, feedback INTRODUCTION Janelle et al., 1997; Dowrick, 1999, 2012; Wulf et al., 2010; Magill, 2011; Ste-Marie et al., 2013). Previous research provided evidence Musculoskeletal disorders (MSDs) are one of the most common on the effectiveness of the feedback derived of self-observation occupational disorders worldwide (Schneider and Irastorza, in the learning of motor skills (Salmoni et al., 1984; Newell, 2010; Hossain et al., 2018; Penkala et al., 2018). Lifting and 1991; Kernodle and Carlton, 1992; Janelle et al., 1997; Silva handling of loads have been associated with an increased risk et al., 2016). Moreover, starting with an hetero-observation to of back disorders, mainly developing low back pain (Bernard, provide information on what is well done, what should be 1997; Hoogendoorn et al., 2000; Thiese et al., 2014). Even improved, and how it should improve (these last two pieces of though the majority of the tasks have been automated, manual information are known as feedforward) has proved to have a material handling (MMH) is still being carried out in numerous positive effect in order to modify different behaviors in different workplaces (e.g., building sites, nursing, or the food industry), settings (Newell, 1991; Kernodle and Carlton, 1992; Buggey, and what is more, there are plenty of daily life activities in 2007; Dowrick, 2012; Mason et al., 2016). As far as we know, which people may perform MMH during non-working hours the combination of feedback based on self-observation (the (e.g., lifting or carrying people or objects such as boxes). observed and the observer are the same person) and hetero- Throughout the last decades, there has been an increase of observation (the observed and the observer are different people, studies focused on developing, implementing and evaluating the e.g., worker and technician) has not been incorporated in the efficacy of MMH training on the reduction of low back pain occupational MMH training. Based on the studies reviewed, or back injury prevention (Punnett et al., 2009; van der Beek these components could be particularly beneficial to improve the et al., 2017). The systematic reviews of Clemes et al. (2010), training of workers with special responsibilities in fostering safety Verbeeck et al. (2011), and Hogan et al. (2014) agreed that most of habits in workplaces (e.g., supervisors). MMH trainings are not effective to reduce low back pain or back On the issue of observational instruments, there are few injury. This lack of effectiveness has been related to the focus on studies specifically evaluating behavioral change in manual a task-specific training instead of a multidimensional approach handling, and they generally rely on observational methods (Clemes et al., 2010). Also, it has been related to the use of developed with the aim of assessing MSD risk (e.g., Snook and interventions not based on a behavioral change theory (Verbeeck Ciriello, 1991; Daltroy et al., 1993; Waters et al., 1993; Best, 1997; et al., 2011; Hogan et al., 2014). Most studies have focused on Lortie and Baril-Gingras, 1998; Nygård et al., 1998; Monnington evaluating the effectiveness of training on long-term results, such et al., 2002; Chang et al., 2003; Marras and Karwowski, 2006; as reduction of MSDs, and there are few rigorous evaluations of Batish and Singh, 2008; Village et al., 2009). However, these kinds the effect of MMH training on key intermediate variables of the of instruments do not allow for the assessment of behavior change changing health behavior theory, such as knowledge or behavior in the way of characterizing body positions adopted during the change (Hogan et al., 2014). The inclusion of these intermediate process of manual handling execution, and this is crucial to create variables is necessary to evaluate the implementation process, feedback on performance instead of only feedback on results which in turn is necessary to identify why an intervention (Salmoni et al., 1984; Kernodle and Carlton, 1992; Janelle et al., worked or not, and under which operating conditions these 1997; Magill, 2011). interventions are likely to be most effective (Pedersen et al., 2012). The general purpose of this study was to develop and justify Another limitation is the lack of transferability of the intervention observational instruments that combine two uses during manual effectiveness from the training situation to other labor and non- handling training on back pain prevention. On the one hand, labor settings; and this lack of transferability has been related change assessment in worker behavioral patterns. On the other to the components of those interventions (Ford et al., 1998; hand, its use as a source of feedback based on systematic self- Clemes et al., 2010; Hogan et al., 2014). Solving some of the observation by workers and the hetero-observation of workers aforementioned constraints cannot be done without tackling a by a technician. The development of observational instruments methodological challenge related to the development of reliable was based on observational methodology (Anguera, 2003) observational instruments to measure manual handling practice and will be reported following the Guidelines for Reporting during and after training (Hogan et al., 2014). Evaluations Based on Observational Methodology (GREOM) On the transferability issue, the way that observation (self- (Portell et al., 2015; included in the EQUATOR library). This observation or hetero-observation) is used in other areas methodological approach allows the systematic analysis of to create feedback addressed to modify motor activities can spontaneous behavior in a natural environment using a set of justify innovative approaches (Kernodle and Carlton, 1992; dimensions and categories to evaluate behavior changes and Frontiers in Psychology | www.frontiersin.org 2 June 2019 | Volume 10 | Article 1247 Portell et al. Systematic Observation at the Workplace temporal patterns using T-pattern analysis (Magnusson, 1996, methodology (Anguera et al., 2001; Portell et al., 2015), the 2000, 2018; Casarrubea et al., 2018), lag sequential analysis design used can be classified as nomothetic, follow-up and (Bakeman and Quera, 2011; Quera, 2018) or polar coordinate multidimensional. It is nomothetic because we conducted a analysis (Sackett, 1980; Morillo et al., 2018; Rodríguez-Medina parallel analysis of a group of workers. We classify as incomplete et al., 2018). Observational methodology is well established follow-up given the observation during two training sessions in several fields (Blanco-Villaseñor et al., 2010; Anguera and and the intra-sessional follow-up (intensive or continuous Hernández-Mendo, 2016; Arias-Pujol and Anguera, 2017; Cerezo monitoring of events throughout observation sessions in order et al., 2017; Escolano-Pérez et al., 2017; Izquierdo and Anguera, to obtain behavioral dynamic indicators or sequential data). 2018; Pérez-Tejera et al., 2018; Santoyo and Mendoza, 2018). The multidimensional nature of the design was determined Nonetheless, their use to provide feedback as an intervention for the multiple criteria included in the purpose-designed component within an occupational MMH training, as well as observation instrument. Additionally, the quantitative approach their use to assess behavioral change derived from training was complemented with data collected from a cross-sectional is innovative. We use the term SsObserWork (Systematic design in order to obtain evidence of usability and usefulness. Self-Observation of Work) to label our application of the observational methodology for workplace health promotion. Observation Instruments Development The SsObserWork design and evaluation can be seamlessly The SsObserWork instruments were developed in three phases. integrated into the research framework for the development In the first phase, a review of the scientific literature was and implementation of interventions preventing work-related conducted to establish the research background. We searched the musculoskeletal disorders proposed by van der Beek et al. Scopus, Web of Science, PubMed and Google Scholar databases (2017). Specifically, the aims of this paper were: (1) justify for studies on the MMH training phases (Hsiang et al., 1998; and develop an hetero-observation instrument with a level of Plamondon et al., 2014), and mostly studies on the effect of analysis convenient to derive a parallel version for systematic the back position (Anderson and Chaffin, 1986; McGill, 1997; self-observation during the training on MMH technique; (2) Larivière et al., 2000; Straker, 2003b; Mörl et al., 2005; Kahrizi provide evidence on the inter-rater reliability of the hetero- et al., 2007; McGill, 2007; Mawston and Boocock, 2012), feet observation instrument; (3) provide evidence on the usability of position (Kirby et al., 1987; Authier et al., 1996; Kingma et al., the self-observational instrument and its usefulness perceived; 2004; Kingma et al., 2006; Demaret et al., 2007; Zhou et al., and (4) provide evidence on the use of the hetero-observational 2013), knee position (Hagen et al., 1993; Straker, 2003a,b; Kingma instrument to create feedback based on T-pattern analysis and et al., 2004; Mörl et al., 2005), arms and load position (Hsiang polar coordinate analysis. et al., 1997; Marras et al., 1999; Demaret et al., 2007; Faber et al., 2009; Colombini et al., 2013) on MMH and their impact on the risk of back disorders. Based on this review and considering METHODS the objective to establish an instrument similar for hetero- observation and systematic self-observation, the granularity of Concerning the methodological approach, this work is enclosed the codes (Bakeman and Quera, 2011; Anguera et al., 2018a) within the mixed methods perspective that are characterized by was established. It was done considering that the procedure of the integration of qualitative and quantitative elements. This modeling movements by observation with the objective of a integration was carried out from the “connect” option (Creswell performance description can vary from a micro to macro (or and Plano Clark, 2011). Systematic observation grounded in molecular to molar) level of granularity or specificity. On one observational methodology (Anguera, 2003) was applied, because side, codes can be created to capture minimum details (e.g., it was suitable in relation to the proposed aim. Recently it repeated measurement of movements during the process based has been considered in scientific literature that observational on sensor-based high-resolution) while on the other end of the methodology studies, in which the QUAL-QUAN-QUAL macro continuum they can be relatively broad (e.g., general assessment stages take place, can also be considered mixed method of balance and coordination at the end of the process). We studies in certain circumstances, as they apply unconventional selected a medium level of granularity for the decomposition into approaches (i.e., not based on frequency counts) to quantitize categories of the MMH process. With this level of granularity, qualitative data (Sánchez-Algarra and Anguera, 2013; Anguera the instrument for worker hetero-observation will be able et al., 2017, 2018b). Researchers have been mixing qualitative to generate sequential data of their performance. Moreover, and quantitative approaches from the last 20 years. Many this level of granularity seems molar enough to allow us to researchers do not mix qualitative and quantitative approaches obtain a parallel version understandable for the worker during in optimal ways, but qualitative techniques can be used to the implementation of the systematic self-observation; we will enhance the development of quantitative instruments and vice refer to this version as self-observation instrument, hereinafter versa. Its potential is very broad, and includes instrument S-O instrument. fidelity, “maximizing the appropriateness and/or utility of In the second phase, the observational context was established, the instruments used, whether quantitative or qualitative” and a pilot testing of a preliminary version of the hetero- (Onwuegbuzie et al., 2010, p. 57). observation (H-O) instrument was performed. The instrument We used systematic and direct observation. According was developed focusing on a standard MMH task in a training to the possible study designs, described in observational context in which attention is paid to lifting, carrying and lowering Frontiers in Psychology | www.frontiersin.org 3 June 2019 | Volume 10 | Article 1247 Portell et al. Systematic Observation at the Workplace actions done spontaneously by the employee from a sagittal scheme used took into account the position of the box within plane. The recording requirements were established based on session and between sessions. This sampling scheme ensured that a previous review (Kilbom, 1994; da Silva et al., 2014). The a minimum of 77% of the participants were represented on this instrument was developed as a combination of field format and box-manipulation sample. category systems, and it was created with nine dimensions. For Regarding the study on usability of the S-O instrument, the first eight dimensions we built category systems to codify the the participants were a subsample of 27 workers who used positions of different body segments and sub-segments: feet, feet the S-O instrument in two separate self-observation sessions. with respect to floor, knees, back, arms with respect to legs, arms Participation in this study was voluntary, but the opportunity with respect to back, elbows, and shoulders. The last dimension to take part in two self-observation sessions was randomly was a catalog of codes to identify the phases. A standardized established. The participant subsample were 10 men (37%) and training manual was developed with definitions and diagrams 17 women (63%), of which 41% were between 29 and 39 years old, for each category and code. This first version was applied in 37% were between 40 and 50 years old and 22% were more than a pilot study implemented in a company in the metallurgical 50 years old; no statistically significant differences were observed sector. Pilot testing of this preliminary version was performed by on these variables regarding the general sample. two observers trained in ergonomics who applied the instrument The Ethics Committee of the Universitat Autònoma de to the observation of 16 workers doing 160 MMH tasks. Data Barcelona approved the study protocol. In accordance with quality was evaluated by an inter-observer reliability analysis, the principles of the Declaration of Helsinki, participants were using conventional Cohen’s Kappa agreement index. The Kappa informed that they were being filmed. They were shown the values of the dimensions ranged from 0.39 to 0.75, with a median location of the video cameras, which were positioned discretely of 0.65, indicating moderate agreement (Landis and Koch, 1977). to minimize reactivity bias. Informed consent was also obtained. In the third phase, the H-O instrument and the manual were optimized and the self-observation (S-O) version was Instruments created. Four external experts in different areas took part in According to the GREOM (Portell et al., 2015), the reporting this process: a physiotherapist specialized in patient transfer, of systematic observation studies must clarify the distinction two specialists in occupational medicine and safety, ergonomics, between observation instruments (i.e., purposed-designed psychology and industrial hygiene, and an expert in physical instruments to analyze a given participant; where development activity and sports science, specifically in physical activity at is the main objective of this paper and it will be presented work. These experts participated in qualitative interviews on as a part of the results) and recording instruments (i.e., the appropriateness and representativeness of the dimensions, those used to record and code data according to the categories and training context. The interviews were organized dimensions established by the observational instrument). as an iterative process whereby each interview informs the next, In this study, the recording instrument was LINCE (v.1.2.1) and subsequent interviews are used to explore the weaknesses (Gabin et al., 2012; Hernández-Mendo et al., 2014). Box raised in previous interviews (Brod et al., 2009). Interviews were manipulation was registered with Sony HD video cameras conducted by MP and AS-M. During this phase, the number and videotapes were transferred to Toshiba Portege of dimensions and categories was reduced, the definitions were laptops for their codification. The following computer improved, the classification criteria to define supra-categories software was used to perform the data analysis: GSEQ (recommended or non-recommended to prevent health risk) (v.5.1) (Bakeman and Quera, 2011), HOISAN (v.1.6.3.3) were established, recording rules were clarified, and the observers (Hernández-Mendo et al., 2012), and THEME (v.6.0 Edu) training process was improved. (Magnusson, 2000). Participants and Samples Procedure A convenience sample was initially composed by 53 blue-collar As a part of a multicomponent training in health promotion workers of a food processing company in Catalonia (Spain), at the workplace, workers were video-recorded while having who were interested in participating in two sessions of a to lift, carry and lower 5 boxes (8 kg each). The video camera multicomponent training on health promotion in the workplace. was positioned at the workers’ hip height and the plane The sample participants were 24 men (45%) and 29 women was sagittal. All observations took place in company spaces (55%), of which 4% were between 18 and 28 years old, 38% adapted to training. After MMH performance and recording, a were between 29 and 39 years old, 30% were between 40 and 50 subsample of 27 workers observed their own performance years old and 28% were more than 50 years old. Participants did using the S-O instrument. During this self-observation not have chronic bone, muscle or joint pathology in the trunk, task the technician provided feedback and feedforward knees, nor chronic or acute pain diagnosed by a specialist. While considering the classification associated to H-O and S-O attending training sessions, workers were video-recorded during instruments (Table 1). These workers repeated the MMH box manipulation (see next section). performance recording and systematic self-observation task On the reliability study of the H-O instrument the sampling 3 weeks later. At the end of this second systematic self- unit was the box-manipulation. From a total number of 530 boxes observation task, workers were required to answer five manipulated, 84 box-manipulations were randomly selected for questions on usability and usefulness perceived of the S-O the reliability study. The hierarchical stratification sampling instrument, which were adapted from previous studies Frontiers in Psychology | www.frontiersin.org 4 June 2019 | Volume 10 | Article 1247 Portell et al. Systematic Observation at the Workplace (Tsakonas and Papatheodorou, 2008; Zapata et al., 2015). the sample). This technician codified the 84 units, acting The usability questions explore the understandability of the as an independent observer, and creating a second block of terminology, images, aesthetic appearance and layout. The codes (block C). usefulness questions explore the worker’s perceptions regarding the instrument’s ability to improve their knowledge on MMH Data Analysis technique and their behavior during MMH. A 10-point response Time-Unit Kappas and Event-Alignment Kappas scale was used for all questions (from 1 – very low – to Considering the purpose of this paper, we used a demanding 10 – very high). approach to the agreement study, emphasizing observer For the reliability study of the H-O instrument we adapted agreement regarding the data collected (not with scores derived a double approach justified by Arana et al. (2016). A first from such data). For each category system, data were recorded block of codes (block AB) was generated by two members as time-event sequential data (Bakeman and Quera, 2011), that (AS-M, MP) using a process of qualitative consensus is, mutually exclusive and exhaustive (ME&E) codes for each agreement in the application of the H-O instrument dimension had been assigned to events as they unfold over time to the 84 units. Parallel to this, a senior technician for (micro-coded). We applied the agreement study algorithms work hazard prevention (who did not participate in the for time-event sequential data proposed by Bakeman et al. instrument development) was trained in the use of the (2009): time-unit kappas and event-alignment kappas. Both H-O instrument for 9 h with specific materials (the box- time-unit and event-based kappas were computed using GSEQ manipulation used for this training was not included in (v.5.1) (Bakeman and Quera, 2011; Quera, 2018). Time-unit TABLE 1 | Dimensions and category systems of SsObserWork instruments: hetero-observation (H-O) and self-observation (S-O) version. H-O instrument Classificationa S-O Category systems Code (Recommended -R- or Non- instrumentb Recommended -NR-) Dimension 1. Feet Symmetric feet behind the load p1 NR, during LF and LW phases. X Asymmetric feet behind the load p2 NR, during LF and LW phases. X Symmetric feet beside the load p3 R, during LF and LW phases. X One foot beside the load and the p4 HR, during LF and LW phases. X other behind it Walking ppv Dimension 2. Knee joints Extension – slight flexion rex X Moderate flexion rmo R, during LF and LW phases. It is just considered NR when X the highest position and the upright (0 cm) position are concurring. Maximum flexion rsq NR, during all MMH phases. X Walking rcv Dimension 3. Back Neutral tne R, during all the MMH phases. X Flexion tF NR, during all the MMH phases. X Maximum flexion thip NR, during all the MMH phases. X Extension tEx NR, during all the MMH phases. X Dimension 4. Elbow joints Extension – slight flexion b1 R, during all the MMH phases. X Flexion b2 NR, during all the MMH phases. X Dimension 5. Load position Close to the body ap R, during all the MMH phases. X Separated from the body se NR, during all the MMH phases. X Dimension 6. Interaction between back tilt and displacement Tilt at 0 cm sin R, during LF and LW phases. Tilt at > 0 cm non NR, during all the MMH phases. Upright at 0 cm f13 R, during the highest position of LF and LW phases. Upright at > 0 cm anda R, only during CA phase. a The classification as recommended (R) or non-recommended (NR) is defined in relation to the MMH code phases: LF-Lifting [with three subphases: initial position (f1a), lowest position (f1b), highest position (f1c)]; CA-Carrying (f2); LW-Lowering [with three subphases: highest position (f3b), lowest position (f3c), final position (f3d)]. b The check mark (X) indicates the presence of the category in the S-O instrument. Frontiers in Psychology | www.frontiersin.org 5 June 2019 | Volume 10 | Article 1247 Portell et al. Systematic Observation at the Workplace kappa examines interrater agreement in time (i.e., how long involves the detection of significant associations between a agreement and disagreement lasted). On this data we consider behavior of interest (referred in polar coordinate analysis as focal it acceptable not to count minor errors of timing on the order behavior) and other behaviors (referred as conditional behaviors). below 0.5 s and we define five levels of tolerance: 0.5, 1, 1.5, Polar coordinate analysis is based on the complementarity and 2 s. Event-alignment kappa examines interrater reliability between two analytical perspectives: prospective and on code order for timed events (i.e., onsets and durations). In retrospective, concerning the focal behavior as zero point. this procedure, GSEQ aligns codes using a predefined algorithm The Zsum statistic provides a representative value for a series of (Bakeman et al., 2009) and examines agreements, omission and independent values (adjusted residuals at different prospective commission errors. We also use the mentioned five levels of or retrospective lags) to produce prospective and retrospective tolerance and an 80% event overlap. Zsum values. Prospective Zsum values are represented in X axis, and Zsum retrospective values in Y axis. The resulting values T-Pattern Detection and Analysis and their sign (positive or negative) determine the quadrant T-pattern analysis was proposed and developed by Magnusson in which the different vectors are located and indicate their (1996, 2000). This analysis allows for detection of hidden or respective lengths and angles (Sackett, 1980). The value of angles nonobvious temporal patterns in behavior that are not always implies a necessary adjustment of the trigonometrical value, visible. The detection algorithm first identifies significant (non- as a function of the quadrant. Vectors provide information random) recurrences of any two events (in T-pattern analysis on the nature of the relationship (prospective/retrospective “event” refers to the configuration of codes from each dimension and activation/inhibition) between a focal behavior, which is of the observational instrument) within a similar temporal equivalent to a given behavior in lag sequential analysis, and each configuration (critical interval) in real-time behavioral data conditional behavior that we have proposed in our study (see and then proceeds to identify hierarchical relationships with results section). We used the genuine retrospective approach. any other antecedent or subsequent events. T-pattern analysis The concept of genuine retrospectivity (Anguera, 1997) was involves the use of an algorithm that calculates temporal distances introduced at a later stage to improve the classic concept of between behaviors and analyzes the extent to which critical retrospectivity (Sackett, 1980). This approach considers negative intervals remain invariant relative to the null hypothesis, that lags from a backwards rather than a forward perspective, i.e., it each behavior is independently and randomly distributed over looks at what happened from lag 0 back to lag −1 rather than time. The algorithm developed by Magnusson (1996, 2000) has from lag −1 to lag 0, and the same in successive lags. Sackett been implemented in the THEMETM software (Patternvision (1980) recommended using the same number of prospective Ltd., Iceland). Data from this study was analyzed using Theme and retrospective lags. Based on experience to date (Sackett, 6.0 Edu. Once T-patterns have been detected it is possible to 1987; Anguera et al., 2018c), five prospective lags and five use this new information in different ways (Casarrubea et al., retrospective lags (−5 to +5) were analyzed. The meaning of 2015, 2018). One approach is based on the analysis of pattern the vectors varies according to the quadrant in which they sets and another approach is to analyze patterns individually (e.g., are located, and the position in one quadrant or another is Castañer et al., 2017). This second approach is the one used on determined by the combination of positive or negative signs this paper. From that approach it is crucial to be transparent on prospective and retrospective Zsum values. In quadrant I regarding the qualitative and quantitative filters used to select (+ +), the focal and conditional behaviors activate each other; T-patterns to suit the objective of the analysis (Amatria et al., in quadrant II (− +), the focal behavior inhibits and is activated 2017). Considering the purpose of our T-pattern selection, which by the conditional behavior; in quadrant III (− −), the focal is to help employees analyze their own MMH in a better way, and conditional behaviors inhibit each other; and in quadrant the filters used were first quantitative and secondly qualitative. IV (+ −), the focal behavior activates and is inhibited by the The quantitative filters were: (1) minimum one occurrence of conditional behavior. Vector length indicates the strength of the pattern on each MMH, (2) frequency of occurrence higher the association between focal and conditional behaviors. The than 3; (b) significance level of 0.005 (0.5% probability of critical HOISAN program (v.1.6.3.3) (Hernández-Mendo et al., 2012, interval being due to chance). The qualitative filters were (applied 2014) was used to calculate adjusted residuals, Z-values, and as a lexicographic decision rule): (1) maximum length (number vector length and angles; the program includes a feature to of event-types in the terminal string of a pattern), (2) maximum produce results in graph form. level (number of hierarchical levels in a pattern), (3) different event-types (configuration of codes), (4) items (codes) related to recommendable position, (4) maximum duration. The results RESULTS were validated by simulation, through data randomization on five occasions, accepting only patterns for which the probability of randomized data coinciding with real data is zero. SsObserWork Systematic Observation Tools: H-O Instrument and Polar Coordinate Analysis S-O Instrument Polar coordinate analysis, which was proposed by Sackett Based on literature review and expert opinion, the H-O (1980), combines adjusted residuals from lag sequential analysis instrument proposed was a combination of a field format and (Bakeman, 1978) and the Zsum statistic (Cochran, 1954). It category systems (Anguera, 2003; Portell et al., 2015). It was Frontiers in Psychology | www.frontiersin.org 6 June 2019 | Volume 10 | Article 1247 Portell et al. Systematic Observation at the Workplace composed of six dimensions (feet, knee joints, back, elbow of the 21 categories included in the H-O instrument; thus, 95% joints, load position, and interaction between back tilt and of the categories presented kappa values indicating good to displacement) and twenty-one categories (Table 1). Additionally, very good agreement. these six dimensions had a formal category null (empty set) Regarding the results obtained from time-unit kappas for which marks off an unobservable action. The set of categories each criterion, the kappa values ranged from 0.90 (for the corresponding to each dimension met the requirements of back criterion at 0.5 tolerance level) to 0.99 (for the knee ME&E. Additionally, there was a structural dimension which joints criterion at 2 tolerance level), indicating very good defined MMH phases. In Table 1, dimensions and categories agreement according to the criteria of Landis and Koch are shown (details in Supplementary Table S1). Based on the (1977) and Altman (1991). Regarding time-unit kappas for the scientific review described in the previous section, the twenty- categories (Supplementary Table S3) and considering the higher one categories were classified according to their effect on health. demanding tolerance level (0.5 s), the kappa values ranged from On the third column of Table 1, the derived classification as 0.53 (for the “symmetric feet beside the load” category included a recommended or non-recommended position to be adopted in the criterion feet) to 0.99 (for the “walking” category included during MMH is summarized. This data was the base to create in the criterion feet), indicating moderate to very good agreement the modified version of the H-O instrument to construct the S-O according to the criteria of Landis and Koch (1977) and Altman instrument and to create feedback. The last column of Table 1 (1991). The values indicating moderate agreement happened in indicates the categories included on the S-O instrument with a just one of the 21 categories included in the H-O instrument; check mark. The S-O instrument kept all dimensions except the thus, 95% of the categories presented kappa values indicating most complex one (interaction between tilt and displacement) good to very good agreement. and all categories were defined graphically. Both instruments can be used to observe the MMH task Usability and Usefulness Perceived of in formative training addressed to improve MMH in diverse the S-O Instrument work environments at a range of industrial sites. The application Table 3 reports the median, mean and standard deviation of the requirements were: (1) a conventional video camera positioned usability and usefulness perceived by workers who apply the S-O at the workers’ hip height; and (2) the instrument to record instrument during two separate sessions. The last column shows observational data. Figure 1 shows space disposition for MMH the results of the one-tailed sign test using 7.5 as a cut point. For in formative context and Figure 2 shows the interface used to all the studied aspects, the median was statistically greater than codify. The observational instruments, the record instrument 7.5 (on a scale of 1–10, being 10 the best). to automatize codification and the training program could be facilitated upon request. Average time required to apply the observation instruments Use of the H-O Instrument to Create were (time by box in the generic setting illustrated in Figure 1): Feedback Based on T-Pattern Analysis 9 s for MMH recording, 6 min for the codification of the H-O and Polar Coordinate Analysis instrument by the technician, and 3 min for the codification of To illustrate the functionality of the H-O instrument to create the S-O instrument by the worker. feedback on performance, two workers were selected (we refer to them as WO1 and WO2), both males, aged between 40 and 50 years old, who have equivalent values in terms of recommended Inter-Rater Reliability Testing of the position when they are examined globally throughout the session H-O Instrument (global values between P40 = 53.3% and P50 = 54.8% of the The values of agreement statistics across observers (AB register distribution of relative durations). Using the GREOM terms vs. C) for each dimension are summarized in Table 2. (Portell et al., 2015; Chacón-Moscoso et al., 2018), duration Supplementary Tables S2, S3 list the values of agreement is a static behavioral indicator, but the granularity of the statistics for each category. In addition to kappa values, Table 2 H-O instrument is be able to define additional and more includes the kappa maximum and percentage of agreement. complex dynamic indicators. Based on these behavioral dynamic Regarding the results obtained from event-based kappas indicators, joined with the possibilities of sequential analysis for each dimension, the kappa values ranged from 0.72 (for techniques, Figures 3–5 illustrate the new information that can elbow joints criterion at 0.5 tolerance level) to 0.9 (for feet emerge from the observational record. criterion at 1.5 tolerance level), indicating good to very good For each worker, Figures 3, 4 show the two most relevant agreement according to the criteria of Landis and Koch (1977) T-patterns detected by applying the selection criteria described and Altman (1991). Regarding event-based kappas for the in the method section. For each T-pattern two diagrams are categories (Supplementary Table S2) and considering the higher presented. Firstly, the tree graph pattern (Figures 3A.1,B.1, demanding tolerance level (0.5 s), the kappa values ranged from 4A.1,B.1) shows which event types (configurations of concurrent 0.5 (for the “symmetric feet beside the load” category included in codes) are included in the pattern and how they are connected. the feet criterion) to 0.97 (for the “walking” category included in Secondly, the instance graph (Figures 3A.3,B.3, 4A.3,B.3) the feet criterion), indicating moderate to very good agreement provides information about the real-time pattern structure (the according to the criteria of Landis and Koch (1977) and Altman time period represented only includes the observation periods (1991). Values indicating moderate agreement happened in one in which the worker loads, displaces, or unloads each load). Frontiers in Psychology | www.frontiersin.org 7 June 2019 | Volume 10 | Article 1247 Portell et al. Systematic Observation at the Workplace FIGURE 1 | Space disposition for MMH in formative context. FIGURE 2 | Sample interface used to codify. Left side shows the recording instrument. Right side shows the visual part of one category definition in the H-O instrument. Written informed consent was obtained from the depicted individual for the publication of these images. On the right side of the tree graph pattern, the pictures I, II, III, or IV). Under each graph, the table with the polar highlight configuration details (for confidentiality reasons the coordinate analysis results is presented for the statistically workers’ images have been replaced by photographs from the significant associations. H-O instrument). The qualitative information which can be deduced from these Figure 5 summarizes the results of polar coordinate analysis analyses is presented below to illustrate how they provide useful for each worker. Each graph represents the statistically significant resources to improve the informative quality of the feedback and associations (activation or inhibition) between the focal and feedforward on worker performance. conditional behaviors. The focal behavior on this analysis was Regarding worker WO1, the first T-pattern analysis was “neutral back position.” The conditional behaviors were the applied at 693 micro-coded observational data which were categories of the dimensions feet, knee joints, elbow joints and structured in 99 configurations (33 different event-types). Taking load position, after applying two kinds of code transformations into account all the dimensions of the H-O instrument, the based on the classification criteria included in Table 1. The most complete pattern identified for 100% of the MMH first was a merger of categories of the feet dimension and the performed only showed a regular structure in event-types second the creation of new superordinate codes combining the related to the lifting phase (Figures 3A.1,A.3). This pattern knee joints and phase dimensions (the new codes have been also showed that knee joints, elbow joints and interaction are created by joining the codes of the original categories that dimensions that remain in a recommended position during were already defined in Table 1). The association is shown this regular pattern of behavior. Theme also allows more both quantitatively (vector length) and qualitatively (quadrant detailed analysis of a part of the dimensions. In this case, Frontiers in Psychology | www.frontiersin.org 8 June 2019 | Volume 10 | Article 1247 Portell et al. Systematic Observation at the Workplace position analysis related to upper body dimensions was applied was about the feet (code P1P2) in quadrant III, where the vector (Figures 3B.1,B.3). The most complex pattern obtained for represents a mutually inhibitory sequential relationship between this subset of configurations, which is significant for 100% a neutral back position and a non-recommended position of the of the MMH performed, connects the way of performing feet during lifting and lowering phases. In quadrants II and IV the lifting with load displacement, with lowering performance two other strong relationships can be seen, both of which are not being structured enough for it to be incorporated into of asymmetric dependence. The first relationship pointed out the pattern. It is also worth noting the regularity of the co- is that the recommended position of the back inhibited load occurrence between the neutral back position and the position position close to the body, although load position close to the close to the body detected by Theme (see the biggest picture in body activated the neutral back position. The relationship shown the Figure 3B.2). in quadrant IV is the mirror image: the recommended position of Polar coordinate analysis showed that the neutral back the back activated the load position separated from the body and position of the worker WO1 (Figure 5A) was sequentially this last inhibited the recommended position of the back. associated with the dimensions: feet, knee joints, elbow joints Regarding worker WO2, the first T-pattern analysis was and load position. From this analysis, the most interesting applied at 525 micro-coded observational data which was relationship that can be presented as feedback to worker WO1 structured in 75 configurations (30 different event-types). Taking TABLE 2 | Agreement between register AB and C for each dimension, presenting kappa based on events and kappa based on time units for each tolerance level, as well as percentage of agreement and kappa maximum. Based on events Based on time units Dimension Tolerance level % agreement Kappa Kappa max. % agreement Kappa Kappa max. Feet 0.5 92 0.87 0.93 97 0.92 0.95 1.0 93 0.89 0.94 97 0.93 0.95 1.5 94 0.90 0.94 97 0.93 0.95 2.0 94 0.90 0.94 97 0.93 0.96 Knee joints 0.5 89 0.86 0.97 99 0.97 0.98 1.0 90 0.87 0.97 99 0.98 0.99 1.5 91 0.87 0.97 99 0.98 0.99 2.0 91 0.87 0.97 99 0.99 0.99 Back 0.5 89 0.83 0.94 94 0.90 0.97 1.0 92 0.88 0.95 96 0.93 0.97 1.5 91 0.87 0.95 97 0.94 0.97 2.0 92 0.88 0.95 97 0.94 0.97 Elbow joints 0.5 86 0.72 0.88 96 0.92 0.94 1.0 89 0.78 0.89 96 0.93 0.94 1.5 91 0.82 0.89 97 0.93 0.94 2.0 91 0.82 0.89 97 0.93 0.94 Load position 0.5 87 0.76 0.95 95 0.91 0.94 1.0 91 0.81 0.94 97 0.94 0.96 1.5 93 0.85 0.93 98 0.97 0.98 2.0 93 0.85 0.91 99 0.98 1 Interaction between back tilt and displacement 0.5 84 0.78 0.92 97 0.96 0.98 1.0 86 0.80 0.92 98 0.97 0.98 1.5 87 0.82 0.92 98 0.97 0.99 2.0 87 0.82 0.92 99 0.98 0.99 TABLE 3 | Usability and usefulness perceived of the S-O instrument. Itema Md M SD Sign test pb Usability: understandability of the terminology 8 8.44 1.50 <0.0001 Usability: understandability of the images 9 8.81 1.52 <0.0001 Usability: aesthetic appearance and easy and clear layout 9 8.78 1.25 <0.0001 Usefulness: behavior improvement during the MMH 9 8.81 1.42 <0.0001 Usefulness: knowledge improvement on MMH technique 10 9.11 1.19 <0.0001 Descriptive statistics and one-tailed sign test results. a 10-point response scale was used for all questions (from 1 – very low – to 10 – very high). b One-tailed sign test using 7.5 as a cut point to define dichotomy. Frontiers in Psychology | www.frontiersin.org 9 June 2019 | Volume 10 | Article 1247 Portell et al. Systematic Observation at the Workplace FIGURE 3 | Schematic representation of two T-patterns (A,B), that occur throughout the MMH process of the worker WO1. (A.1,B.1) Include the pattern tree graph. (A.2,B.2) Include pictures highlighting details of the configurations (the recommended positions in larger sizes than those non-recommended); for confidentiality reasons, the workers’ images have been replaced by pictures from the H-O instrument manual. (A.3,B.3) Includes the instance graph. Written informed consent was obtained from the depicted individual for the publication of these images. into consideration all the dimensions of H-O instrument, only back tilted due to an anterior hip flexion, in the final position one common pattern was detected in 100% of the MMH of the lowering phase and with the rest of the dimension in performed. This very simple pattern structures behavior during a non-recommended position) and another event-type which the carrying phase (Figures 4A.1,A.3). This pattern detects includes two very recommended co-occurrences (load position regular associations between an event-type and especially non- close to the body and neutral back position). Figures 4B.1,B.2 recommended co-occurrences (starting displacement with the show the results focusing on the search of T-patterns among Frontiers in Psychology | www.frontiersin.org 10 June 2019 | Volume 10 | Article 1247 Portell et al. Systematic Observation at the Workplace FIGURE 4 | Schematic representation of two T-patterns, that occur throughout the MMH process of the worker WO2. (A.1,B.1) Includes the pattern tree graph. (A.2,B.2) Includes pictures highlighting details of the configurations (the recommended positions in larger sizes than those non-recommended); for confidentiality reasons, the workers’ images have been replaced by pictures from the H-O instrument manual. (A.3,B.3) Includes the instance graph. Written informed consent was obtained from the depicted individual for the publication of these images. the subset of dimensions back, elbow joints, load position and When the polar coordinate analysis was applied to worker interaction. The most complex T-pattern detected in 100% WO2 data (Figure 5B), it was observed that the neutral back of the MMH performed, structures the temporary association position adopted by this worker is sequentially associated with previously detected (Figure 4A.1) with other configurations, other dimensions such as feet, knee joints and elbow joints. all of which are not recommended, during the loading and The strongest sequential relationships were observed in quadrant lowering phases. IV and this indicates the following asymmetric dependence Frontiers in Psychology | www.frontiersin.org 11 June 2019 | Volume 10 | Article 1247 Portell et al. Systematic Observation at the Workplace FIGURE 5 | Maps of polar coordinate analysis for worker WO1 (A) and worker WO2 (B). Under each graph the table with the polar coordinate analysis results is presented for the statistically significant associations. The focal category is the neutral back position (tne). The conditional behaviors are: non-recommended feet position (p1p2), recommended feet position (p3p4), knees flexed moderately during lifting and lowering (rmo), knees extended or slightly flexed during lifting and lowering (rex), knee joints in a non-recommended position in initial/lifting-lowest/lowering-lowest/final position (rex1a/rex1b/rex3c/rex3d), knees are flexed severely (rcv), elbow joint is extended or slightly flexed (b1), one or both elbow joints are flexed (b2), load close to the body (ap), load separated to the body (se). relationship: the recommended back position activates non- change assessment and the creation of feedback under a training recommended feet and knee joint positions during the middle component that we have labeled SsObserWork. Following the part of the lowering, and non-recommended feet and knee research framework for the development and implementation joint positions inhibit the recommended position of the back. of interventions preventing work related MSD established by The previous relationship is not particularly useful to use as van der Beek et al. (2017), we assume that the etiology of back an informative feedback for the worker. From this analysis, pain is multifactorial meaning that the SsObserWork approach the most interesting relationship that can be presented as is suggested as a piece for use in multicomponent occupational feedback for worker WO2 is about the mutually inhibitory intervention, which in turn should be designed according to sequential relationship (quadrant III) between a recommended the “risk control hierarchy model.” We organized the discussion back position and a non-recommended knee joint position at the around two aspects. On the one hand, the novelties associated to end of the lowering phase (the knees were bent less than what SsObserWork instruments, and on the other, the evidence of their should be recommended). reliability, usefulness and usability. The H-O instrument has been justified and developed for observing worker behavior during the MMH training process. DISCUSSION The H-O instrument is composed of the dimensions feet, knee joints, back, elbow joints, load position, and interaction Our study shows the design and implementation of two between back tilt and displacement. For each dimension an systematic observation instruments of worker behavior during independent category system has been created; additionally, the MMH training for back-pain prevention in the workplace. These H-O instrument includes a structural dimension related to the observation instruments address a double purpose: the behavioral MMH phase. This instrument has been developed in such a Frontiers in Psychology | www.frontiersin.org 12 June 2019 | Volume 10 | Article 1247 Portell et al. Systematic Observation at the Workplace way that it allows us to derive the simplified S-O instrument, There are some instruments focused on the evaluation of whereby workers can apply self-observation during MMH. Both changes in worker behavior using video observation (Gattinger instruments are highly structured, and they were developed for et al., 2014), but another important peculiar feature of the H-O video observation, as well as for the generation of feedback. instrument is its level of granularity. The H-O medium level Literature review provided the theoretical models to underpin of granularity requires video observations, because this allows the link between dimensions, categories and classification criteria a large number of items to be assessed, as video films can to provide feedback. The expert judgment also provided support be replayed several times in order to observe the dimensions on the appropriateness of the dimensions and categories for the separately. A more decreased granularity could require the use of objectives established as well as their representativeness regarding monitoring instruments such as sensors that are attached directly the MMH training task selected. to the worker for the measurement of exposure variables at work. The reliability results obtained by the H-O instrument were The use of these direct measurement systems can provide large highly satisfactory. We have presented a thorough assessment of quantities of highly accurate data, however the feasibility of this the agreement between the codes assigned by two research team approach in the workplace has been highly controversial (David, observers and the codes assigned by a technician in occupational 2005; Mgbemena et al., 2017). We have selected a medium level health and safety, previously trained in the use of H-O (Denis of granularity, which has been used for the modeling of complex et al., 2002). The levels of agreement obtained provide strong behavior in the work setting (Neumuth et al., 2010), but as far evidence on the reliability of the H-O instrument. The small as we know it has not been applied for training in the field changes observed between kappa values when the analysis is of occupational risk prevention and health promotion in the based on time or event units are explained by the overestimation workplace. The rationale for this medium level of granularity lies, and underestimation associated with the respective algorithms firstly, in the purpose of the instrument which is the assessment of used for obtaining these results, hence the recommendation has behavior change during the training intervention process, instead been followed to present them jointly (Bakeman et al., 2009). The of the MSD risk assessment. Secondly, in the interest to enable results support the use of H-O (by a technician in occupational the feasibility of the H-O assessment in different workplaces by health and safety) as a reliable instrument to evaluate changes in occupational health professionals. Thirdly, in the interest to use a worker behavior during an MMH training. level of granularity that may also be used for the S-O instrument The use of observational techniques in ergonomics is not new without hindering the worker’s understandability. at all (David, 2005). The novelty of the procedures being proposed The usability and usefulness perceived of the S-O instrument here lies in the function of observation, in the granularity level has been highly valued by workers. Among the various concepts used by the instruments, and in the connection between the under investigation on usefulness and usability (Tsakonas and instruments and the provision of feedback for the worker. Papatheodorou, 2008; Akin et al., 2013; Zapata et al., 2015), The H-O instrument has not been created for the assessment our questions explored understandability, aesthetic appearance of exposure to MSD risk factors associated with a task and layout, worker perception regarding the instrument’s ability or a workplace, it has been created for the assessment of to improve their knowledge and their behavior during MMH. changes in worker behavior during MMH in a formative Workers reported very favorable assessments toward all the context. This characteristic that we could call worker-centered usability and usefulness attributes explored and it can be evaluation, singularizes the H-O instrument regarding many of interpreted as support for the selected level of granularity. the observational techniques used in ergonomics (David, 2005; In addition to being convenient for the S-O instrument, Takala et al., 2010). the granularity selected allows the H-O instrument to generate Another remarkable characteristic of the H-O instrument suitable data to make sequential analysis able to uncover “hidden arises when evaluating the effectiveness of different behavior- time patterns” (Magnusson, 2000) in the behavior during the change-intervention approaches. Behavioral training evaluations training process. This paper shows how it can be used to using assessment instruments exist which focus on distal improve the informative quality of the feedback and feedforward outcomes such as back injuries or wage-loss claims due to back on worker performance. To illustrate the usefulness of the injuries (Clemes et al., 2010). In other cases, the assessment instrument for this purpose, we present the data of two workers instruments are focused on intermediate outcomes measured who would receive very similar feedback if only summary through self-report, such as knowledge about safe lifting and indicators of the global results of their MMH are taken into posture (Haslam et al., 2007). Back injuries are a consequence account (e.g., the global proportion of time with a recommended of the behavior that the training hopes to change, but it is position of the back). The results show the wealth of information an outcome further down the causal chain of a behavioral that can be extracted to inform workers of different regular intervention. The assessment of a behavioral intervention should aspects of their execution. The feedback/feedforward build from include behavioral outcomes (Michie and Johnston, 2012). The this information combined with the self-observation made by H-O instrument can remedy the detected lack of instruments the worker himself or herself is the base of the SsObserWork addressed to the assessment of the proximal effect of manual intervention component. This approach is in line with previous handling training on intermediate variables that link training studies which established the importance of different forms of to distal changes in employee behavior (Hogan et al., 2014; feedback used in combination with video feedback (Wulf et al., Kay et al., 2014). Moreover, the H-O instrument can contribute 2010; Ste-Marie et al., 2013; Lim et al., 2015). Systems already to the understanding of the processes in training implementation. exist for generating feedback that informs on awkward postures Frontiers in Psychology | www.frontiersin.org 13 June 2019 | Volume 10 | Article 1247 Portell et al. Systematic Observation at the Workplace adopted by workers in real time, as well as to evaluate the training interventions. Thirdly, the results on usability and ergonomic risk (Mgbemena et al., 2017). However, SsObserWork utility are only based on self-reported data from a small is presented here as a different approach (although it would sample. It would be convenient to replicate this analysis with not be incompatible). Here the feedback is raised as a training larger samples and complement what is presented here with component that gives the worker material to engage in systematic qualitative techniques (e.g., in-depth interviews) in order to gain self-observation and reflection on his or her non-recommended a deeper understanding of worker perception as well as technician position within his or her “work gesture.” Results show how the perceptions on the application of these instruments. combined use of the H-O instrument with sequential analysis techniques allows feedback to be completed based on static indicators (e.g., relative duration of trunk in non-recommendable CONCLUSION position) with dynamic indications that describe the interaction of the worker with the load. Using the complementary methods In conclusion, this work introduces the SsObserWork approach of T-pattern analysis and polar coordinate analysis (Anguera and it focuses on the development, implementation and et al., 2017) it is possible to detect how co-occurrences and evaluation of new instruments included in SsObserWork. This recommended position sequences interact within gestures that detailed study has been extremely useful for the next phase of globally can be deficient, thereby contributing toward the this project consisting in the evaluation of the effect of the new creation of positive feedback and feedforward (Ste-Marie et al., SsObserwork component using a randomized parallel group trial. 2011). The contribution of these analyses can overcome the The results presented in this work provide empirical evidence in limitations of the positive self-review modeling (Dowrick, 1999) favor of the appropriateness and reliability of these instruments. and it allows progress to be made in the delivery of feedforward As far as we know, the approach we propose is novel in the that helps the worker construct a previously unachieved but field of MDS disorder prevention from MMH training at the possible future pattern of movements. This sort of individualized workplace. A number of convergent research areas support the feedback can open new possibilities especially addressed to use of participatory approaches to health training which allow influential workers because of their basic structure position (e.g., individuals to control features of their own learning environment supervisors) as well as in parallel structures (e.g., members of the enhancing motor learning as well as the transfer of training to health and safety committees, or older employees participating as different settings (Burke et al., 2006; Ste-Marie et al., 2013; Portell mentors in intergenerational learning programs). et al., 2014). The combined use of the S-O instrument with This study has both strengths and limitations. The main external feedback based on data provided by the H-O instrument strength of this study is to provide two new connected opens new possibilities for these participatory approaches, always instruments for systematic observation that make possible the thinking on the SsObserWork approach as a piece to integrate in assessment and implementation of new components to enrich multicomponent occupational intervention. interventions in the prevention of back pain associated with material handling tasks. The mixed method design based on systematic observation (QUAL-QUAN-QUAL phases) makes it possible to tackle the methodological complexity of the DATA AVAILABILITY development, assessment and integration of the results provided The datasets for this manuscript are not publicly available for by these instruments. We want to highlight the demanding confidentiality reasons. Requests to access the datasets should be reliability study presented and its very satisfactory results; this directed to mariona.portell@uab.cat. point seems remarkable considering the lack of reliability data in several assessment instruments noted by different authors (Village et al., 2009; Kadikon and Rahman, 2016). Regarding the limitations and future directions of the research ETHICS STATEMENT we will mention three points. Firstly, the number of companies and participants has been low. In the pilot study, a company The Ethics Committee of the Universitat Autònoma de Barcelona from the metallurgical sector participated and the final version approved the study protocol (Reference Number 1742). In was applied in a food processing company. An area in which accordance with the principles of the Declaration of Helsinki, we consider that the SsObserWork approach could be of great participants were informed that they were being filmed. All interest is in training for the mobilization of patients in hospitals. participants signed the informed consent form. The structure of the H-O instrument (combined field format and category systems) facilitates adaptations and extensions to include interaction dimensions that are essential to address the AUTHOR CONTRIBUTIONS complexity of manual handling in healthcare settings (Kay et al., 2014). A second limitation has to do with the high time and MP conceived and supervised the development of the project, material cost of the coding and sequential analysis process. drafted the manuscript and she also contributed to the design We hope that future studies will advance in the automation of the instruments, the data collection, the data codification, of procedures since this will facilitate the application to more and the data analysis. AS-M implemented the self-observation companies, as well as their integration into multicomponent intervention, collected the data, codified the data, and she also Frontiers in Psychology | www.frontiersin.org 14 June 2019 | Volume 10 | Article 1247 Portell et al. Systematic Observation at the Workplace contributed to the development of the project, the design of DISSENYS [GRID]). Tecnología i aplicació multimedia i digital the instruments, and the drafting of the manuscript. MTA als dissenys observacionals [Grants 2014 SGR 971 and 2018 contributed to the development of the project, the data analysis, SGR 1405]. Lastly, MP also acknowledges the support of the and the drafting of the manuscript. GKJ and JLL contributed to Autonomous University of Barcelona. the data analysis and the drafting of the manuscript. All authors have made a substantial, direct and intellectual contributions to the work, and approved it for publication. ACKNOWLEDGMENTS We are very grateful for the cooperation of NOEL Alimentaria FUNDING SAU (http://www.noel.es/), the cooperation of the cluster INNOVACC (http://www.innovacc.cat/) and the cooperation We gratefully acknowledge the support of the Spanish of MADEL Air Technical Diffusion (http://www.madel.com). Government (Ministerio de Ciencia, Innovación y Universidades, We are also extremely grateful to the group of experts for Programa Estatal de Generación de Conocimiento y their feedback on the preliminary versions of the instruments Fortalecimiento Científico y Tecnológico del Sistema I+D+i) and the contribution of an expert to the external codification: with the subproject Integration ways between qualitative and Rui Garganta, Eduard Gaynés, Joan Pons, Albert Prat, and quantitative data, multiple case development, and synthesis Carles Salvadó. review as main axis for an innovative future in physical activity and sports research [PGC2018-098742-B-C31], that is part of the coordinated project New approach of research SUPPLEMENTARY MATERIAL in physical activity and sport from mixed methods perspective (NARPAS_MM) [SPGC201800X098742CV0]. We gratefully The Supplementary Material for this article can be found acknowledge the support of the Generalitat de Catalunya online at: https://www.frontiersin.org/articles/10.3389/fpsyg. Research Group (GRUP DE RECERCA E INNOVACIÓ EN 2019.01247/full#supplementary-material REFERENCES Anguera, M. T., Camerino, O., Castañer, M., Sánchez-Algarra, P., and Onwuegbuzie, A. J. (2017). The specificity of observational studies in physical Akin, B. A., Bryson, S. A., Testa, M. F., Blase, K. A., McDonald, T., and Melz, activity and sports sciences: moving forward in mixed methods research and H. (2013). Usability testing, initial implementation, and formative evaluation proposals for achieving quantitative and qualitative symmetry. Front. Psychol. of an evidence-based intervention: lessons from a demonstration project to 8:2196. doi: 10.3389/fpsyg.2017.02196 reduce long-term foster care. Eval. Program Plann. 41, 19–30. doi: 10.1016/j. Anguera, M. T., and Hernández-Mendo, A. (2016). Avances en estudios evalprogplan.2013.06.003 observacionales en ciencias del deporte desde los mixed methods [Advances in Altman, D. G. (1991). Practical Statistics for Medical Research. London: observational studies in sport sciences from a mixed methods approach]. Cuad. Chapman & Hall. Psicol. Deporte 16, 17–30. Amatria, M., Lapresa, D., Arana, J., Anguera, M. T., and Jonsson, G. K. (2017). Arana, J., Lapresa, D., Anguera, M. T., and Garzón, B. (2016). Ad hoc procedure for Detection and selection of behavioral patterns using Theme: a concrete example optimizing agreement between observational records. Anal. Psicol. 32, 589–595. in grassroots soccer. Sports 5:E20. doi: 10.3390/sports5010020 doi: 10.6018/analesps.32.2.213551 Anderson, C. K., and Chaffin, D. B. (1986). A biomechanical evaluation of five Arias-Pujol, E., and Anguera, M. T. (2017). Observation of Interactions in lifting techniques. Appl. Ergon. 17, 2–8. doi: 10.1016/0003-6870(86)90186-9 adolescent group therapy: a mixed methods study. Front. Psychol. 8:1188. doi: Anguera, M. T. (1997). “From Prospective Patterns in Behavior to Joint Analysis 10.3389/fpsyg.2017.01188 with a Retrospective Perspective,” in Colloque sur invitation “Méthodologie Authier, M., Lortie, M., and Gagnon, M. (1996). Manual handling techniques: d’Analyse des Interactions Sociales”. Paris: Sorbonne Université. comparing novices and experts. Int. J. Ind. Ergon. 17, 419–429. doi: 10.1016/ Anguera, M. T. (2003). “Observational methods (general),” in Encyclopedia of 0169-8141(95)00005-4 Psychological Assessment, Vol. 2, ed. R. Fernández-Ballesteros (London: Sage), Bakeman, R. (1978). “Untangling streams of behavior: sequential analysis of 632–637. observation data,” in Observing Behavior: Data Collection and Analysis Methods, Anguera, M. T., Blanco-Villaseñor, A., and Losada, J. L. (2001). Diseñ;os Vol. 2, ed. G. P. Sackett (Baltimore, MD: University Park Press), 63–78. Observacionales, cuestión clave en el proceso de la metodología observacional. Bakeman, R., and Quera, V. (2011). Sequential Analysis and Observational Methods [Observational designs, a key question in the observational methodology for the Behavioral Sciences. Cambridge: Cambridge University Press. process]. Metodol. C Comport. 3, 135–161. Bakeman, R., Quera, V., and Gnisci, A. (2009). Observer agreement for timed-event Anguera, M. T., Blanco-Villaseñor, A., Losada, J. L., and Portell, M. (2018a). Pautas sequential data: a comparison of time-based and event-based algorithms. Behav. para elaborar trabajos que utilizan la metodología observacional [Guidelines Res. Methods 41, 137–147. doi: 10.3758/BRM.41.1.137 for designing and conducting a study that applies observational methodology]. Batish, A., and Singh, T. P. (2008). MHAC- An assessment tool for analysing Annu. Psicol. 48, 9–17. doi: 10.1016/j.anpsic.2018.02.001 manual material handling tasks. Int. J. Occup. Saf. Ergon. 14, 223–235. doi: Anguera, M. T., Blanco-Villaseñor, A., Losada, J. L., Sánchez-Algarra, P., and 10.1080/10803548.2008.11076754 Onwuegbuzie, A. J. (2018b). Revisiting the difference between mixed methods Bernard, B. P. (1997). Musculoskeletal Disorders and Workplace Factors. A Critical and multimethods: is it all in the name? Qual. Quant. 52, 2757–2770. doi: Review of Epidemiologic Evidence for Work-Related Musculoskeletal Disorders 10.1007/s11135-018-0700-2 of the Neck, Upper Extremity and Low Back. Publication Num. Cincinnati, OH: Anguera, M. T., Portell, M., Chacón-Moscoso, S., and Sanduvete-Chaves, National Institute for Occupational Safety and Health, 97–141. S. (2018c). Indirect observation in everyday contexts: concepts and Best, M. (1997). An evaluation of manutention training in preventing back strain methodological guidelines within a mixed methods framework. Front. and resultant injuries in nurses. Saf. Sci. 25, 207–222. doi: 10.1016/S0925- Psychol. 9:13. doi: 10.3389/fpsyg.2018.00013 7535(97)00003-9 Frontiers in Psychology | www.frontiersin.org 15 June 2019 | Volume 10 | Article 1247 Portell et al. Systematic Observation at the Workplace Blanco-Villaseñor, A., Sastre, S., and Escolano, E. (2010). Desarrollo ejecutivo executive function assessment and association with later academic skills. Front. temprano y teoría de la generalizabilidad: bebés típicos y prematuros [Executive Psychol. 8:2031. doi: 10.3389/fpsyg.2017.02031 function in early childhood and generalizability theory: typical babies and Faber, G. S., Kingma, I., Bakker, A. J. M., and van Dieën, J. H. (2009). Low-back preterm babies]. Psicothema 22, 221–226. loading in lifting two loads beside the body compared to lifting one load in Brod, M., Tesler, L. E., and Christensen, T. L. (2009). Qualitative research and front of the body. J. Biomech. 42, 35–41. doi: 10.1016/j.jbiomech.2008.10.013 content validity: developing best practices based on science and experience. Ford, J. K., Smith, E. M., Weissbein, D. A., and Gully, S. M. (1998). Relationship Qual. Life Res. 18, 1263–1278. doi: 10.1007/s11136-009-9540-9 of goal orientation, metacognitive activity, and practice strategies with learning Buggey, T. (2007). A picture is worth: video self-modeling applications outcomes and transfer. J. Appl. Psychol. 83, 218–233. doi: 10.1037/0021-9010. at school and home. J. Posit. Behav. Interv. 9, 151–158. doi: 10.1177/ 83.2.218 10983007070090030301 Gabin, B., Camerino, O., Anguera, M. T., and Castañer, M. (2012). Lince: Burke, M. J., Sarpy, S. A., Smith-Crowe, K., Chan-Serafin, S., Salvador, R. O., and multiplatform sport analysis software. Procedia Soc. Behav. Sci. 46, 4692–4694. Islam, G. (2006). Relative effectiveness of worker safety and health training doi: 10.1016/j.sbspro.2012.06.320 methods. Am. J. Public Health 96, 315–324. doi: 10.2105/ajph.2004.059840 Gattinger, H., Stolt, M., Hantikainen, V., Köpke, S., Senn, B., and Leino-Kilpi, H. Casarrubea, M., Jonsson, G. K., Faulisi, F., Sorbera, F., Di Giovanni, G., Benigno, (2014). A systematic review of observational instruments used to assess nurses’ A., et al. (2015). T-pattern analysis for the study of temporal structure of animal skills in patient mobilisation. J. Clin. Nurs. 24, 640–661. doi: 10.1111/jocn. and human behavior: a comprehensive review. J. Neurosci. Methods 239, 34–46. 12689 doi: 10.1016/j.jneumeth.2014.09.024 Hagen, K. B., Hallen, J., and Harms-Ringdahl, K. (1993). Physiological and Casarrubea, M., Magnusson, M. S., Anguera, M. T., Jonsson, G. K., Castañer, M., subjective responses to maximal repetitive lifting employing stoop and squat Santangelo, A., et al. (2018). T-pattern detection and analysis for the discovery technique. Eur. J. Appl. Physiol. Occup. Physiol. 67, 291–297. doi: 10.1007/ of hidden features of behaviour. J. Neurosci. Methods 310, 24–32. doi: 10.1016/ bf00357625 j.jneumeth.2018.06.013 Haslam, C., Clemes, S., McDermott, H., Shaw, K., Williams, C., and Haslam, Castañer, M., Barreira, D., Camerino, O., Anguera, M. T., Fernandes, T., and R. (2007). Manual Handling Training: Investigation of Current Practices and Hileno, R. (2017). Mastery in goal scoring, T-pattern detection, and polar Development of Guidelines. Research Report RR No. 583. Norwich: HSE Books. coordinate analysis of motor skills used by Lionel Messi and Cristiano Ronaldo. Hernández-Mendo, A., Castellano, J., Camerino, O., Jonsson, G., Blanco- Front. Psychol. 8:741. doi: 10.3389/fpsyg.2017.00741 Villaseñor, A., Lopes, A., et al. (2014). Observational software, data quality Cerezo, M. A., Sierra-García, P., Pons-Salvador, G., and Trenado, R. M. (2017). control and data analysis. Rev. Psicol. Deporte 23, 111–121. Parental and infant gender factors in parent–infant interaction: state-space Hernández-Mendo, A., López-López, J. A., Castellano, J., Morales-Sánchez, V., dynamic analysis. Front. Psychol. 8:1724. doi: 10.3389/fpsyg.2017.01724 and Pastrana, J. L. (2012). HOISAN 1.2 programa informático para uso Chacón-Moscoso, S., Sanduvete-Chaves, S., Anguera, M. T., Losada, J. L., Portell, en metodología observacional [HOISAN 1.2: computer program for use in M., and Lozano-Lozano, J. A. (2018). Preliminary checklist for reporting observacional methodology]. Cuad. Psicol. Deporte 12, 55–78. doi: 10.4321/ observational studies in sports areas: content validity. Front. Psychol. 9:291. s1578-84232012000100006 doi: 10.3389/fpsyg.2018.00291 Hogan, D. A., Greiner, B. A., and O’Sullivan, L. (2014). The effect of manual Chang, C. C., Hsiang, S., Dempsey, P. G., and McGorry, R. W. (2003). A handling training on achieving training transfer, employees’s behaviour computerized video coding system for biomechanical analysis of lifting tasks. change and subsequent reduction of work-related musculoskeletal disorders: a Int. J. Ind. Ergon. 23, 239–250. doi: 10.1016/S0169-8141(03)00065-9 systematic review. Ergonomics 57, 93–107. doi: 10.1080/00140139.2013.862307 Clemes, S. A., Haslam, C. O., and Haslam, R. A. (2010). What constitutes effective Hoogendoorn, W. E., Bongers, P. M., de Vet, H. C., Douwers, M., Koes, B. W., manual handling training? A systematic review. Occup. Med. 60, 101–107. doi: Miedema, M. C., et al. (2000). Flexion and rotation of the trunk and lifting 10.1093/occmed/kqp127 at work are risk factors for low back pain. Spine 25, 3087–3092. doi: 10.1097/ Cochran, W. G. (1954). Some methods for strengthening the common χ2 tests. 00007632-200012010-00018 Biometrics 10, 417–451. doi: 10.2307/3001616 Hossain, M. D., Aftab, A., Al Imam, M. H., Mahmud, I., Chowdhury, I. A., Colombini, D., Occhipinti, E., Alvarez-Casado, E., and Waters, T. (2013). Manual Kabir, R. I., et al. (2018). Prevalence of work related musculoskeletal disorders Lifting. A Guide to the Study of Simple and Complex Lifting Tasks. Boca Raton, (WMSDs) and ergonomic risk assessment among readymade garment workers FL: CRC Press. of Bangladesh: a cross sectional study. PLoS One 13:e0200122. doi: 10.1371/ Creswell, J. W., and Plano Clark, V. L. (2011). Designing and Conducting Mixed journal.pone.0200122 Methods Research, 3rd Edn. Thousand Oaks, CA: Sage. Hsiang, S. M., Brogmus, G. E., and Courtney, T. K. (1997). Low back paing (LBP) da Silva, G. P., Ferreira, B., da Silva, Onety, G. C., Verri, E. D., Siéssere, S., and lifting technique - A review. Int. J. Ind. Ergon. 19, 59–74. doi: 10.1016/0169- et al. (2014). Comparative analysis of angles and movements associated with 8141(95)00086-0 sporting gestures in road cyclists. Open Sports Med. J. 8, 23–27. doi: 10.2174/ Hsiang, S. M., Brogmus, G. E., Martin, S. E., and Bezverkhny, I. B. (1998). Video 1874387001408010023 based lifting technique coding system. Ergonomics 41, 239–256. doi: 10.1080/ Daltroy, L. H., Iversen, M. D., Larson, M. G., Ryan, J., Zwerling, C., Fossel, A. H., 001401398187017 et al. (1993). Teaching and social support: effects on knowledge, attitudes, and Izquierdo, C., and Anguera, M. T. (2018). Movement notation revisited: syntax of behaviors to prevent low back injuries in industry. Health Educ. Q. 20, 43–62. the Common Morphokinetic Alphabet (CMA) system. Front. Psychol. 9:1416. doi: 10.1177/109019819302000106 doi: 10.3389/fpsyg.2018.01416 David, G. C. (2005). Ergonomic methods for assessing exposure to risk factors for Janelle, C. M., Barba, D. A., Frehlich, S. G., Tennant, L. K., and Cauraugh, work-related musculoskeletal disorders. Occup. Med. 55, 190–199. doi: 10.1093/ J. H. (1997). Maximizing performance feedback effectiveness through videotape occmed/kqi082 replay and a self-controlled learning environment. Res. Q. Exerc. Sport 68, Demaret, J. P., Gavray, F., Tassignon, M. C., and Willems, F. (2007). Prévention des 269–279. doi: 10.1080/02701367.1997.10608008 Maux de dos Dans le Secteur de L’aide à Domicile. Brussels: Direction générale Kadikon, Y., and Rahman, M. N. A. (2016). Manual material handling Humanisation du travail. risk assessment tool for assessing exposure to risk factor of work-related Denis, D., Lortie, M., and Bruxelles, M. (2002). Impact of observers’ experience and muskuloskeletal disorders: a review. J. Eng. Appl. Sci. 11, 2226–2232. doi: 10. training on reliability of observations for a manual handling task. Ergonomics 3923/jeasci.2016.2226.2232 45, 441–454. doi: 10.1080/00140130210136044 Kahrizi, S., Parnianpour, M., Firoozabadi, S. M., Kasemnejad, A., and Karimi, E. Dowrick, P. W. (1999). A review of self modeling and related interventions. Appl. (2007). Evaluation of spinal internal loads and lumbar curvature under holding Prev. Psychol. 8, 23–39. doi: 10.1016/S0962-1849(99)80009-2 static load at different trunk and knee positions. Pak. J. Biol. Sci. 10, 1036–1043. Dowrick, P. W. (2012). Self modeling: expanding the theories of learning. Psychol. doi: 10.3923/pjbs.2007.1036.1043 Sch. 49, 30–41. doi: 10.1002/pits.20613 Kay, K., Glass, N., and Evans, A. (2014). It’s not about the hoist: a narrative Escolano-Pérez, E., Herrero-Nivela, M. L., Blanco-Villaseñor, A., and Anguera, literature review of manual handling in healthcare. J. Res. Nurs. 19, 226–245. M. T. (2017). Systematic observation: relevance of this approach in preschool doi: 10.1177/1744987112455423 Frontiers in Psychology | www.frontiersin.org 16 June 2019 | Volume 10 | Article 1247 Portell et al. Systematic Observation at the Workplace Kernodle, M. W., and Carlton, L. G. (1992). Information feedback and the learning Mörl, F., Wagner, H., and Blickhan, R. (2005). Lumbar spine intersegmental of multiple-degree-of-freedom activites. J. Mot. Behav. 24, 187–196. motion analysis during lifting. Pathophysiology 12, 295–302. doi: 10.1016/j. Kilbom, A. (1994). Assessment of physical exposure in relation to work-related pathophys.2005.09.003 musculoskeletal disorders- what information can be obtained from systematic Neumuth, T., Kaschek, B., Neumuth, D., Ceschia, M., Meixensberger, G., Strauss, observations? Scand. J. Work Environ. Health 20, 30–45. G., et al. (2010). An observation support system with an adaptive ontology- Kingma, I., Bosch, T., Bruins, L., and van Dieën, J. H. (2004). Foot positioning driven user interface for the modeling of complex behaviors during surgical instruction, initial vertical load position and lifting technique: effects on low interventions. Behav. Res. Methods 42, 1049–1058. doi: 10.3758/BRM.42.4.1049 back loading. Ergonomics 47, 1365–1385. doi: 10.1080/00140130410001714742 Newell, K. M. (1991). Motor skill acquisition. Annu. Rev. Psychol. 42, 213–237. Kingma, I., Faber, G. S., Bakker, A. J. M., and van Dieën, J. H. (2006). Can low Nygård, C. H., Merisalo, T., Arola, H., Manka, M. L., and Huhtala, H. (1998). back loading during lifting be reduced by placing one leg beside the object to be Effects of work changes and training in lifting technique on physical strain: a lifted? Phys. Ther. 86, 1091–1105. doi: 10.1093/ptj/86.8.1091 pilot study among female workers of different ages. Int. J. Ind. Ergon. 21, 91–98. Kirby, R. L., Price, N. A., and MacLeod, D. A. (1987). The influence of foot doi: 10.1016/S0169-8141(97)00028-0 position on standing balance. J. Biomech. 20, 423–427. doi: 10.1016/0021- Onwuegbuzie, A. J., Bustamante, R. M., and Nelson, J. A. (2010). Mixed research 9290(87)90049-2 as a tool for developing quantitative instruments. J. Mix. Methods Res. 4, 56–78. Landis, J., and Koch, G. (1977). The measurement of observer agreement for doi: 10.1177/1558689809355805 categorical data. Biometrics 33, 159–174. doi: 10.2307/2529310 Pedersen, L. M., Nielsen, K. J., and Kines, P. (2012). Realistic evaluation as a new Larivière, C., Gagnon, D., and Loisel, P. (2000). The effect of load on the way to design and evaluate occupational safety interventions. Saf. Sci. 50, 48–54. coordination of the trunk for subjects with and without chronic low back pain doi: 10.1016/j.ssci.2011.06.010 during flexion-extension and lateral bending tasks. Clin. Biomech. 15, 407–416. Penkala, S., El-Debal, H., and Coxon, K. (2018). Work-related musculoskeletal doi: 10.1016/s0268-0033(00)00006-1 problems related to laboratory training in university medical science students: Lim, S., Ali, A., Kim, W., Kim, J., Choi, S., and Radlo, S. J. (2015). Influence of a cross sectional survey. BMC Public Health 18:1208. doi: 10.1186/s12889-018- self-controlled feedback on learning a serial motor skill. Percept. Mot. Skills 120, 6125-y 462–474. doi: 10.2466/23.pms.120v13x3 Pérez-Tejera, F., Valera, S., and Anguera, M. T. (2018). Using systematic Lortie, M., and Baril-Gingras, G. (1998). Box handling in the loading and unloading observation and polar coordinates analysis to assess gender-based differences of vans. Int. J. Occup. Saf. Ergon. 4, 3–18. doi: 10.1080/10803548.1998.11076376 in park use in Barcelona. Front. Psychol. 9:2299. doi: 10.3389/fpsyg.2018. Magill, R. A. (2011). Motor Learning and Control. Concepts and Applications. 02299 New York, NY: McGraw-Hill. Plamondon, A., Delisle, A., Bellefeuille, S., Denis, D., Gagnon, D., Larivière, C., Magnusson, M. S. (1996). Hidden real-time patterns in intra- and inter-individual et al. (2014). Lifting strategies of expert and novice workers during a repetitive behavior: description and detection. Eur. J. Psychol. Assess. 12, 112–123. doi: palletizing task. Appl. Ergon. 45, 471–481. doi: 10.1016/j.apergo.2013.06.008 10.1027/1015-5759.12.2.112 Portell, M., Anguera, M. T., Chacón-Moscoso, S., and Sanduvete-Chaves, S. (2015). Magnusson, M. S. (2000). Discovering hidden time patterns in behavior: t-patterns Guidelines for reporting evaluations based on observational methodology. and their detection. Behav. Res. Methods Instrum. Comput. 32, 93–110. doi: Psicothema 27, 283–289. doi: 10.7334/psicothema2014.276 10.3758/BF03200792 Portell, M., Gil, R. M., Losilla, J. M., and Vives, J. (2014). Characterizing Magnusson, M. S. (2018). “Temporal patterns in interactions. T-Patterns and their occupational risk perception: the case of biological, ergonomic and detection with THEME,” in The Cambridge Handbook of Group Interaction organizational hazards in Spanish healthcare workers. Span. J. Psychol. Analysis, eds E. Brauner, M. Boos, and M. Kolbe (Cambridge: Cambridge 17:E51. doi: 10.1017/sjp.2014.55 University Press), 323–353. doi: 10.1017/9781316286302.017 Punnett, L., Cherniack, M., Henning, R., Morse, T., Faghri, P., Cph-New Marras, W. S., Granata, K. P., Davis, K. G., Allread, W. G., and Jorgensen, Research, et al. (2009). A conceptual framework for integrating workplace M. J. (1999). Effects of box features on spine loading during warehouse order health promotion and occupational ergonomics programs. Public Health Rep. selecting. Ergonomics 42, 980–996. doi: 10.1080/001401399185252 124(Suppl. 1), 16–25. doi: 10.1177/00333549091244S103 Marras, W. S., and Karwowski, W. (2006). Fundamentals and Assessment Tools for Quera, V. (2018). “Analysis of interaction sequences,” in The Cambridge Occupational Ergonomics. Boca Raton, FL: Taylor & Francis Group. Handbook of Group Interaction Analysis, eds E. Brauner, M. Boos, and Mason, R. A., Davis, H. S., Ayres, K. M., Davis, J. L., and Mason, B. A. (2016). M. Kolbe (Cambridge: Cambridge University Press), 295–322. doi: 10.1017/ Video self-modeling for individuals with disabilities: a best-evidence, single 9781316286302.016 case meta-analysis. J. Dev. Phys. Disabil. 28, 623–642. doi: 10.1007/s10882-016- Rodríguez-Medina, J., Rodríguez-Navarro, H., Arias, V., Arias, B., and Anguera, 9484-2 M. T. (2018). Non-reciprocal friendships in a school-age boy with autism: the Mawston, G. A., and Boocock, M. G. (2012). The effect of lumbar posture on spinal ties that build? J. Autism Dev. Disord. 48, 2980–2994. doi: 10.1007/s10803-018- loading and the function of the erector spinae: implications for exercise and 3575-0 vocational rehabilitation. NZJP 40, 135–140. Sackett, G. P. (1980). “Lag sequential analysis as a data reduction technique in McGill, S. M. (1997). Biomechanics of low back injury: implications for the social interaction research,” in Exceptional Infant. Psychosocial Risks in Infant- workplace and clinic. J. Biomech. 30, 465–475. doi: 10.1016/s0021-9290(96) Environment Transactions, eds D. B. Sawin, R. C. Hawkins, L. O. Walker, and 00172-8 J. H. Penticuff (New York, NY: Brunner), 300–340. McGill, S. M. (2007). Low Back Disorders. Evidence-Based Prevention and Sackett, G. P. (1987). “Analysis of sequential social interaction data: Some Rehabilitation. Champaign, IL: Human Kinetics. issues, recent developments, and a causal inference model,” in Handbook Mgbemena, C. E., Oyekan, J., Hutabarat, W., Xu, Y., and Tiwari, A. (2017). of Infant Development (2nd ed.), ed J. D. Osofsky (New York: Wiley), Design and implementation of ergonomic risk assessment feedback system 855–878. for improved work posture assessment. Theor. Issues Ergon. Sci. 19, 431–455. Salmoni, A. W., Schmidt, R. A., and Walter, C. B. (1984). Knowledge of results doi: 10.1080/1463922x.2017.1381196 and motor learning: a review and critical reappraisal. Psychol. Bull. 95, 355–386. Michie, S., and Johnston, M. (2012). Theories and techniques of behaviour change: doi: 10.1037//0033-2909.95.3.355 developing a cumulative science of behaviour change. Health Psychol. Rev. 6, Sánchez-Algarra, P., and Anguera, M. T. (2013). Qualitative/quantitative 1–6. doi: 10.1080/17437199.2012.654964 integration in the inductive observational study of interactive behaviour: impact Monnington, S. C., Pinder, A. D., and Quarrie, C. (2002). Development of an of recording and coding predominating perspectives. Qual. Quant. 47, 1237– Inspection Tool for Manual Handling Risk Assessment. Sheffield: Health and 1257. doi: 10.1007/s11135-012-9764-6 Safety Laboratory. Santoyo, C. V., and Mendoza, B. G. (2018). Behavioral patterns of children involved Morillo, J. P., Reigal, R. E., Hernández-Mendo, A., Montaña, A., and Morales- in bullying episodes. Front. Psychol. 9:456. doi: 10.3389/fpsyg.2018.00456 Sánchez, V. (2018). Decision-making by handball referees: design of an ad hoc Schneider, E., and Irastorza, X. (2010). OSH in Figures: Work-Related observation instrument and polar coordinate analysis. Front. Psychol. 8:1842. Musculoskeletal Disorders. Luxembourg: European Agency for Safety and doi: 10.3389/fpsyg.2017.01842 Health at Work. Frontiers in Psychology | www.frontiersin.org 17 June 2019 | Volume 10 | Article 1247 Portell et al. Systematic Observation at the Workplace Silva, L. C. D., Pereira-Monfredini, C. F., and Teixeira, L. A. (2016). Improved Verbeeck, J. H., Martimo, K. P., Karppinen, J., Kuijer, P. P. F. M., Viikari-Juntura, children’s motor learning of the basketball free shooting pattern by associating E., and Takala, E. P. (2011). Manual material handling advice and assistive subjective error estimation and extrinsic feedback. J. Sports Sci. 35, 1825–1830. devices for preventing and treating back pain in workers. Cochrane Database doi: 10.1080/02640414.2016.1239025 Syst. Rev. 15:CD005958. doi: 10.1002/14651858.CD005958.pub3 Snook, S. H., and Ciriello, V. M. (1991). The design of manual handling tasks: Village, J., Trask, C., Luong, N., Chow, Y., Johnson, P., Koehoorn, M., et al. (2009). revised tables of maximum acceptable weights and forces. Ergonomics 34, Development and evaluation of an observational Back-Exposure Sampling Tool 1197–1213. doi: 10.1080/00140139108964855 (Back-EST) for work-related back injury risk factors. Appl. Ergon. 40, 538–544. Ste-Marie, D. M., Vertes, K., Rymal, A. M., and Martini, R. (2011). Feedforward doi: 10.1016/j.apergo.2008.09.001 self-modeling enhances skill acquisition in children learning trampoline skills. Waters, T. R., Putz-Anderson, V., Garg, A., and Fine, L. J. (1993). Revised NIOSH Front. Psychol. 2:155. doi: 10.3389/fpsyg.2011.00155 equation for the design and evaluation of manual lifting tasks. Ergonomics 36, Ste-Marie, D. M., Vertes, K. A., Law, B., and Rymal, A. M. (2013). Learner- 749–776. doi: 10.1080/00140139308967940 controlled self-observation is advantageous for motor skills acquisition. Front. Wulf, G., Shea, C., and Lewthwaite, R. (2010). Motor skill learning and Psychol. 3:556. doi: 10.3389/fpsyg.2012.00556 performance: a review of influential factors. Med. Educ. 44, 75–84. doi: 10.1111/ Straker, L. M. (2003a). A review of research on technique for lifting low-lying j.1365-2923.2009.03421.x objects: 2. Evidence for a correct technique. Work 20, 83–96. Zapata, B. C., Fernández-Alemán, J. L., Idri, A., and Toval, A. (2015). Empirical Straker, L. M. (2003b). Evidence to support using squat, semi-squat and stoop studies on usability of mHealth Apps: a systematic literature review. J. Med. techniques to lift low-lying objects. Int. J. Ind. Ergon. 31, 149–160. doi: 10.1016/ Syst. 39:1. doi: 10.1007/s10916-014-0182-2 S0169-8141(02)00191-9 Zhou, J., Dai, B., and Ning, X. (2013). The assessment of material handling Takala, E. P., Pehkonen, I., Forsman, M., Hansson, G. A., Mathiassen, S. E., strategies in dealing with sudden loading: influences of foot placement on trunk Neumann, W. P., et al. (2010). Systematic evaluation of observational methods biomechanics. Ergonomics 56, 1569–1576. doi: 10.1080/00140139.2013.822568 assessing biomechanical exposures at work. Scand. J. Work Environ. Health 36, 3–24. doi: 10.5271/sjweh.2876 Conflict of Interest Statement: The authors declare that the research was Thiese, M. S., Hegmann, K. T., Wood, E. M., Garg, A., Moore, J. S., Kapellusch, conducted in the absence of any commercial or financial relationships that could J. M., et al. (2014). Low back pain ratings for lifetime, 1-month period, and be construed as a potential conflict of interest. point prevalences in a large occupational population. Hum. Factors 56, 86–97. doi: 10.1177/0018720813493641 The reviewer AH-M and handling Editor declared their shared affiliation. Tsakonas, G., and Papatheodorou, C. (2008). Exploring usefulness and usability in the evaluation of open access digital libraries. Inf. Process. Manag. 44, Copyright © 2019 Portell, Sene-Mir, Anguera, Jonsson and Losada. This is an open- 1234–1250. doi: 10.1016/j.ipm.2007.07.008 access article distributed under the terms of the Creative Commons Attribution van der Beek, A. J., Dennerlein, J. T., Huysmans, M. A., Mathiassen, S. E., Burdorf, License (CC BY). The use, distribution or reproduction in other forums is permitted, A., van Mechelen, W., et al. (2017). A research framework for the development provided the original author(s) and the copyright owner(s) are credited and that the and implementation of interventions preventing work-related musculoskeletal original publication in this journal is cited, in accordance with accepted academic disorders. Scand. J. Work Environ. Health 43, 526–539. doi: 10.5271/sjweh. practice. No use, distribution or reproduction is permitted which does not comply 3671 with these terms. Frontiers in Psychology | www.frontiersin.org 18 June 2019 | Volume 10 | Article 1247