Cost-efficient measurement strategies for posture observations based on video recordings Svend Erik Mathiassen a, * , Per Liv a , Jens Wahlström a, b a Centre for Musculoskeletal Research, Department of Occupational and Public Health Sciences, University of Gävle, SE-80176 Gävle, Sweden b Department of Public Health & Clinical Medicine, Occupational and Environmental Medicine, Umeå University, SE-90185 Umeå, Sweden article info Article history: Received 1 August 2012 Accepted 20 December 2012 Keywords: Cost-efficiency Exposure assessment Research budget Resource consumption abstract Assessment of working postures by observation is a common practice in ergonomics. The present study investigated whether monetary resources invested in a video-based posture observation study should preferably be spent in collecting many video recordings of the work and have them observed once by one observer, or in having multiple observers rate postures repeatedly from fewer videos. The study addressed this question from a practitioner’s perspective by focusing two plausible scenarios: doc- umenting the mean exposure of one individual, and of a specific occupational group. Using a data set of observed working postures among hairdressers, empirical values of posture variability, observer varia- bility, and costs for recording and observing one video were entered into equations expressing the total cost of data collection and the information (defined as 1/SD) provided by the resulting estimates of two variables: percentage time with the arm elevated <15  and >90  . Sixteen measurement strategies involving 1e4 observers repeating their posture ratings 1e4 times were examined for budgets up to V2000. For both posture variables and in both the individual and group scenario, the most cost-efficient strategy at any specific budget was to engage 3e4 observers and/or having observer(s) rate postures multiple times each. Between 17% and 34% less information was produced when using the commonly practiced approach of having one observer rate a number of video recordings one time each. We therefore recommend observational posture assessment to be based on video recordings of work, since this allows for multiple observations; and to allocate monetary resources to repeated observations rather than many video recordings. Ó 2012 Elsevier Ltd and The Ergonomics Society. All rights reserved. 1. Introduction In ergonomics, working postures are often assessed by obser- vation, either on location in the field or from video recordings (Li and Buckle, 1999; Denis et al., 2000; Takala et al., 2010). Modern technology offers technical measurement devices allowing long- term, continuous recordings of postures (Svendsen et al., 2004; Leijon et al., 2005; Unge et al., 2007; Wahlström et al., 2010; Fethke et al., 2011), but so far collection and processing of data using this equipment requires competence and resources beyond the reach of most practitioners (Trask et al., 2007 , 2012). In contrast, observation procedures are easier to learn and do not entail any particular capital costs, while at the same time offering data of an acceptable quality for many purposes (Winkel and Mathiassen, 1994; van der Beek and Frings-Dresen, 1998). Thus, posture assessment by observation is an attractive approach, in particular to practicing ergonomists. Posture assessment is often carried out for the purpose of estimating the mean value of some posture variable believed to be associated with risk of contracting musculoskeletal disorders. Thus, the occurrence of both extreme and neutral postures has been addressed in observational studies, since they are com- monly accepted to indicate risk and recovery, respectively (Denis et al., 2000; Takala et al., 2010). The practicing ergono- mist may perform an assessment at the level of an individual worker, as when examining the effect of a subject-specific er- gonomics intervention (Mathiassen and Paquet, 2010), or at the level of groups, as when documenting awkward postures at an office or a production line (Juul-Kristensen et al., 2001; Lindegård et al., 2003). In both cases, posture assessments, as any collection of occu- pational exposure data, face resource limits, usually in terms of * Corresponding author. Tel.: þ46 706 788158; fax: þ46 26 648686. E-mail addresses: svenderik.mathiassen@hig.se (S.E. Mathiassen), perliv@hig.se (P. Liv), jens.wahlstrom@envmed.umu.se (J. Wahlström). Contents lists available at SciVerse ScienceDirect Applied Ergonomics journal homepage: www.elsevier.com/locate/apergo 0003-6870/$ e see front matter Ó 2012 Elsevier Ltd and The Ergonomics Society. All rights reserved. http://dx.doi.org/10.1016/j.apergo.2012.12.003 Applied Ergonomics 44 (2013) 609e617