Transport Research Laboratory Creating the future of transport

Transcription

1 Transport Research Laboratory Creating the future of transport CLIENT PROJECT REPORT CPR1306 Removing road danger lamps at relaxation road works: impact on driver behaviour and road worker risk S Helman, S Cynk, M Palmer, C Fowler and I Rillie Prepared for: Project Ref: Highways Agency Task 546(1308)HALC Quality approved: Siggi Clark (Project Manager) Iain Rillie (Technical Referee) Transport Research Laboratory 2012

2 Disclaimer This report has been produced by the Transport Research Laboratory under a contract with Highways Agency. Any views expressed in this report are not necessarily those of Highways Agency. The information contained herein is the property of TRL Limited and does not necessarily reflect the views or policies of the customer for whom this report was prepared. Whilst every effort has been made to ensure that the matter presented in this report is relevant, accurate and up-to-date, TRL Limited cannot accept any liability for any error or omission, or reliance on part or all of the content in another context. When purchased in hard copy, this publication is printed on paper that is FSC (Forest Stewardship Council) and TCF (Totally Chlorine Free) registered. Contents amendment record This report has been amended and issued as follows: Version Date Description Editor Technical Referee /03/12 Initial draft SH IMR /03/12 Draft report for client review SH IMR /04/12 Draft version incorporating client comments SC IMR /04/12 Final version accepted by client SC IMR

3 Contents Disclaimer 2 Contents amendment record 2 1 Introduction 4 2 Effect of road danger lamps on lateral position of traffic passing relaxation works Introduction Method Equipment Procedure for data collection Design and analysis Results Unlit offside closures Lit offside closures Lit nearside closures Qualitative data from the on-road trial Summary of results of on-road trial 19 3 Ergonomics assessment of road danger lamp use Method Results 21 4 Calculation of injury risk sub-value Risk exposure calculation Additional risks associated with use of road danger lamps Potential risk from debris from impacts 26 5 Discussion of findings 28 6 Conclusions and recommendations 30 Acknowledgements 31 References 31 1 CPR1306

4 Executive Summary The Highways Agency is committed to driving down the risks to which its road workers are exposed, while at the same time finding more efficient ways of working to deliver its service provision to road users. Road workers put themselves at risk of injury when placing longitudinal cones to mark out the edge of road works. One possible contributory factor to the overall risk presented by this activity is the use of road danger lamps which increases the time taken to set out the cones and hence increases the time spent in close proximity to live traffic. This study used three methods to examine the likely impact of removing road danger lamps from relaxation works: 1. An on-road trial examined the impact on road user behaviour by measuring the lane distribution of vehicles passing relaxation works with and without road danger lamps on the longitudinal cones. The position of vehicles in the adjacent lane was also coded. 2. An ergonomics assessment of the activities involved in laying and retrieving cones with and without road danger lamps was conducted. 3. The Measurement of Injury Risk (MIRi) Index (Fowler, Clark, Rillie, Cuerden & Smith, 2011) was used to calculate a risk sub-value associated with deployment and retrieval of longitudinal coning with and without road danger lamps. The on-road trial provided no firm evidence for or against the removal of road danger lamps from longitudinal cones. There was some evidence in some conditions that vehicles in the lane adjacent to the longitudinal coning drove slightly closer to the cone line when road danger lamps were absent (lit offside and lit nearside closures), although in unlit offside closures there was a trend in the data for some vehicles moved to lanes and positions further away (unlit offside closures). The ergonomics assessment supported the case for the removal of road danger lamps; in almost all cases when laying and retrieving cones, scores on the Rapid Entire Body Assessment (REBA) coding tool (Hignett & McAtamney, 2000) were higher when road danger lamps were used, meaning the road workers have a higher risk of musculoskeletal injuries. The MIRi work indicated that the removal of road danger lamps would be likely to be associated with a safety benefit for road workers, mainly through reducing their exposure (reduced time) to the activity of laying and retrieving cones. This has the potential to achieve a reduction in risk equivalent to 10% of the risk reduction benefit achieved from introduction of temporary traffic management sign simplification (IAN150/12). Such a decrease in risk has the potential to deliver a valuable contribution to the HA s Aiming for Zero road worker safety improvement programme. Based on the results from this study, it is therefore considered that there is no categorical evidence that indicates continued use of road danger lamps is essential to the safety of road workers or road users. 2 CPR1306

5 Given the positive impact of removing lamps suggested by the ergonomics and MIRI analyses, and the uncertainties surrounding the driver response, the following recommendations are made: 1. The trial of omitting road danger lamps should be continued with monitoring of the driver response (possibly through incident and near miss reporting from road workers on the ground) over a much longer period of time than that covered by this study 2. This should be rolled out via an Area Management Memo, permitting the relaxation of the requirement to use road danger lamps provided monitoring data are provided (following the precedent of AMM125/10 for the innovative taper). 3. The Supply Chain should be involved with the development of the roll-out process and be consulted on the option to have a staged approach to removing the lamps. 4. Once sufficient data are received and analyses are completed indicating the safety case is favourable, the HA should consider as to whether omission of road danger lamps from longitudinal coning at relaxation scheme traffic management is incorporated into formal guidance. A final recommendation of this study is that musculoskeletal injury rates be monitored in those operatives who spend time laying and retrieving longitudinal coning on all road works (even without road danger lamps) to confirm the occupational injury risk from this task. 3 CPR1306

6 1 Introduction The Highways Agency (HA) is committed to reducing the risks faced by road workers, while continuing to find more efficient ways of working to reduce its costs and carbon footprint. Road workers put themselves at risk of injury when placing longitudinal cones to mark out the edge of road works; one possible contributory factor to the overall risk presented by this activity is the use of road danger lamps. These lamps come in a number of types (two of which are shown below) but all must be fitted onto or removed from the cones manually prior to placing the cone on the road, which increases the time taken to set out the cones. This in turn increases the time spent in close proximity to live traffic, with the associated increase in risk of serious or fatal injury. Guidance in the Traffic Signs Manual Chapter 8 is that when it is dark, road danger lamps should be used on longitudinal cones for relaxation works. Paragraph O4.7.1 of Chapter 8 (Part 2) states that: The outer boundary of the works site, temporary fences, barriers etc. should always be marked by the addition of warning lights during the hours of darkness The full width and length of the obstruction shall be clearly indicated by lights. Table A1.3 of the same document gives values for the maximum spacing of lamps for relaxation works; these are nine metres on the taper, and 18m on the longitudinal coned section. It has been suggested that the risk to road workers could be reduced if the use of road danger lamps on longitudinal cones in relaxation works was stopped. In addition, because the lamps contribute to the HA s costs and carbon footprint, savings in these 4 CPR1306

7 areas would also be expected from discontinuing their use. If this change is to be implemented nationally however, its impact needs to be evaluated. This is because since the purpose of road danger lamps (as defined in Regulation 55 of the Traffic Signs Regulations and General Directions 2002) is to provide road users with an indication of the extent of a hazard, it is possible that their removal will lead to undesirable changes in road user behaviour. The impact of this on injury risk for other road users and for those road workers working within the relaxation works area needs to be understood to advise any decision. This document reports the findings from a study to evaluate the impact of the removal of road danger lamps. The overall aim of the study is to establish whether or not there is a safety case to support the removal of road danger lamps from relaxation works. The study has the following research questions: 1. Does the removal of road danger lamps from longitudinal cones used in relaxation road works have any impact on how vehicles are driven past the road works? Do drivers drive any closer to (or further away from) the longitudinal cones when road danger lamps are removed, compared with when they are present? 2. What are the ergonomic risks to the road workers, associated with placing longitudinal cones in relaxation works, with and without road danger lamps? 3. How do road danger lamps on longitudinal cones affect risk to road workers laying and retrieving longitudinal cones at relaxation works, as quantified using the Measurement of Injury Risk (MIRi) Index (see Fowler, Clark, Rillie, Cuerden & Smith, 2011)? The remainder of this report is structured as follows: Section 2 presents the method and findings from the observational study that examined the effect of road danger lamps on road user behaviour. Section 3 presents the ergonomics assessment of the tasks of placing and retrieving longitudinal cones with and without road danger lamps. Section 4 presents calculations of the MIRi scores for placing and retrieving longitudinal cones with and without road danger lamps. Section 5 discusses the findings from the three technical sections in relation to the removal of road danger lamps from longitudinal coning at relaxation works and examines their implications. Section 6 then draws conclusions and makes recommendations on the basis of the findings. 5 CPR1306

8 2 Effect of road danger lamps on lateral position of traffic passing relaxation works 2.1 Introduction The aim of the observational component of the study was to assess the effect of removing road danger lamps (RDLs) from longitudinal cones on the lane position of vehicles passing relaxation road works. The purpose of RDLs at relaxation works is to provide road users with an indication of the extent of the works during hours of darkness. It is likely that their removal might lead to the longitudinal cone line being less conspicuous to road users, which may lead to road users changing their behaviour. No firm predictions are possible as to what this change in behaviour might be. One possible outcome is that road users might drift closer to the cone line, due to its relative inconspicuousness. Another possible outcome is that drivers might notice that the cone line is not as conspicuous, and may choose to leave a larger gap between them and the line of cones 2.2 Method Equipment A camera system was designed and manufactured by TRL to gather video recordings to allow traffic flow, lane occupancy and driver behaviour to be determined and analysed. The system allowed for data to be collected on lit and unlit carriageways. The system (see Figure 2-1) comprised a recorder and battery within a case fitted with a power switch and loop for security. Figure 2-1: Recorder case, cover removed, showing battery and recorder The external camera connected to the recorder box via a lead and was fitted to a mounting clamp. This is shown in Figure CPR1306

9 Figure 2-2: Camera and mounting clamp Each system was placed at the relevant location (see Figure 2-3) on a sign frame within the road work longitudinal coning for the duration of each set of road works and removed before the cones were retrieved. Figure 2-3: Camera system mounted on sign frame TRL uniquely identified each camera system at each site corresponding to its position. This was either at least 300m after the work zone (Camera A) or 100m after this (Camera B) 1. This layout is shown in Figure Two camera positions were used for redundancy. Analysis confirmed that there was no difference between the lane choice of vehicles passing the two cameras, and therefore footage from camera 1 was used in the analysis. 7 CPR1306

10 Figure 2-4: Camera layout for data collection (nearside/offside closure on three-lane carriageway) Typical camera images are shown in Figure 2-5. a) b) c) d) Figure 2-5: Typical camera views: a) lit nearside closure b) lit offside closure c) unlit nearside closure d) unlit offside closure 8 CPR1306

11 2.2.2 Procedure for data collection The camera systems were deployed by road workers setting out relaxation works; data collection began on 16/01/12 in HA Area 5 and on 26/01/12 in HA Area 9, and ended on 08/03/12 in both Areas. The following restrictions were placed on those sites that could be used for data collection: 1. Drivers to be alongside a longitudinal run of cones (either with or without cone lamps, depending on condition) at a relaxation works 2. This run of cones to be sufficiently away from the works area so that drivers are not influenced by the presence of any works vehicles or crew 3. The works should be away from junctions or dedicated slip roads, so that driver lane choice is not influenced 4. Works not within major road improvement scheme traffic management (TM) 5. Minimum of two running lanes remaining open alongside the cone line 6. Hard shoulder present and not used as one of the two running lanes Labelled storage and despatch envelopes were provided for each memory card to assist with the accurate return of camera data to TRL. Upon receipt by TRL, the footage was analysed manually using bespoke screen overlays on PCs playing the footage back through VLC media player; these screen overlays permitted coding of the position of each vehicle passing the cones into the zones described in Section Design and analysis The research questions were as follows: 1. Is there a difference in lane distribution of vehicles between the two conditions (experimental or control)? 2. Is there a difference in the distribution of vehicle positions within the lane adjacent to the works between the two conditions? These questions were addressed separately for combinations of lit/unlit and nearside/offside closures, and separately for cars and HGVs. Generalized linear models were used to analyse the data; this statistical technique allows a research question to be addressed while controlling for potential confounding variables (such as traffic flow). The key outcome measure in the study was the lateral position of vehicles as they passed the longitudinal cones. This was expressed in two ways. Firstly, the proportion of vehicles in the lane adjacent to the closure was measured. Secondly, for those vehicles in the lane adjacent to the closure, position within the lane was coded. In practice this was achieved through coding a single position measure from the video analysis for the wheel (nearest to the works) of each vehicle that passed the camera. Five zones were defined according to the following criteria: Zone 1 was 0cm to 50cm from the cones Zone 2 was 51cm to 100cm from the cones Zone 3 was 101cm to 150cm from the cones Zone 4 was the remaining portion of the lane next to the closure 9 CPR1306

12 Zone 5 was always the lane(s) further away from the closure The zones were named in this way regardless of whether closures were nearside or offside. Thus for nearside closures the zones ran left to right on the road in ascending number order, while for offside closures they ran right to left. The proportion of vehicles in each zone served as the outcome measure (dependent variable) in the design. In most cases Zone 1 covered a portion of the adjacent lane and a very small part of the closed lane, if the cones were placed nearly up against the line markings. However because there was no standardised measuring of the lateral position of the cone line, the extent to which Zone 1 encroached into the closure lane varied; on a very small number of occasions, Zone 2 also encroached into the closure lane for the same reason. The independent variables were experimental condition (control, experimental) and zone (either 1 4 and 5 for the first research question, or 1, 2, 3 and 4 for the second research question). For the condition variable, the control condition reflected current practice, with road danger lamps used on the longitudinal coning. In the experimental condition, no road danger lamps were used on the longitudinal cones. In both conditions, sequentially flashing road danger lamps were used on the taper. Additional variables were measured so that they could be controlled for or balanced as well as possible in the analysis. These variables were traffic flow 2, and the distance of the cone from the line of the closure lane 3. Where it has not been possible to balance completely for these variables, this is noted in the text. 2.3 Results In all analyses, traffic flow was taken into account at the different sites and data weighted to account for any differences observed between the control and experimental sites. In all, data were collected from 23 experimental and 25 control sites, split mainly between unlit offside, lit offside, and lit nearside closures 4. For each dataset, the generalised linear models tested for interactions between zone and condition; namely, they asked whether the proportion of vehicles across zones varied according to the condition (control, or experimental) Unlit offside closures Data were collected from seven experimental and three control sites with unlit offside closures. After balancing for cone distance, data were used from five experimental and 2 Note that one limitation of the study is that findings should not be generalised to traffic flows that have not been tested. This is because as traffic flow goes up there may be other constraints on lane and positionwithin-lane arising from the proximity of other road users. The highest flow seen in a trial site was 1301 vehicles per hour. 3 Variation in the distance between the cones and line could not be fully controlled in the data collection. It was controlled in the analysis by balancing as far as possible the numbers of sites of each type (control or experimental) with each cone distance from the following categories which were measured from the film footage for each site: On line, <33cm, 33-66cm, cm, >100cm. 4 Due to only receiving data for one unlit nearside closure, no comparisons were possible for this setting. 10 CPR1306

13 Proportion of cars in zone Removing road danger lamps at relaxation road works two control sites, all with a cone distance category of <33cm. Table 2-1 shows the total number of car and HGV observations used in the analysis for zones 1 to 5 (lane distribution) and zones 1 to 4 only (position in adjacent lane). Table 2-1: Sample size for cars and HGVs at unlit offside closures (zones 1 5 combined, and zones 1 4 only) Zone Vehicle Control Experiment All Car HGV Car HGV Lane distribution of vehicles in control and experimental conditions unlit offside closures The analysis for cars revealed a statistically significant (at the 5% level) interaction between zone and condition. The data shows that when RDLs were absent, there was an increase in the proportion of cars moving to the lane further away from the longitudinal cones. Figure 2-6 shows these data. It should be noted that this pattern of findings could be partly due to the fact that control sites had higher traffic flows than experimental sites; as traffic flow increases, drivers in the left hand lane are more likely to try and overtake vehicles ahead of them, which increases the number of vehicles in the right hand lane. This effect was observed in the control data but it was not possible to deal with this completely through the weighting of the data. 100% 80% 60% 40% 20% 0% Zone 5 Zone 1-4 Control(2 sites) Experiment (5 sites) Figure 2-6: Proportion of cars in each lane (zone 5 = left lane, zones 1 4 = right lane) at unlit offside closures 5 The corresponding analysis for HGVs revealed no difference between conditions; in experimental and control sites, around 96% of HGVs were in zone 5. 5 In all graphs, error bars show the entire range of scores. 11 CPR1306

14 Proportion of HGVs in zone Removing road danger lamps at relaxation road works Position of vehicles in adjacent lane in control and experimental conditions unlit offside closures The analysis for cars revealed no difference in zone distribution between conditions; the absence of RDLs did not change the position of cars in the adjacent lane for unlit offside closures. Conversely, the analysis for HGVs revealed a statistically significant interaction between zone and condition, with a greater proportion of HGVs in zone 4 (i.e. further away from the cones) when the RDLs were absent. These findings should be treated with caution as the sample of HGVs in the adjacent lane was quite small. These data are shown in Figure % 80% 60% 40% 20% 0% Zone 4 Zone 3 Zone 2 Zone 1 Control (2 sites) Experiment (5 sites) Figure 2-7: Proportion of HGVs in zones 1 to 4 at unlit offside closures (zone 1 is closest to cone line) The slight tendency for vehicles to move further away from the cones when RDLs were absent needs to be treated with caution, due to the fact that traffic flow could not be fully controlled in the analysis, and due to the small sample size of HGVs. In summary, therefore, the data from unlit offside closures do not permit any firm conclusions to be drawn regarding the removal of RDLs Lit offside closures Data were collected from 10 experimental and 16 control sites with lit offside closures. After balancing as much as possible for cone distance, data were used from all of the experimental sites, and 10 of the control sites. Sites with cone distances in categories on line, <33cm and >100cm were used. It was not possible to completely balance control and experimental sites for cone distance; therefore the analyses were also run using only cone distance <33cm (the most numerous category) to see if this made any difference to the findings. Where it did this is noted in the text. 12 CPR1306

15 Table 2-2 shows the total number of car and HGV observations used in the analysis for zones 1 to 5 (lane distribution) and zones 1 to 4 only (position in adjacent lane). 13 CPR1306

16 Proportion of HGVs in zone Removing road danger lamps at relaxation road works Table 2-2: Sample size for cars and HGVs at lit offside closures (zones 1 5 combined, and zones 1 4 only) Zone Vehicle Control Experiment All Car HGV Car HGV Lane distribution of vehicles in control and experimental conditions lit offside closures The analysis for cars revealed no difference between conditions; in control and experimental sites, around 55 60% of cars were in zone 5. For HGVs, the analysis revealed a statistically significant interaction between zone and condition. The data show that when RDLs were absent, there was an increase in the proportion of HGVs moving to the lane further away from the longitudinal cones. This finding should be treated with caution however, since when cone distance was controlled by selecting only category <33cm sites, the effect was not statistically significant, although the pattern overall remained the same. 100% 80% 60% 40% 20% 0% Zone 5 Zone 1-4 Control (10 sites) Experiment (10 sites) Figure 2-8: Proportion of HGVs in each lane (zone 5 = left lane, zones 1 4 = right lane) at lit offside closures Position of vehicles in adjacent lane in control and experimental conditions lit offside closures Figure 2-9 and Figure 2-10 show the distributions of cars and HGVs respectively across zones 1 to 4 of the adjacent lane to the lit offside closures, by condition. The analysis for cars showed that the interaction between zone and condition did not quite reach statistical significance at the 5% level, while for HGVs it reached statistical significance at the 1% level. The pattern of data is the same in both cases, in that when RDLs were absent there was a tendency for both cars and HGVs to be more likely to be in zones 2 and 1. It should be noted however that the proportion of vehicles in zone 1 was extremely small in both cases. Displacement of vehicles from zones 3 and 4 to zone 2 (the zone 51 to 100cm from the cone line) appears to be responsible for the effect. 14 CPR1306

17 Proportion of HGVs in zone Proportion of cars in zone Removing road danger lamps at relaxation road works However the result for cars should be treated with extreme caution due to not reaching significance, and also because when cone distance was controlled (by selecting only category <33cm sites) the effect was not statistically significant, although the pattern overall remained the same. The fact that the pattern for HGVs and cars was the same suggests that vehicles that are in the adjacent lane do tend, if anything to drive closer to the cones when RDLs are absent. 100% 80% 60% 40% 20% 0% Zone 4 Zone 3 Zone 2 Zone 1 Control (10 sites) Experiment (10 sites) Figure 2-9: Proportion of cars in zones 1 to 4 at lit offside closures (zone 1 is closest to cone line) 100% 80% 60% 40% 20% 0% Zone 4 Zone 3 Zone 2 Zone 1 Control (10 sites) Experiment (10 sites) Figure 2-10: Proportion of HGVs in zones 1 to 4 at lit offside closures (zone 1 is closest to cone line) In summary, although the data from lit offside closures do not permit any firm conclusions to be drawn regarding the removal of RDLs, there is a trend in the data for a small shift in vehicle position in the adjacent lane towards the cone line when RDLs are removed. If any decision is made by the HA to remove RDLs from relaxation works, further monitoring is strongly recommended. 15 CPR1306

18 2.3.3 Lit nearside closures Data were collected from five experimental and six control sites with lit nearside closures. After balancing as much as possible for cone distance, data were used from all of the experimental sites, and five of the control sites. Sites with cone distances in categories on line and <33cm were used. It was not possible to completely balance control and experimental sites for cone distance; therefore the analyses were also run using only cone distance <33cm (the most numerous category) but this made no difference to any of the findings. Table 2-3 shows the total number of car and HGV observations used in the analysis for zones 1 to 5 (lane distribution) and zones 1 to 4 only (position in adjacent lane). Table 2-3: Sample size for cars and HGVs at lit nearside closures (zones 1 5 combined, and zones 1 4 only) Zone Vehicle Control Experiment All Car HGV Car HGV Lane distribution of vehicles in control and experimental conditions lit nearside closures The analyses for cars and HGVs revealed no differences in lane distribution between the control and experimental sites for lit nearside closures. Around 65% of cars and around 96% of HGVs were in zones 1 4 (corresponding to the left-hand lane, nearest the nearside closure) regardless of whether RDLs were used Position of vehicles in adjacent lane in control and experimental conditions lit nearside closures Figure 2-11 and Figure 2-12 show the distributions of cars and HGVs respectively across zones 1 to 4 of the adjacent lane to the lit nearside closures, by condition. Both analyses revealed a statistically significant interaction between zone and condition, with the pattern of data the same for cars and HGVs; in both cases when RDLs were absent vehicles were more likely to be in zones closer to the cone line. For cars this seemed to be due to vehicles moving from zone 4 to zones 3 and 2. For HGVs vehicles were mainly moving from zone 3 to zone CPR1306

19 Proportion of HGVs in zone Proportion of cars in zone Removing road danger lamps at relaxation road works 100% 80% 60% 40% 20% 0% Zone 1 Zone 2 Zone 3 Zone 4 Control (5 sites) Experiment (5 sites) Figure 2-11: Proportion of cars in zones 1 to 4 at lit nearside closures (zone 1 is closest to cone line) 100% 80% 60% 40% 20% 0% Zone 1 Zone 2 Zone 3 Zone 4 Control (5 sites) Experiment (5 sites) Figure 2-12: Proportion of HGVs in zones 1 to 4 at lit nearside closures (zone 1 is closest to cone line) In summary, although as before the data from lit nearside closures do not permit any firm conclusions to be drawn regarding the removal of RDLs, there is a significant trend in the data for a small shift in vehicle position in the adjacent lane towards the cone line when RDLs are removed. As before, if any decision is made by the HA to remove RDLs from relaxation works, further monitoring is strongly recommended Qualitative data from the on-road trial The on-road trial also involved monitoring for any specific incidents (such as cone strikes) during the data collection. Although the time frame for data collection would be 17 CPR1306

20 unlikely to permit formal analysis of any such incidents (due to their rarity) it was important these were captured and assessed in order to monitor the safety of the trial. Only two incidents were reported. In one, a cone (with RDL) was blown over by an HGV during control data collection at an unlit offside site. This incident was caught on camera, and is shown in Figure Figure 2-13: Cone fitted with a RDL toppling following passage of a HGV This incident occurred at an unlit site where the lane closure was on the offside. Turbulence from a passing HGV that was overtaking another HGV in the inside lane appeared to cause the cone to topple over. The cone can be seen beginning to lift in the left frame (due presumably to turbulence from the passing HGV) and is seen on its side in the right frame. The other incident was a report from an operative in Area 9 that a cone (without a lamp) had been displaced laterally (not toppled) from the cone line during experimental data collection. It was not possible to verify the incident from the video data for the site in question. 18 CPR1306

21 2.4 Summary of results of on-road trial The results from the on-road trial are not straightforward, and probably reflect the complex nature of the driving environment and how this interacts with the vehicles and types of relaxation work under investigation. Table 2-4 summarises the differences seen by closure type, for cars and HGVs. Table 2-4: Summary of changes seen between control and experimental conditions in the on road trial. Closure type Lane choice: when RDLs are absent Position of vehicles in adjacent lane: when RDLs are absent Unlit offside cars move to more distant lane* HGVs remaining in this lane move slightly further away** Lit offside HGVs move to more distant lane*** cars*** and HGVs remaining in this lane move slightly closer Lit nearside no change cars and HGVs remaining in this lane move slightly closer * Treat with caution possibly due to traffic flow differences ** Treat with caution based on very small sample size *** Treat with caution possibly due to cone distance differences The pattern of results seems to indicate that when RDLs are absent at lit sites this leads to a greater proportion of vehicles that are in the lane adjacent to the closure driving slightly closer to the line of cones (generally moving from over 100cm away to between cm away). There is also evidence that some HGVs move to the lane further away from the cone line when RDLs are absent. For unlit sites (offside closures only) the data show only evidence of vehicles moving further away from the cone line. Several of these findings need to be treated with caution due to possible confounds with traffic flow and cone distance. Overall, the on-road trials did not produce clear evidence either in favour of the continued use of road danger lamps on the longitudinal cones at relaxation works or in favour of their omission. 19 CPR1306

22 3 Ergonomics assessment of road danger lamp use This section summarises an assessment of the ergonomics issues relating to the tasks of deploying and retrieving cones at relaxation road works; the purpose of the assessment is to describe the level of risk of developing musculoskeletal problems from this task. In particular, the assessment compares the tasks of deploying and retrieving the cones with and without road danger lamps (RDLs). The assessment is based on an observational study of cone deployment, carried out in a controlled environment that provides a representation of the physical actions required to carry out cone deployment and retrieval. 3.1 Method Three traffic management operatives from RenTeQ Traffic Management were recruited to demonstrate the tasks of laying and retrieving cones with and without RDLs, as they would in real operations. Each operative was filmed to capture accurately their relative trunk, neck and limb positions. The three operatives also demonstrated the task of attaching RDLs to cones in order for force measurements to be taken. Forces were measured by a Kistler dynamometer situated under the cone used during the demonstrations. Short interviews were also conducted with the operatives to obtain information about their experience of conducting the tasks and their opinions of it, including parts of the tasks that they find easy, parts that they find difficult, and where they feel the risks lie. These short interviews highlighted specific parts of the tasks that warranted attention during analysis. Following the observations, the video was reviewed and key postures identified. Fortyeight postures in total were selected for analysis. The postures chosen include those relating to the task of attaching RDLs to cones, and to deploying and retrieving cones with RDLs attached. Postures associated with the deployment and retrieval of cones only were also included as a comparison. Each posture was subjected to review using the Rapid Entire Body Assessment (REBA) coding tool (Hignett & McAtamney, 2000). This is a rigorous and accepted method for coding and scoring postures according to: The relative positioning of different component parts of the body A broad categorisation of the actions being performed, including: o o o How much load or force is being supported How this load is applied How good the hand-holds are on the object being interacted with The scoring system works by combining scores for individual body parts to give an overall score for the posture. This score is initially out of 12 (with increasing score corresponding to increasing risk) with an additional three points available depending on: Whether or not the posture is held static for long periods The involvement of repetitive movements Whether the action causes large range changes in posture or an unstable base 20 CPR1306

23 The final score is associated with a risk level which provides guidance on how important it is to take action to reduce the level of musculoskeletal risk. The risk levels are categorised as negligible, low, medium, high and very high. The higher risk levels indicate which activities should be tackled as a priority. 3.2 Results Nine key activities were observed where use of RDLs caused operatives to perform tasks differently or to adopt postures different from those seen with cones free of RDLs. These were: Retrieving an RDL (only applicable when using RDLs) Attaching an RDL to a cone (only applicable when using RDLs) Lifting a cone off the flatbed Holding a cone, ready to place it on the carriageway Placing a cone on to the carriageway Catching a cone from the carriageway Swinging the cone up from the carriageway to return it to the flatbed Swinging the cone over the threshold of the flatbed Placing the cone on the flatbed During the interviews, the operatives stated that they typically lay out a 2km enclosure during one shift which takes around minutes. Not including the taper, this would require the operatives to lay out around 100 cones but they may lay out as many as 200 cones in a shift. Each action is therefore repeated several times every minute and this was taken into account during the REBA analysis. Anecdotally, operatives reported that RDLs make the task of laying out cones more difficult. This is because operatives must position their hands in such a way that RDLs do not fall onto the carriageway. The RDLs must also face in the right direction when the cone is in position and so operatives must ensure that the RDL is stable when the cone is being placed. Table 3-1 presents the headline findings from the REBA analysis and identifies the levels of risk from the postures that were adopted during the activities. More detailed data (including raw REBA scores for each activity) can be found in Appendix A. 21 CPR1306

24 Table 3-1: REBA level for each activity associated with laying and retrieving cones, with and without RDLs. Green shading denotes that the REBA level was lower without RDLs, red that it was higher REBA scores for. Operative 1 Operative 2 Operative 3 Without RDL With RDL Without RDL With RDL Without RDL With RDL Retrieving an RDL - High - High - Medium Attaching an RDL to a cone - Medium - Medium - Medium Lifting a cone off the flatbed High High Medium Medium High 6 Medium Holding a cone before placing it on the carriageway Placing a cone on the carriageway Catching a cone from the carriageway Swinging the cone up to return it to the flatbed Swinging the cone over the threshold Placing the cone on the flatbed Medium High Medium Medium Activity eliminated High Medium Medium High High Medium Medium Medium High Very high Very high High High High Very high Medium High High High High High High High High High Medium Medium Medium High Medium High Results from the analysis confirm that the use of RDLs is associated with increased risk of developing musculoskeletal disorders 7. In general, REBA scores were higher when operatives performed activities using a cone with an RDL attached than when using a cone only, although in some cases the differences were not large enough to change the risk level. Clear differences were seen in a number of activities where omission of RDLs was associated with lower risk levels, as illustrated by the green shading in the table. The risk level for the step of holding a cone, ready to place it on the carriageway decreased for one operative and for another the risk was eliminated completely as the activity was no longer necessary. Placing the cone on the flatbed also achieved a reduction in risk for two operatives. However, the greatest potential benefits in terms of risk level change were associated with catching a cone from the carriageway and swinging the cone up. These activities were associated with all the very high risk levels seen in the assessment; omission of the road danger lamps resulted in significant or very significant decreases in risk. As the highest risk levels from the REBA scoring process indicate the areas which should be targeted (as described above), these two activities would represent the areas that would be a priority for action. 6 This anomalous result was caused by the operative adopting a less favourable neck position when lifting the cones without RDLs. 7 It is also worth noting that the REBA levels for the baseline condition of laying cones without RDLs are generally medium to high. Monitoring of these activities through standard occupational safety measures is advised. 22 CPR1306

25 The key factors that brought about the differences in REBA scores were: The position of RDLs, when they were used, meant that operatives were only able to gain a pinch grip on a cone instead of the power grip that a cone without an RDL permitted Poor grip when using a cone with an RDL attached led to operatives using poorer wrist positions than when conducting the activities with a cone only Two extra activities were identified when operatives laid out cones using RDLs. They were retrieving an RDL and attaching an RDL to a cone The forces that were measured at the bottom of the cone when RDLs were attached were high (between 10kg and 80kg). There is likely to be a moderating effect of the lamp and cone on how much of this force is transferred to the operative s hand; nevertheless, operatives who do not let go of the RDL before the impact would experience a shock force as the lamp hits the cone and comes to a stop. For repetitive activities, hand forces much smaller than those measured during this study increase the risk of developing musculoskeletal disorders affecting the hands, arms, wrists, elbows and shoulders. One limitation of the study is that due to operational constraints, only a small number of individuals were studied. Since there is no defined policy or manual handling process for these activities, the operatives unsurprisingly showed differences in the methods that they used to deploy and retrieve cones. Although it is not possible to confirm that the methods used by these participants and the postures they adopted would be the same for all operatives, the results indicate that the risk of developing musculoskeletal disorders is higher when using RDLs. The results of the ergonomics review would therefore support a case for the removal of RDLs as this would reduce the risk of operatives developing musculoskeletal disorders. A more detailed description of the ergonomics assessment is available in Appendix A. 23 CPR1306

26 4 Calculation of injury risk sub-value Omitting road danger lamps from longitudinal coning of works areas may reduce risk for road workers in three main areas: Exposure during deployment and retrieval Routine maintenance of the lamps on the cone line Debris from impacts 4.1 Risk exposure calculation The Measurement of Injury Risk (MIRi) Index has been used to calculate a risk sub-value associated with deployment and retrieval of longitudinal coning, with and without road danger lamps. The Measurement of Injury Risk (MIRi) Index (as described in Fowler et al., 2011) was developed for the Agency by TRL for the purpose of determining the risk associated with carrying out carriageway crossings in the advance signing zone on the approach to road works. It allows investment decisions to be prioritised based on risk benefit and reduction in carriageway crossings. The MIRi Index itself is a score which takes into account the specifics of a given temporary traffic management operation, including the activities which comprise it, the operatives involved and the associated risks. The use of the MIRi Index and its component scores thus provides a means of assessing the effects of changes in site work methodology and practice on risk (safety initiatives), in order to demonstrate safer working practice and progress towards Aiming for Zero (AfZ) targets. Assumptions for the calculations relating to the use of road danger lamps are given in A.1.1.1Appendix B and are based on similar assumptions for the calculation of the MIRi Index values associated with laying the taper and advance signing. The calculation of the MIRi Index sub-value resulted in the scores given in Table 4-1 below: Table 4-1: Calculation of MIRi index sub-values Closure type Speed of operation Average MIRi Index sub-value (with RDLs) Average MIRi Index sub-value (without RDLs) Reduction in subvalue achieved from removal of road danger lamps Nearside Closure High estimate Average Low estimate Offside Closure High estimate Average Low estimate CPR1306

27 The MIRi Index sub-value for installation of the longitudinal coning decreases by between 25% and 17% when RDLs are not used. This occurs as a consequence of the increase in speed at which cones without RDLs can be placed or retrieved, which in turn reduces exposure to a hazardous condition (a vehicle striking the traffic management installation vehicle) and the associated injury consequences for the road workers. Detailed results for the range of speed changes examined for this calculation are given in A.1.1.1Appendix B. The risks associated with the use of road danger lamps cannot be directly compared to those associated with changes to the advance signing or lane change zone (for example the introduction of sign simplification via IAN 150/12). This is because this work has derived a MIRi Index sub-value specifically applicable to one element of the deployment of the longitudinal (namely setting and retrieving cones), making comparison of its absolute value with the total MIRi Index value that is associated with all elements of advance sign and taper installation and retrieval inappropriate. However, the relative change in risk achievable by the omission of road danger lamps can be compared with the relative change in risk achieved via sign simplification / IAN150/12. When compared in this way, the elimination of road danger lamps provides a per-kilometre risk reduction equivalent to 2.5% of the total risk reduction achieved from sign simplification via IAN150/12. This means on a typical 4km lane closure, the omission of road danger lamps could achieve a risk reduction to road workers equivalent to 10% of the total risk reduction achieved from the introduction of IAN150/12, although it should be noted the omission of road danger lamps would not achieve any reduction in carriageway crossings. 4.2 Additional risks associated with use of road danger lamps When a lane closure is being installed, the road danger lamps are preplaced on the cone before it is placed on the carriageway. As has been seen from the ergonomics review, the lamps require that the cones are placed at a specific alignment when dropped onto the carriageway in order to ensure they are facing oncoming traffic. In addition, when the cone is dropped on to the carriageway the lamp may be jolted off the cone and fall onto the carriageway surface. In this case, the TM crew must stop, enter the live lane, retrieve the lamp and replace it on the cone before continuing to place the longitudinal. It is not currently possible to quantify how often this occurs, but unarguably it places the operatives who must enter the live lane to retrieve the lamp at significantly increased risk compared to their relative place of safety within the coning well on the traffic management vehicle. Elimination of the need to deploy road danger lamps would eliminate this intermittent yet significant risk. Once the cones have been placed on the carriageway, the traffic management is regularly inspected to ensure it remains effective. As part of this routine activity, maintenance work should be undertaken to ensure road danger lamps are correctly aligned and are lit. It is unlikely that every single road danger lamp deployed will operate correctly and it is likely that during a typical eight-hour work period some lamps that were initially operational may become defective. Maintenance requires that the lamps are re-aligned and any lamps with spent batteries are either replaced or the battery changed on-site. While this activity can be carried out from within the closed lane, it requires the operative maintaining the traffic management to enter the safety zone (this is permitted within Chapter 8 paragraph O3.2.5) and thus 25 CPR1306

28 be in close proximity to fast-moving traffic. This places the operative at risk which can be eliminated if the requirement to maintain the lamps is removed by their omission from the traffic management layout. Regarding this point, several crews commented during the on-road trials that maintenance time was reduced when road danger lamps were not used: All ok tonight. No negative remarks from anyone. TM crew said installation & removal was much easier & took less time. Also commented no lamps to check made Maintenance run easier. All went ok last night got some feedback from the TM crew, very happy put the closure like this maintenance runs easier faster to pick up. Anecdotal evidence from operatives suggests that in a typical 1Km closure, five lamps out of the 55 used will need battery replacement as part of a routine maintenance check. This is usually carried out as described above, but if a large number of lamps require maintenance then the crew would work from the vehicle, with an operative in the well changing the lamps for new (and presumably operational) ones. It has been suggested that the number of serviceable road danger lamps on any given closure may be lower than previously thought, which in turn suggests that any benefit gained from their use may also be reduced. While this cannot be put forward as an argument in favour of omitting lamps, the lower numbers of inoperative lamps does not seem to cause the industry significant concerns nor cause increased numbers of accidents. This is further confirmed by informal indications that having up to 25% lamps inoperative on any given closure may be considered acceptable by traffic management crews, with maintenance only being carried out when additional lamps fail. Indeed, some hard shoulder closures have been seen where 100% of the road danger lamps were inoperative. Such evidence suggests that loss of the benefit from the use of lit road danger lamps is not perhaps perceived by the industry as of significant concern. 4.3 Potential risk from debris from impacts Reports from road workers suggest that cone strikes within the longitudinal sections of lane closures are a common occurrence, even with functioning road danger lamps fitted. Opinions gained from across the industry suggest that most incidents where cones are struck will go unreported but that as many as 1 in 10 closures will suffer from some sort of damage, either from vehicle impacts or with cones having been moved or toppled by vehicle turbulence as noted during a control deployment during on-road trials (see Section 2.3.4). It is also claimed (though not confirmed) that the presence of road danger lamps may exacerbate the tendency for cones to topple, as the road danger lamp presents a larger surface area for the vehicle turbulence to act upon as well as placing a weight (the lamp battery) at top of the cone where it can have a significant influence on the centre of gravity of the cone/lamp combination. Regardless of how a cone is knocked over, each instance requires corrective action, whether simply replacing the cone and lamp upright or the collection and replacement of damaged cones and broken lamps, such as those shown in Figure CPR1306

29 Figure 4-1: Damaged road danger lamps. Upper lamp has suffered vehicle impact During an impact, it is probable that the road danger lamp will become separated from the cone, either by falling off or being knocked off. If displaced from the cone during an impact, the lamp has the potential to be thrown from the cone with some force and so become a missile. While the displaced lamps may enter live lanes, it is reported that they are more likely to be propelled into the work zone and in some cases be thrown over the central reserve barrier into the oncoming traffic. This may be a particular problem if HGV drivers deliberately attempt to flip cones by driving over the edge of their bases; as the cones are thrown sideways the lamps will be ejected from the cones into the work zone or potentially onto the opposing carriageway. If road danger lamps are propelled into or through the work zone, road workers are placed at additional risk of injury from the debris. In addition, where lamps enter live lanes they or the debris they produce may damage fast moving vehicles and so must be retrieved to ensure the safety of road users. Retrieval of this debris places road workers (whether ISU crews or Traffic Officers) at risk, which could be eliminated by removing the requirement for the use of the road danger lamps. These additional considerations are impossible to quantify as the types of incidents involved are ephemeral. However, they represent a real qualitative risk to road workers that could be eliminated if the requirement for the use of road danger lamps was relaxed. Overall, the calculation of MIRi values, along with consideration of the other anecdotal issues reported in this chapter suggest that the removal of RDLs would result in a safety benefit to road workers. 27 CPR1306

30 5 Discussion of findings The findings from the three phases of the project regarding the removal of road danger lamps from longitudinal coning can be summarised as follows: 1. The on-road study showed that in lit relaxation works (offside and nearside) there is some evidence that when RDLs are absent from longitudinal coning vehicles that are in the lane adjacent to the line of cones are more likely to drive in lateral positions that are slightly closer to the cones than when RDLs are present. The picture in unlit relaxation works is less clear, but if anything goes in the other direction for HGVs. Overall there is no firm indication that the removal of RDLs would be beneficial or detrimental, based solely on the on-road trial data. 2. The ergonomics assessment showed that almost without exception when compared with RDL-free cones, the tasks of laying and retrieving cones with RDLs resulted in postures that are associated with a higher risk of musculoskeletal injury. 3. The MIRi analysis showed that there is likely to be a safety benefit to roadworkers if RDLs are no longer used at relaxation works. The aim of the study was to establish if there is an overall safety case to support the removal of road danger lamps from relaxation works. The data do not permit an unequivocal answer to this question; this is due to the fact that although the ergonomics and MIRi elements of the study support the case, the findings from the on road element of the study are less clear cut. The responses seen in drivers were variable as is often the case when attempting to quantify human behaviour, suggesting that a number of underlying psychological mechanisms are likely to have been at play. In some conditions there was a suggestion that drivers were moving further away from cones when road danger lamps were absent, through shifting to a further lane. Those drivers remaining in the adjacent lane tended to drift slightly closer to the cones, although HGV drivers in the unlit offside closures actually drifted further away if anything (although this finding was based on a small sample). It is tempting to speculate on the psychological mechanisms that in combination can explain the findings. For example perhaps some drivers notice that the cone line is difficult to make out and make a conscious decision to move away from it, while others do not notice this and drift closer to the cone line due to its lack of perceptual conspicuity. Another interesting consideration is the position drivers adopt within a lane when they only have the white lines to guide them. As part of this work, video data from another study were used to measure the position of vehicles in the middle lane of a three-lane motorway 800m from the point at which the left lane is closed for road works. This suggested that without the presence of cones on either side of the vehicle, the majority of vehicles were in what would be classed as Zone 4 in the current study, when measured from the left hand line of the middle lane. Conversely, the majority of cars in the current study drove in Zone 3 when passing the cone line at nearside works, that is to say closer to the cones irrespective of whether or not road danger lamps were fitted. This raises the question (which cannot be answered from the current data) as to the magnitude of the effect of the white lining on the road and whether the presence of cones themselves causes a change in the baseline lane position adopted by drivers. 28 CPR1306

31 Such speculation, while potentially interesting, does not seem necessary within the context of this study however. The key conclusion regarding the behaviour data from the on-road trials would seem to be that there are a number of behavioural responses seen in drivers and that understanding the net safety benefit of these would require further investigation. Another reason why it is difficult to draw firm conclusions related to safety from the onroad study is that although the measurement of lateral position has validity as a proxy for safety (driving closer to the cone line gives less of a margin for error), the actual link between this measure and risk is not established. That is to say, if a vehicle drives a given distance closer to the cone line (all other things being equal) there is no way of knowing the corresponding increase in the likelihood of the vehicle actually striking the cone line. In addition, even if this happens, there is currently no data on the risk that such an incident would present to road users or road workers. Overall, the changes in driver lane position were small 8 ; no obvious, large effects that would signal a gross change in behavioural response were apparent. Nonetheless, the changes may be indicative of behavioural effects that carry with them a possible risk of incidents in the future. For example the fact that drivers in the adjacent lane seem to drift closer to the cones (especially for HGVs) may make cone strikes more likely (even if very rare), and this has some implications for safety especially if it happens adjacent to the works. Further monitoring of these issues seems warranted in any continuation of this approach. Taken together, the findings from the on-road trial do not offer conclusive evidence for or against the removal of RDLs from relaxation works. The ergonomics and MIRi analyses, on the other hand, firmly support the removal of RDLs. 8 One reason why this might not be surprising is that the performance of retroreflective markings on traffic cones, especially in the rain, is likely to have improved since road danger lamps were first prescribed in 1975 (Lyus, personal communication). 29 CPR1306

32 6 Conclusions and recommendations The findings from the ergonomics and MIRi analyses support the safety case for the removal of road danger lamps from relaxation works. The MIRi analysis suggests that a risk reduction equal to % of the benefit from sign simplification could be achieved through the omission of road danger lamps. In addition, the ergonomics analysis shows that almost without exception the use of road danger lamps on longitudinal coning leads to road workers adopting postures that are associated with a greater risk of musculoskeletal injury, when compared with those postures they adopt when laying and retrieving cones without road danger lamps. The findings from the on road element of the study do not permit an unequivocal conclusion either for or against the removal of road danger lamps from relaxation works. Instead they suggest that drivers respond to the removal of road danger lamps in a variety of ways, at least some of which may be interpreted as resulting in a slightly higher risk of striking the cone line (if we assume that driving closer to the cone line leaves less margin for error). This could result in at least a slight increase in accident risk which nonetheless may be broadly tolerable. Based on the results from this study, it is therefore considered that there is no categorical evidence that indicates continued use of road danger lamps is essential to the safety of road workers or road users. Given the positive impact of removing lamps suggested by the ergonomics and MIRI analyses, and the uncertainties surrounding the driver response, the following recommendations are therefore made: 1. The trial of omitting road danger lamps should be continued with monitoring of the driver response (possibly through incident and near miss reporting from road workers on the ground) over a much longer period of time than that covered by this study 2. This should be rolled out via an Area Management Memo, permitting the relaxation of the requirement to use road danger lamps provided monitoring data are provided (following the precedent of AMM125/10 for the innovative taper). 3. The Supply Chain should be involved with the development of the roll-out process and be consulted on the option to have a staged approach to removing the lamps. 4. Once sufficient data are received and analyses are completed indicating the safety case is favourable, the HA should consider as to whether omission of road danger lamps from longitudinal coning at relaxation scheme traffic management is incorporated into formal guidance. A final recommendation of this study is that musculoskeletal injury rates be monitored in those operatives who spend time laying and retrieving longitudinal coning on all road works (even without road danger lamps) to confirm the occupational injury risk from this task. 30 CPR1306

33 Acknowledgements The work described in this report was carried out in the Road Safety Group of the Transport Research Laboratory. The authors are especially grateful to the following organisations for their full involvement and support of the on-road trial: Amey, Connect Plus, RenTeQ Traffic Management Limited, HW Martin (Traffic Management) Ltd, and Chevron Traffic Management Limited. Special thanks are also due to Leon Ingram (TSCO) and Sandra Cusick at Amey, and Colin Rudd (TSCO) Stuart Baker (TSCO) and Simon Taylor at Connect Plus, as well as the many individuals who facilitated data collection. References Fowler C, Clark S, Rillie I, Cuerden R & Smith L (2011). Development of the Measurement of Injury Risk (MIRi) Index. Unpublished Project Report. CPR1106. Crowthorne: Transport Research Laboratory. Barr, E. & Barbe, M.F. (2002). Pathophysiological tissue changes associated with repetitive movement: a review of the evidence, Physical Therapy, 82(2), pp Hignett S, & McAtamney, L. (2000). Rapid Entire Body Assessment (REBA). Applied Ergonomics, 31, pp Lyus, B. (2012 personal communication in to Siggi Clark, TRL) TSO (2009) The Traffic Signs Manual Chapter 8 Part 1: Design. London: The Stationery Office TSO (2002) The Traffic Signs Regulations and General Directions London: The Stationery Office 31 CPR1306

34 Appendix A Ergonomics assessment of road danger lamp use This appendix documents an assessment of the ergonomics issues relating to the tasks of deploying and retrieving cones at relaxation road works; the purpose of the assessment is to describe the level of risk of developing musculoskeletal problems from this task. In particular, the assessment compares the tasks of deploying and retrieving the cones with and without road danger lamps (RDLs). The assessment is based on an observational study of cone deployment, carried out in a controlled environment that provides a representation of the physical actions required to carry out cone deployment and retrieval. A.1 Method Operatives were observed as they created and removed lines of cones with and without RDLs attached. Force testing was conducted to measure the forces applied when RDLs were attached to cones and operatives participated in semi-structured interviews to provide information about their experience of the tasks. A.1.1 Participants Three traffic management operatives from RenTeQ Traffic Management were recruited to demonstrate the tasks of laying and retrieving cones with and without RDLs; each operative had at least 5 years experience of carrying out these activities on the HA network. Operatives heights were taken into consideration as this can have a significant influence on how a person holds their body and interacts with equipment. One operative was shorter and two operatives were taller than the average British male. Operatives heights and percentile values (the percentage of the population that the operative is taller than) are shown in Table A1. Table A1: Operatives heights Operative Height Percentile A.1.2 Test location This observation could not be undertaken in operational conditions since the work being studied is conducted at night and thus suitable footage for the analysis could not be obtained. All the demonstrations were therefore carried out on the motorway section of TRL s track facility. This section simulates a real motorway environment complete with three running lanes, a hard shoulder and gantries. Figure A1 shows the layout of the test track, with the motorway section marked near the bottom. 32 CPR1306

35 Figure A1: TRL test track layout A.1.3 Equipment A Traffic Management vehicle was used during the observations and was loaded with the materials required, including: Traffic cones approximately 8kg each and 1m tall RDLs approximately 1kg in weight The equipment used during the observations was the same as equipment that is used operationally. A.1.4 Trial procedure A Observation of the tasks The three operatives were told the purpose of the study and that the intention was to observe them carrying out activities as they would on a normal carriageway. For the purpose of this report, Operative A was the person standing on the flatbed removing the cone from a stack or returning the cone to the stack; Operative A was also responsible for attaching and removing RDLs if required. Operative B was the person standing in the well placing cones on the carriageway or removing them. The third operative drove the vehicle. All operatives performed the roles of Operative A and Operative B during the observations. Approximately 20 cones were deployed and retrieved by each operative. 33 CPR1306

36 The procedure was: A. Setting out 1. Operative A removes a cone from a stack 2. Operative A picks up an RDL (if required) 3. Operative A attaches the RDL to the cone (if required) 4. Operative A passes the cone (with RDL if attached) to Operative B 5. Operative B places the cone (with RDL if attached) onto the carriageway This process was repeated for each cone. B. Clearing away 1. Operative B picks up the cone (with RDL if attached) from the carriageway 2. Operative B passes the cone (with RDL if attached) to Operative A 3. Operative A removes the RDL from the cone (if required) 4. Operative A throws the RDL aside (if required) 5. Operative A returns the cone to a stack This process was repeated for each cone. Using a video camera, operatives were recorded demonstrating the tasks of deploying and retrieving cones twice; once with the use of RDLs and once without. Each run was filmed from a different angle (approximately as shown in Figure A2) to capture accurately the relative trunk, neck and limb positions of the operatives. Figure A2: Relative positioning of items within the trial environment 34 CPR1306

37 A Measurement of forces The three operatives also demonstrated the task of attaching RDLs to cones in order for force measurements to be taken. Forces were measured by a Kistler dynamometer situated under the cone used during the demonstrations; a bespoke plate was manufactured that ensured a good fit between the cone and the device to minimise the chances of the cone slipping off the dynamometer during the trials. During the trials, each of the operatives attached the RDLs to the cone five times and forces were measured at the bottom of the cone. In addition, the masses of the cone and the lamp used during the trials were measured. A Interviews Short interviews were also conducted with the operatives to obtain information about their experience of conducting the tasks and their opinions of it, including parts of the tasks that they find easy, parts that they find difficult, and where they feel the risks lie. These short interviews highlighted specific parts of the tasks that warrant some attention during analysis. A.1.5 Data analysis Following the observations, the video was reviewed and key postures identified; these were chosen to be representative of the main activities carried out during the tasks. These postures typically represent the most extreme body position adopted in order to complete each activity, although for some activities more than one posture was selected, either because there were clearly different movements within an activity or because multiple ways of conducting an activity were observed. Forty-eight postures in total were selected for analysis. The postures chosen include those relating to the task of attaching RDLs to cones, and to deploying and retrieving cones with RDLs attached. Postures associated with the deployment and retrieval of cones only were also included as a comparison. Each posture was subjected to review using the Rapid Entire Body Assessment (REBA) coding tool (Hignett & McAtamney, 2000). This is a rigorous and accepted method for coding and scoring postures according to: The relative positioning of different component parts of the body A broad categorisation of the actions being performed, including: o o o How much load or force is being supported How this load is applied How good the hand-holds are on the object being interacted with The scoring system works by combining scores for individual body parts to give an overall score for the posture. This score is initially out of 12 (with increasing score corresponding to increasing risk) with an additional three points available depending on: Whether or not the posture is held static for long periods The involvement of repetitive movements Whether the action causes large range changes in posture or an unstable base 35 CPR1306

38 The final score is associated with a risk level which provides guidance on how important it is to take action to reduce the level of musculoskeletal risk. The risk levels are categorised as negligible, low, medium, high and very high. The higher risk levels indicate which activities should be tackled as a priority. Table A2 shows how the risk level categories correspond to REBA scores. Table A2: Risk levels for REBA REBA Score Risk Level 1 Negligible 2-3 Low 4-7 Medium 8-10 High Very high The coding was carried out by one researcher. A random subset was coded independently by another researcher to check for consistency in the application of the coding scheme. Overall, the scoring was deemed to be consistent. A.2 Results and Discussion Nine key activities were observed where use of RDLs caused operatives to perform tasks differently or to adopt postures different from those seen with cones free of RDLs. These were: Retrieving an RDL (only applicable when using RDLs) Attaching an RDL to a cone (only applicable when using RDLs) Lifting a cone off the flatbed Holding a cone, ready to place it on the carriageway Placing a cone on to the carriageway Catching a cone from the carriageway Swinging the cone up from the carriageway to return it to the flatbed Swinging the cone over the threshold 9 of the flatbed Placing the cone on the flatbed During the interviews, the operatives stated that they typically lay out a 2km enclosure during one shift which takes around minutes. Not including the taper, this would require the operatives to lay out around 100 cones but they may lay out as many as 200 cones in a shift. Each action is therefore repeated several times every minute and this was taken into account during the REBA analysis. 9 Threshold is used to refer to the edge of the flatbed. It is the point at which if the operative lets go of the cone, it lands on the flatbed rather than the foot-well. 36 CPR1306

39 This section presents the results from the REBA analysis and identifies the levels of risk from the postures that were adopted during the activities. Information gleaned during the interviews is also discussed. A.2.1 Retrieving an RDL This activity is only carried out when RDLs are used. When using an RDL, operatives have to retrieve one from a pile by bending down. The pile of RDLs is typically stowed in the middle of a surrounding wall of cones on the flatbed, and diminishes over time as RDLs are retrieved. Figure A3 shows the posture that an operative adopted when conducting the activity. Figure A3: Posture adopted for retrieving an RDL Table A3 presents REBA scores for each operative for this activity. The task of retrieving RDLs produced medium and high risk postures as operatives flexed and twisted their trunks, and fully extended their upper and lower arms. Operative 3 produced a lower REBA score than the other operatives because he did not flex his trunk to the same extent. This may have been because Operative 3 was the shortest of the three to demonstrate the activity and his height meant he did not need to bend down as far. In addition, the pile of RDLs was plentiful at the time he carried out the activity and he was able to take an RDL from the top of the pile. As the pile of RDLs diminishes, it is likely that Operative 3 would flex his trunk to a greater extent, potentially increasing the level of postural risk, assuming he adopted a similar position to the other operatives. Table A3: REBA scores for retrieving an RDL Operative REBA Score Risk Level 1 9 High 2 9 High 3 7 Medium A.2.2 Attaching an RDL to a cone Again this activity is only required when laying cones with RDLs attached. Figure A4 shows the posture that an operative adopted when attaching an RDL to a cone. 37 CPR1306

40 Figure A4: Posture adopted for attaching an RDL Attaching RDLs to cones produced medium risk postures. Although operatives were able to adopt relatively good postures, the factors that drove the risk level from low to medium were the neck positions that operatives adopted, the repetitive nature of the activity and the rapid build-up of force (known as shock force) experienced when attaching the RDL. The REBA scores and risk levels for attaching an RDL to a cone are presented in Table A4. Table A4: REBA scores for attaching an RDL to a cone Operative REBA Score Risk Level 1 5 Medium 2 5 Medium 3 5 Medium At the bottom of the cone, the forces recorded by the dynamometer ranged from 10kg to 80kg. Although the forces on the operative s hand would be lower than 80kg (due to factors such as absorption of some force by the cone and lamp), some previous studies have defined hand forces over 4kg to be high (Barr & Barbe, 2002) for repetitive activities. The forces, repetition and speed of build-up taken together give rise to concerns about operatives developing musculoskeletal disorders affecting their hands, wrists, arms, elbows or shoulders. During one of the demonstrations; however, an operative let go of the RDL before it fully connected with the cone. This posture was not recorded for coding as it was only observed once while force measurements were being taken; however, it would have most likely resulted in a lower risk level from the activity. 38 CPR1306

41 A.2.3 Lifting a cone off the flatbed Lifting a cone off the flatbed was identified as a key activity for laying cones. The operatives performed this activity slightly differently when the cone was fitted with an RDL than when they were deploying cones only. The use of an RDL resulted in operatives altering the position of their hands. Figure A5 shows an operative demonstrating the activity. Figure A5: Lifting a cone off the flatbed without an RDL attached (left) and with an RDL attached (right) With an RDL attached, operatives were not able to position their hands to get a good grip on the cone, and this in turn increased the score attributed to the wrist position (see Figure A5). With RDLs attached, operatives lifted the cones using a pinch grip, where operatives used their fingertips only to lift the cone. A pinch grip is considered to be weaker than the power grip that the operatives were able to use when handling a cone only. A power grip is where the fingers and thumb are used to clamp the cone against the palm. With the RDL attached to the cone, REBA scores for Operatives 1 and 2 increased, but the risk levels did not change. Surprisingly the REBA score for Operative 3 was lower when using a cone with an RDL attached. The risk level decreased from high in the cone only condition to medium when using a cone with an RDL attached. This difference occurred because the operative adopted a less favourable neck position when lifting the cone without an RDL off the flatbed. The REBA scores for lifting a cone off the flatbed, with and without an RDL, are presented in Table A5. 39 CPR1306

42 Table A5: REBA scores for lifting a cone off the flatbed Without RDL With RDL Operative REBA Score Risk Level REBA Score Risk Level 1 9 High 10 High 2 5 Medium 7 Medium 3 8 High 7 Medium A.2.4 Holding a cone, ready to place on the carriageway When deploying cones, operatives tended to hold the cone for a few seconds, ready to place it in the correct position on the carriageway. The operatives had different methods for doing this activity. One operative held the cone close to his body and relatively high up from the carriageway while another used one hand to hang the cone very close to the carriageway before dropping it into position. Operative 3 was able to eliminate this step completely when using cones without RDLs attached (and hence provides no data for this step). Figure A6 shows an operative carrying out the activity with and without an RDL attached. Figure A6: Holding a cone without an RDL attached (left) and with an RDL attached (right) before placing it on the carriageway The activity of holding a cone without an RDL attached, ready to place it on the carriageway produced medium risk postures. One medium risk posture and two high risk postures were adopted when using a cone with an RDL attached. The position of the RDL on the cone meant that operatives were less able to obtain a good grip than when using a cone only. To be able to hold the cone with the RDL attached, operatives flexed their wrists and one operative flexed his upper arm past 90 and extended it outwards (see Figure A6). During the interviews, operatives noted that placing the cones with RDLs attached on the carriageway is difficult because the RDL must face in the right direction. Along with poor 40 CPR1306

43 grip, this requirement could have contributed to the high risk postures that were assumed since operatives had to adapt their postures in order to ensure that the RDL would face the right way when the cone was placed on the carriageway. It could also explain why one operative was able to eliminate the activity of holding the cone in position when no RDL was used. Without an RDL, he was able to remove the cone from the flatbed and place it onto the carriageway without pausing to ensure the RDL would point in the correct direction. The REBA scores for holding the cone in position are presented in Table A6. Table A6: REBA scores for holding a cone before placing it on the carriageway Without RDL With RDL Operative REBA Score Risk Level REBA Score Risk Level 1 5 Medium 9 High 2 5 Medium 5 Medium 3 (activity eliminated) 9 High A.2.5 Placing a cone on the carriageway Placing a cone on a carriageway was identified as a key activity for the task of laying cones. Figure A7 shows an operative demonstrating the activity. Figure A7: Placing the cone on the carriageway without an RDL (left) and with an RDL (right) With and without an RDL attached, operatives adopted medium and high risk postures when placing a cone on the carriageway. There were no changes in the risk levels achieved by each of the operatives when placing a cone with an RDL attached when compared with placing a cone without an RDL. The REBA score only increased slightly for 41 CPR1306

44 one operative when placing a cone with an RDL attached. Despite small or no changes in the REBA scores, it was clear that the RDL made it more difficult for operatives to grip the cone. The operatives stated that placing a cone with an RDL attached onto a carriageway is difficult because they must ensure that every RDL is facing in the correct direction when the cone is in position. Operatives also highlighted a problem that RDLs can swing around the cone and so they must stabilise the lamp. Indeed, one of the operatives was seen to use his free hand to ensure the RDL was facing the right way and to prevent it from spinning around the cone (see Figure A7). Table A7 presents the REBA scores for each of the operatives who demonstrated this activity. Table A7: REBA scores for placing a cone on the carriageway Without RDL With RDL Operative REBA Score Risk Level REBA Score Risk Level 1 6 Medium 7 Medium 2 8 High 8 High 3 7 Medium 7 Medium A.2.6 Catching a cone from the carriageway One key posture that was identified for clearing the equipment away was catching a cone from the carriageway. Figure A8: Operative 1 catching the cone without an RDL (left) and with an RDL (right) With an RDL attached to a cone, operatives adopted high and very high risk postures when picking up the cone from the carriageway. When performing the activity with a cone only, operatives adopted medium and high risk postures. The REBA scores were lower for all operatives when performing the activity with a cone only, though the 42 CPR1306

45 difference was not enough to reduce the risk level for one of these operatives. The REBA scores produced for this activity are presented in Table A8. Table A8: REBA scores for catching a cone from the carriageway Without RDL With RDL Operative REBA Score Risk Level REBA Score Risk Level 1 6 Medium 12 Very high 2 8 High 10 High 3 8 High 11 Very high The risk levels reduced from very high for Operatives 1 and 3 when picking up a cone with an RDL attached to medium for Operative 1 and high for Operative 3 when picking up a cone only. This was because the operatives had to use both hands when picking up the cone with the RDL attached to prevent the RDL from falling off and onto the carriageway, as stated during the interviews, and also because the position of the RDL makes it difficult to obtain a power grip on the cone. However, when picking up the cone only, the operatives used just one arm and Operative 1 used his free hand to lean on part of the vehicle which helped to stabilise his posture. In addition, the operatives bent their knees to a further extent when catching the cone with an RDL attached. Figure A8 shows the different postures that Operative 1 adopted when carrying out the activity. For similar reasons, REBA scores were lower for Operative 2 when picking up a cone only than when picking up a cone with an RDL attached, but his posture was not improved enough to change the risk level from high. A.2.7 Swinging the cone up to return it to the flatbed Swinging the cone up from the carriageway to return it to the flatbed was identified as a key posture for clearing away the equipment. Figure A9 shows an operative swinging a cone up from the carriageway with and without an RDL attached. 43 CPR1306

46 Figure A9: Swinging the cone up from the carriageway without an RDL (left) and with an RDL (right) Swinging the cone up from the carriageway to return it to the flatbed resulted in operatives adopting high and very high risk postures when using a cone with an RDL attached and medium and high risk postures when using a cone only. With an RDL attached, REBA scores were higher for every operative when swinging the cone up from the carriageway than when using a cone only. When an RDL is attached to a cone, operatives were only able to obtain a pinch grip. Interestingly, all of the operatives used their left hands to hold the top of the cone when performing the activity with a cone only but used their right hands to hold the top of the cone when an RDL was attached. The different technique allowed operatives to use both hands during the activity with the RDL attached to ensure that it did not fall onto the carriageway. When swinging up a cone only from the carriageway, however, all operatives used only one hand, demonstrating that the activity was made easier without an RDL attached to a cone. One operative even used his free hand to lean on part of the vehicle, creating a more stable posture. REBA scores achieved by each of the operatives are presented in Table A9 for this activity. Table A9: REBA scores for swinging the cone up to return it to the flatbed Without RDL With RDL Operative REBA Score Risk Level REBA Score Risk Level 1 8 High 11 Very high 2 6 Medium 9 High 3 8 High 10 High 44 CPR1306

47 A.2.8 Swinging the cone over the threshold Clearing away the equipment required that operatives swing the cone over the threshold. Figure A10 shows an operative demonstrating the activity of swinging the cone over the threshold. Figure A10: Swinging a cone over the threshold without an RDL (left) and with an RDL (right) With and without an RDL attached to a cone, swinging the cone over the threshold resulted in high risk postures for all operatives. In contrast to what was expected, the REBA scores decreased slightly for two of the operatives when they performed the activity using a cone with an RDL attached. However, a higher REBA score was achieved by one of the operatives when using a cone with an RDL attached. Due to the postures adopted by the operatives for this activity, it was difficult to determine the position of operatives left arms and wrists and so REBA scores may not exactly reflect the position that was held. REBA scores for each operative are presented in Table A10. Table A10: REBA scores for swinging the cone over the threshold Operative Without RDL With RDL REBA Score Risk Level REBA Score Risk Level 1 9 High 8 High 2 9 High 10 High 3 9 High 8 High A.2.9 Placing the cone on the flatbed The final activity that was identified as a key posture for clearing away the equipment was placing the cone on the flatbed. Figure A11 shows an operative placing the cones with and without an RDL attached on the flatbed. 45 CPR1306

48 Figure A11: Placing a cone on the flatbed without an RDL (left) and with an RDL (right) With an RDL attached, the risk level was medium for one operative but high for the other operatives when placing a cone on the flatbed. Without an RDL attached to a cone, all operatives achieved medium risk postures. The REBA scores were higher when placing a cone with an RDL attached on to the flatbed for all operatives. When compared with using a cone only, operatives were less able to obtain a good grip on the cone when an RDL was attached. Poor grip on the cone meant that operatives adopted poor wrist positions when demonstrating the activity with an RDL attached. Both of these factors resulted in higher REBA scores and risk levels when placing a cone on the flatbed with an RDL attached than when placing a cone only. Table A11 presents the REBA scores for each operative. Table A11: REBA scores for placing the cone on the flatbed Operative Without RDL With RDL REBA Score Risk Level REBA Score Risk Level 1 6 Medium 7 Medium 2 7 Medium 9 High 3 5 Medium 9 High 46 CPR1306