Investigation of injury data at a detonator facility

JCHAS-836; No of Pages 6

FEATURE

Investigation of injury data at a detonator facility Los Alamos National Laboratory (LANL) has been the design agency for nuclear weapon detonators for over 70 years. Detonator development, in support of the U.S. Department of Energy’s (DOE) nuclear weapons program, account for most activities performed by the Detonator Production Agency (DET). Management of work-related injuries includes in-field monitoring of injury/illness reports. Employing Lean Manufacturing and Six Sigma business practices (LSS), statistically significant variations (trends) have been identified in DET injury reports. An output metric has been developed that measures DET management progress toward meeting its operational safety objectives and goals. An I-chart format has been chosen to validate the variation of DET Occupational Safety and Health Administration (OSHA) recordable and subrecordable injury cases because this information is tracked as changes in the number of days between DET recordable and sub-recordable injury cases. Using a timeline, the primary injury types have been tracked. Using a Pareto Chart, the primary injury factors have been prioritized. This paper focuses on the collection of injury data; incorporation of this information into a visual format that DET management uses to make decisions to improving operations. Results from this study include of the following: chemical exposure cases have declined because the Hazard Assessment of each DET operation has been formally reviewed; Slip/ Trip/Fall factors have decreased due to Slip Simulator training; and work station evaluations have led to fewer injuries with Lift/Push/Pull factors. Rotation of employees, ergonomically friendly balances, automatic powder dispensers, and other equipment procurements will lower ergonomic injuries.

By Michael E. Cournoyer, Cindy M. Lawton, Marylou Apodaca, Robert A. Bustamante, Mark A. Armijo INTRODUCTION

Los Alamos National Laboratory (LANL) has been the design agency Michael E. Cournoyer is affiliated with the Los Alamos National Laboratory, Los Alamos, NM 87545, United States (Tel.: +505 665 7616; e-mail: [email protected]). Cindy M. Lawton is affiliated with the Los Alamos National Laboratory, Los Alamos, NM 87545, United States. Marylou Apodaca is affiliated with the Los Alamos National Laboratory, Los Alamos, NM 87545, United States. Robert A. Bustamante is affiliated with the Los Alamos National Laboratory, Los Alamos, NM 87545, United States. Mark A. Armijo is affiliated with the Los Alamos National Laboratory, Los Alamos, NM 87545, United States.

1871-5532 http://dx.doi.org/10.1016/j.jchas.2015.08.001

for nuclear weapon detonators for over 70 years.1 A detonator is a device used to initiate the high explosive (HE) charge surrounding an implosion-type nuclear weapon. Detonators for the stockpile were built at LANL before the Picatinny Arsenal and Mound production agencies took over the manufacturing mission. With the closure of the Mound facility in 1993, detonator production for the nuclear weapons stockpile returned to LANL. Detonator development, in support of the U.S. Department of Energy’s (DOE) nuclear weapons program, account for most activities performed by the Detonator Production Agency (DET). Management of work-related injuries includes in-field monitoring of injury/illness reports. As previously reported in this journal, statistically significant variations (trends) have been identified in a National Security Laboratory injury reports through the use of recognized statistical tools (Output Metrics, Control Charts, Timelines, and Pareto Charts).2 Investigation of National Security Laboratory injury data demonstrated that the management expectation that workers seek medical attention

before an injury or illness becomes serious has been effectively implemented.3 DET’s defense-in-depth is the site’s built-in capacity to detect or prevent errors without suffering undesirable consequences, i.e., DET’s ‘‘safety envelope.’’4,5 Redundant defenses improve safety margins, but also increase complexity. Flawed defenses and safety hazards become more difficult to detect. Redundant defenses make improvements more difficult to identify as well. Quality trending prevents defenses from degrading or being eliminated over time. Positive trends that reflect DET management commitment to operational safety may be overlooked. In this regard, quality trending continues through the use of Lean Manufacturing and Six Sigma business practices (LSS).6 In this study, trends have been identified in DET injury cases. An output metric is developed to measures DET management progress toward meeting its operational safety objectives and goals. Timelines are used to determine trends in injury types. Using Pareto Charts, the injury causal factors have been prioritized.

ß Division of Chemical Health and Safety of the American Chemical Society Elsevier Inc. All rights reserved.

Please cite this article in press as: Cournoyer, M.E., et al., Investigation of injury data at a detonator facility, J. Chem. Health Safety (2015), doi:http://dx.doi.org/10.1016/j.jchas.2015.08.001

1

JCHAS-836; No of Pages 6

METHODOLOGY

 Injury Output Metric Some injuries result in a recordable incident, as defined by the Occupational Safety and Health Act (OSHA), e.g., significant diagnosed injuries or those that require medical treatments such as wound closing that requires adding stitches.2,3 Sub-recordable (First Aid) injuries are those that do not meet these criteria, such as the use of bandages to cover a wound or removing foreign bodies from the eye using only irrigation or a cotton swab. The LANL Injury/Illness database is the primary repository for injury and illness information, including all demographic information about the incident, employee statement, medical record, investigation report, primary injury factors, body parts, and all OSHA classification information. This study consists of all injuries that have occurred to programmatic employees at DET between June of 2006 and May 2015. An output metric for DET injury/illness data has been developed using four sets of data: – Recordable outcomes (OSHA) are represented by light blue bars – Sub-recordable outcomes (First Aid) are represented by light yellow bars – 12 Quarter Rolling Average (QRA) – Linear Trend Line The 12 QRA is a method of calculating central tendency over time, an attempt to even out short-term oscillations and thus identify trends. The average is calculated over a 12 quarter period. For each quarter after this, the earliest value is dropped from the calculation and the most recent one is added, again to make an average over a 12 quarter period. The linear trend line (depicted in the metric as Linear) is a best-fit straight line. The trendline was calculated using a linear least-squares method. The linear trend line shows whether something is increasing or decreasing since the time that data had been first collected. As stated above, data has been collected since June 2006. Thus, the 12 QRA and linear trend line represent shortand long-term trends respectively. Only 2

the data from the last twelve quarters is displayed for the output metric. The trend line gives a good indication of past years performance in the output metric. An ideal output metric shows both recordable and sub-recordable data steadily decreasing both in the shortand long-term. The number of sub-recordable cases should be an order of magnitude higher than recordable ones.4 The ratio of sub-recordable to recordable cases can only be used in time periods where recordable cases have occurred, or else the ratio goes to infinity.  Control Chart An I (Individuals) chart format has been chosen to validate the variation of DET recordable and sub-recordable injury cases because this information is tracked as changes in the number of days between DET recordable and subrecordable injury cases. The individuals control chart is used for data that is collected and plotted one point at a time.7 1. First, the average of the individual values, i.e., the control limit (CL) is calculated: Pm xi x¯ ¼ i¼1 m where xi = individual values, m = total number of events 2. The difference between an individual value, xi, and its predecessor, xi1, is calculated as MRi ¼ jxi  xi1 j 3. Next, the average moving range (the arithmetic mean of these values) is calculated as ¯ ¼ MR

Pm

i¼2 MRi m1

4. The upper control limit (UCL) and lower control limit (LCL) for the individual values are calculated by adding or subtracting 2.66 times the average moving range to the process average: ¯ UCL ¼ x¯ þ 2:66MR ¯ LCL ¼ x¯  2:66MR

The value of 2.66 is obtained by dividing 3 by the sample-size-specific d2 anti-biasing constant for n = 2, 1.128.8 d2 is a function only of the sample size n. 5. The upper sigma limits (1UCL, 2UCL) and lower sigma limits (1LCL, 2LCL) for the individual values are calculated by adding or subtracting one-third or two-thirds of 2.66 times the average moving range to the process average:   ¯ 2:66MR 1UCL ¼ x¯ þ 3   ¯ 2:66MR 2UCL ¼ x¯ þ 2 3   ¯ 2:66MR 1LCL ¼ x¯  3   ¯ 2:66MR 2LCL ¼ x¯  2 3 In general, 25 or more data points create a statistical Baseline; the more data points, the sounder the statistical analysis. A trend is defined by the relationship of data points plotted on a control chart and are detected through preset rules. If a trend is detected, the special cause of the trend is determined. Trends in the existing data are determined by the following criteria6: — One point outside the UCL or LCL (Definitive) — Two out of three points; two standard deviations above/below average (Sigma Zone) — Four out of five points; one standard deviation above/below average (Sigma Zone) — Seven points in a row; all above/ below the CL (Pattern) — Seven points in a row; all increasing/ decreasing (Pattern) — Ten out of eleven points in a row; all above/below the CL (Pattern) Trends serve as a notice a special cause variation likely exists and adjustments to DET operations may be necessary. When a trend is identified, one must determine if the trend is Definitive, Sigma Zone, or Pattern. A Definitive trend result occurs when one or more points fall outside the UCL or LCL. This is the case when outliers are

Journal of Chemical Health & Safety, September/October 2015

Please cite this article in press as: Cournoyer, M.E., et al., Investigation of injury data at a detonator facility, J. Chem. Health Safety (2015), doi:http://dx.doi.org/10.1016/j.jchas.2015.08.001

JCHAS-836; No of Pages 6

 Injury Types Similar injuries types were binned together. For example, abrasions (scratches), bites/stings were included in lacerations (Cut). As discussed above, the main types of injuries reported for DET are Bruise, Chemical Exposure, Cut, Strain/Sprain, and Other. Detailed descriptions of Bruise, Chemical Exposure, Cut, Strain/Sprain, and Other types have been previously discussed in this journal and will be iterated.2,3 Chemical exposures involve the risk presented by a toxic chemical if there is evidence that acute or chronic health effects occur in workers coming in contact with that chemical. Abrasions are skin wounds that rub or tear off skin. Bruises develop when small blood vessels under the skin tear or rupture, most often from a bump or fall. Cuts are open wounds through the skin. A cut is a forceful injury to the skin. A strain occurs when the muscle tendon unit is inflamed. Injury to the ligament is called a sprain. Other injury types were primarily Possible ExposureNot Diagnosed and No Injury Detected. A timeline was plotted of the injury types over the June 2006 and May 2015 time period. The higher up the chart, the more often the DET injury type cases occurred. No distinction between recordable and sub-recordable injury cases was made.  Injury Factors Causal factors associated with DET job-incurred injuries consist of Ergonomic, Chemical Agents, Slips/Trips/Falls, Struck against/by, Lift/Push/Pull, and

Other. Sprains that develop after many repetitions of a motion and strains that develop due to awkward posture involve Ergonomic factors. Chemical Agents involve the risk presented by a toxic chemical if there is evidence that acute or chronic health effects occur in workers coming in contact with that chemical. A slip occurs when there is too little friction or traction between feet (footwear) and walking/ working surface, resulting in loss of balance. A trip occurs when the foot or lower leg contacts an object, resulting in loss of balance or when one steps down to lower surface and loses balance. A fall occurs when one is too far off center of balance. A fall can occur at the same level, i.e., when one falls to same walking or working surface, or fall into or against objects above same surface. A fall can also occur at a lower level, i.e., when one falls to a level below walking or working surface. Sprains or strains that occur from a single stressful incident are categorized as lift/push/pull factors. Struck against/by injuries are produced by forcible contact or impact between the injured person and an object or piece of equipment. Using a Pareto Chart, the primary causal factors were determined. The

left vertical axis is the frequency of injury factors for DET. The right vertical axis is the cumulative frequency of injury factors for DET. Once again, no distinction between recordable and sub-recordable cases was made.

RESULTS

Summary of DET Injuries From June 2006 to May 2015, there were 69 injuries cases reported for DET. Of those reported, 21 of them were recordable cases and 48 were sub-recordable cases. The recordable and sub-recordable cases in the last twelve quarters are shown in Figure 1. The number of recordable cases peaked in second quarter of 2013 with 2. The number of sub-recordable cases peaked in last quarter of 2012 with 6. The average numbers of cases per quarter during last twelve quarters were 0.5 and 1.0 recordable and subrecordable cases, respectively. The ratio of sub-recordable to recordable cases is 2.0. Short- and long-term trends for recordable cases are flat. Short- and long-term trends for subrecordable cases are declining. A Baseline of 22 points was established between June 2006 and May 2015 for DET recordable cases. The average number of days between DET recordable injury cases is 149

DET Injury/Illness Metric Jun 06 to May 15

7 6 5

Cases

present, and is the first test for an outof-control process. A Sigma Zone trend occurs when a certain number of data points can be located in specific control chart zones. The zone category of indicators focuses on the location of the data points rather than on the existence of a pattern. A certain number of points falling within one of the zones or beyond serve as a notice that special cause variation exist and that process adjustments maybe necessary. Pattern trend occurs when a consecutive string of data show a pattern. While all trends must be analyzed, patterns may warrant a recalculation of the Baseline and UCL and LCL.

4 3 2 1 0

Jul-12 Oct-12 Jan-13 Apr-13 Jul-13 Oct-13 Jan-14 Apr-14 Jul-14 Oct-14 Jan-15 Apr-15

OSHA 12 QRA (OSHA) Linear (OSHA)

First Aid 12 QRA (First Aid) Linear (First Aid)

Figure 1. DET injury/illness metric.

Journal of Chemical Health & Safety, September/October 2015

Please cite this article in press as: Cournoyer, M.E., et al., Investigation of injury data at a detonator facility, J. Chem. Health Safety (2015), doi:http://dx.doi.org/10.1016/j.jchas.2015.08.001

3

JCHAS-836; No of Pages 6

Days between Cases

700

DET Recordable Injury Cases Jun 06 to May 15

600 500 OSHA

400

CL

300

UCL

200 100 0 Jun-06 Jun-07 Jun-08 Jun-09 Jun-10 Jun-11 Jun-12 Jun-13 Jun-14

Figure 2. DET recordable injuries I-chart.

350

DET Sub-Recordable Injury Cases Jun 06 to May 15

Days between Cases

300 250

First Aid

200

CL

150

UCL

100 50 0 Jun-06 Jun-07 Jun-08 Jun-09 Jun-10 Jun-11 Jun-12 Jun-13 Jun-14

Figure 3. DET Sub-recordable injuries I-chart.

days with an UCL of 512 days, as shown in Figure 2. One Definitive trend occur during this period, since this data point is part of the Baseline, no analysis is made. Twenty-six days have passed, since the last recordable case. A Baseline of 25 points was established between June 2006 and February 2011 for DET sub-recordable injury cases. The average number of days between DET sub-recordable injury cases is 67 days with an UCL of 244 days, as shown in Figure 3. One Definitive trend occur during this period, since this data point is part of the Baseline, no analysis is made. A Pattern trend with the sub-recordable injury cases occurs between August and November 2012; seven points in 4

a row all below average (three cases occurred on the same day). After correcting the Pattern trend, the average number of days between DET sub-recordable injury cases is 15 days with an UCL of 192 days. This trend, due to workers seeking medical attention before injuries became serious, is not considered adverse. A Pattern trend with the sub-recordable injury cases occurs between December 2012 and 2014; 7 points in a row all above average. After correcting the Pattern trend, an average number of days between DET sub-recordable injury cases is 111 days with an UCL of 289 days. This trend, due to the introduction of an ergonomics program, is not considered adverse. 159 days have passed, since the last sub-recordable case. The variations of injury types between June 2006 and May 2015 for DET are shown in Figure 4. For Strains/Sprain type cases, no significant trends were seen. Beginning in November/December 2012 timeframe, a decline in Chemical Exposure and Other type cases was observed. Between February 2008 and May 2013, there was a significant drop in Cut type cases. Between February 2007 and November 2013, there was a significant drop in Bruise type cases. The number of injury causal factors documented between June 2006 and May 2015 are shown in Figure 5. In this study, an injury can have only one causal factor. The three main caus-

DET In njury Types Jun 06 0 to Ma ay 15 Strain/ Spra ain Chemical osure Expo er Othe

Cut

Bruisse Jun n-06 Jun-07 7 Jun-08 Ju un-09 Jun-1 10 Jun-11 Jun-12 J Jun--13 Jun-14

Figure 4. DET injuries type timeline.

Journal of Chemical Health & Safety, September/October 2015

Please cite this article in press as: Cournoyer, M.E., et al., Investigation of injury data at a detonator facility, J. Chem. Health Safety (2015), doi:http://dx.doi.org/10.1016/j.jchas.2015.08.001

JCHAS-836; No of Pages 6

Injury Factors Jun 06 - May 15

Factors

20

100%

15

75%

10

50%

5

25%

0

Chemical Agent

Ergonomic

Slip/Trip/Fall

Struck against/by

Frequency 06 - 14 Cum. Frequency 06 - 14

Other

Lift/Push/Pull

0%

Frequency Last 12 Months Cum. Frequency Last 12 Months

Figure 5. DET injury factor Pareto chart. Table 1. Comparison of DET Causal Factors.

Factors

6/06–6/14a

6/14–5/15

Chemical Agent Ergonomic Slip/Trip/Fall Struck against/by Other Lift/Push/Pull

2.1 1.9 1.3 1.0 1.0 0.9

0 2 1 1 0 0

The bold value represents a factor that has increased during the last twelve months. Ergonomic factors have increased. Chemical Agent, Other, and Lift/Push/Pull factors have decreased or been eliminated. a Annualized.

al factors for DET were Chemical Agents, Ergonomic, and Slip/Trip/ Fall; they account for over 60% of the injuries between June 2006 and 2014. In the last twelve months, Chemical Agent, Other, and Lift/Push/Pull factors have been eliminated. Ergonomic factors replaced Chemical Agents factors as a primary causal factor for DET. The number of injury causal factors documented between June 2006 and May 2014 can be directly compared to the number of injury causal factors documented between June 2014 and May 2015 by dividing the values of the former by 8 (96 months), see Table 1.

DISCUSSION

As the work environment becomes more hazardous or more hazards are recognized, more controls are put in place to protect the worker. Injury cases result when these controls have

failed. Concerning injury types and factors no distinction between recordable and sub-recordable cases was made because it is DET management expectation that all injuries are preventable. The probability of injuries occurring is independent of the consequences. The differences between recordable and sub-recordable cases are the extent of the consequences. DET management expectations for recordable cases are neutral in both the short- and long-term, as shown in Figure 1. This was due in part because DET management campaigned to identify, report, and treat all ergonomic injuries. DET management expectations for sub-recordable cases are being met in both the short- and longterm because more sub-recordable cases are being reported than recordable cases. Excellence will be achieved when the ratio of sub-recordable to recordable cases approaches ten.3 Since July 2014, resources are being focused on reducing Ergonomic factors.

DET workers have been encouraged to self-report ergonomic injuries. After an ergonomic program is implemented, recordable injuries increase in the short-term and decrease in the longterm. Sub-recordable injuries decrease in the short-term and increase in the long-term. This pattern was seen with glovebox glove injuries, as reported in this journal.3 Eventually, recordable injuries disappear and sub-recordable injuries are infrequent. Thus, the ratio of sub-recordable to recordable cases becomes favorable. As 22 data points were used to create a statistical baseline, the results of the DET Recordable Injuries I-Chart analysis are not significant (see Figure 2). DET Recordable Injury Cases Metric performance is stable. Until a CL of 365 average number of days between DET recordable injury cases is observed, performance need improvement. As 25 data points were used to create a statistical baseline, the results of the DET Sub-Recordable Injuries I-Chart analysis are significant (see Figure 3). The Pattern trend with the sub-recordable injury cases between August and November 2012 required no DET management action to reverse the trend. The Pattern trend with the sub-recordable injury cases between December 2012 and 2014 required no DET management action to reverse the trend. Nevertheless, DET Management will iterate to First Line Mangers the following expectation: Workers seek medical attention before injuries become serious. Favorable declines in Chemical Exposure and Other Type cases had occurred in DET, as shown in Figure 4. Chemical Exposure cases have declined because the Hazard Assessment of each DET operation has been formerly reviewed. Other Type cases have declined because the medical doctors entering information into the LANL Injury/Illness database are more experienced in describing injuries and illnesses types. Plans need to be developed that lower number of Strain/Sprain, Cut, and Bruise type cases. With the use of the Pareto Chart, critical injuries causal factors were prioritized in relation to the number of

Journal of Chemical Health & Safety, September/October 2015

Please cite this article in press as: Cournoyer, M.E., et al., Investigation of injury data at a detonator facility, J. Chem. Health Safety (2015), doi:http://dx.doi.org/10.1016/j.jchas.2015.08.001

5

JCHAS-836; No of Pages 6

cases, as shown in Figure 5. Management efforts to reduce Chemical Agent, Slip/Trip/Fall, Other, and Lift/Push/Pull factors have been successful. There appeared to be an unfavorable increase in Ergonomic Injury types between April 2014 and May 2015, as shown in Table 1. This was due in part because DET management campaigned to identify, report, and treat all ergonomic injuries. Resources are being focused on reducing Ergonomic factors. For example, employee tasks are being rotated. Ergonomically friendly balances, automatic powder dispensers, other equipment procurements have been procured. As chemical agents are the leading factors behind Chemical Exposure Type cases, Chemical Agent factors have been reduced due to the Hazard Assessment of each DET operation. Slip/Trip/Fall factors have decreased due to Slip Simulator training. The kinetic learning experience with the Slip simulator training has taught workers how to walk safely in adverse conditions.3 ‘‘Other’’ factors have declined for the same reason ‘‘Other’’ type cases have declined; the medical doctors entering information into

6

the LANL Injury/Illness database are more experienced in describing injuries and illnesses factors. Work station evaluations have led to fewer injuries with Lift/Push/Pull factors.

National Nuclear Security Administration, and LANL’s Plutonium Science & Manufacturing Directorate for support of this work.

REFERENCES CONCLUSIONS

The LANL Injury/Illness database documents and records injuries accurately and objectively. A better understanding of injury data is enhanced by LSS tools. Collecting injury data gives DET management information they need to concentrate on vulnerabilities that require their support. As DET Recordable and Sub-Recordable Injuries I-Charts deviates from the optimum, programmatic operations in DET can be improved to bring them back into control. Data generated from analysis of injury data help identify and reduce number of corresponding causal factors. This increases technical knowledge and augments operational safety. ACKNOWLEDGEMENTS The authors would like to acknowledge the U.S. Department of Energy,

1. Johnson, M. LA-UR-04-6924; Los Alamos National Laboratory: Los Alamos, NM, 2004, October. 2. Cournoyer, M. E.; et al. Investigation of injury data at a nuclear facility. J. Chem. Health Saf. 2012, 19(5), 17–25. 3. Cournoyer, M. E.; et al. Investigation of injury data at a nuclear facility: part II. J. Chem. Health Saf. 2014, 21(6), 1–7. 4. Cournoyer, M. E.; et al. Safety observation contributions to a glovebox safety program. J. Chem. Health Saf. 2011, 18(5), 43–55. 5. Cournoyer, M. E.; et al. Investigation of radiation protection observation data at a nuclear facility. J. Chem. Health Saf. 2012, 19(2), 20–24. 6. Cournoyer, M. E.; et al. Application of lean six sigma business practices to an air purifying respirator program. J. Chem. Health Saf. 2013, 20(2), 34–39. 7. Montgomery, D. Introduction to Statistical Quality Control; John Wiley & Sons, Inc.: Hoboken, NJ, 2005, p. 294. 8. https://www.qimacros.com/control-chart/ control-chart-constants/ [accessed 15.07.15]

Journal of Chemical Health & Safety, September/October 2015

Please cite this article in press as: Cournoyer, M.E., et al., Investigation of injury data at a detonator facility, J. Chem. Health Safety (2015), doi:http://dx.doi.org/10.1016/j.jchas.2015.08.001