Minor injuries, cognitive failures and accidents at work: incidence and associated features

Transcription

1 Occupational Medicine 2005;55: doi: /occmed/kqi035 Minor injuries, cognitive failures and accidents at work: incidence and associated features Sharon A. Simpson, Emma J. K. Wadsworth, Susanna C. Moss and Andrew P. Smith Background There is a dearth of information examining minor injuries and cognitive failures in the workplace. Aims This study aims to describe the incidence of, and characteristics associated with, minor injuries, cognitive failures and accidents in the workplace. Methods The study was a community based postal questionnaire survey of people selected at random from the Electoral Registers of Cardiff and Merthyr Tydfil; 7980 questionnaires were returned. Results Frequent/very frequent minor injuries were reported by 9.8% of the workers and frequent/very frequent cognitive failures (e.g. problems of memory, attention or action) by 10.5%. Work accidents requiring treatment were reported by 5.6% of the workers. Minor injuries, cognitive failures and accidents were associated with a number of demographic and health factors. In addition, minor injuries, cognitive failures and accidents were found to be related to each other. Conclusions A large number of workers experience accidents and minor injuries at work. Minor injuries are not recorded by official sources but could have implications in productivity and worker health. Factors identified in this study, as associated with injuries or accidents, could be addressed by specific workplace policies or by raising individual awareness of the risks. Key words Accidents; associations; injuries; work. Introduction Accidents and injuries in the workplace are a cause for concern for employees and employers alike. The Labour Force Survey 1999/2000 [1] reported that 1.03 million workers suffered work related injuries in the preceding 12 months. In 2000 the UK government announced that by the year 2010 they aim to reduce the incidence of people suffering from work related ill-health by 20% [2]. In the UK the main sources of information relating to injuries in the workplace are injury reports made to RIDDOR (Reporting of Injury, Diseases and Dangerous Occurrences Regulations) and from questions included in the Labour Force Survey. However, there is a dearth of research examining the epidemiology of minor injuries [3] or cognitive failures at work. Although the impact of these in the workplace is less obvious than accidents, they still have implications on productivity and worker health. Indeed, Di Lorenzo [3] identified that minor injuries represent the greater part of occupational injuries. Centre for Occupational and Health Psychology, Cardiff University, 63 Park Place, Cardiff, UK. Correspondence to: Sharon A. Simpson, Department of General Practice, University Wales College of Medicine, Llanedeyrn Health Centre, Maelfa, Llanedeyrn, Cardiff, CF23 9PN, UK. Tel: ; fax: ; simpsonsa@cardiff.ac.uk Cognitive failures can be defined as cognitive-based errors on simple tasks that a person should normally be able to complete without fault; these mistakes include problems with memory, attention or action and will be considered in this study for two reasons. First, they are in effect human errors and human error is associated with many accidents. Second, this sort of slip occurs more frequently than accidents; cognitive failures may occur in many contexts, but only in some situations will they lead to accidents. There is some evidence to suggest that there is a link between cognitive failures and accidents [4,5]. Other research also suggests a relationship between accidents and distractibility, poor selective attention and mental error [6,7], in effect cognitive failures. In the literature, associations have been noted between cognitive failures and depression [8], stress and anxiety [9]. The only other relevant published study examining minor injuries and cognitive failures is a study completed in Bristol [10]. This study sent a questionnaire, which included questions on health, injuries and occupational characteristics, to people. Minor injuries at work were associated with smoking and total negative job characteristics, as well as sleeping problems and low income. Cognitive failures at work were associated with sleeping problems, anxiety, work stress and symptoms in Occupational Medicine, Vol. 55 No. 2 q Society of Occupational Medicine 2005; all rights reserved 99

2 100 OCCUPATIONAL MEDICINE the last 14 days. Accidents at work were associated with smoking, male sex and total negative job characteristics. Due to the lack of literature on factors associated with cognitive failures and minor injuries, and since they may be related to accidents, the following sections will review the available literature on accidents. In the present study, the definition of accident is an incident where a person is injured and requires medical treatment from someone else. Other studies described here have used varied definitions of accidents, but they usually involve an injury. Research has identified that one of the most crucial factors related to accidents in the workplace is the job sector or the jobs that people do [11 15]. The riskiest jobs include metal making, construction, machine operating, quarrying and agriculture. Some studies have found consistently higher rates of accidents among the self-employed [16]. However, Driscoll et al. [17] using data adjusted for occupation and industry found no evidence for an increased risk in this group. Other factors associated with increased risk of accidents include stress [18,19], male gender [1,11,20], risk taking [21,22], neuroticism [23], alcohol [3,24 27], smoking [3] and anxiety [28]. Also, fatigue or sleep problems have been implicated in workplace accidents [29 32]. Time on the job seems to be a significant factor in research on accidents in the workplace [33 35]. The Labour Force Survey found that workers in their first few months with a new employer have the highest rate of injury [1]. Age has also been associated with accidents at work, independent of the job type [1,11,12]. Chi and Wu [12] found age as a significant factor in accidents at work, although this varied with the industry sector, as those over 55 had the highest rates in construction and transport, but the youngest had very high rates in mining. Studies describing accident rates and associated factors tend to concentrate on fatal accidents [12,29, 36], or specific industries [13,22,37], accident types [38], or age groups [39]. In addition, studies have often failed to take into account the range of possible confounding factors on work accidents [20,36,37]. Finally, although there are a number of papers examining serious or fatal injuries among workers [12,29,40] there has been little research examining the epidemiology of minor injuries [3] or cognitive failures in the workplace. This paper is one of a series examining minor injuries, cognitive failures and accidents. This study by nature is hypothesis generating rather than hypothesis testing as it examines a number of different variables, in combination, in a general working population that may have been implicated in other studies in isolation, or identified as significant within specific industries. Attempts have been made to remedy some of the weaknesses of previous studies including the Bristol Stress and Health at Work study, by covering a larger geographic area and including more detailed questions about accidents and possible associated factors. Method Thirty thousand individuals were selected from the electoral registers of Cardiff and Merthyr Tydfil ( and 7500 residents, respectively), using a Microsoft Access script written to randomly select individuals. Cardiff and Merthyr Tydfil were chosen for their differing social and economic conditions. Cardiff is a modern, post-industrial city with high employment, whereas Merthyr Tydfil suffers from high levels of unemployment, poverty and associated social problems. Cardiff University, School of Psychology, Ethics Committee approved the study and the application was submitted to the Local Research Ethics Committee administrator, however formal approval was not considered necessary. Questionnaires and covering letters were posted to individuals. They were returned anonymously, there were no identifiers attached and therefore no reminder or follow-ups were completed. The questionnaire was based on that used in the Bristol Stress and Health at Work project [41]. The three main outcome variables were Minor injuries at work (this includes minor injuries as a result of car accidents): respondents were asked; In the last 12 months have you had minor injuries (e.g. cuts and bruises) at work that did not require medical attention from someone else? This was measured on a single item 5-point scale (not at all, rarely, occasionally, quite frequently and very frequently). Those reporting quite or very frequent minor injuries were compared with those reporting none, rare or occasional minor injuries. Cognitive failures at work: respondents were asked; How frequently do you find that you have problems of memory (e.g. forgetting where you put things), attention (failures of concentration), or action (doing the wrong thing) at work? No time interval was specified for this measure as it was thought that people would have difficulty remembering slips of memory over long time periods. This was measured, as with minor injuries, on a single item 5-point scale. This single question was used, instead of the more extensive Cognitive Failures Questionnaire [42], to limit the length of the questionnaire. However, this single item was correlated at 0.7 with Broadbent et al. s original questionnaire [43]. Those reporting quite or very frequent cognitive failures were compared with those reporting none, rare or occasional cognitive failures. Accidents at work: respondents were asked; Thinking about the last 12 months have you had any accidents

3 S. A. SIMPSON ET AL.: MINOR INJURIES, COGNITIVE FAILURES AND ACCIDENTS AT WORK 101 while you were working that required medical attention from someone else (e.g. first aider, GP, nurse or hospital doctor)? This includes traffic accidents where the respondent was the driver and the accident happened while travelling to or from work, or while travelling as part of work. This was measured on an 8-point scale from none to 7+. Those reporting one or more accidents were compared with those reporting none. The independent variables included in the logistic regression analyses are listed in Table 1. These were suggested by previous studies and analyses of the Bristol Stress and Health study data [41]. The combined incident variable was included to examine whether there was an independent association between accidents, minor injuries and cognitive failures. The total negative score variable is a sensitive overall measure of occupational characteristics [44,45]. Single item measures of stress, risk taking and health were used instead of the longer measures that have already been developed. There were several reasons for this; first, this allows comparison with previous work and other phases of the study. Second, these items provide comparable measures as established elsewhere [41]. Third, the length of the questionnaire would have been prohibitive if the full-length measures had been included. There are overlaps in some of the variables included, for example, anxiety, neuroticism and depression. All three of these were entered in the logistic regression, as the authors wanted to examine the possible impact of these separately. Studies have shown that the Hospital Anxiety and Depression Scale differentiates between anxiety and depression effectively [46]. Also the Eysenck personality questionnaire, which measures neuroticism, is one of the best validated and most commonly used instruments in the study of personality [47]. Since neuroticism measures negative affect the overlap between depression and neuroticism is not surprising. However, there is evidence to suggest that neuroticism is likely to have trait-like properties [48], as well as state-like properties that have been suggested by Farmer, et al. [49]. Also, tests of multicolinearity were carried out; these showed no evidence of high correlations between independent variables; thus the predictors were independent from each other. The Cardiff study was designed to examine factors associated with the occurrence of accidents, injuries and cognitive failures; not to test any specific hypotheses. Thus, factors identified in previous studies were included in the logistic regression models as independent variables, to try and ascertain the most significant characteristics associated with the three outcomes (see Table 1). In order to reduce injury rates in the workplace, these variables need to be identified and their relationships explored. Data were analysed using Pearson chi-square, Mann Whitney and logistic regression modelling techniques. Results Seven thousand nine hundred and eighty questionnaires were returned completed, giving an overall response rate of 27%. Fifty-eight per cent of respondents were female, 97% were white, 57% were working (75% were full-time and 25% were part-time). The mean age was years and the range was years. There were some respondents aged under 18 because individuals can register to vote from the age of 16. Assessments of the representativeness of the study data with the Census data were made separately for Cardiff and Merthyr Tydfil. Study data were also compared with the UK as a whole. These demographic comparisons show that the study sample and the general population were broadly similar. The accident data reported here are based on the working respondents only. These were identified by including a simple question asking if they had a job at the moment. In the last year (among 58% of the working respondents) frequent or very frequent work related minor injuries were reported by 9.8%, quite or very frequent cognitive failures by 10.5% and one or more work accidents by 5.6% of the workers. Rates standardized for age and gender were very similar, but more females reported cognitive failures than males (P, 0.002). This pattern was the same for minor injuries, but the difference was not significant. Work accident rates were lower for females than males (P, 0.001). Of those who reported cognitive failures, 18% also reported minor injuries and 7% also reported accidents at work. Of those reporting minor injuries 12% also reported an accident. Three different combined incident variables were developed to include as independent variables in the analyses. Age was examined to see whether there were any differences in accident rates across age bands. There did not appear to be an age effect for work accidents or cognitive failures. However, more respondents in the and age groups reported minor injuries (18% and 10%, respectively) than in the and age groups (8% and 3%, respectively). More than double the number of accidents and minor injuries were reported among those who had a manual job, however similar numbers in the manual/non-manual groups reported cognitive failures at work. The differences evident between manual and non-manual jobs were significant using Pearson chi-square (P, 0.001). The number of respondents in different occupational classifications who had minor injuries, cognitive failures or accidents was examined. Those with the highest rates of

4 102 OCCUPATIONAL MEDICINE Table 1. Description of variables included in the analyses Health Anxiety and depression Neuroticism Sleeping problems Job type General health over the past 12 months was measured on a 5-point scale taken from the Bristol Stress and Health study [41]. Respondents were asked over the last 12 months how would you say your general health has been? Very good, good and fair were compared with bad and very bad These were measured using the hospital anxiety and depression scale [61]. The case cut off point was a score greater than 11 Neuroticism was measured using the Eysenck personality inventory [62]. Scores were compared across quartiles This question, asking about sleep problems in the last 14 days, was taken from the Whitehall II study [63] This was taken from the Bristol Stress and Health study [41]. Manual and non-manual jobs were compared Employment status Questions on employment status were taken from the Bristol Stress and Health study [41]. Self-employed and employee (including managers) were compared Alcohol consumption The questions about alcohol consumption were taken from the Bristol Stress and Health study [41]. Weekly consumption, in units, at or above the 90th percentile (calculated separately for men and women) was compared with lower consumption Smoking Risks at work and outside work The questions about smoking were taken from the Bristol Stress and Health study [41]. Current cigarette smokers were compared with those who did not currently smoke cigarettes Risk taking behaviour was measured on a 5-point scale. The respondents were asked How frequently do you take risks at work? Not at all, rarely, and occasionally were compared with quite frequently and very frequently Work related stress Stress was measured on a 5-point scale taken from the Bristol Stress and Health study [41]. Respondents were asked, In general, how stressful do you find your job? Not at all, mildly, and moderately were compared with very and extremely General stress Income Measured as for work stress The 10-point income scale was from the Bristol Stress and Health study [41]. Those earning less than per annum were compared with those earning , per annum, , per annum, and per annum or more Gender Males and females were compared Age Age was split into four groups: 16 24, 25 39, and 60+ and were compared Total negative job characteristics Questions about working hours and patterns, physical hazards and job characteristics came from the Bristol Stress and Health study [41]. These questions also included the job content instrument [64]. Total negative score is derived by adding responses across all these scores [44,45]. Scores were compared across quartiles Chronic symptoms This question, asking about health problems ever, was taken from the Whitehall II study [63], Symptoms in the last year Symptoms in the last 14 days Experience in the job Combined incident variable the depression item was excluded because this is measured elsewhere This question, asking about health problems in the last year, was taken from the Whitehall II study [63]. The depression, backache and other items were excluded a This question, asking about health problems in the last 14 days, was taken from the Whitehall II study [63]. The dizziness, depression, sleep problems, backache and other were excluded a This was derived from questions from the Bristol Stress and Health study [41]. Those respondents who had worked in their jobs less than 6 months were compared with those who had been in the post for 6 months or more This was derived using the questions on accidents, minor injuries and cognitive failures that were taken from the Bristol Stress and Health study [41]. a These items were excluded because they could be the result of an accident at work. minor injuries were in craft and related occupations (15%) and personal and protective occupations (14%). For cognitive failures, professional occupations had the highest rates (20%) followed by managers and administrators (19%). However, the pattern for accidents was different, with rates reported being fairly uniform across the categories of occupation, the highest being associated with professional and technical (14%) and craft and related (12%) occupations. More detailed analyses of accidents in different occupations or industry sectors were not completed, as the numbers were too small. Work related accidents most commonly occurred when lifting or carrying (21%), followed by slips, trips and falls at the same level (18%). Univariable analyses were carried out to test for associations between minor injuries, cognitive failures and accidents and the possible factors that are related to these outcomes. Patterns of association (Table 2) were similar across minor injuries, cognitive failures and accidents with the following factors being associated with increased risk for all three outcomes: anxiety and depression, difficulty in sleeping, higher levels of risk taking, high general stress, combined incident reporting, neuroticism and total

5 S. A. SIMPSON ET AL.: MINOR INJURIES, COGNITIVE FAILURES AND ACCIDENTS AT WORK 103 Table 2. Summary table for univariable results (chi-square) Accidents at work (P levels) Minor injuries at work (P levels) Cognitive failures at work (P levels) Clinical anxiety 0.009, 0.001,0.001 Clinical depression, ,0.001 Difficulty sleeping,0.001, 0.001,0.001 Risk taking at work,0.001, Risk taking outside work,0.001, 0.001,0.001 Work stress 0.20, 0.001,0.001 General stress ,0.001 Experience Health in last 12 months, ,0.001 Gender, Income 0.12, Job type,0.001, Self-employed Cigarette smoking Total alcohol units 0.03, Neuroticism score 0.005, 0.001,0.001 Age 0.05, Total negative score,0.001, 0.001,0.001 Chronic symptoms,0.001, 0.001,0.001 Symptoms in last 14 days 0.001, 0.001,0.001 Symptoms in last year 0.66, 0.001,0.001 Combined incident,0.001, 0.001,0.001 negative job characteristics. Chronic symptoms and symptoms in the last 14 days were also associated with accidents, injuries and cognitive failures. Those with more symptoms tended to report accidents and injuries more often. Backward stepwise binary logistic regression models were used to assess the association of potential risk factors with work related minor injuries, cognitive failures and accidents. The covariates included in the models are listed in Table 1. Hosmer Lemeshow goodness of fit statistics showed that each of the models fitted the data well. Work related minor injuries were associated with risk taking outside work, total negative score, manual jobs, young age, lower income, more symptoms, higher levels of work stress, combined incidents (cognitive failures and accidents), sleep problems and being self-employed (Table 3). The pattern of association for cognitive failures at work was comparable with minor injuries, however there were also associations with female gender, combined incidents (minor injuries and accidents), higher levels of stress, higher levels of neuroticism, depression, anxiety, not smoking and less experience in the job. Accidents were independently associated with greater risk taking at work, male gender, total negative job characteristics score, greater intake of alcohol (approaches significance), manual jobs, symptoms, poor general health in the last year and combined incidents (cognitive failures and minor injuries). Discussion This study identified the rates of accidents, injuries and cognitive failures. These rates were similar to those reported in other studies [10,20]. Factors associated with minor injuries, cognitive failures and accidents were also identified and will be discussed below in relation to other published work. Workers in craft related occupations, or associate professional and technical categories had the highest rate of accidents. The pattern of work related accidents was similar to that reported by the Health and Safety Executive [40], where accidents occuring while lifting or carrying were the most common, followed by slips, trips or falls at the same level. In the present study, accidents appear to be related to minor injuries and minor injuries to accidents and cognitive failures (P, 0.01), which has also been reported in previous studies [4,5,10]. This lends some support to the notion of a human error pyramid with accidents at the top, followed by more common minor injuries and the most numerous cognitive failures at the bottom. Thus, it may be dependent on the context whether minor injury or cognitive failure results in an accident. Previous studies [12] identified job characteristics or job sector as an important factor in relation to injuries and accidents. Using a composite score, similar to

6 104 OCCUPATIONAL MEDICINE Table 3. Summary table of logistic regression results Covariate Odds ratio 95% confidence limits P levels Minor injuries at work in the last year Hosmer Lemeshow goodness of fit P, 0.30, model chi-square P, 0.001, n ¼ 3214 Risk taking outside work None/rarely P, 0.02 Frequently 1.64 Age P, Income, P, Work stress Not at all/mild/moderate 1.00 P, Very/extremely Total negative score 1st quartile 1.00 P, nd quartile rd quartile th quartile Job type Non-manual 1.00 P, Manual Sleep problems No 1.00 P, Yes Self-employed No 1.00 P, 0.02 Yes Chronic symptoms No symptoms 1.00 P, or more symptoms Symptoms in the last 14 days 0, 1 or 2 more symptoms 1.00 P, symptoms Combined incidents No accidents or cogn. failures 1.00 P, Accidents only Cognitive failures only Cognitive failures and accidents Cognitive failures at work Hosmer Lemeshow goodness of fit P, 0.64 Model chi-square P, n ¼ 3214 Risk taking outside work None/rarely 1.00 P, Frequently Gender Male 1.00 P, 0.02 Female Age P, Income, P, Work stress Not at all/mild/moderate 1.00 P, 0.02 Very/extremely General stress Not at all/mild/moderate 1.00 P, 0.04 Very/extremely Neuroticism 1st quartile 1.00 P, nd quartile rd quartile th quartile Clinical anxiety No 1.00 P, 0.01 Yes Clinical depression No 1.00 P, Yes Smoking No 1.00 P, 0.04 Yes

7 S. A. SIMPSON ET AL.: MINOR INJURIES, COGNITIVE FAILURES AND ACCIDENTS AT WORK 105 Table 3. (continued) Covariate Odds ratio 95% confidence limits P levels Experience in job In job $6 months 1.00 P, 0.03 In job, 6 months Symptoms in the last 14 days 0 2 symptoms 1.00 P, symptoms Combined incidents No accidents or minor injuries 1.00 P, Accidents only Minor injuries only Minor injuries & accidents Work Accidents in the last 12 months Hosmer Lemeshow goodness of fit P, 0.62 Model chi-square P, n ¼ 3214 Risk taking at work None/rarely 1.00 P, 0.02 Frequently Gender Male 1.00 P, 0.02 Female Total negative score 1st quartile 1.00 P, nd quartile rd quartile th quartile Alcohol,40 pw for men and 24 pw women 1.00 P, 0.06 $ 40 pw men and 24 pw women Job type Non-manual 1.00 P, 0.03 Manual Health in the last year Good to very good 1.00 P, Bad to very bad Chronic symptoms No symptoms 1.00 P, or more symptoms Symptoms in the last year 0 or 1 symptoms 1.00 P, or more symptoms Combined incidents No minor injuries or cognitive failures 1.00 P, Minor injuries only Cognitive failures only Cognitive failures and minor injuries Melamed et al. s [15] Ergonomic Stress Level, which focused on the negative aspects of individual jobs rather than general job characteristics, this study found that higher levels of negative job characteristics were related to increased risk of minor injuries and accidents at work. It was felt that this composite score was a more sensitive measure of the negative aspects of individual jobs, because within broad occupational groups there are different levels of risk. Previous research has indicated the association between long hours [50] and shift work and the increased risk of accidents [51]. It is unclear which aspects of negative job characteristics are predictive of accidents, minor injuries and cognitive failures (analysis of this is ongoing), but the authors feel that it is the combined effects of these variables that is important in identifying those at risk. Minor injuries were also associated with work stress, age, frequent risk taking outside work, symptoms, income, job type, sleep problems, combined incidents (cognitive failures and accidents) and being self-employed. Studies have shown an effect for stress on the risk of occupational injuries [18,19,52] and for perceived stress (caused by environmental factors) and emotional wellbeing on the likelihood of being injured [53 55]. In the present study, the age group was more likely to report minor injuries than the older age groups. Other studies [40] have noted higher accident rates among the youngest people, even when job type is taken into consideration. This could be explained to some extent by the increasing experience and skill of older workers. However, Chi and Wu [12] suggested that the relationship between age and fatality rate was dependent on the industry and accident type, as older workers were more prone to certain types of accidents. As with other studies [16,17,56], those who were selfemployed reported more minor injuries, which could be explained by lower standards of health and safety due to cost restrictions and differing regulatory control. The self-employed also tend to work in riskier industries like farming or mining [36] and may work longer hours and set deadlines that could increase the risk of accidents, e.g. long distance lorry drivers. Disturbed sleep or tiredness has been associated with increased risk of accidents [29 31] and linked to performance decrements [50].

8 106 OCCUPATIONAL MEDICINE In the present study sleep problems were associated with minor injuries at work. Chronic health symptoms and symptoms in the last 14 days were associated with an increase in minor injuries. Somatic symptoms and minor ailments such as cold and influenza have been associated with decreased work performance [57,58] but the direction of causality is unclear, as poor health could lead to accidents but the reverse could also be the case. The pattern for cognitive failures is similar to minor injuries, however gender, neuroticism, anxiety, depression, being a non-smoker and time on the job are all associated with cognitive failures. There is little direct evidence on the factors associated with cognitive failures in the workplace but cognitive failures have been associated with depression and psychiatric morbidity [59], anxiety [60] and stress [50]. Examination of the accident literature revealed similar patterns [3,23,28]. Such disturbed cognitive functioning could lead to a decline in the performance or an increased lack of attention or vigilance, which in turn could increase the likelihood of accidents. There was a significant effect for gender independent of risk taking behaviour, job type (manual/non-manual) and working hours. Research has shown that [29], men generally have a greater accident risk than women [40]. This may be because even within job sectors the risk of injury to men is greater, e.g. in a white collar job men are more likely to get involved in lifting and carrying than women. The pattern for accidents is somewhat similar to minor injuries and cognitive failures, but the differences evident between the results for cognitive failures and accidents may reflect the fact that accidents are context dependent; only in some cases will cognitive failures or errors lead to accidents and injury. Also, accidents and cognitive failures differ in terms of frequency and consequences, and the perception of cognitive failures is very subjective, whereas the recall of an accident is somewhat less so. There are a number of limitations of the study. Selfreported data will be influenced by a number of factors including the individual differences in recall, definitions, health perceptions and the saliency of events. The study was anonymous and no follow-ups were initiated, the response rate of 27% was quite low, although normal for this type of postal survey, so there is a possibility of response bias. Workers may be less likely to respond and those who did may have been more likely to have had an accident, thus giving an inflated estimate of the accident rate. However, the study reports result from many types of accidents and we feel that there is no reason to believe that these factors will vary systematically, thus the impact of the low response rate on the factors associated with accidents may be considerably less. Also, despite problems with the cross-sectional design, this approach enabled us to examine minor injuries, cognitive failures and accidents in a cross-section of the population with a variety of occupations and demographic characteristics. In conclusion, the study has identified the incidence of work-related minor injuries, cognitive failures and accidents among a sample of the working population in South Wales. Factors associated with minor injuries, cognitive failures and accidents were also identified and although similar to those highlighted in previous work, some new associations have been noted. These may have implications for worker health and productivity. Although the lack of research examining minor injuries and cognitive failures precludes direct comparisons, this work provides a valuable starting point for future research in this area. Finally, given that the cost of accidents and injuries to the individual, the employer and the economy per annum is enormous ( 558 million, 2.5 billion and between 2.9 and 4.2 billion, respectively) this study s findings should prove useful to those seeking to identify patterns of work related incidents and risk factors and to develop policies to reduce occupational injury, regardless of the severity. Acknowledgements This project was funded by the Health and Safety Executive. The authors would like to thank the Steering Committee: Professor Mick Bloor, Cardiff University; Dr Dick Pates, Community Addiction Unit, Cardiff; Dr Peter Kelly, Health and Safety Executive; Mr Peter Richmond, University Hospital Wales; Professor Ronan Lyons, University of Swansea; Professor Dave Nutt, University of Bristol; Ann Harrington, Health and Safety Executive and Professor Tim Peters, University of Bristol, for their advice and comments. References 1. Health and Safety Executive. Levels and Trends in Workplace Injury: reported Injuries and the Labour Force Survey, Health and Safety Commission, Department of the Environment. Revitalising Health and Safety Strategy, Health and Safety Executive, Di Lorenzo L, Zocchetti C, Platania A, et al. Minor and major industrial accidents in a southern Italian (Apulia) food factory: a ten-year study. Med Lav 1998; 89: Larson G, Alderton D, Neideffer M, Underhill E. Further evidence on the dimensionality and correlates of the cognitive failures questionnaire. Br J Psychol 1997;88: Larson G, Merritt C. Can accidents be predicted? An empirical test of the cognitive failures questionnaire. Applied Psychology: An International Review 1991;40: Hansen C. A causal model of the relationship among accidents, biodata, personality and cognitive factors. J Appl Psychol 1989;74:81 90.

9 S. A. SIMPSON ET AL.: MINOR INJURIES, COGNITIVE FAILURES AND ACCIDENTS AT WORK Arthur W, Barrett G, Alexander R. Prediction of vehicular accident involvement: a meta-analysis. Hum Perform 1991; 4: Wagle A, Berrios G, Ho L. The cognitive failures questionnaire in psychiatry. Compr Psychiatry 1999;40: Mahoney, Dalby J, King MC. Cognitive failures and stress. Psychol Rep 1998;82: Wadsworth E, Simpson S, Moss S, Smith A. The Bristol Stress and Health Study: accidents, minor injuries and cognitive failures at work. Occup Med (Lond) 2003;53: Moss S, Wadsworth E, Simpson S, Lyons R, Richmond P, Smith A. Identifying risk factors for accidents within a population of Accident and Emergency attendees in Wales. J Emerg Med (in review). 12. Chi C, Wu M. Fatal occupational injuries in Taiwan a relationship between fatality rate and age. Saf Sci 1997;27: Dell T, Berkhout J. Injuries at a metal foundry as a function of job classification, length of employment and drug screening. J Saf Res 1998;29: Blasi B, Caccavari R. Risk of accidents to hospital personnel. Acta Biomed Ateneo Parmense 1983;54: Melamed S, Luz J, Najensen T, Jucha E, Green M. Ergonomic stress levels, personal characteristics, accidents occurrences and sickness absence. Ergonomics 1989;32: Feyer AM, Langley J, Howard M, Horsburgh S, Wright C, Alsop J, et al. Work Related Fatal Injuries in New Zealand 1985 to Dunedin: University of Otago, Driscoll T, Healey S, Mitchell R, Mandryk J, Leigh Hendrie A, Hull B. Are the self-employed at higher risk of fatal work-related injury? Saf Sci 2003;41: Murata K, Kawakami N, Amari N. Does job stress affect injury due to labour accident in Japanese male and female blue collar workers? Ind Health 2000;38: Nolting H, Berger J, Schiffhorst G, Genz H, Kordt M. Job strain as a risk factor for occupational accidents among hospital nursing staff. Gesundheitswesen 2002;64: Lindqvist K, Schelp L, Timpka T. Gender aspects of work related injuries in a Swedish municipality. Saf Sci 1999;31: Evans L, Wasielewski P. Do accident-involved drivers exhibit riskier everyday driving behaviour? Accid Anal Prev 1983;15: Mearns K, Flin R, Gordon R, Fleming M. Human and organizational factors in offshore safety. Work Stress 2001; 15: Kirkcaldy B, Furnham A. Positive affectivity, psychological well being, accidents and traffic deaths and suicide: an international comparison. Stud Psychol 2000;42: Lewis R, Cooper S. Alcohol, other drugs, and fatal workrelated injuries. J Occup Med 1989;31: Baker S, O Neill B, Ginsberg M, Li G. The Injury Fact Book. New York: Oxford University Press, Brewer R, Morris P, Cole P, Watkins S, Patetta M, Popkin C. The risk of dying in alcohol-related automobile crashes among habitual drunk drivers. N Engl J Med 1994;331: Trucco M, Rebolledo P, Gonzalez X, Correa A, Bustamente M. Recent alcohol and drug consumption in victims of work related accidents. Rev Med Chil 1998;126: Murray M, Fitzpatrick D, O Connell C. Fishermen s blues: factors related to accidents and safety among Newfoundland fishermen. Work Stress 1997;11: Åkerstedt T, Fredlund P, Gillberg M, Jansson BJ. A prospective study of fatal occupational accidents: relationship to sleeping difficulties and occupational factors. J Sleep Res 2002;11: Melamed S, Oksenberg A. Excessive daytime sleepiness and risk of occupational injuries in non-shift daytime workers. Sleep 2002;25: Ohayon MM, Caulet M, Philip P, Guilleminault C, Priest RG. How sleep and mental disorders are related to complaints of daytime sleepiness. Arch Intern Med 1997; 157: Li C, Chen K, Wu C, Sung F. Job stress and dissatisfaction in association with non-fatal injuries on the job in a crosssectional sample of petrochemical workers. Occup Med (Lond) 2001;51: Cellier J, Eyrolle H, Bertrand A. Effects of age and level of work experience on occurrence of accidents. Percept Mot Skills 1995;80: Butani S. Relative risk analysis of injuries in coal mining by age and experience at present company. J Occup Accid 1988;10: Helmkamp J, Bone C. The effect of time in a new job on hospitalization rates for accidents and injuries in the US navy, 1977 through J Occup Med 1987;29: Driscoll T, Healey S, Mitchell R, Mandryk J, Leigh Hendrie A, Hull B. Are the self-employed at higher risk of fatal work-related injury? Saf Sci 2003;41: Nag P, Patel V. Work accidents among shift workers in industry. Int J Ind Ergon 1998;21: Connor J, Whitlock G, Norton R, Jackson R. The role of driver sleepiness in car crashes: a systematic review of epidemiological studies. Accid Anal Prev 2001;33: Zwerling C, Sprince NL, Wallace RB, Charles S. Risk factors for occupational injuries among older workers: an analysis of the health and retirement study. Am J Public Health 1996;89: Health and Safety Executive. Health and Safety Statistics, Health and Safety Executive, Smith A, Johal SS, Wadsworth E, Davey Smith G, Peters T. The Scale of Occupational Health: The Bristol Stress and Health at Work Study. London: Health and Safety Executive, 2000b. 42. Broadbent DE, Cooper PF, Fitzgerald P, Parkes KR. The Cognitive Failures Questionnaire (CFQ) and its correlates. Br J Clin Psychol 1982;21: Smith A, McNamara R, Wellens B. Combined effects of occupational health hazards. Research Report. HSE Books, Smith A, MacKay C (eds) The Combined Effects of Occupational Health Hazards. London: Taylor and Francis, Smith A, Brice C, Sivell S, Wellens B (eds) The Combined Effects of Occupational Factors on Subjective Reports of Health. London: Taylor and Francis, 2001.

10 108 OCCUPATIONAL MEDICINE 46. Bjelland I, Dahl A, Haug T, Neckelman D. The validity of the hospital anxiety and depression scale. An updated literature review. J Psychosom Res 2002;52: Jones J, Janman D, Payne R. Some determinants of stress in psychiatric nurses. Int J Nurs Stud 1987;24: Santor D, Bagby R, Joffe R. Evaluating stability and change in personality and depression. J Pers Soc Psychol 1997;73: Farmer A, Redman K, Harris T, Mahmood A, Sadler S, Pickering A, McGuffin P. Neuroticism, extraversion, life events and depression. Br J Psychiatry 2002;181: Spurgeon A, Harrington JM, Cooper C. Health and safety problems associated with long working hours: a review of the current position. Occup Environ Med 1997;54: Garbarino S, Beelke M, Costa G, et al. Brain function and effects of shift work: implications for clinical neuropharmacology. Neuropsychobiology 2002;45: Johnston J. Occupational injury and stress. J Occup Environ Med 1995;37: Caplan R. Person environment fit: past, present and future In: Cooper C (ed.) Stress Research. New York: John Wiley, 1983; Sauter S, Murphy L. The changing face of work and stress. Organizational Risk Factors for Job Stress. Washington: American Psychological Association, Li C. Job stress and dissatisfaction in association with non-fatal injuries on the job in a cross-sectional sample of petrochemical workers. Occup Med (Lond) 2001;51: HSE HaSE. Revitalising Health and Safety: Consultation Document. United Kingdom: HSE, Houston D, Allt S. Psychological distress and error making among junior house officers. Br J Health Psychol 1997;2: Smith A. A review of the effects of colds and influenza on human performance. J Soc Occup Med 1989;39: Van den Bosch R, Rombouts R, Van Asma M. Subjective cognitive dysfunction in schizophrenic and depressed patients. Compr Psychiatry 1993;34: Broadbent DE. Individual differences in reponse to low discretion jobs. Ergonomics 1994;37: Zigmond A, Snaith R. The hospital anxiety and depression scale. Acta Psychiatr Scand 1983;67: Eysenck HJ. Eysenck Personality Inventory/Questionnaire. In: Bellack AS (ed.) Dictionary of Behavioural Assessment Techniques. Oxford: Pergamon Press, 1988; Marmot M, Davey-Smith G, Stanfield S. Health inequalities among British civil servants: the Whitehall II Study. Lancet 1991;337: Karasek R, Brisson C, Kawakami N, Huntman I, Bougers P, Amick B. The job context questionnaire (JCQ): an instrument for internationally comparitive assessments of psychosocial job characteristics. J Occup Health Psychol 1998;3: