Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Evaluation of safety training for manual handling of people with disabilities in specialised group homes in Australia

Evaluation of safety training for manual handling of people with disabilities in specialised... Abstract Objective: This paper attempted to examine the question of whether 351 disability services workers in an Australian State Government agency given manual handling training had significantly lower injury rates from manual handling injuries a year post-training than a control group (n=351) without the training. Timothy Ore Department of Justice, Victoria Methods: Control group members were selected by stratified random sampling to closely match the intervention group. The measures employed were the number of reported manual handling injuries per 100 full-time equivalents by certain characteristics, frequency of incidents, mean workers’ compensation cost, and mean compensated days. Results: Training in manual handling methods significantly reduced the risk by as much as 42%, with an average rate of 49.6 per 100 full-time equivalents (95% confidence interval 44.4-55.0) among the intervention group compared with 84.8 per 100 full-time equivalents (95% confidence interval 76.0-94.1) among controls. The risk differential was consistent across gender, age group, length of service, and job classification. However, in two injury categories (client lift/transfer and general manual handling), the intervention group had a marginally higher risk. The average workers’ compensation cost in the control group was 4.2 times that in the intervention group. Conclusion: While the study has some weaknesses, overall, there is little evidence to suggest that the outcome observed was due to threats to internal validity. It is expected that the findings can encourage further evaluation studies of workplace safety training programs. (Aust N Z J Public Health 2003; 27: 64-9) recent study1 of manual handling injuries among disability services workers in the Australian State Government agency examined in this paper shows that over the three years to June 2000 an average of 905 injuries occurred annually (25.7 injuries per 100 full-time equivalents), with approximately 11% involving workers’ compensation payments. Nearly half (47.7%) the injuries were associated with providing support to clients, with the most common cases being assisting unsteady clients into baths and in the shower, dressing clients, assisting clients into and out of bed, and holding clients during epileptic seizures. About 9 in 10 of the injuries resulted in sprains and strains of joints and adjacent muscles. High-risk groups for manual handling injury among the disability services workers were identified to include women, particularly older women, young employees, employees with less than one year of service, full-time employees, and employees with low job classifications.1 The workers accounted for a disproportionate share of the agency’s workers’ compensation payments involving manual handling injuries, an annual average of 80% compared with their share of the workforce (50%). A small proportion (13%) of the costliest claims represented a majority (76%) of the total compensation payments. The pattern of manual handling injury among the workers is consistent with those in the published literature on health care workers.2,3 The preponderance of compensation payments borne by a fraction of the injuries is also similar to those reported in overseas studies.4-6 In Australia, the manual handling policy of most jurisdictions requires disability services workers to be trained in manual handling techniques. In the setting examined in this paper, training in manual handling is provided to some disability services staff caring for clients exhibiting challenging behaviour, such as very aggressive behaviour towards others, self or the environment.7 The primary purpose of the training is to reduce the risk for manual handling. Although training for staff at risk of manual handling injury and working with clients exhibiting challenging behaviour is frequently recommended, sometimes implemented, it is rarely adequately evaluated.8,9 This paper examines the question of whether disability services workers with manual handling training a year post-training had significantly lower injury rates from manual handling than their counterparts without the training. Training, defined as formal or informal methods to assist individuals in acquiring knowledge, changing attitudes, or executing safe work behaviours, continues to be considered a practical approach to injury control.10-12 In the United States, more than 100 Occupational Safety and Health Administration (OSHA) standards for controlling workplace hazards contain requirements for worker training to reduce risk factors for injury.13 Other standards limit certain tasks to workers considered competent by virtue of special training.13 Indicators of safety training effectiveness Submitted: July 2002 Revision requested: November 2002 Accepted: November 2002 Correspondence to: Dr Timothy Ore, 25 Kelson Street, Coburg, Victoria 3058. Fax: (03) 9627 6611; e-mail: timothy.ore@justice.vic.gov.au AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2003 VOL. 27 NO. 1 Illness Prevention Safety training for manual handling of people with disabilities are an important way for improving the planning, funding and delivery of disability services for many reasons. First, quality of staff well-being is a critical contributor to the quality of service provided to people with disabilities. Second, with the number of people with disabilities estimated to increase 4% annually in Australia,14 demand for community-based disability accommodation services will continue to rise while at the same time it will become more difficult to attract and retain direct care staff.15 Third, to prove due diligence, employers need to demonstrate that employees have been provided with safety training and that recipients understood the training and successfully applied it to their assigned tasks.16 Fourth, the public sector and community services are one of the four industries with the poorest safety performance, with manual handling being the leading cause of injury.17 Methodology Disability services workers are primarily responsible for the daily care of people with a range of disabilities, including intellectual disability, acquired brain injury, and psychiatric disability. Women represent over two-thirds (69%) of the workers. Clients live in small (about five in a house) group homes offering varying levels of support to people with disabilities.18 An average of 44% of residents have specialised needs regarding challenging behaviour. Depending on the skills and abilities of each client, the residential facilities are primarily geared at permitting clients to progress from the 24-hour intensive support model to one involving greater independence.19 Physical aggression is one of the most common forms of challenging behaviour experienced by staff supporting people with disabilities.20,21 Indeed, the presence of aggressive behaviour is the most frequent reason for referral to specialist intervention services.22 Based mainly at participants’ worksites or group homes, the 35-hour manual handling training is provided by an ergonomist, involving on-site assessment of manual handling tasks, training on specific techniques, and equipment design and correct use. Other components of the training include selection and purchase of manual handling equipment, design of manual handling tasks, how hazardous manual handling affects the body, factors that can increase the risk for musculoskeletal disorders, and prevention of the disorders. Throughout the training, participants are expected to recognise the risks of manual handling and the safest way for carrying out their tasks without risking their health. Participants are also expected to practise the techniques, under supervision, before being required to use them. The training, which is offered at different periods in the group homes, is available to those nominated by their house supervisors. As part of the induction course to the agency, all disability services workers receive a short (four hours), non-task specific training in manual handling. Unlike the formal training evaluated here, this one is insufficient to prepare the workers for the hazards on the job. A common problem in training evaluation is attributing 2003 VOL. 27 NO. 1 observed changes in injury experience in the intervention group to the training program being assessed.11,12 This is an important issue as there are numerous factors other than training that may influence such changes. Examples include client profile, psychosocial, demographic, environmental, incident reporting procedures, and work organisational factors. This paper uses a post-test control design to minimise the impact of some of these problems. Post-test control designs are used primarily when time for data collection is limited or if pre-testing trainees would facilitate learning and thus inflate measures of training effectiveness.23 Many studies have employed post-test only control, including Feldstein et al.,24 Parkinson et al.,25 Tan et al.,26 Videman et al.27 and Jansson.28 The most stringent evaluation of training, however, is randomised pre-test/post-test design with a control group or a controlled randomised intervention trial.11,12 However, as noted by several researchers,11,12,23 it is questionable as to how practical it is to implement this type of study in the workplace, including the present setting. When randomisation is not possible, it is important to determine if the two groups are similar with respect to any measurable and available characteristics. A search of the occupational health and safety training records maintained by the agency and the worksites showed that a total of 351 disability services workers completed the manual handling training between 1 July 1998 and 30 June 1999. The intervention group for the study comprised all the 351 workers, with no need for sampling. Controls (n=351) were selected from all the 4,500 Table 1: Percentage distribution of sample characteristics. Characteristics Gender Male Female Age (in years) < 30 30-49 50+ Length of service <2 years 2 years and over Job classification HSWsa IDSOsb MRNsc Other Work-site/residence CRUsd Congregate care Number 74 26 351 76 24 351 27 50 15 8 25 50 15 10 58 42 60 40 28 42 30 30 40 30 36 64 35 65 Percentage distribution (%) Intervention group Control group Notes: (a) Human services workers. (b) Intellectual disability services officers. (c) Mental retardation nurses. (d) Community residential units. AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH Ore Article full-time disability services workers in the agency, during the same period, who had not been given the training to closely match the intervention group by stratified random sampling, with respect to age, sex, job classification, length of service, and worksite/ residential type (see Table 1). These parameters were selected to minimise a selection threat to internal validity.12 The threat results when the observed impact of an intervention could be due to differences in subject characteristics in the groups being compared rather than in the intervention itself. The characteristics have also been shown to be associated with the risk for manual handling injury among disability services workers.1 Owing to a lack of data from the training or personnel records, other potentially relevant matching variables, such as job satisfaction, attitude, motivation, values, and socio-economic status, could not be used. Stratified random sampling generally has more statistical power than simple random sampling. The method ensures that for a nonhomogenous population, the variability within groups is lower than the variability for the population, thereby yielding more accurate inferences on a question of interest. The measures employed were the number of reported manual handling injuries (from 1 July 1999 to 30 June 2000) per 100 full-time equivalents (FTEs) by age, sex, injury type, length of service in the agency, and job category; frequency of incidents; mean workers’ compensation cost; and mean compensated days (time off work as a result of injury). An FTE is defined as a person working 76 hours per fortnight. The data were not analysed by type of employment because most (92%) of the sample were full-time employees. Using personnel identification numbers, the reported incidents of manual handling were extracted from the agency’s injury surveillance system. In compliance with the State’s Accident Compensation Act 1985, which requires all employers to operate a register of injuries in each workplace, employees at the state agency must fill, sign and date the front page of an injury reporting form as soon as possible. Information provided includes type of incident, cause of injury, and nature of injury. Where an employee is unable to report the injury, someone else can give notice on the employee’s behalf. For injuries that result in time off work, the Table 2: Statistical analysis of manual handling. Parameter Gender (cases/100 FTEs) Male Female Total Age (cases/100 FTEs) <30 30-49 50+ Total Length of service (cases/100 FTEs) <2 years 2 years and over Mean claims cost ($) Mean compensated days Injury type (cases/100 FTEs) Client support/assistance Client lift/transfer MH general MH while handling objects Static postural loading Frequency (%) Injured once Injured more than once Job category (cases/100 FTEs) HSWsc IDSOsd MRNse Notes: (a) Highly significant. (b) Significant. (c) Human services workers. (d) Intellectual disability services officers. (e) Mental retardation nurses. Intervention group (95% CI) 48.96 (40.6-58.0) 49.67 (43.3-56.5) 49.56 (44.4-55.0) 52.75 (42.6-63.9) 48.42 (40.7-56.7) 49.74 (40.5-59.9) 49.56 (44.4-55.0) 48.54 (42.0-55.6) 49.88 (42.0-58.4) 2,658.45 (n=12) 52.75 (n=4) 17.13 (14.2-20.4) 8.57 (6.5-10.9) 7.95 (5.9-10.2) 3.06 (1.9-4.5) 0.61 (0.2-1.3) 100.0 0.0 52.54 (42.3-63.9) 49.36 (42.2-57.1) 38.30 (28.5-49.6) Control group (95% CI) 81.69 (67.8-96.9) 85.69 (74.6-97.5) 84.83 (76.0-94.1) T-test p-value 0.0004a 8.45 90.79 (73.4-110.1) 83.52 (70.3-97.8) 83.56 (68.0-100.7) 84.83 (76.0-94.1) 8.51 79.80 (69.0-91.4) 87.61 (73.8-102.6) 11,354.33 (n=31) 103.63 (n=16) 30.30 (25.0-36.1) 7.27 (5.5-9.2) 7.88 (5.9-10.1) 11.51 (7.1-16.9) 9.09 (4.1-16.1) 1.84 94.0 6.0 6.44 107.8 (86.8-131.1) 81.5 (69.7-94.2) 75.17 (55.9-97.3) 2.80 5.32 3.94 0.0004a 0.0004a 0.0174b 0.0335b 0.0034b 0.0013a AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2003 VOL. 27 NO. 1 Illness Prevention Safety training for manual handling of people with disabilities employee also completes a workers’ compensation claim form, accompanied with a medical certificate and the injury notification form. The information is then forwarded to the agency’s insurer within 10 days.1 Cost effectiveness was estimated as the mean workers’ compensation payments of the control group minus that of the intervention group divided by the estimated training cost per trainee. As the numerator does not include indirect benefits, the costeffectiveness ratios should be considered the conservative minimum estimate. Indirect costs are real costs and they increase as direct costs increase, but are subjective and difficult to measure.29 Some indirect costs have been estimated at between 20% and 30% of premium payments.29 It should be noted that not all the difference in claims cost between the control and intervention groups are due to training alone. The incident data were entered and analysed using SAS. T-tests were computed with 95% confidence intervals (CI). Results Details of the risk variations for manual handling injury between the intervention and control groups are presented in Table 2. Training in manual handling methods significantly reduced the risk by as much as 42%, with an average rate of 49.6 per 100 FTEs (95% CI 44.4-55.0) among the intervention group compared with 84.8 per 100 FTEs (95% CI 76.0-94.1) among controls. The risk differential was consistent across gender, age group, length of service, and job classification. However, in two injury categories (client lift/transfer and general manual handling), the intervention group had a marginally higher risk. The control group had an average claims cost of $11,354, 4.2 times that of the intervention group ($2,658). Manual handling training was highly cost effective, with an estimated return of $5.80 per dollar invested. Discussion This paper has attempted to evaluate the impact of training in manual handling techniques on injury rates from manual handling in two comparison groups of disability services workers in an Australian State Government agency. The training was found to be effective, with the intervention group having a significantly lower risk for injury than the control group. The large differences in workers’ compensation claims cost between the two groups also indicate that the training was a good investment. Three major weaknesses are evident in this paper. One concerns the use of post-test design with a non-randomised control group. An experimental design provides the strongest evidence of a causal link between a specific intervention and outcome primarily because of an unbiased random allocation of people or workplaces to intervention and control groups.12 The principal reason for not using randomisation relates to the requirements of the Occupational Health and Safety (Manual Handling) Regulations 1999, which came into operation in the State on 1 July 1999, with the objective of reducing the number and severity of 2003 VOL. 27 NO. 1 musculoskeletal disorders associated with manual handling. An employer must ensure that any task undertaken by an employee involving manual handling of live persons or loads is identified and a risk assessment carried out to determine whether there is any risk of a musculoskeletal disorder affecting an employee as a result of that task. Where there is a risk, an employer is required to either eliminate the risk or reduce it by altering the workplace environment, changing the systems of work or provide training in safe manual handling techniques. Thus, random assignment of workers to intervention and control groups could constitute a potential breach of the regulations. A second limitation is that because an after-only design was used, the pre-intervention injury experience of the intervention and control groups was not taken into account. Injury records before July 1998 were checked to establish differences in manual handling injury rates between the two groups. However, as most (over half) of the workers were either new, with only a few months on the job, reliable rates could not be generated. The disadvantage of not obtaining ‘before’ measurements is that it is not possible to know if the groups differed initially with respect to the outcome measure. Another weakness is that the injury rates did not account for whether the training changed the practices targeted, that is, correct and safe manual handling. For instance, if injury rates fell post intervention while unsafe manual handling practices remained unchanged, it raises the possibility that the reduction in injury rates could be due to other factors. If, however, a decrease occurred in both injury rates and unsafe practices, it could be assumed that the improvement was associated with the intervention. Could the results of the study be due to factors other than the intervention itself? Alternative interpretations of the results can be attempted by examining a number of threats to internal validity. These include selection threat, regression-to-the-mean, testing threat, drop-out threat, selection interaction, and diffusion threat.12 The more these threats are present in an evaluation the less likely it is that the results are actually due to the intervention. The possibility of a selection threat was minimised by closely matching control group members to the intervention group on important characteristics, such as demographics and job classification. However, lack of data precluded using other relevant matching variables. The credibility of evaluation findings may be threatened by selection interaction effects, that is, dissimilarity occurring in the characteristics of the groups over the duration of the analysis. Examples of sources of dissimilarity include history, incident reporting, and maturation. Common history threats such as changes in supervisor or management personnel, worker-management relations, legislation, structure or pace of work were not evident. There was little evidence of under-reporting of manual handling injuries. Robson et al.12 have suggested using the ratio of major to minor injuries to help detect reporting threat with regard to injury statistics. The greater the suppression of reporting, the higher the ratio will be since minor injuries are easier to underreport than major injuries, that is, those involving compensation AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH Ore Article payments. If the ratio is constant, then the likelihood of a reporting threat is reduced.12 The ratio of major injuries to minor injuries (not involving compensation payments) for the first half of the period of analysis was similar to that for the second half (approximately 1:11) for both the control and intervention groups. A regression-to-the-mean threat to internal validity occurs if an intervention is delivered to a group of workers with high injury rates and the rates for the group are compared with groups with lower injury rates. Without intervention, there is a tendency for the high injury group to have a decrease in injury rates. This is mainly because of the tendency for workplace injury rates to sometimes fluctuate from year to year, with the effect that any group with a lower-than-usual injury rate in any given year is more likely to have their rates increase than decrease in the following year, assuming workplace conditions remain unchanged. Similarly, any group with a higher-than-usual injury rate in any given year is more likely to have their rates decrease than increase in the subsequent year.12 Because selection for the manual handling training is not systematic across the organisation, there is little evidence to indicate that particular workers were nominated for the training because of their elevated risk. A drop-out threat occurs if one group has a greater quit rate than the other, particularly if it results in the control and intervention groups having different characteristics, particularly in age, gender, and service length. Membership of both groups remained stable over the one-year period of analysis. There were only five known cases of turnover (two in the control group and three in the intervention group) and these were insufficient to significantly change the results. A contamination effect, alternatively known as a diffusion effect, relates to circumstances in which the intervention provided diffuses to the control group. Given the tendency for workers to share information with one another through learning and observation, the present study was unable to ascertain the extent to which this threat was present. One way to minimise the possibility of a diffusion threat is by keeping the control group separate from the intervention group, but this was rarely the case in the setting. A testing threat results when the practice of taking the before measurement could affect the safety outcome used to evaluate an intervention. This threat is only a problem where outcomes are assessed via questionnaires, interviews or observation to determine employee safety knowledge, practices and attitudes. Because injury outcomes were measured only post intervention without interaction with the workers, a testing threat had no influence on the results. Where an intervention is targeted at reducing workplace injuries, as in the current setting, the ideal outcome measurement is an unbiased measure of injuries. One problem with this is that the frequency of lost time events in a given month or year in most workplaces may be too low to show clearly the effect of an intervention.12 There was little evidence that this was the case in this study. With an average of more than 900 manual handling injuries per year,1 the frequency of lost time events was sufficient to establish the impact of the manual handling training. Misclassification errors can lead to problems in injury statistics.30 This may help explain the slightly higher injury rates (but not statistically significant) in the intervention group than the control group in only two injury categories, client lift/transfer and general manual handling. Given the consistently higher injury rates among the control group in the remaining parameters, there is little basis for assuming that the training was unsuccessful. The analysis of manual handling injuries among the workers in the current setting by Ore1 found that there was limited information on injury description, job classification, and time of incidents. One-fifth of the 2,714 injuries identified over a three-year period were classif ied as general manual handling. Misclassification can be attributed to several sources, including disagreement or misinterpretation of incidents on the part of individual coders and the classification scheme itself.12 The International Classification of Diseases, 9th revision (ICD-9) has been shown to perform poorly for soft-tissue disorders, as several different ICD-9 codes can be assigned for the same condition.30 The ability to code incident descriptions consistently has been reported to depend on which aspect of the incident is being coded (Robson et al. 2001). For two trained people coding the same 100 injury reports using a standardised coding scheme, the degree of agreement in coding was 44% for event precipitating injury and 61% for contributory factors, well below the reliability threshold of 70%.12 By contrast, the agreement on gender was 98%, type of injury 92%, part of body injured 92%, and agent of injury 79%.12 Conclusion Training in manual handling techniques was successful in reducing the risk for injury among the intervention group of disability services workers in the Australian State Government agency. While the study has some weaknesses, overall, there is little evidence to suggest that the outcome observed was due to threats to internal validity. It is expected that the findings can encourage further evaluation studies of workplace safety training in particular and occupational injury intervention programs more generally. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Australian and New Zealand Journal of Public Health Wiley

Evaluation of safety training for manual handling of people with disabilities in specialised group homes in Australia

Download PDF
6 pages

Loading next page...
 
/lp/wiley/evaluation-of-safety-training-for-manual-handling-of-people-with-pel1VWcE4b

References (32)

Publisher
Wiley
Copyright
Copyright © 2003 Wiley Subscription Services, Inc., A Wiley Company
ISSN
1326-0200
eISSN
1753-6405
DOI
10.1111/j.1467-842X.2003.tb00382.x
Publisher site
See Article on Publisher Site

Abstract

Abstract Objective: This paper attempted to examine the question of whether 351 disability services workers in an Australian State Government agency given manual handling training had significantly lower injury rates from manual handling injuries a year post-training than a control group (n=351) without the training. Timothy Ore Department of Justice, Victoria Methods: Control group members were selected by stratified random sampling to closely match the intervention group. The measures employed were the number of reported manual handling injuries per 100 full-time equivalents by certain characteristics, frequency of incidents, mean workers’ compensation cost, and mean compensated days. Results: Training in manual handling methods significantly reduced the risk by as much as 42%, with an average rate of 49.6 per 100 full-time equivalents (95% confidence interval 44.4-55.0) among the intervention group compared with 84.8 per 100 full-time equivalents (95% confidence interval 76.0-94.1) among controls. The risk differential was consistent across gender, age group, length of service, and job classification. However, in two injury categories (client lift/transfer and general manual handling), the intervention group had a marginally higher risk. The average workers’ compensation cost in the control group was 4.2 times that in the intervention group. Conclusion: While the study has some weaknesses, overall, there is little evidence to suggest that the outcome observed was due to threats to internal validity. It is expected that the findings can encourage further evaluation studies of workplace safety training programs. (Aust N Z J Public Health 2003; 27: 64-9) recent study1 of manual handling injuries among disability services workers in the Australian State Government agency examined in this paper shows that over the three years to June 2000 an average of 905 injuries occurred annually (25.7 injuries per 100 full-time equivalents), with approximately 11% involving workers’ compensation payments. Nearly half (47.7%) the injuries were associated with providing support to clients, with the most common cases being assisting unsteady clients into baths and in the shower, dressing clients, assisting clients into and out of bed, and holding clients during epileptic seizures. About 9 in 10 of the injuries resulted in sprains and strains of joints and adjacent muscles. High-risk groups for manual handling injury among the disability services workers were identified to include women, particularly older women, young employees, employees with less than one year of service, full-time employees, and employees with low job classifications.1 The workers accounted for a disproportionate share of the agency’s workers’ compensation payments involving manual handling injuries, an annual average of 80% compared with their share of the workforce (50%). A small proportion (13%) of the costliest claims represented a majority (76%) of the total compensation payments. The pattern of manual handling injury among the workers is consistent with those in the published literature on health care workers.2,3 The preponderance of compensation payments borne by a fraction of the injuries is also similar to those reported in overseas studies.4-6 In Australia, the manual handling policy of most jurisdictions requires disability services workers to be trained in manual handling techniques. In the setting examined in this paper, training in manual handling is provided to some disability services staff caring for clients exhibiting challenging behaviour, such as very aggressive behaviour towards others, self or the environment.7 The primary purpose of the training is to reduce the risk for manual handling. Although training for staff at risk of manual handling injury and working with clients exhibiting challenging behaviour is frequently recommended, sometimes implemented, it is rarely adequately evaluated.8,9 This paper examines the question of whether disability services workers with manual handling training a year post-training had significantly lower injury rates from manual handling than their counterparts without the training. Training, defined as formal or informal methods to assist individuals in acquiring knowledge, changing attitudes, or executing safe work behaviours, continues to be considered a practical approach to injury control.10-12 In the United States, more than 100 Occupational Safety and Health Administration (OSHA) standards for controlling workplace hazards contain requirements for worker training to reduce risk factors for injury.13 Other standards limit certain tasks to workers considered competent by virtue of special training.13 Indicators of safety training effectiveness Submitted: July 2002 Revision requested: November 2002 Accepted: November 2002 Correspondence to: Dr Timothy Ore, 25 Kelson Street, Coburg, Victoria 3058. Fax: (03) 9627 6611; e-mail: timothy.ore@justice.vic.gov.au AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2003 VOL. 27 NO. 1 Illness Prevention Safety training for manual handling of people with disabilities are an important way for improving the planning, funding and delivery of disability services for many reasons. First, quality of staff well-being is a critical contributor to the quality of service provided to people with disabilities. Second, with the number of people with disabilities estimated to increase 4% annually in Australia,14 demand for community-based disability accommodation services will continue to rise while at the same time it will become more difficult to attract and retain direct care staff.15 Third, to prove due diligence, employers need to demonstrate that employees have been provided with safety training and that recipients understood the training and successfully applied it to their assigned tasks.16 Fourth, the public sector and community services are one of the four industries with the poorest safety performance, with manual handling being the leading cause of injury.17 Methodology Disability services workers are primarily responsible for the daily care of people with a range of disabilities, including intellectual disability, acquired brain injury, and psychiatric disability. Women represent over two-thirds (69%) of the workers. Clients live in small (about five in a house) group homes offering varying levels of support to people with disabilities.18 An average of 44% of residents have specialised needs regarding challenging behaviour. Depending on the skills and abilities of each client, the residential facilities are primarily geared at permitting clients to progress from the 24-hour intensive support model to one involving greater independence.19 Physical aggression is one of the most common forms of challenging behaviour experienced by staff supporting people with disabilities.20,21 Indeed, the presence of aggressive behaviour is the most frequent reason for referral to specialist intervention services.22 Based mainly at participants’ worksites or group homes, the 35-hour manual handling training is provided by an ergonomist, involving on-site assessment of manual handling tasks, training on specific techniques, and equipment design and correct use. Other components of the training include selection and purchase of manual handling equipment, design of manual handling tasks, how hazardous manual handling affects the body, factors that can increase the risk for musculoskeletal disorders, and prevention of the disorders. Throughout the training, participants are expected to recognise the risks of manual handling and the safest way for carrying out their tasks without risking their health. Participants are also expected to practise the techniques, under supervision, before being required to use them. The training, which is offered at different periods in the group homes, is available to those nominated by their house supervisors. As part of the induction course to the agency, all disability services workers receive a short (four hours), non-task specific training in manual handling. Unlike the formal training evaluated here, this one is insufficient to prepare the workers for the hazards on the job. A common problem in training evaluation is attributing 2003 VOL. 27 NO. 1 observed changes in injury experience in the intervention group to the training program being assessed.11,12 This is an important issue as there are numerous factors other than training that may influence such changes. Examples include client profile, psychosocial, demographic, environmental, incident reporting procedures, and work organisational factors. This paper uses a post-test control design to minimise the impact of some of these problems. Post-test control designs are used primarily when time for data collection is limited or if pre-testing trainees would facilitate learning and thus inflate measures of training effectiveness.23 Many studies have employed post-test only control, including Feldstein et al.,24 Parkinson et al.,25 Tan et al.,26 Videman et al.27 and Jansson.28 The most stringent evaluation of training, however, is randomised pre-test/post-test design with a control group or a controlled randomised intervention trial.11,12 However, as noted by several researchers,11,12,23 it is questionable as to how practical it is to implement this type of study in the workplace, including the present setting. When randomisation is not possible, it is important to determine if the two groups are similar with respect to any measurable and available characteristics. A search of the occupational health and safety training records maintained by the agency and the worksites showed that a total of 351 disability services workers completed the manual handling training between 1 July 1998 and 30 June 1999. The intervention group for the study comprised all the 351 workers, with no need for sampling. Controls (n=351) were selected from all the 4,500 Table 1: Percentage distribution of sample characteristics. Characteristics Gender Male Female Age (in years) < 30 30-49 50+ Length of service <2 years 2 years and over Job classification HSWsa IDSOsb MRNsc Other Work-site/residence CRUsd Congregate care Number 74 26 351 76 24 351 27 50 15 8 25 50 15 10 58 42 60 40 28 42 30 30 40 30 36 64 35 65 Percentage distribution (%) Intervention group Control group Notes: (a) Human services workers. (b) Intellectual disability services officers. (c) Mental retardation nurses. (d) Community residential units. AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH Ore Article full-time disability services workers in the agency, during the same period, who had not been given the training to closely match the intervention group by stratified random sampling, with respect to age, sex, job classification, length of service, and worksite/ residential type (see Table 1). These parameters were selected to minimise a selection threat to internal validity.12 The threat results when the observed impact of an intervention could be due to differences in subject characteristics in the groups being compared rather than in the intervention itself. The characteristics have also been shown to be associated with the risk for manual handling injury among disability services workers.1 Owing to a lack of data from the training or personnel records, other potentially relevant matching variables, such as job satisfaction, attitude, motivation, values, and socio-economic status, could not be used. Stratified random sampling generally has more statistical power than simple random sampling. The method ensures that for a nonhomogenous population, the variability within groups is lower than the variability for the population, thereby yielding more accurate inferences on a question of interest. The measures employed were the number of reported manual handling injuries (from 1 July 1999 to 30 June 2000) per 100 full-time equivalents (FTEs) by age, sex, injury type, length of service in the agency, and job category; frequency of incidents; mean workers’ compensation cost; and mean compensated days (time off work as a result of injury). An FTE is defined as a person working 76 hours per fortnight. The data were not analysed by type of employment because most (92%) of the sample were full-time employees. Using personnel identification numbers, the reported incidents of manual handling were extracted from the agency’s injury surveillance system. In compliance with the State’s Accident Compensation Act 1985, which requires all employers to operate a register of injuries in each workplace, employees at the state agency must fill, sign and date the front page of an injury reporting form as soon as possible. Information provided includes type of incident, cause of injury, and nature of injury. Where an employee is unable to report the injury, someone else can give notice on the employee’s behalf. For injuries that result in time off work, the Table 2: Statistical analysis of manual handling. Parameter Gender (cases/100 FTEs) Male Female Total Age (cases/100 FTEs) <30 30-49 50+ Total Length of service (cases/100 FTEs) <2 years 2 years and over Mean claims cost ($) Mean compensated days Injury type (cases/100 FTEs) Client support/assistance Client lift/transfer MH general MH while handling objects Static postural loading Frequency (%) Injured once Injured more than once Job category (cases/100 FTEs) HSWsc IDSOsd MRNse Notes: (a) Highly significant. (b) Significant. (c) Human services workers. (d) Intellectual disability services officers. (e) Mental retardation nurses. Intervention group (95% CI) 48.96 (40.6-58.0) 49.67 (43.3-56.5) 49.56 (44.4-55.0) 52.75 (42.6-63.9) 48.42 (40.7-56.7) 49.74 (40.5-59.9) 49.56 (44.4-55.0) 48.54 (42.0-55.6) 49.88 (42.0-58.4) 2,658.45 (n=12) 52.75 (n=4) 17.13 (14.2-20.4) 8.57 (6.5-10.9) 7.95 (5.9-10.2) 3.06 (1.9-4.5) 0.61 (0.2-1.3) 100.0 0.0 52.54 (42.3-63.9) 49.36 (42.2-57.1) 38.30 (28.5-49.6) Control group (95% CI) 81.69 (67.8-96.9) 85.69 (74.6-97.5) 84.83 (76.0-94.1) T-test p-value 0.0004a 8.45 90.79 (73.4-110.1) 83.52 (70.3-97.8) 83.56 (68.0-100.7) 84.83 (76.0-94.1) 8.51 79.80 (69.0-91.4) 87.61 (73.8-102.6) 11,354.33 (n=31) 103.63 (n=16) 30.30 (25.0-36.1) 7.27 (5.5-9.2) 7.88 (5.9-10.1) 11.51 (7.1-16.9) 9.09 (4.1-16.1) 1.84 94.0 6.0 6.44 107.8 (86.8-131.1) 81.5 (69.7-94.2) 75.17 (55.9-97.3) 2.80 5.32 3.94 0.0004a 0.0004a 0.0174b 0.0335b 0.0034b 0.0013a AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2003 VOL. 27 NO. 1 Illness Prevention Safety training for manual handling of people with disabilities employee also completes a workers’ compensation claim form, accompanied with a medical certificate and the injury notification form. The information is then forwarded to the agency’s insurer within 10 days.1 Cost effectiveness was estimated as the mean workers’ compensation payments of the control group minus that of the intervention group divided by the estimated training cost per trainee. As the numerator does not include indirect benefits, the costeffectiveness ratios should be considered the conservative minimum estimate. Indirect costs are real costs and they increase as direct costs increase, but are subjective and difficult to measure.29 Some indirect costs have been estimated at between 20% and 30% of premium payments.29 It should be noted that not all the difference in claims cost between the control and intervention groups are due to training alone. The incident data were entered and analysed using SAS. T-tests were computed with 95% confidence intervals (CI). Results Details of the risk variations for manual handling injury between the intervention and control groups are presented in Table 2. Training in manual handling methods significantly reduced the risk by as much as 42%, with an average rate of 49.6 per 100 FTEs (95% CI 44.4-55.0) among the intervention group compared with 84.8 per 100 FTEs (95% CI 76.0-94.1) among controls. The risk differential was consistent across gender, age group, length of service, and job classification. However, in two injury categories (client lift/transfer and general manual handling), the intervention group had a marginally higher risk. The control group had an average claims cost of $11,354, 4.2 times that of the intervention group ($2,658). Manual handling training was highly cost effective, with an estimated return of $5.80 per dollar invested. Discussion This paper has attempted to evaluate the impact of training in manual handling techniques on injury rates from manual handling in two comparison groups of disability services workers in an Australian State Government agency. The training was found to be effective, with the intervention group having a significantly lower risk for injury than the control group. The large differences in workers’ compensation claims cost between the two groups also indicate that the training was a good investment. Three major weaknesses are evident in this paper. One concerns the use of post-test design with a non-randomised control group. An experimental design provides the strongest evidence of a causal link between a specific intervention and outcome primarily because of an unbiased random allocation of people or workplaces to intervention and control groups.12 The principal reason for not using randomisation relates to the requirements of the Occupational Health and Safety (Manual Handling) Regulations 1999, which came into operation in the State on 1 July 1999, with the objective of reducing the number and severity of 2003 VOL. 27 NO. 1 musculoskeletal disorders associated with manual handling. An employer must ensure that any task undertaken by an employee involving manual handling of live persons or loads is identified and a risk assessment carried out to determine whether there is any risk of a musculoskeletal disorder affecting an employee as a result of that task. Where there is a risk, an employer is required to either eliminate the risk or reduce it by altering the workplace environment, changing the systems of work or provide training in safe manual handling techniques. Thus, random assignment of workers to intervention and control groups could constitute a potential breach of the regulations. A second limitation is that because an after-only design was used, the pre-intervention injury experience of the intervention and control groups was not taken into account. Injury records before July 1998 were checked to establish differences in manual handling injury rates between the two groups. However, as most (over half) of the workers were either new, with only a few months on the job, reliable rates could not be generated. The disadvantage of not obtaining ‘before’ measurements is that it is not possible to know if the groups differed initially with respect to the outcome measure. Another weakness is that the injury rates did not account for whether the training changed the practices targeted, that is, correct and safe manual handling. For instance, if injury rates fell post intervention while unsafe manual handling practices remained unchanged, it raises the possibility that the reduction in injury rates could be due to other factors. If, however, a decrease occurred in both injury rates and unsafe practices, it could be assumed that the improvement was associated with the intervention. Could the results of the study be due to factors other than the intervention itself? Alternative interpretations of the results can be attempted by examining a number of threats to internal validity. These include selection threat, regression-to-the-mean, testing threat, drop-out threat, selection interaction, and diffusion threat.12 The more these threats are present in an evaluation the less likely it is that the results are actually due to the intervention. The possibility of a selection threat was minimised by closely matching control group members to the intervention group on important characteristics, such as demographics and job classification. However, lack of data precluded using other relevant matching variables. The credibility of evaluation findings may be threatened by selection interaction effects, that is, dissimilarity occurring in the characteristics of the groups over the duration of the analysis. Examples of sources of dissimilarity include history, incident reporting, and maturation. Common history threats such as changes in supervisor or management personnel, worker-management relations, legislation, structure or pace of work were not evident. There was little evidence of under-reporting of manual handling injuries. Robson et al.12 have suggested using the ratio of major to minor injuries to help detect reporting threat with regard to injury statistics. The greater the suppression of reporting, the higher the ratio will be since minor injuries are easier to underreport than major injuries, that is, those involving compensation AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH Ore Article payments. If the ratio is constant, then the likelihood of a reporting threat is reduced.12 The ratio of major injuries to minor injuries (not involving compensation payments) for the first half of the period of analysis was similar to that for the second half (approximately 1:11) for both the control and intervention groups. A regression-to-the-mean threat to internal validity occurs if an intervention is delivered to a group of workers with high injury rates and the rates for the group are compared with groups with lower injury rates. Without intervention, there is a tendency for the high injury group to have a decrease in injury rates. This is mainly because of the tendency for workplace injury rates to sometimes fluctuate from year to year, with the effect that any group with a lower-than-usual injury rate in any given year is more likely to have their rates increase than decrease in the following year, assuming workplace conditions remain unchanged. Similarly, any group with a higher-than-usual injury rate in any given year is more likely to have their rates decrease than increase in the subsequent year.12 Because selection for the manual handling training is not systematic across the organisation, there is little evidence to indicate that particular workers were nominated for the training because of their elevated risk. A drop-out threat occurs if one group has a greater quit rate than the other, particularly if it results in the control and intervention groups having different characteristics, particularly in age, gender, and service length. Membership of both groups remained stable over the one-year period of analysis. There were only five known cases of turnover (two in the control group and three in the intervention group) and these were insufficient to significantly change the results. A contamination effect, alternatively known as a diffusion effect, relates to circumstances in which the intervention provided diffuses to the control group. Given the tendency for workers to share information with one another through learning and observation, the present study was unable to ascertain the extent to which this threat was present. One way to minimise the possibility of a diffusion threat is by keeping the control group separate from the intervention group, but this was rarely the case in the setting. A testing threat results when the practice of taking the before measurement could affect the safety outcome used to evaluate an intervention. This threat is only a problem where outcomes are assessed via questionnaires, interviews or observation to determine employee safety knowledge, practices and attitudes. Because injury outcomes were measured only post intervention without interaction with the workers, a testing threat had no influence on the results. Where an intervention is targeted at reducing workplace injuries, as in the current setting, the ideal outcome measurement is an unbiased measure of injuries. One problem with this is that the frequency of lost time events in a given month or year in most workplaces may be too low to show clearly the effect of an intervention.12 There was little evidence that this was the case in this study. With an average of more than 900 manual handling injuries per year,1 the frequency of lost time events was sufficient to establish the impact of the manual handling training. Misclassification errors can lead to problems in injury statistics.30 This may help explain the slightly higher injury rates (but not statistically significant) in the intervention group than the control group in only two injury categories, client lift/transfer and general manual handling. Given the consistently higher injury rates among the control group in the remaining parameters, there is little basis for assuming that the training was unsuccessful. The analysis of manual handling injuries among the workers in the current setting by Ore1 found that there was limited information on injury description, job classification, and time of incidents. One-fifth of the 2,714 injuries identified over a three-year period were classif ied as general manual handling. Misclassification can be attributed to several sources, including disagreement or misinterpretation of incidents on the part of individual coders and the classification scheme itself.12 The International Classification of Diseases, 9th revision (ICD-9) has been shown to perform poorly for soft-tissue disorders, as several different ICD-9 codes can be assigned for the same condition.30 The ability to code incident descriptions consistently has been reported to depend on which aspect of the incident is being coded (Robson et al. 2001). For two trained people coding the same 100 injury reports using a standardised coding scheme, the degree of agreement in coding was 44% for event precipitating injury and 61% for contributory factors, well below the reliability threshold of 70%.12 By contrast, the agreement on gender was 98%, type of injury 92%, part of body injured 92%, and agent of injury 79%.12 Conclusion Training in manual handling techniques was successful in reducing the risk for injury among the intervention group of disability services workers in the Australian State Government agency. While the study has some weaknesses, overall, there is little evidence to suggest that the outcome observed was due to threats to internal validity. It is expected that the findings can encourage further evaluation studies of workplace safety training in particular and occupational injury intervention programs more generally.

Journal

Australian and New Zealand Journal of Public Health – Wiley

Published: Feb 1, 2003

There are no references for this article.