Assessing Occupational Safety
and Health Training

With Technical Assistance from:
Raymond Sinclair
Jerry Newman
Ronald Schuler

June 1998

Mention of company names or products does not constitute endorsement by the National Institute for Occupational Safety and Health (NIOSH), Centers for Disease Control and Prevention (CDC).

Copies of this and other NIOSH documents are available from the
National Institute for Occupational Safety and Health
Publications Dissemination
4676 Columbia Parkway
Cincinnati, Ohio 45226-1998

To receive information on other occupational safety and health issues, call
1-800-35-NIOSH (1-800-356-4674), or
visit NIOSH Home Page on the World Wide Web at
http://www.cdc.gov/niosh
Email: pubstaft@cdc.gov

Occupational safety and health training remains a fundamental element in workplace hazard control programs. As training objectives, recognition of job hazards, learning safe work practices and appreciating other preventive measures are expected to contribute to the goal of reducing occupational risk of injury and disease. This report reviews data found in the literature reflecting the significance of training in meeting these kinds of objectives and outcomes. As will be seen, there is much positive evidence but the results seem very selective and highly qualified. An analysis to identify factors underlying a successful training experience is also presented and does confirm basic principles of learning. Here too, however, important gaps are noted in the available data. The document proceeds to offer an agenda for addressing outstanding needs and ways for strengthening the role that training can play in improving workplace safety and health. The reader audience concerned with these kinds of issues should find this report to be most informative.

Linda Rosenstock, M.D., M.P.H.
Director, NIOSH

Abstract

More than 100 Occupational Safety and Health Administration (OSHA) standards for hazard control in the workplace contain requirements for training aimed at reducing risk factors for injury or disease; others limit certain jobs to persons deemed competent by virtue of special training. A literature review was undertaken to assess the merits of such training rules to achieve this objective and to sort out factors of consequence. The review focussed heavily on published reports, primarily drawn from the period 1980 through 1996, wherein training was used as an intervention effort to reduce risk of work-related injury and disease. Eighty (80) such reports were found and gave overwhelming evidence to show the merits of training in increasing worker knowledge of job hazards, and in effecting safer work practices and other positive actions in a wide array of worksites. Reports from select surveys and investigations of worker injuries and workplace fatalities were also accessed with many implicating lack of training as a contributing factor to the mishaps. In still other studies, workplace training devoted to first aid instruction showed linkage to reduced worker injury rates, suggesting that even this kind of training has benefits to job safety overall.

A critical analysis of the above findings found certain qualifications in viewing training impacts and successes with regard to current workplace standards. For example, most of the reported training intervention studies did not address OSHA training rules per se, and knowledge gain and safe behavior measures were used in many evaluations as opposed to actual injury/disease indicators. Also, in some instances, the training was coupled with other forms of intervention to make attribution difficult. Training deficits noted in some surveys of work injury cases lacked for confirmation and no information was available on the quality of the instruction if given at all.

Despite the above reservations and uncertainties, training's role as a necessary element in developing and maintaining effective hazard control activities remained firmly supported by the available literature. What did emerge from this review and analysis was an appreciation of meaningful training procedures and the recognition of factors both within and beyond the training process that could greatly affect its impact. In this regard, the OSHA voluntary training guidelines were described along with illustrations from the reports to show how the various steps contained within them can be met in realistic ways and have merit in framing an effective program. In addition, factors both within and beyond the training process were assessed for their effects on training outcomes based on data found in the reviewed literature. Variables such as size of training group, length/frequency of training, manner of instruction, and trainer credentials were each shown to be significant determinants to the training process. Equally important were extra-training factors such as goal setting, feedback and motivational incentives along with managerial actions to promote the transfer of learning to the jobsite.

Based on the literature review, follow-on efforts to address outstanding issues and needs regarding effective occupational safety and health training were noted.

Back to the Table of Contents

Executive Summary This review sought evidence from the literature bearing on two questions: Are occupational safety and health training (OS&H) requirements, as cited in many Federal standards governing workplace conditions and operations, effective in reducing work related injury and illness? Does the available evidence show certain training factors or practices to be more important than others in having positive effects on these outcome measures? The literature search focussed on reports of training intervention efforts designed in whole or in part to enhance worker knowledge of workplace hazards, effect behavior changes to ensure compliance with safe work practices, or prompt other actions aimed at reducing the risk of occupational injury or disease. Eighty (80) such reports met criteria for inclusion. They were products of two literature searches. The first search covered the period up to 1993; the second extended the first search through 1996. The included work came mainly from the period 1980 through 1996 and, by intent, addressed five types of hazardous agents. These were: traumatic injury forces, toxic chemicals/materials, harmful physical factors, ergonomic stressors, and biologic/infectious agents as encountered in an array of work settings. The search also examined data from select surveys and investigative reports where training factors were either implicated in the etiology of workplace injury or disease incidents or, alternatively, were considered a key element to the success of worksite hazard control programs that showed exemplary safety and health records. Still another source for information was reports of worksite training directed to other needs (e.g., first aid) but that had apparent positive effects on worksite safety and health indicators as well. With regard to answering the first question, the literature accessed and reviewed in this report offered much direct and indirect evidence to show the benefits of training in establishing safe and healthful working conditions. The intervention studies in particular were especially supportive. Findings here were near unanimous in showing how training can attain objectives such as increased hazard awareness among the workers at risk, knowledge of and adoption of safe work practices, and other actions that improve workplace safety and health protection. Data from other types of studies suggested too that lack of required training may have contributed to events where workers were injured or killed. Although affirming the effectiveness of training to meet hazard control objectives, this review also drew attention to some shortcomings in the supportive data. For example, the training intervention work that offered the most positive evidence did not address OSHA training requirements per se. Rather, the training interventions targeted site-specific problems, and while showing success in resolving such problems, i.e., improving safety performance, their exact relationship to OSHA mandated training rules was unclear. Moreover, where studies reported lower injury rates, reduced lost time or medical costs after training, analyses were lacking to show how the improvements could be accounted for by the positive results from training when measured in terms of knowledge gain or behavior indicators. One study that attempted such an analysis found that the targeted training could only account for 25% of the observed reduction in injury rate. Also, in some instances, the training was coupled to other forms of intervention (engineering, ergonomic) so as to make attribution even more difficult. Still another tempering fact was that successful training results appeared greatly influenced by "extra-training" considerations. Management's role/support of safety training and its transfer to the jobsite, setting goals and providing feedback to motivate use of the knowledge gained, and offering incentives or rewards for reinforcing safe performance all seemed crucial to attaining a positive result. These types of factors are not acknowledged in OSHA training requirements. It is noteworthy that a proposed OSHA safety and health program standard does recognize the need for management actions to support OS&H training among other critical components. Some reports that suggested training deficits as factors contributing to injury/health problems lacked for confirmatory information. In other reports, workers who had received training to protect them against certain job hazards were nevertheless afflicted. Missing in these cases was information about the quality of the training offered; whether it met OSHA requirements, or took note of any of the extra-training factors noted above. Even with the above reservations and uncertainties, however, the role of training as a necessary element in developing and maintaining effective hazard control activities appeared firm. Indeed, the issue was not so much whether OS&H training could make a difference in reducing risks from workplace hazards. Clearly, it can. But rather ascertaining the conditions for maximizing these training effects. This was the second question posed in this review and two approaches were used to offer a response. One effort focussed on the OSHA voluntary training guidelines wherein examples were extracted from the set of intervention studies to highlight the different steps that had to be taken. Various exhibits in this exercise showed how the steps could be met in realistic ways and could have merit in framing and implementing an effective training program. A second approach considered factors not covered in the OSHA guidelines but more specific to the training process itself or the attainment of its objectives. Variables included were the size of training group, length and/or frequency of training, manner of instruction, trainer credentials, and training/transfer conditions. Addressed among the latter conditions were extra-training factors such as management/supervisory roles and motivational techniques for reinforcing the learning at the jobsite. The intervention studies were examined for data that could justify statements as to the significance of these factors and/or conditions and where they had the greatest potential for effecting successful training outcomes. A set of statements, tying together evaluative information from the different reports, emerged from this exercise. Some were more supportable than others owing to limitations in the data contained in the reports under review. Most statements about specific factors agreed with concepts in the general learning and motivation literature, i.e., increases in training time per unit group of workers or use of more frequent and shorter sessions suggested more favorable outcomes as did active learning experiences stressing jobsite applications. Some unique extensions or refinements were also noted such as the added benefits of having supervisors or foremen assume a more active trainer role in workplace safety and health training as opposed to others who may have that special responsibility. Particular attention was drawn to the length and frequency of training because of its implications for training schedules, both for initial and for refresher instruction. Needs to develop a decision logic for this purpose were mentioned along with some of the variables to be addressed. Suggestions for follow-on work were offered to take account of the gaps in the reviewed literature or other shortcomings in responding to the two questions posed. Among those recommended were: 1) Undertaking studies to ascertain how industry is responding to OSHA training rules and the quality of such efforts. The major data set used in this literature review were researcher directed efforts and, for that reason were not the norm. Focussing the efforts on the most prevalent types of injuries and illnesses and selecting industries or work operations where they are most recurrent would be ideal. Differences in how the mandated training rules were met at the various selected sites and links between the training undertaken and specific injury and disease risk factors would be analyzed. The extent to which the operant practices followed OSHA training guidelines, and the resulting experiences could offer an important reference in gauging their utility. 2) Conducting in-depth studies of training practices and their interrelationship with other elements in an establishment's hazard control program. Directing this effort at companies showing exemplary safety and health records could offer program models for effective training that can best complement or enhance other workplace measures aimed at maximizing risk management. 3) Using case-control or cohort studies to compare differences in the level of training of workers injured or afflicted by occupational disease against those not so affected. The intent here would be to get a better assessment of how training deficits can lead to such problems. Such an analysis would require measures to separate out many nontraining factors that could also be responsible for apparent differences in these cases. 4) Convene workshops to discuss issues concerned with the effectiveness of worksite OS&H training both now and in the future. Invitees would include experts and practitioners conversant with OS&H training, job skills training, health education, organizational behavior and evaluation subject areas. The workshops would seek to pool ideas bearing on the questions posed in this report and added concerns such as the adequacy of current regulatory language in OSHA training rules, future training challenges due to changing workplace technologies, worker demographics, measurement outcomes for assessing the effectiveness of training, the merit of merging different workplace training domains (i.e., OS&H training, job skills training, worksite health promotion), and other issues.

Back to the Table of Contents

Table of Contents

Abstract

I.     Introduction
II.   Occupational Safety and Health Training in Broad Perspective

A.  Basic Program Thrusts
B.  Worksite Training on Health Protection/Health Promotion

III.  Occupational Safety and Health Training in Relation to Other Worksite Activities
IV.  General Training Considerations

A.  Some Definitions
B.   Critical Training Elements

1.  Needs Assessment
2.  Establishing Training Objectives
3.  Specifying Training Content and Media
4.  Accounting for Individual Differences
5.  Specifying Learning Conditions
6.  Evaluating Training
7.  Revising the Training

V.   Occupational Safety and Health Training Rules as Found in Current Workplace Standards

A.  Nature of Existing OSHA Training Rules

1.  Content
2.  Frequency/Duration
3.  Documentation/Assurance
4.  Trainer Qualifications and Specialty Training
5.  Methods

B.  OSHA Voluntary Training Guidelines

1.  Determining If Training Is Needed
2.  Identifying Training Needs
3.  Identifying Goals and Objectives
4.  Developing Learning Activities
5.  Conducting the Training
6.  Evaluating Program Effectiveness
7.  Improving the Program

C.  Summary

VI.  Data on Effectiveness of Occupational Safety/Health Training

A.  Nature of the Literature Review
B.  Data Sources Tapped
C.  Training Intervention Studies

1.  Data Set and General Observations
2.  Intervention Data on OSHA Requirements/Guidelines

D.  Training Data from Studies of Afflicted Workers

1.  Bureau of Labor Statistics Work Injury Reports
2.  Investigative Studies Acknowledging Training Needs and Effectiveness
3.  Appraisal of Training Data on Afflicted Workers

E.  Training Factors in the Context of Other Worksite Programs

1.  Training in Successful Occupational Safety and Health Programs
2.  Implications of First Aid Training
3.  Worksite Physical Fitness/Exercise Training

VII. Verifying Critical Elements/Program Factors for Effective Occupational Safety and Health Training

A.  Examples Supporting OSHA Voluntary Training Guidelines
B.  Evidence for Distinctive Training and Extra-Training Factors
VIII. Summary

A.  General Conclusions on Training Effectiveness
B.  General Conclusions on Critical Training Factors
C.  Needs for Follow-on Work
IX. References


APPENDIX A:


Training Intervention Studies as Found in the Literature
Addressing Various Types of Occupational Hazards

APPENDIX B:


Extracts from Bureau of Labor Statistics (BLS) Work Injury Reports
for Discerning Real/Possible Gaps in Job Safety/Health Training

I

Introduction
azard control and prevention strategies to assure every working man and woman safe, healthful workplace conditions regard training as an axiomatic part of all such efforts. To reinforce this point, requirements for worker safety and health training are found in more than 100 occupational safety and health standards promulgated by the Occupational Safety and Health Administration (OSHA); others limit certain jobs assignments to persons judged competent by way of special training. Yet, the merits of training provisions in worksite hazard control regulations and for hazard control efforts in general have not been without question. Indeed, instances where training has been shown to be ineffective in reducing work related disabilities have been reported (Tan et al., 1991, Snook, Campanelli & Hart, 1978, Linnemann, et al., 1991), and at least one review has raised concerns about the worth of workplace safety training programs (Hale, 1984). In its defense, training shortcomings could reflect use of inadequate instructional techniques or situational factors that confound the learning process or its objectives. More importantly, however, the notion that training is somehow exempt from the accountability demands of business operations is no longer tenable. Moreover, in appreciating that adherence to training rules imposes added costs on responsible employers and extra burdens on workers too, knowledge of factors that can influence success in training efforts would be especially important. Indeed, one could argue that faulty or bad training may have worse consequences than no training at all.

It was these issues that prompted this literature review. The originally conceived purpose was to address two questions:

Q-1: Can the occupational safety and health (OS&H) training requirements cited in numerous Federal standards governing workplace conditions and operations be shown to be truly effective in reducing work related injuries and illness?
Q-2: Does the evidence single out certain elements or characteristics of training programs as having the most positive effects on worker safety and health?

Approaches to answering these questions would seem straightforward. Among the more obvious:

(1) Search the literature for reports on how employers implemented applicable OSHA training requirements and the associated change in their work injury and disease experience.
(2) Locate data sources comparing the extent and nature of OS&H training given workers who incurred work related injury or disease with those who have not but who are in similar occupations or work situations.
(3) Analyze the safety/health training practices of employers having exemplary injury/disease control programs and contrast them with others having worse performance but comparable job operations, workforce size/makeup, and other workplace factors.

As noted in the course of this review, the existing literature only partially satisfies these approaches. For example, reports of efforts to comply with prescribed OSHA training rules are relatively few; more common in the literature are training studies targeting site-specific problems that may or may not be covered by regulation. Knowledge gain, attitude change, indications of safer work procedures being adopted by the workers are most frequently used to measure the training impact. Fewer studies note reduced injuries or disease outcomes from training, and if they do, there are cautions about drawing cause-effect conclusions. Results of epidemiology studies of occupational injury and disease may suggest training shortcomings as contributing factors but lack for baseline data or other evidence needed to support such claims. Assessments of OS&H training practices or the merits of certain instructional methods as found in the literature may not always be able to separate training-specific effects from other workplace factors. Also, in many instances, data collection for worksite training evaluation purposes may not always follow sound study design principles.

Clearly, then, the current literature cannot supply complete answers to the questions posed. Definitive responses require more research and some suggestions for such work are described at the conclusion of this report. Even with the above reservations, the benefits of a literature review at this time are believed twofold. First, to offer a preliminary appraisal of the merits of training directed to workplace safety and health concerns. Second, to offer a conceptual framework for treating OS&H training so as to clarify the issues that should be addressed in follow-on efforts.

Although the literature has much theory on learning having relevance to training, this review is not theory driven, i.e., intended to serve as a test of specific hypotheses derived from specific theoretical formulations. As will be noted, the treatment in various places does take account of suggested training models or guidelines for purposes of organizing and analyzing the material under review.

It is envisioned that this report can serve different audiences. For one, policy-makers engaged in setting OS&H standards need to be assured that training requirements are essential to achieving improved workplace safety and health. Finding such supportive evidence for training rules would offer this assurance. Also, information on whether certain rule formulations have more merit than others in producing successful training results would be of benefit to policy-makers. A second audience would be researchers. This review reveals both strengths and weaknesses in the database for effective training procedures and factors affecting OS&H training. Especially challenging and needed are studies to definitively tie immediate training results, e.g., increased knowledge of hazards and safer work behaviors, to outcome indicators such as reduced worker injuries and illness. A third and last audience for this report would be those who conduct OS&H training. Indeed, the wealth of training studies appended to this report is an information resource on instructional methods and evaluation techniques. By addressing different types of hazards in a variety of work settings and offering measures for indicating effects, they offer important lessons for OS&H trainers.

Back to the Table of Contents

II

Occupational Safety and Health Training in
Broad Perspective

A. Basic Program Thrusts

OS&H training embodies instructing workers in recognizing known hazards and using available methods for protection. Worker education, in contrast, prepares one to deal with potential hazards or unforeseen problems; guidance is given in ways to become better informed and to seek actions aimed at eliminating the hazard. As explained in a 1985 Office of Technology Assessment (OTA) report on preventing illness and injury in the workplace, the distinction between worker training and education programs is often blurred and depends on the role that the worker is expected to assume in the process. "The narrower the role, the more the instruction is training; the broader the role, the more the instruction is education" (OTA, 1985, Pg.189). Much of the information in this review draws on the narrower training reference, but worker education approaches are also acknowledged.

The OTA report included an analysis of 40 worker training and worker education programs conducted by business firms, trade associations, unions, hospitals and universities, and coalitions of OS&H groups. The analysis of activities and objectives suggested four types of programs: fundamental, recognition, problem-solving, and empowerment programs.

Fundamentals Programs: These programs involve instruction in prevention of work-related injury and illness through proper use and maintenance of tools, equipment, materials; knowledge of emergency procedures; personal hygiene measures; needs for medical monitoring; and use of personal protective equipment for non-routine operations or as an interim safeguard until engineering controls can be implemented.

Training interventions having these objectives permeate the literature and comprise most of the reports in Appendix A, which offers an inventory of studies aimed at evaluating the effectiveness of OS&H training for controlling workplace hazards.

Recognition Programs: These programs include instruction emphasizing awareness of workplace hazards; knowledge of methods of hazard elimination or control; understanding right-to-know laws and ways for collecting information on workplace hazards; recognizing symptoms of toxic exposures; and observing and reporting hazards or potential hazards to appropriate bodies. Training activities of this type were spurred largely by the OSHA Hazard Communication Standard (OSHA, 1983). The standard required employers to inform workers of chemical hazards found in their work areas and of ways to reduce apparent risk. Use of Material Safety Data Sheets and labels along with training are the three means prescribed for communicating the essential information. Several reports evaluating training efforts of this type are found in Appendix A.

Problem-Solving Programs: Instruction is aimed at giving workers the information and skills enabling them to participate in hazard recognition and control activities; to help identify/solve problems through teamwork, to use union and management means, and to exercise rights to have outside agencies investigate workplace hazards when warranted. Inviting worker input in company planning or in design of new operations or processes is recognized as a viable means for improving productivity, quality of products, and worker motivation. Extending this approach to hazard control seems reasonable especially since workers, owing to their everyday job work experience, possess an intimate knowledge of the hazards connected with their jobs and could be a rich source for corrective ideas. One report in assessing a worker participation approach to hazard control at a worksite (Lin & Cohen, 1983) found the overall effort to be successful but at the same time took note of some gaps of worker knowledge for which added worker instruction would have been of benefit. This report is listed in Appendix A along with other reports of union-sponsored efforts to sharpen worker skills in addressing the requirements of specific OS&H standards so that they could play a more effective role in responding to their workplace needs.

Empowerment Programs: These programs provide instruction to build and broaden worker skills in hazard recognition and problem-solving skills much like that noted above. Emphasis, however, is on worker activism with the goal of ensuring their rights to an illness-and injury-free workplace (Wallerstein & Baker, 1994). Hence, the program aims at enabling workers to effect necessary control measures through educating co-workers and supervisors, and through use of committee processes or in health/safety contract negotiations. This approach is in accord with the current "Total Quality Management" philosophy - having rank-and-file workers along with their supervisors share greater roles in and be more accountable for addressing workplace hazard control needs. Several reports of union and university coalitions engaged in furnishing such training and some first attempts to provide outcome measures are noted in Appendix A.

The above types of training suggest a progression from a workforce learning basic forms of protection to known hazards, through instruction aimed at enhancing their awareness of potential problems and problem-solving skills, and then learning how to make it all happen in their workplaces. Although treated separately, any given training program may contain elements of these approaches in varying degrees.

B. Worksite Training on Health Protection/Health Promotion

The aforementioned training and education activities are all directed to worksite health protection, that is, to controlling occupational/environmental risk factors for disease or injury. They should not be confused with worksite health promotion programs that also involve training/education activities but whose objectives are to alter personal lifestyle factors that may pose risks to one's health and well-being. Instruction here targets smoking, substance abuse, inadequate diet, poor physical fitness among other problems and the intent is to effect behavior change for risk reduction.

Personal lifestyle and occupational risk factors may interact in ways that can heighten the potential for adverse outcomes. For example, asbestos workers who smoke may have a 10-fold greater risk for lung cancer (Hammond, et al., 1979); alcohol or illicit drug use has been implicated in work accidents in high risk jobs (Holcomb, Lehman & Simpson, 1993). Alternatively, exercise training for enhancing physical fitness has been suggested as an added means to limit strains from jobs imposing undue stress on the musculoskeletal system (Genaidy, Gupta, & Alshedi, 1990; Hilyer et al., 1990; Shi, 1993). For these reasons, training and education activities addressing worksite health protection and health protection goals in combination may have mutually reinforcing effects. While health promotion studies are outside the scope of this literature review, some reports of these worksite activities have been evaluated from the standpoint of reducing occupational risks and will be so noted.

Back to the Table of Contents

III

Occupational Safety and Health Training in
Relation to Other Worksite Activities

OS&H training as implemented at the workplace rarely has a stand-alone status. For example, OS&H training may be a natural part of job skills training or a simple add-on. On-the-job type of training, of necessity, would have to cover both objectives. In some cases, work methods to be learned and safe work practices prescribed by OSHA standards are much the same. In occupations such as logging, "poor technique" is widely reported as a cause of injury with inadequate training cited as the underlying problem. This shortcoming refers to a failure to develop (learn) proper work methods and the safety precautions that derive from them (Slappendel et al., 1993).

OS&H training is also an element of hazard control programming. Instruction in hazard recognition and control methods, knowledge of emergency procedures, and use of personal protective equipment may or may not be distinctive-the degree depends on what OSHA requirements may dictate. The Hazard Communication Standard (OSHA, 1983) for example, requires a written training plan describing the nature of instruction to ensure workers understand the chemical hazards to which they might be exposed, recognition of symptoms of overexposure, safeguards to be taken Other standards merely acknowledge the need for training but are less explicit as to requiring evidence of a plan for its implementation. Although not always recognized, OS&H training may also be needed to cover operational aspects of engineering or physical hazard control systems so that their benefits are realized to the fullest extent.

OS&H instruction as a component of both regular job training and worksite hazard control program practices is depicted in Figure 1 as nested in and representing the overlapping parts of these two types of activities. Encircling the job training and occupational safety and health program efforts are factors which can shape the success of these activities and their respective impacts on productivity and safety/health experience. Indeed, the transferability of training to actual jobsite demands, use of promotional or motivational efforts to drive the training exercise as well as boost post-training performance in both productivity and hazard control terms, and management's commitment to these activities as communicated to the workers can do much to affect these kinds of outcomes. Encompassing all of the above and shaping the ultimate results are an employer's goals and objectives. Presumably, management policies and actions that do not put company productivity and worker safety and health in conflict but rather stress their positive interrelationships would be ideal.

Viewing OS&H training in this context underscores the difficulty in attempts to treat or evaluate its effects separate from other workplace considerations. This is especially true if "bottom line" outcomes such as work related injuries and illness are used in the assessment. As will be noted, many evaluations of OS&H training use measures more immediate to the learning process itself (e.g., knowledge gained); others may take account of the instruction plus certain extra- or post-training factors in assessing on-the-job safety performance. In these instances, surrogate indicators for injury or disease reduction may be used (e.g., increased compliance with safe work practices and/or associated changes in exposure levels); these indicators can offer more opportunity for discerning effects than injury or illness measures. It must be remembered that work related injury or disease cases at least at the individual employer level are rare events which can complicate efforts at evaluating change.



Figure 1. Depiction of occupational safety/health (OSH) training as nested in other employer programs which are affected by numerous extra- or post-training factors. Interactions here complicate efforts to assess training-specific impacts on safety/health experience or productivity/performance measures. Employer goals and objectives and related policies ultimately determine the priorities or trade-offs between the two outcomes.

Back to the Table of Contents

IV

General Training Considerations

A. Some Definitions

In general, training refers to instruction and practice for acquiring skills and knowledge of rules, concepts, or attitudes necessary to function effectively in specified task situations. With regard to OS&H, training can consist of instruction in hazard recognition and control measures, learning safe work practices and proper use of personal protective equipment, and acquiring knowledge of emergency procedures and preventive actions. As noted in the OTA report (1985), training could also provide workers with ways to obtain added information about potential hazards and their control; they could gain skills to assume a more active role in implementing hazard control programs or to effect organizational changes that would enhance worksite protection.

Performance represents observable actions or behaviors reflecting the knowledge or skill acquired from training to meet a task demand. With regard to OS&H, performance can mean signs of complying with safe work practices, using protective equipment as prescribed, demonstrating increased awareness of hazards by reporting unsafe conditions to prompt corrective efforts, and executing emergency procedures should such events occur.

Motivation refers to processes or conditions that can energize and direct a person's behaviors in ways intended to gain rewards or satisfy needs. Setting goals for performance coincident with learning objectives and use of feed-back to note progress have motivational value. With regard to OS&H, motivation can mean one's readiness to adopt or exhibit safe behaviors, take precautions, or carry out self-protective actions as instructed. Bonuses, prizes, or special recognition can act as motivational incentives or rewards in eliciting as well as reinforcing these behaviors when they are displayed.

Knowledge or skills acquired in training may not always result in improved performance in actual work situations. This may indicate 1) lack of suitable motivation, 2) training content does not fit job demands (i.e., a problem in defining suitable training objectives, or 3) dissimilarity or conflicts between the instruction/practice in training conditions when compared to actual job conditions (i.e., a problem in transfer of training). More is said about this in the sections to follow.

B. Critical Training Elements

Different authoritative reviews of the general training literature (Goldstein & Buxton (1982), Campbell (1988), Tannenbaum & Yukl (1992)), and job training in particular, emphasize the importance of certain elements as critical to an effective program. They are noted below. The OSHA set of voluntary training guidelines to assist employers in furnishing safety and health information and instruction to workers (OSHA, 1988) mimics most of the same elements within an OS&H context. They are treated in a later section.

1.       Needs Assessment

According to the general literature, training goals presuppose: 1) consistency with organizational goals, 2) the presence of jobs designed to yield performance outputs that meet the organization's goals, and 3) performance levels dependent on knowledge of the job tasks, skill, attentiveness to the work or factors where training can make a difference. On the last point, expecting training to solve problems related to internal organizational conflicts or to overcome deficiencies in equipment or work methods is unrealistic. Job analyses determine which of the relevant performance factors comprise the highest priority training needs either now or in the future. The process includes defining the tasks involved, their order of importance (in terms of frequency, criticality, complexity), and details of the steps necessary to accomplish them.

2.       Establishing Training Objectives
According to the general training literature, the needs assessment provides the information to establish the objectives of the training program. These are stated as observable behaviors expected of the trainee after the instruction, and they may acknowledge the conditions under which they should be performed and the required level of proficiency.

3.       Specifying Training Content and Media
According to the general training literature, content represents the knowledge or skill that the trainee must master to be able to meet the behavioral objectives. The judgement of those who know the job demands is the most common approach to specifying training contents. Other approaches may be the products of problem-solving exercises, or be based on mistakes people make in using a skill such as to design corrective learning measures. Evidence that one teaching method such as lectures, televised instruction, computer-aided instruction, or interactive video methods is superior to another is not that clear (Kearsley, 1991). Much depends on the specific training needs, makeup of trainee group and other factors. Why or how a particular method facilitates learning and how it can be made more effective are issues requiring further study.

4.       Accounting for Individual Differences
According to the general literature, effective training should take account of the characteristics or attributes of the trainees. Aside from differences in aptitude, literacy, or pretraining skill levels, how trainees view the training program in terms of improving their job performance or self-efficacy may dictate variable approaches. The kind and level of training for new job applicants versus long-term or older workers reassigned to the same tasks also has to be addressed.

5.       Specifying Learning Conditions
In general, instructional events comprising the training method should not inhibit, conflict with, or be unrelated to the processes that lead to mastery. If the learning is to develop capabilities in problem-solving techniques, the instructional approach should stress thinking/reasoning approaches not rote memorization. Training methods should require the trainee to use the training content in active or productive ways, e.g., restating or applying principles rather than just recalling them, or adapting the information to new situations rather than mere repetition in the same one. The current literature suggests that using learning events that require productive behavior or that provide appropriate feedback (positive/accurate/credible) and opportunities for practice under conditions that promote transfer to the actual job are ideal.

6.       Evaluating Training
According to Kirkpatrick (1967), training evaluations in the general literature can take four forms which are viewed as a series of steps or levels. They are:

Step #1: Reaction—How did the trainees like the program? Typically this is done through evaluation sheets completed at the end of the training. Typical items inquire as to whether the material was well organized, relevant to the trainees needs, made interesting through the instructor's manner of presentation or use of visual aids, demonstrations, etc.

Step #2: Knowledge Gain (or Skills Acquired)—What principles, facts and techniques were learned? Knowledge of facts and principles is usually evaluated via pre/post paper-and-pencil tests or quizzes. Assessment of skills may be done through performance tests before and after training. An untrained or control group can be similarly tested to indicate any differences resulting from just the test-retest experience

Step #3: Behavior Change—What changes in behavior occurred as a result of the program? For this purpose, reports by the trainees themselves (self-appraisals) of their on-the-job performance, or observations by their peers, supervisors, instructors can be used. A time interval between the end of training and the observations may be necessary to allow for the training to be put into practice. Post-training measures taken at different time points are also suggested to determine if the training effect is sustained or needs refreshment. Again similar observations for a control group are recommended to acknowledge any effects from repeated testing. These control data also provide an added reference for gauging the significance of the apparent behavior changes in the training group.

Step #4: Results—What were the tangible results of the program in terms of its objectives or goals for the organization? Did it result in reduced injuries or illness, lower medical costs, improved productivity? As noted in Figure 1, extra- or post-training factors can affect these types of outcomes, and it is not always possible to design evaluations that can isolate the specific training contribution. Undertaking evaluations where these "extra-training factors" are held constant during the pre-and post stages of the training assessment or can be segregated as to their influence through use of suitable control groups are ideal. Needless to say, training impacts at the organization level can require an extended time line especially in using injury/illness outcomes owing to their infrequency.

Criteria for rating training effects are the focus of much discussion in the literature. Several points that deserve mention or added emphasis in light of the subject of this report are:

(1) Past surveys have shown that most in-house assessments of training programs measure only trainee reactions of how well they liked the instruction (Smeltzer, 1979; Smith, 1980; Parker, 1984, Alliger & Janak, 1989). Efforts to determine the extent to which the training content was absorbed or resulted in changes in actual on-the-job behaviors, or had impacts on organizational measures (e.g., quantity/quality of production, sales, absences/turnover, injury/illness rates) were rarer. Among reasons offered for the lack of more intensive efforts at evaluating training were the unquestioned beliefs that training works, that workplace conditions do not readily lend themselves to systematic assessments of training, and that more rigorous attempts will entail high costs. Increasingly, however, there is the call for more extensive training evaluations to verify the benefits as witness this exercise (Blomberg, et al., 1988).

(2) Reinforcing the above statement, trainee reactions to instruction may bear little relationship to the extent of actual learning. (Liking the instruction does not imply learning). Hence, it should not be used as the sole criterion to gauge effectiveness. Similarly, pre- and post-training quizzes or tests of skill showing the gains from instruction may or may not be related to improved on-the-job performance. Needs for multiple measures of effectiveness are apparent.

(3) As already noted in Figure 1, the effect of training is greatly affected by other workplace factors both in the training and post-training environment. Evaluations will need to account for these factors in terms of their influence on the training outcome.

(4) Simple performance outcome measures representing various levels of achievement may be critical to determining the validity of the instruction but may not indicate the factors that influenced these results. Provision of "process" measures, reflecting various amounts of training time, modes of training, trainer attitude/competency, can indicate why the overall results were or were not achieved. This can be important in efforts to revise the training to improve its efficacy.

7.       Revising the Training
The evaluation of training as noted by Goldstein and Buxton (1982) offers information as to whether the instruction has had its intended effect on the measures set out for that purpose. Seldom do the data indicate a program was a complete success or a failure, given multiple criteria for gauging the results. Rather, the data may indicate better understanding, retention or application of some course material as compared with others. Gaps or variations in knowledge or competencies resulting from the training may reflect needs to consider more training time, alternative instructional techniques, or more capable instructors.

Back to the Table of Contents

V

Occupational Safety and Health Training Rules as
Found in Current Workplace Standards

A. Nature of Existing OSHA Training Rules

Because of the agency's hazard-by-hazard approach to rule-making, OSHA training requirements in current safety and health standards number in the hundreds and vary greatly in nature. In 1992, OSHA excerpted and collected the various training provisions into a single report to ease the difficulty in locating them in the different standards. OSHA also contracted for a second report to identify and classify their common elements (Meridian Research, 1993). Underlying the latter effort is the thought of developing a stand-alone generic training standard offering possibilities for portraying the existing rules in a more integrated and easier-to-follow format. But this categorization aside, excerpts of the training requirements listed in the OSHA 1992 report show the wide variation in the manner and detail of the applicable training requirements - requirements befitting differences in the scope and complexity of the conditions addressed by the standards. The discussion below elaborates on the variable nature of the training rules with regard to factors such as content, frequency/duration, documentation/assurance, trainer qualifications, and methods used.

1.       Content

A number of standards are quite explicit about what safe practices should be taught. Training rules for pulpwood logging and materials handling operations are of this nature. For example, the pulpwood logging standard lists the details of chainsaw instruction. A sample item: "Chainsaw operators shall be instructed to start the saw at least 10 feet away from the fueling area" (29 CFR 1910.266(c)(5)(v)).¹ Similar exposition occurs in a materials handling standard dealing with the servicing of single rim wheels. In this case the instruction must cover safe work practices so as to ensure ". . . that tires shall be completely deflated by removal of the valve core before demounting; mounting and demounting of the tire shall only be done from the narrow ledge side of the wheel; tires shall not be inflated when any flat solid surface is in the trajectory and within one foot of the sidewall" (29 CFR 1910.177(g)(1)(2)(7). In contrast, other standards are more general as to the content of the training. For example, the powered industrial truck standard states: "Only trained and authorized operators shall be permitted to operate a powered industrial truck. Methods shall be devised to train operators in the safe operation of powered industrial trucks" (29 CFR 1910.178). Nothing more is specified. Still other standards acknowledge topics to be covered (e.g., recognition of hazardous conditions, risk factors and potential outcomes, needs for and means for hazard control) but do not go further, thus leaving the specific content up to the employer.

---

¹ Excerpts of the training standards cited in Chapter V are taken from the OSHA (1992) report. More complete statements are found in Collert (1996).

---

2.       Frequency/Duration

Standards covering exposures to toxic agents dictate that employees receive training before an initial job placement and repeat training on some periodic basis. Reflecting more explicit requirements, the Hazardous Waste Operations and Emergency Standard (29 CFR 1910.120(e)) indicates minimum durations for initial training offsite and supervised onsite instruction plus yearly refresher training. At the other extreme, a number of standards make no reference either to the duration of required training or to the need for repeat or follow-up instruction.

3.       Documentation/Assurance

Certain standards require evidence of a formal training plan or training materials, or both as well as written documents certifying successful completion of the necessary training. The standards covering powered platform operations (29 CFR 1910.66) and hazardous waste operations and emergency response (29 CFR 1910.120) contain such provisions. So too do various standards governing exposures to toxic chemicals and harmful dusts. The regulation on asbestos abatement work first drafted by the Environmental Protection Agency and adopted by OSHA (29 CFR 1915 1001(k)(9)) even requires written examinations of the trainees and attainment of a minimum score to assure competency. The language in most other standards is less specific about either the needs for a formalized program, records of employee participation or achievement. A number of standards indicate that employees shall demonstrate proficiency following instruction but do not actually require certification.

4.       Trainer Qualifications and Specialty Training

Some standards indicate teaching requirements for those slated to instruct employees or stipulate tasks to be undertaken by competent persons. Competent persons are defined as those having acquired necessary skills by virtue of attending training schools, holding academic degrees, or possessing specialty experience. Related to this point, the hazardous waste operations and emergency response standard acknowledges separate training requirements for waste site workers versus those having management or supervisor roles. The revised OSHA asbestos standard mentioned above (29 CFR 1915.1001(k)(9)) dictates different training for asbestos abatement workers, supervisors, inspectors, management planners, and project designers.

5.       Methods

With few exceptions, current standards do not dictate methods to be used in meeting the required training objectives. One standard on powered platform operations (29 CFR 1910.66)) mentions that pictorial methods can be used instead of written work procedures in the training activity. Several others stipulate a portion of the training time to be a "hands-on" learning experience. More generally, however, the "how to" of training is left to the discretion of the employer.

Foreign Occupational Safety and Health Training Rules

Heath in a series of four papers (1982(a), 1982(b), 1982(c), 1982(d) summarized OS&H training as practiced in six industrialized western nations which included Canada, Great Britain, Federal Republic of Germany, Sweden, France and the U.S.. Like the U.S., legislation requiring employers to provide information and instruction on workplace risks were also found in the other five countries. There was much similarity too in how rules on training content and informational materials pertaining to use of equipment, exposures to agents were developed by governmental agencies with inputs from health and safety professionals, trade associations, union and insurance groups. Heath noted two major differences between the U.S. and the other countries in the rule-making area. One was that foreign country legislation was far more precise than that in the U.S. in designating who was responsible for carrying out the training requirements in a work establishment. An excerpt from one country reads: "Every employer shall ensure the adequate direction and instruction of workers in the safe performance of their duties. Every supervisor shall be responsible for the proper instruction of workers under his direction or control and for ensuring that the work is performed without undue risk." (cited in Heath 1982(a), pg. 8). Second, the U.S. was the lone country providing central government funding to support job safety and health training programs. In the foreign countries, this burden was assumed by employers, insurance companies, private safety and health organizations. But more pertinent to the subject of this literature review, Heath found that the evidence in these countries to show that training was an effective countermeasure to reducing worksite injury and illness ranged from poor to nonexistent (Heath, 1982(b)).

B. OSHA Voluntary Training Guidelines

In the matter of how best to implement required training, OSHA has training guidelines to assist employers in furnishing safety and health information and instruction to workers (OSHA, 1988a). The guidelines are voluntary and are meant to enhance or supplement other employer training activities. Tailoring their application to meet individual worksite needs or local working conditions is encouraged.

The OSHA voluntary training guidelines follow a model whose elements reiterate most of those in the general job training literature reviewed earlier. The seven guidelines below makes this readily apparent.

1.       Determining If Training Is Needed

Are the needs for hazard control more readily solvable by training, i.e., increased knowledge of a work process or adoption of safe work practices as opposed to engineering or physical control alternatives?

2.       Identifying Training Needs

Job hazard analyses plus examinations of company health/safety records and worker perceptions of job risks are suggested as means for identifying what training is needed and where improvements can be made in hazard control. Obviously, reference to applicable federal/state standards will also shape the training content.

3.       Identifying Goals and Objectives

The OSHA guidelines call for identifying what the instruction is intended to achieve and defining evidence for it being met in explicit, observable terms. OSHA indicates that a specific objective (e.g., "An employee will be able to describe how a respirator works, how to ensure an effective fitting, and when it should be used" ) is preferable to a vague goal (e.g., "The employee will understand the use of a respirator")(Pg. 5, OSHA, 1992).

4.       Developing Learning Activities

The OSHA guidelines suggest learning activities be aimed at well-defined objectives and in substance take account of mental and/or physical skill factors as may be required to meet specified needs. The actual content or coverage of topics may be dictated by OSHA regulations. Instruction that employs task sequences and situations to simulate the actual job conditions are suggested to ensure the transfer of this training to the work situation. Like the general training literature, the OSHA guidelines acknowledge that training materials and techniques can vary; the important point is that the activities allow the employees to demonstrate that they have acquired the desired knowledge.

5.       Conducting the Training

This OSHA guideline, like the one in the general training literature, stresses the need for an instructional format that invites worker inputs into the training process, and provides for hands-on experiences and exercises promoting active learning. It also makes reference to other means of motivating and maintaining worker interest. Relating the training to their current skill levels and experiences and emphasizing the benefits (increased worker knowledge and skills, more marketable attributes as an employee who is informed and safety conscious) are among the ideas offered.

6.       Evaluating Program Effectiveness

Each program should determine whether the training has accomplished its goal. Trainee opinions, supervisor observations and workplace improvements resulting in reduced injury or illness are among the means recognized for this purpose. As already mentioned, incidents of illness/injuries for rating the impact of OS&H training programs or other intervention activities are such rare events that surrogate measures may be needed. Frequencies of "near miss" incidents, evidence of reduced exposure levels to a hazard, measures of compliance with safe work practices offer possibilities. Also, reduced injury and disease as outcomes of training would have to account for other factors as well. (See Figure 1).

7.       Improving the Program

If the evaluation proved that the training was deficient, efforts to revise aspects of the training or to offer periodic retraining may be in order. Repeating the steps in the training model may help determine where course revision is needed.

The OSHA voluntary training guidelines also contain suggestions for identifying those workers who may be at higher levels of risk and thus have the greatest need for training. Occupations posing known exposure hazards or otherwise shown to be associated with excess injury/illness are one determinant. The age and job service of the worker group in question can be another. (Young, new workers show a disproportionate number of injuries and illnesses.) Still another may be the size of the establishment. (Though the pattern may vary with industry, medium size companies (50 to 249 workers) tend to have higher incident rates than the rates for smaller or larger firms (Bureau of Labor Statistics, 1997).

C. Summary

Upon reflection, the following points summarize the current state of affairs regarding the regulatory language covering training in most OSHA standards.
(1) Requirements appear fragmented without a sense of a whole plan. Steps to realize training objectives are rarely given. (As the OSHA voluntary training guidelines are a recent development, older standards lacked for an adequate frame of reference). The most detailed requirements for training are found in the Hazardous Waste Operations and Emergency Response standard (1910.120) and the OSHA adoption of the EPA asbestos abatement work rule (40 CFR 763(e)Appendix (C). These provisions may be a prototype for addressing training needs in future rule formulations.
(2) Without explicit requirements for a plan embodying more of the elements to meet the training objectives cited in the standards, employers may opt for minimal efforts whose results are marginal at best.
(3) Perhaps most critical is the need to ascertain whether the OSHA voluntary plan has merit. Reports of field efforts to implement various aspects of the guidelines and the resultant indications of success or failure in terms of impacting workplace safety/health problems can offer evidence. Further, a review and analyses of this work could define factors that would realize the training objectives and improve safety/health outcomes. The balance of this review is devoted to this task.

Back to the Table of Contents

VI

VI. Data on Effectiveness of Occupational
Safety/Health Training

A. Nature of the Literature review

This review sought empirical information relevant to assessing whether OS&H training as dictated by OSHA standards had any beneficial effects in reducing the risk of work related injury and illness, and factors that were especially critical to successful training efforts. Specifically, the following types of documentation were sought:
(1) Reports of studies where training interventions were used for purposes of reducing apparent worker risks of workplace injury or disease and evaluative data obtained to indicate their significance. References took account of training programs undertaken at specific jobsites, laboratory studies with training simulators, and efforts to implement OSHA training rules by user/affected groups (i.e., unions/trades organizations) and indications of results.
(2) Surveys or investigative reports offering data on training (or the lack thereof), as well as other factors contributing to work related injuries, fatalities, and health problems.
(3) Reports on occupational safety and health program practices for employers having exemplary safety/health performance to isolate training factors that may have contributed to their success.
(4) Other studies in the education/learning field or ancillary areas that deal with issues especially pertinent to effective OS&H training.

Regarding items (1) and (4) above, the literature review was confined to those reports that described training plans and objectives, manner of implementation, methods for evaluation, and data/results reflecting effectiveness. Articles offering testimonials to various training approaches, or anecdotal references were not considered. Item (2) dealt with results from questionnaire surveys of persons having certain types of work injuries and investigative reports of workplace incidents where workers were injured, killed or found to exhibit health problems. The intent in both cases was to examine any references to training in appreciating why and how the injury producing event or health problems occurred. Item (3) sought analytical information on company program practices, especially aspects of worker training that were relatable to their success in hazard control.

B. Data Sources Tapped

The search strategy made use of NIOSHTIC, the National Institute for Occupational Safety and Health database which covers a broad range of OS&H literature, plus other on-line computerized reference systems and abstract listings for specialty areas available through DIALOGUE Information Services and the National Institute of Medicine MEDLARS system. Searches for relevant citations in specialty areas tapped databases in applied psychology (e.g., PsychScan, PsychINFO), education (e.g., ERIC), social science (e.g., Social Scisearch), health planning and administration (e.g., HEALTH/MEDLARS), trade/industry/occupation issues (e.g., AGRICOLA, Trade & Industry ASAP), specific occupational health topics/problems (e.g., CANCERLIT, TOXLINE, TOXLIT), among others.

A total of 25 different reference systems were scanned in the course of searching for documentation. The uniqueness of the indexing systems used with the various information sources coupled with the multidisciplinary nature and use of training concepts made a single list of keywords for searching impractical. In many instances, a customized list of database specific terms made up of a root or common word in combination with free-text search terms having proximal connection to the keyword had to be devised for the search. Some sample terms for one database (MEDLARS) included program-evaluation, risk-management, attitude-to-health, evaluation-studies, dangerous-behavior, inservice-training, patient compliance. Searches for titles through the various databases were conducted while constantly varying the mix of keywords and free-text in essentially a round-robin fashion. This yielded a listing of over 2000 citations which after deleting duplicates, eliminating false hits, was reduced to approximately 700 titles. Abstracts and references to these reports in other reviews found less than 150 to meet one of the four selection requirements noted earlier. The literature base for this review was the product of two searches. The one just described was concluded in 1993 and included reports published no later than that year. A second, subsequent search extended coverage through 1996 and yielded 188 additional titles. Screening abstracts of these reports found no more than 15 to be worthy of further consideration with regard to the purposes of this review.

The references drawn from the two search efforts were mostly from the period 1980 to 1996, though supplemented in some cases by earlier studies believed cogent to training evaluation issues. Final selection ensured representation in five different hazard agent categories which were:

•  Injury-producing forces,
•  Toxic chemicals or materials,
•  Harmful physical agents,
•  Ergonomic stressors,
•  Biologic/infectious agents.


The 5 agent conditions were chosen to reflect work related exposure risks recognized in various ways by current OSHA standards or those under development. (NOTE: Ergonomic agents were included in the data set in light of OSHA plans to develop a standard for controlling work related risks of musculoskeletal disorders (OSHA, 1992).

C. Training Intervention Studies

1.       Data Set and General Observations
APPENDIX A to this report summarizes 80 studies in which training, in whole or part, constituted an intervention effort aimed at hazard control. For this reason, these reports are viewed as offering the pivotal data to this assessment exercise. The studies are classified by five different agent/hazard conditions, namely, injury-producing hazards, chemical agent hazards, physical agent hazards, ergonomic agent hazards, and biologic agent hazards.

To varying degrees, the listed reports met the requirements stated above in describing training objectives, some form of training, efforts at implementation, and data collection on certain measures to evaluate the results of the instruction. Where noted extra-training factors of consequence to the evaluation are also entered in the listing. Although the 80 reports do not exhaust the available literature, they represent a large and diverse sample of empirical work which offers the reader a sense of the types of training efforts that have been conducted in addressing workplace hazards, the approaches used in their evaluation, and the results obtained.

Some overall impressions can be gained from viewing Tables 1 and 2 whose entries are culled from the studies listed in APPENDIX A.²

Table 1 tallies the evaluative measures found in the various training intervention studies using a modified version of Kirkpatrick's (1967) classification and the results reported on the measures. Specifically, under the heading of "Subjective/Self-Reports" are columns for "Reaction" (measures of whether the trainee thought the training interesting, worthwhile or relevant); "Knowledge Test" (scores on a quiz or other inquiry to determine their knowledge gain); and "Application" (reports from the trainees that they have changed their work practices as a result of the training or applied the learning in some other way). The heading "Objective/Surrogates" refers to more independent indicators, listing direct observations of "behavior change" or "other markers" (biological, environmental) that can be a product of the behavior changes. Positive changes from training on either of these measures can serve as surrogates for improved health/safety outcomes. The last columns of the table refer to actual measures of safety/health experience reflecting the organizational impact of the training. Note that some studies used more than one evaluative measure so the total number of entries in Table 1 exceeds the 80 reports.

---

² Where the text cites a study summarized in APPENDIX A, the notation (A-I, A-II, A-III, A-IV, or A-V) in bold will appear with the author(s) name(s), year citation to aid the reader in locating the material by section. Specifically, Roman numerals indicate the reference by hazard section: A-I=Injury; A-II= Chemical; A-III= Physical; A-IV= Ergonomic; and A-V= Biologic, and the entries found therein are listed alphabetically by senior author. Example: The work cited as Komaki, Heinzman & Lawson, 1980 [A-I] is listed in APPENDIX A in the Injury section under Komaki.

---

Table 1: Nature of Evaluative Measure Used in Training Interventions and Effects Reported

Table 2 collates the studies in terms of key elements of the methodology employed and related factors in the interventions. The headings and descriptor terms found therein require some explanation. Under "Basic Design" are entries defining whether the data collection for assessment involved only post-training measures on trainees (Post Test), or comparisons between pre- and post-training measures on a trainee group (Pre/Post), or repeated measures on a trainee group before, during and after training referred to as a time series design (TimeSer). In none of these instances was an untrained or control group compared with those receiving the training. The remaining designs did so: One form comparing post-test measures on a trained group versus a non-trained control group (Post/Ctr); another before and after training measures for a trained group compared with similar measures for a non-trained group (Pre/Post/Ctr); and last, a multiple baseline method (MultBsl) where repeated measures were taken on different groups before, during and after training and upon the introduction of other factors which may influence the training outcome. Through staggering the schedule of treatments, the latter method enabled the measures on one group to serve as an added control for measuring training plus other factors affecting other groups.

The training target column in Table 2 uses the OTA categorization (1985) for defining the objectives of the study. As noted earlier, these were defined as learning fundamental work practices (FndtWkPract), training in hazard recognition-awareness (HazRecog), worker directed or participative efforts in hazard recognition and control (WkrPartic), and worker empowerment training for the same purpose (WkrEmpwr). The variables column describes the nature of the training conditions and/or other factors manipulated during training or in the follow-on evaluation. Those noted refer to training only (TrngOnly); feedback (FdBk) with and without goal setting (GlSet); use of incentives (Trng&Incentives); and where the training plan called for specific manipulations of training content/delivery variables (TrngMode).

The post-training measures column takes note of the frequency and time span of the data collected on the evaluative measures. The basic categories were short- versus long-term (ShTm, LgTm), and whether the data were collected one or more times (One, Rep). Long-term (LgTm) refers to a post-training period exceeding 3 months,

The last column, headed "Other Considerations", identifies other conditions described in the studies that deserved special mention in light of their likely effect on the results. Nature of management support, workplace constraints or aids to facilitating the training or its application in the post-training environment were among the factors noted.

The following observations elaborate upon the summary data contained in
Tables 1 and 2:

(1) First and foremost, the number of entries indicating a positive effect attributed to the training or training plus other factors in the intervention is near overwhelming. In Table 1, 119 of 138 entries show a positive change from the training or instructional effort on the measures indicated and 10 others show partial success. In only nine cases do the findings display no effect which includes one that sought to convey safety instructions to workers via slogans in an informational campaign (Saarela, Saari & Alltonen, 1989 [A-I]). Admittedly, one could question whether this was a true training effort. Since the bulk of reports come from scientific journals, it could be alleged that these impressive results supporting a training effect could be a product of a bias toward publishing work showing positive findings. The consistency of the findings, however, given fairly similar training approaches for addressing the same or different agent hazards, is reassuring as is the fact that other types of reports (government documents, thesis studies) yield similar results.


(2) As displayed in Table 1, for the few cases where training plans failed to yield positive results or achieved only partial success, most occurred in the ergonomic agent category. Training interventions for back injury prevention are primarily responsible for this result which is not surprising given the complex etiology of back disorders, lingering questions as to what constitutes safe lifting positions, and apparent constraints posed by workplace conditions to using learned lifting techniques. Studies by Carlton (1987 [A-IV]) and Scholey (1983 [A-IV]) in particular illustrate how situational factors can complicate the transfer of safe lifting instruction. In Carlton's study, for example, kitchen workers in training sessions held away from their jobsites, learned and demonstrated ways for lifting trays that imposed less strain on the low back. Subsequent observations found that these types of lifts could not be performed in their work areas because the layout and obstacles made them assume awkward postures in handling loads of trays and the work pace precluded time to follow through on the acts required for risk reduction. Scholey reported that nurses could not transfer methods learned to ease the burden imposed in patient handling/lifting tasks because some patients refused to cooperate in moving to the edge of the bed or chair to facilitate the move. But even with these difficulties, the ergonomics agent category also shows some of the more substantial reductions in the frequency and cost of work related injuries reported as outcomes from training. Illustrative are reports of Hilyer et al., (1990 [A-IV]); McKenzie et al., (1985 [A-IV]); Schwartz (1987 [A-IV]) and Lepore, Olson & Tomer (1984 [A-IV]). As noted in Table 2 and APPENDIX A, because these studies did include ergonomic enhancements (i.e., improved workstation layout, tool redesign) along with training as part of the intervention, it is difficult to determine the specific training effect on these measures.

(3) Self-reports, especially knowledge gain and expressions of use in jobs, and objective behavioral measures predominate as outcome indicators in the training interventions. Positive findings of behavioral change from training such as adoption of safe work practices were in a number of instances coupled with reductions in injury, illness, and lost-days cost figures. Although suggestive, the reader is cautioned not to draw conclusions about a training specific effect on the latter indicators. One reason is that the instruction may have focussed on select (high risk) groups whereas the injury reports are for the whole organization. (Work reported by Lepore, Olson & Tomer, 1984 [A-IV] is an example). Some studies lacked proper control groups to rule out other factors that may have been responsible for the result on these types of outcome measures. Also, as one report notes (Reber & Wallin, 1984 [A-I]}), the findings may reflect changes in the manner of reporting mishaps during the intervention period. One deliberate effort to tie successful training outcomes such as improved work practices with reduced injury rates found that the adoption of the recommended work practices could only account for 25% of the actual reduction (Saari & Nasanen, 1989 [A-I]). In this study, worker training focussed on correcting housekeeping conditions believed responsible for an excess number of injuries based on accident reports. Post training observations indicated reduced injury rates far greater than those that could be accounted for by increased compliance with the better housekeeping practices also observed. The authors speculated that the gains in housekeeping left increased capacity for workers to notice other potential hazards. Another effort (Sulzer-Azaroff et al., 1990 [A-I]) suggested that attaining certain targeted safe work practices was responsible for reductions in lost-time injuries but furnished no definitive analysis to show an actual cause-effect relationship.

(4) As noted in Table 2, the work settings/occupational groups as training targets for the intervention studies were diverse. Most major industry categories are represented by the entries- agriculture, mining, construction, manufacturing, transportation, public services. Some entries are in several agent categories owing to the variety of hazards encountered in such work environs. As examples, there are reports of training interventions for health care workers (public services) at risk to injury, ergonomic and biologic hazards; metal fabricators (manufacturing) at risk to injury, chemical, ergonomic and physical agent hazards; warehousemen (transportation) at risk to injury and ergonomic hazards, etc. Employing the OTA classification of training emphases, most of the intervention work reported here is directed to learning fundamental work practices, but the more recent reports show worker empowerment approaches to be popular as well. This is especially true in hazardous waste site training where labor groups in concert with universities are conducting such activities. (Examples are McQuiston et al., 1994 [A-II]; Luskin et al., 1992 [A-II]; Cole and Brown, 1996 [A-II]). Hazard recognition training is also noted as are efforts at worker participation in problem solving. A one-mode training approach has been used in most studies with pre/post types of evaluations performed on the trainee group to assess the effects. Actual training objectives and training approaches, however, are much more varied than this summary would suggest. For example, fundamental type training objectives range from appreciating the more typical safeguards, e.g., use of personal protective equipment, proper operation of machines, housekeeping needs in factory environments, to the more exotic techniques such as defusing potentially violent patients in a mental hospital. The training methods also depict an assortment of approaches-classroom lecture, on-the-job instruction, audio-visual techniques, simulators, demonstrations/role play- with sessions of varying length and frequency. (See APPENDIX A for details.) This wide array of training situations and conditions combined with the positive results underscores the versatility of the learning process. At the same time, it invites efforts to examine whether some factors or factor combinations are more critical than others to attaining successful training outcomes.

2.       Intervention Data on OSHA Training Requirements/Guidelines

Ideally, to answer the question of whether OSHA training rules and guidelines have the intended benefit of reducing work related injury or disease, available documentation should offer evidence of:
(1) Training objectives presumably keyed to meeting the requirements of a specified OSHA standard.
(2) A training plan following the training guidelines mentioned earlier that can be reasonably implemented.
(3) A sound evaluation strategy that relates attainments of the training objectives to improved safety/health outcomes.
(4) Evaluative data reflecting increased knowledge/skills in hazard recognition and preventive measures, or commensurate worker actions resulting in reduced worker injuries or illnesses, or both.

The intervention studies reported in APPENDIX A and summarized above, though representing deliberate efforts to use training for hazard control purposes, provide only approximations to the ideal. Key limitations are:

Training Objectives: Many of the reports show training to be successful in improving protective actions or reducing a potentially hazardous condition but reference no particular OSHA standard or requirement pertaining to the situation posing a problem. Indeed, much of the work has been conducted by researchers who seek opportune targets for training that can yield results within a reasonably short time frame. Thus, positive results from training in these cases may or may not be considered as cogent support for any OSHA mandated rule. But having said this, the training plan in many reports does follow the OSHA guidelines in using a hazard analysis or injury reports to target the training needs and objectives, and in undertaking other steps for putting the plan in place. In this regard, the training effort may even be more relevant to the conditions under study. Perhaps OSHA rulemaking language calling for a training plan that defines and addresses site specific needs as opposed to one that dictates what they are a priori is worthy of discussion.

Training Effects/Outcome Measures: Whether the positive training effects reported such as knowledge gained, self-reports and/or actual observations of compliance with safe work practices can account for the reduced accidents, injury and lost days reported in several studies remain debatable. As already mentioned, the presence of other forms of hazard control introduced with training and uncertainties surrounding the basis for the accident/injury statistics argue against any such conclusions. Assessing the effects of training to reduce the risk of work-related chronic disease is even more problematic because of the long latency period needed to observe these kinds of outcomes. The reports on training plans undertaken in response to the OSHA foundry and coke standard show benefits in terms of increased worker knowledge of hazards and their professing greater adherence to safe work practices and protective behaviors (e.g., Parkinson et al., 1989 [A-II]). Although plausible, evidence that these training benefits have or will result in reduced lung/cancer diseases or other occupational illnesses for these work groups may require 10 or more years of surveillance. Similarly, it is too soon to say whether hazardous waste site workers and emergency responders who report that they were better able to manage chemical spills after taking required OSHA training (McQuiston et al., 1994 [A-II]) will also show fewer cases of diseases owing to better control of the exposure hazards.

Evaluation Design: In almost half of the reported studies, training effects were determined by evaluations of post-training measures for a given group or through pre- and post-training differences or comparisons before, during, and after training again on the same group. Since many of the interventions took the form of research projects, there is no way of separating out elements of novelty and researcher effects, which could have influenced the outcomes in these subject groups apart from any training effect. Some studies used comparable but untrained groups as ways to control these kinds of factors. Still, as noted in Table 2, other factors were present during the course of the evaluation, factors whose effects could not be accounted for in the results reported. In this regard, a variety of management actions were noted that deserve particular mention. For example, in one set of studies, they played roles in reinforcing and sustaining the learned behaviors (e.g., Zohar & Fussfeld, 1981 [A-III]). In other cases, supervisors were themselves the trainees and used to spearhead and effect the hazard control practices subject to evaluation (McKenzie et al., 1985 [A-IV]; Maples et al., 1982 [A-II]). As part of the training effort reported in other studies, supervisors were directed to increase their surveillance (Millican et al., 1981 [A-III]), or consider staff compliance in performance evaluations (Lynch et al.,1990 [A-V]). In still other reports, the authors indicate management's indifference to the training objectives (Fox & Sulzer-Azaroff, 1987 [A-I]). Being linked with management, these actions and other more subtle ones probably had profound effects on the evaluations which could not be sorted out because of the lack of suitable control groups or other confounding conditions.

There were other design weaknesses as well. Table 2 shows that more than half of the studies measured post-training effects less than 3 months after the instruction ended or only once after a longer interval of time. Thus, questions as to the durability of the reported effects, or possible intervening events affecting the longer term measures can be raised. Several studies did consider these issues with the results being somewhat gratifying (Zohar, Cohen & Azar, 1980 [A-III]; Hopkins 1983 [A-II]). In a few cases the subject groups were too small to make generalizable conclusions, and others assigned workers to training conditions on a non-random basis which could question the representativeness of the findings. Nevertheless, the sheer number of positive results found in so many different settings suggests that training effects are real even though the appraisals do not meet the more rigorous scientific standards.

Practicality Considerations: Given the success of the many intervention efforts noted in this report, there is the issue of whether the training plans under evaluation would be doable as a general practice at the worksite. The techniques employed in some studies to effect success were most elaborate. One example was the University of Kansas work (1982 [A-II]) in defining fundamental safe work practices and housekeeping measures as training objectives, and proceeding through formal instruction, practice trials plus incentives to establish worker compliance. But in targeting similar needs, the training plans in other reports seem almost incidental, confined to a brief session or two to define safe and unsafe practices. Having workers establish goals for conformance with safe practices when at the jobsite and provide feedback as to progress being made to the workers appear to be the more critical elements in the success of this approach. Studies by Reber & Wallin (1984 [A-II]), Fox & Sulzer-Azaroff (1987 [A-I]), Komaki, Barwick & Scott (1978 [A-I]) offer such examples as does an extensive review by Sulzer-Azaroff, Harris and McCann (1994). Provisions for goal setting and feedback as part of worksite training would not appear to be that formidable. Reports of training beyond learning fundamental work practices accent worker participation, active learning-type experiences using problem-solving exercises or other forms of instruction that show cogency to conditions found at one's workplace. In this regard, the merits of a learner-centered plan to address safety and health problems along with learning experiences that promote worker activism to effect improvements at their worksites have become emphasized. And, indeed, first results from evaluating this approach in hazardous waste site training programs appear promising [McQuiston et al., 1994 [A-II], Luskin et al., 1992 [A-II]; Cole & Brown, 1996 [A-II]). However, these findings are based on self-reports and thus have some limitation. Feasibility problems have also been noted. For example, needs to cover the subject matter of the course as well as to infuse empowerment ideas within usual time periods allotted to training mean that some topics get little treatment (Luskin et al., 1992 [A-II]). Another problem is the growing diversity of the workforce; this imposes added challenges in training approaches which stress worker inputs, group discussions and shared experiences. Cultural/language differences, among other factors, may be inhibiting (Cole & Brown, 1996 [A-II]). Two intervention studies reported in this review suggest ways for accommodating to language differences (Barnett et al., 1984 [A-II]; Weinger & Lyons, 1992 [A-II]). But perhaps the more important point is that training programs must be flexible in order to adapt methods and expectations to different working conditions and worker groups.

In summary, evidence for the effectiveness of OS&H training based upon the intervention studies reviewed above suggest the following observations:

(1) Taken as a whole, there is substantial documentation showing how training can meet objectives of knowledge gain, behavior change for improving worker health and safety. Reductions in work injuries and medical costs may also be noted in conjunction with these changes but evidence to show actual linkage or dependency remains to be ascertained.

(2) Because the intervention work does not address specific OSHA rules, the merits of specific training requirements can't be judged. On the other hand, the literature illustrates the benefits of training in meeting a variety of site-specific health and safety needs and objectives.

(3) Successful training outcomes in enhancing worker knowledge and control of workplace hazards and conformance with safe work practices depend greatly on extra-training or post-training environmental factors. Included in the latter are management's role/actions in favoring conditions which enable the knowledge gained from training or learned behaviors to be readily transferred to the jobsite. A proposed safety and health program standard as currently drafted by OSHA (OSHA, 1996) recognizes the merit of using a systematic approach to workplace safety and health. In doing so, management actions aimed at reinforcing training objectives along with other elements are duly noted.

The intervention literature on training in OS&H appears to be a collection of demonstration studies, some stronger in design than others in making a case for the benefits of a training approach (or in a few instances its shortcomings). As noted, more definitive analyses are needed to show how the outcomes of training relate to observed changes in injury or illness incidence. Also, and to be mentioned in a later section, certain factors critical to the learning process and post-training results, such as the length and frequency of the instruction, remain to be addressed in a more systematic way.

Even with all of the above uncertainties and qualifications, the array of evidence showing training-type interventions to effect positive changes in workplace safety and health remains formidable. Clearly, OS&H training can make a difference in reducing risks from workplace hazards. Rather than debate the issue of whether training is worthwhile, it would appear far more productive to determine what training procedures and related considerations are most crucial to optimizing its effects. Further support for this position comes from reviewing other sources of data bearing on the role of workplace training which are described below.

D. Training Data from Studies on Afflicted Workers

1.       Bureau of Labor Statistics Work Injury Reports
A second level of information pertinent to answering questions about the value of job safety and health training may be derived from worker injury record data or investigations of occupational accidents resulting in injury or fatalities. Indications that significant numbers of those affected didn't recognize hazards or lacked for knowledge of fundamental safe work practices could suggest training shortcomings as a factor in the occurrences. During the years 1978 to 1990, the Bureau of Labor Statistics conducted 20 surveys of workers injured on the job. The surveys focussed on specific injuries and/or on those resulting from high risk job operations. The workers sampled were drawn from workmens' compensation cases in 26 states where awards were made during select reporting periods for the injuries in question.

Using mail questionnaires, information was gathered on demographic factors, working conditions at the time of the injury event, use of personal protective equipment and extent or nature of safety training.

APPENDIX B describes the survey samples and includes extracts of the findings from 19 of 20 BLS survey reports reflecting worker responses to questions dealing with training issues. (One of the BLS reports [falls on stairs] did not include any training questions). The column entries are self-explanatory and cover the essential training-related data found in the surveys. The shaded entries are meant to suggest a serious training deficit for sizeable percentages of the afflicted workers. The authors hasten to note that this is purely judgmental; other readers may have different views about what is and is not noteworthy.

Table 3 summarizes the findings of APPENDIX B, reproducing in part those entries which suggest real or possible gaps in the job safety/health training as reported by the injured workers. For each of the 19 surveys listed, worker responses reveal at least one form of an apparent training deficit or a related problem. All but one injury category have at least two gaps and nine have three or more. "Limited coverage", meaning the percentage of those injured who lacked for any safety training, or had no training specific to their job needs, is the most frequent entry followed by "policy/action lapses" used to address training gaps after injury occurrence. "Content lacking" entries in Table 3 mainly referred to insufficient instruction in fitting, use, or limits of personal protective equipment issued workers. The "inexperience factor" was included in the listing insofar as it warrants even greater concern for adequate training. Entries here show that from 22% to as many as 78% of the affected workers had no more than 1 year's experience at the time of injury; for several categories more than 15% were injured within the first 6 months at the job. Fewer entries are noted in the "refresher needs" column because 13 of the 19 surveys did not include items about the time of the last safety training. Of the six surveys that did, percentage of injured workers with training more than 1 year old exceeded 50% in four of the six cases, and was above 25% in another.



Taken as a whole, Table 3 suggests many training inadequacies that may have contributed to the injury occurrences. However, the self-report nature of the data set, the possible bias of workers in avoiding any semblance of blame, and the lack of any confirmatory observations raises questions about drawing such a conclusion. Other data sources were sought to provide added clarification. These are summarized in the next section.

2.       Investigative Studies Acknowledging Training Needs and Effectiveness

Several studies have focussed on training issues directly or indirectly in the course of investigating workplace hazards and related risk factors for disease and injury. Those offering training-specific data are reviewed here to furnish added information bearing on the question of the efficacy of training in this context. The studies reference the injury, chemical, and ergonomic hazard agent categories and are presented in that order.

Injury Investigations: As part of a NIOSH-supported epidemiological study of workplace fatalities, Manwaring & Conroy (1990) reported the results of on-site investigations of 55 confined-space incidents where 88 workers lost their lives. Through interviews with co-workers and company officials, data were obtained on the conditions surrounding the events, and applicable company safety policy and employee training. This was augmented by information contained in reports from the medical examiners, OSHA compliance officer, responding emergency medical services personnel. Analyses of the 55 incidents to establish possible patterns to the occurrences or common factors revealed that in only three events did workers receive any training in confined-space safety. In these three cases, two supervisors and two workers died, three of whom had received the training. Further testifying to an apparent lack of training among other factors, no testing of the confined-space atmosphere was done before entry in any of the events, nor were confined spaces labeled with appropriate warning signs. Also few events gave evidence of confined-space ventilation prior to entry and no formal space entry authorization procedure was in place. Indications of the wrong type or improper use of respirators were also noted. The authors used the findings to stress the need to increase worker understanding and awareness of confined space entry through development and implementation of confined-space entry procedures and worker training. Poor implementation of training procedures was still evident in a later NIOSH report which summarized the data gained from a greater number of confined-space fatality investigations (NIOSH, 1994).

A similar NIOSH investigation of 201 electrocution-type incidents with 217 worker fatalities also suggested training deficits as a contributing factor but the available data were not as convincing as that shown above for confined spaces (Casini, 1993). In these cases, laborers, who typically received minimal training, showed the highest number of fatalities (42 of the 217 victims). But not far behind were linemen (40 of the 217 victims) who generally received extensive training in electrical safety. A total of 180 victims did receive some typical on-the-job training, in many cases from small employers. This investigation raised questions as to the adequacy of the training, and mentioned needs for more structure and ways to demonstrate that the workers understand the hazards and can carry out recommended safety measures.

Tan et al. (1991) interviewed 41 hospitalized patients who were being treated for hand injuries sustained at their workplaces. The interview took place shortly after the patients were admitted and gathered information on the nature and extent of the patient's training plus factors such as length of job service, description of the circumstances of the injury event. Regarding the latter, rollers, guillotines and chain saws were the machines commonly involved; most workers were unable to give a specific reason for the injury. Twenty-one patients had no job training and 20 had either formal or supervised on-the-job training of variable length. Three weeks of training was noted for the majority of workers. To determine whether training could have prevented the injury, patients having training were compared with those who had none in terms of the time each spent on the job before the injury occurred. The results showed little differences between the two groups. Indeed, 3 workers with training were injured on the first day and 7 were hurt within 12 weeks of starting their jobs as compared with 8 untrained workers who injured their hands during the same period. The authors questioned the adequacy of training in light of these results.

Chemical Hazard Investigation: Bryant, Visser & Yoshida (1989) collected questionnaire data from 165 hospital workers involved in ethylene oxide (EtO) sterilizing work. They found from 20% to 40% of the respondents to suffer from headaches, eye/skin irritations, and sore throats attributed to the exposures. Other symptoms reported were nausea (19%), running nose (16%), shortness of breadth (15%), and drowsiness (20%). Included in the questionnaire were items asking about the amount of training, which for the sample ranged from less than 1 hour to more than 1 day. In more than one-third of the cases first aid was included, and use of protective equipment (i.e., gloves, gowns, masks) while working with EtO. Environmental samples of EtO were collected during each sterilizer task for the 18 hospitals which employed the 165 hospital workers. Correlational analyses showed amount of training time and use of protective equipment to each bear an inverse relation to the prevalence of reported symptoms; however, only a few of the symptoms showed a significant decrease. Moreover, the expected decrease in exposure levels from the use of protective clothing did not cause users to report fewer symptoms of short-term irritation. Indeed, 80% of the workers still complained of one or more symptoms despite exposures within current OSHA regulatory limits.

Ergonomic Hazard Investigations: Snook, Campanelli & Hart (1978) analyzed questionnaire returns from insurance agents who provided data on the latest compensible occupational back injury case in their workload. A total of 191 cases were described; the data included information as to onset of back pain, previous back injuries, act at time of injury along with selection and training procedures the employer was using to reduce the risk of back injury. Lifting/pushing tasks when implicated in the injury were rated in terms of percentage of population who could safely perform the same act without overexertion, and were the tasks used to supply job design or ergonomic reference data in the evaluation. The cases were separated in terms of the presence or absence of various techniques of selection (medical history, low back X-rays), whether or not training in safe lifting techniques was given, and whether the jobs rated below or above the 75% limits of overexertion risk. Neither training nor any of the selection techniques were found to have any significant differential effect on the numbers of reported cases. Only the job load variable proved significant. The authors concluded that selection or training approaches were not effective controls for low back injuries.

Green and Briggs (1989) conducted a questionnaire survey of 514 keyboard operators in a university workforce to determine the benefits of adjustable work station furniture-furniture used to alleviate postural discomfort as well as other ergonomic problems. The respondents included those who were classified as sufferers or nonsufferers of repetitive strain from keyboard work based upon symptoms previously associated with overuse injury. The questionnaire items sought information related to the adjustability of different components of the user's workstation (i.e., desk, chair, monitor), instruction received on how to make such adjustments, and the adequacy of both the instruction and the adjustability features for ensuring comfort. In addition to questionnaire data, anthropometric measures were taken on a subsample of the sufferer and nonsufferer groups on seated postures, specifically popliteal and elbow-rest height. Related to training and though recommended, a full-scale educational program to inform operators on use of the workstation and strategies for relieving discomfort was not followed; rather advice on use of adjustable workstations was handled by issuing circulars diagramming proper positions to the users, and by small seminars and demonstrations run by the safety officer at the request of individual departments.

Respondent data indicated that sufferers had more negative perceptions of the equipment. Significantly more sufferers constantly readjusted the workstations (70% vs 43% for nonsufferers), could not get comfortable (35% vs 18% for nonsufferers), and were more likely to report insufficient adjustability (50% vs 20% for nonsufferers). The anthropometric data found chair heights not in accord with guides distributed. The authors believed that these problems were due to the lack of appropriate information being given to the operators. Almost 40% of the respondents could not recall receiving any guidance on how to adjust their workstations. Only 12% noted the literature distributed by the university; the most common source of information (64%) was from other persons with no formal ergonomic training. The authors concluded that there is a great need for training and/or information on workstation adjustment among keyboard operators, and that the preference for verbal instruction is greater than for printed forms of instruction. Given the widespread use of fellow workers as sources of information, it was also suggested that supervisors as well as keyboard operators be trained to cover new employee needs.

3.       Appraisal of Training Data of Afflicted Workers

The BLS work injury reports combined with the NIOSH findings on confined space fatalities and the problems seen in the postural discomfort/workstation adjustment survey strongly suggest that training deficiencies can contribute to these outcomes. This further supports the need for OS&H training. Yet, some workers who received training in these and the other studies cited in this section were still afflicted. Could the quality of training account for this result? That is, were sound training practices used? Did they, for example, follow the OSHA guidelines noted earlier? Even if they did, could other workplace factors (see Figure 1) have been present to nullify efforts to reduce injury or disease risk through training? Only two of above reviewed reports mention the instruction process. One found distinct weaknesses (Green & Briggs, 1989), and the other described nothing more than the length of training and whether it was formal or supervised "on-the-job" instruction. (Tan et al., 1991). Details of the training plan are sparse or non-existent in most of the above work.

In summary, workers without any OS&H training would appear at high risk for workplace injury or illness. Where training is given, the adequacy of the procedures in use becomes the issue. For this purpose, and like the intervention studies reviewed earlier, factors and conditions associated with effective OS&H training need to be ascertained.

In amplifying this point, OS&H training practices presently conducted throughout U.S. work establishments have not been examined to determine their adequacy in addressing known hazards and compliance with regulatory standards. One effort, limited to a small sample of companies engaged in waste site management, did in fact find many to be deficient in meeting existing OSHA training rules (Cole et al., 1994). The intervention studies summarized earlier cannot offer representative data since they were primarily demonstration efforts and special programs; they are not the norm. Although a new, independent survey would be ideal, much information characterizing OS&H training is already collected but remains buried in various databanks and report files that exist within NIOSH (e.g., Health Hazard Evaluation Reports, National Occupational Exposure Survey databank) and OSHA (e.g., compliance officer inspection reports). Systematic efforts to extract and assemble training related data from these sources could do much to provide a status statement on OS&H training practices and critical points.

E. Training Factors in the Context of Other Worksite Programs and Activities

1.       Training in Successful Occupational Safety and Health Programs
As already mentioned, hazard control programs include a variety of activities, training being one that is interwoven with others in efforts to minimize risk of work related injury and disease. Attempts have been made to isolate factors in such programs-factors which are important to achieving these goals. For this purpose, the safety program practices of companies differing significantly in their injury experience have been compared and other studies have analyzed hazard control efforts among employers who have achieved exemplary safety performance records. Of relevance to this literature review was whether there were any data to show if the training practices noted in these programs made a difference in safety performance or contributed in a unique way to the success experiences reported. Of the literature in this area, the studies noted below were most notable in terms of examining training differences, among other practices, in contrasting high- and low-accident workplaces.

NIOSH Studies: During the period 1975-1979, NIOSH published several reports resulting from a project aimed at defining factors in successful occupational safety programs. The project comprised three phases. The first was a questionnaire survey of the safety program practices of 42 pairs of companies in one state that were matched in type of industry, workforce size and locale but differed by more than two-to-one in recorded injury rate (Cohen, Smith & Cohen, 1975). The second phase comprised site visits to a sub-sample of the above group to verify and observe more closely apparent differences between the pairs that could explain the differential injury rate (Smith, Cohen & Cohen, 1978). The third phase used both mail questionnaires and site visits in collecting information from five companies recognized as having outstanding safety records based on total numbers of hours worked without a disabling injury (Cleveland et al., 1979). This third effort sought added confirmation of the findings from the previous two phases. Data collection for all phases focussed on such factors as management's commitment to the program, job safety training, safety incentives, hazard control measures, accident investigation/reporting procedures, and workforce characteristics. The general finding concerning training was that early indoctrination of new workers in safe job procedures with follow-up instruction to reinforce the learning was most frequently linked with successful safety performance. Formal classroom instruction versus on-the-job training or the use of varied instructional techniques were less notable considerations. The significance of training as compared with other program practices in accounting for safe performance could not be established. It is important to note, however, that management commitment factors both in these studies and others appeared to be the dominant or controlling element (Cohen, 1977).

Bureau of Mines Work: Peters (1989) summarized a number of studies supported largely by the Bureau of Mines dealing with organizational and behavioral factors associated with mine safety. Included were descriptions contrasting training practices and related miner knowledge in several surveys comparing high and low accident rate mines. The following were among the more notable observations:

(1) New miners in high-accident mines were less informed as to how to do their jobs than those in low accident mines.

(2) Lack of training in proper use of safety and health protective equipment was more frequently cited as being an important reason for miners not using the devices in high-accident mines.

(3) Having specific training in how the electrical system works; dealing with hazards such as coal dusts, gases and noise; and how to use tools, equipment was especially prevalent in the low-accident mines.

The Peters report noted that training for managers and supervisors produced significant improvements in mine safety and cited intervention studies demonstrating its effectiveness. One such study (Fiedler et al., 1984 [A-I]) is described in the collection of intervention work listed in APPENDIX A. This study found that the introduction of a structured supervisor training program accenting leadership style and skills in human relations or one stressing team-building and group problem-solving were each linked with reduced rates of injury and reduced MSHA citations at the mines under study.

Three observations about training derive from these NIOSH and BOM studies. The first is that training differences do exist between workplaces with good and poor safety records, but their overall importance remains to be ascertained. The second is that the differences seem relative, i.e., greater or more deliberate efforts are made to train, and to commit supervisor time and resources in the workplaces with better safety records. The third and related to the second is that supervisor training in how best to deliver and reinforce safe work practices seems crucial to the overall training effort and the success of the hazard control program. Somewhat at variance to this last point, one intervention study found that using external instructors to directly train employees was superior to a train-the-trainer approach in implementing aspects of a hazard materials information system in a large sample of companies (Saari et al., 1994 [A-II]). The differences between the two approaches, however, were least significant for companies with other well-established safety and health program practices. This suggests again, as shown in Figure 1, that training effectiveness depends greatly on other variables which complicates efforts at its assessment.

2.       Implications of First Aid Training.

Several OSHA rules (e.g., 29CFR Part 1910.151(b); Part 1926.21(b)) require that persons be trained to render first aid in the event that workers are injured and in need of treatment to maintain life, reduce suffering, or prevent the condition from becoming worse until more expert help arrives. Although one could argue that even these forms of instruction need to be evaluated as to their efficacy, there is evidence to show that workers who have first aid training tend to have fewer workplace injuries than those who lack for this type of training. Miller & Agnew (1973) first reported these findings which were confirmed later in more thorough evaluations by McKenna & Hale (1981; 1982). The importance of these results for this exercise is twofold: First, it documents that certain types of workplace training, though having other objectives, can apparently generalize and benefit workplace health and safety performance. Second, it suggests that melding first aid training with other requirements for OS&H instruction may have a reinforcing effect on the desired hazard control objective. The Miller and Agnew and McKenna and Hale studies offer speculations as to interactions between first aid and regular job safety training. These are contai