Occupational Safety and Health: Efforts to Obtain Establishment-Specific
Data on Injuries and Illnesses (Letter Report, 05/22/98,
GAO/HEHS-98-122).

Pursuant to a congressional request, GAO reviewed the Occupational
Safety and Health Administration's (OSHA) efforts to collect
establishment-specific data on injuries and illnesses.

GAO noted that: (1) with about $2.6 million available annually in 1996,
1997, and 1998 for its data collection surveys, OSHA determined it could
survey about 80,000 establishments each year; (2) within that
constraint, OSHA used mainly Bureau of Labor Statistics data to select
industries with high rates of injuries and illnesses; (3) OSHA used size
of establishment as a determining factor for the number of
establishments to survey; (4) in addition, OSHA knew some of the
industries had high numbers of work-related fatalities; (5) OSHA
surveyed establishments in these industries with 60 or more employees in
both years; (6) employers surveyed were not required to develop new sets
of injury and illness data to respond to OSHA surveys; (7) instead,
these employers were already required by OSHA to keep at their
establishments records of specific information on work-related injuries
and illnesses; (8) OSHA also required surveyed establishments to provide
information on employees' total hours worked and on the average number
of employees who worked during the year; (9) OSHA planned to use the
data collected to better identify establishments with the highest injury
and illness rates so that it could more accurately target on-site
compliance inspections to establishments with safety and health
problems; (10) in addition, OSHA planned to use the data to better
target its technical assistance and consultation efforts and to measure
its performance under the Government Performance and Results Act of 1993
in meeting its goals of reducing establishment injuries and illnesses;
(11) as of April 1998, however, OSHA had made only limited use of the
data collected in its 1996 and 1997 surveys mainly because of two
lawsuits; (12) a federal court ordered OSHA to halt implementation of a
new program that, using the 1997 survey data, identified specific
establishments with the highest lost workday injury and illness rates;
(13) employers who declined to participate in this new program would
remain on OSHA's list of employers most likely to be inspected; (14) the
program has been suspended until the court issues a decision; (15) as a
part of its data collection effort, OSHA gave no assurances about
privacy or confidentiality when it requested establishment information
from employers; and (16) OSHA has received many Freedom of Information
Act requests for the names and addresses of the 12,000 establishments
with high injury and illness rates that it invited to participate in the
new program.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  HEHS-98-122
     TITLE:  Occupational Safety and Health: Efforts to Obtain 
             Establishment-Specific Data on Injuries and Illnesses
      DATE:  05/22/98
   SUBJECT:  Freedom of information
             Occupational health/safety programs
             Occupational safety
             Law enforcement
             Safety regulation
             Data collection
             Surveys
             Industrial safety
             Statistical methods
             Litigation

             
******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO report.  Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved.  Major          **
** divisions and subdivisions of the text, such as Chapters,    **
** Sections, and Appendixes, are identified by double and       **
** single lines.  The numbers on the right end of these lines   **
** indicate the position of each of the subsections in the      **
** document outline.  These numbers do NOT correspond with the  **
** page numbers of the printed product.                         **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
** A printed copy of this report may be obtained from the GAO   **
** Document Distribution Center.  For further details, please   **
** send an e-mail message to:                                   **
**                                                              **
**                                            **
**                                                              **
** with the message 'info' in the body.                         **
******************************************************************


Cover
================================================================ COVER


Report to the Chairman, Subcommittee on Workforce Protections,
Committee on Education and the Workforce, House of Representatives

May 1998

OCCUPATIONAL SAFETY AND HEALTH -
EFFORTS TO OBTAIN
ESTABLISHMENT-SPECIFIC DATA ON
INJURIES AND ILLNESSES

GAO/HEHS-98-122

OSHA Data Collection

(205351)


Abbreviations
=============================================================== ABBREV

  BLS - Bureau of Labor Statistics
  CCP - Cooperative Compliance Program
  FOIA - Freedom of Information Act
  GPRA - Government Performance and Results Act
  LWDII - lost workday injury and illness
  OSHA - Occupational Safety and Health Administration
  SIC - standard industrial classification

Letter
=============================================================== LETTER


B-278544

May 22, 1998

The Honorable Cass Ballenger
Chairman, Subcommittee on Workforce Protections
Committee on Education and the Workforce
House of Representatives

Dear Mr.  Chairman: 

The mission of the Occupational Safety and Health Administration
(OSHA) is to save lives, prevent injuries, and protect the health of
America's workers.  OSHA, an agency in the Department of Labor,
administers the Occupational Safety and Health Act of 1970.  OSHA's
responsibilities include establishing mandatory safety and health
standards, rules, and regulations; enforcing employers' compliance
with the act and its standards, rules, and regulations; and reaching
out to employers and employees through technical assistance and
consultation programs. 

A major part of OSHA's enforcement program is conducting on-site
inspections of employers' facilities.  OSHA targets establishments
for inspection using a priority system.  First priority is given to
unprogrammed, or unscheduled, inspections initiated because of (1)
imminent danger situations, (2) fatalities or catastrophes, (3)
formal complaints, and (4) referrals.  Second priority is given to
programmed, or scheduled, inspections of establishments OSHA
considers more likely than others to be hazardous. 

To schedule inspections of establishments considered more likely to
be hazardous, as well as target other intervention activities, OSHA
relies on industrywide occupational injury and illness data from the
Bureau of Labor Statistics (BLS) rather than specific data for
individual establishments, for which no nationwide data are
available.  BLS, another Labor Department agency, is the federal
government's principal agency in the field of labor economics and
statistics.  It collects, processes, and analyzes statistical data
and publishes reports that are used by other federal agencies, the
Congress, state and local governments, businesses, unions, and the
public. 

BLS conducts an annual survey of occupational injuries and illnesses. 
For this survey, each year BLS selects a scientific probability
sample of all U.S.  establishments, regardless of size, in all
private industries.\1 Employers selected to participate in the survey
are required to report to BLS, among other information, data on
occupational injuries and illnesses that occurred at the selected
establishments in the previous year and information on the average
number of employees and the average number of hours worked during the
year at the establishment.  On the basis of the survey, BLS tabulates
and publishes aggregate statistical data at the national and state
levels on occupational injuries and illnesses for all U.S. 
industries and for each industry sector under the standard industrial
classification (SIC) of industries.\2

To ensure full participation in its survey, BLS pledges
confidentiality of the information provided by respondents.  Because
of this pledge, BLS provides neither to OSHA nor to any other user
the injury and illness and employment data (or any other data) it
collects from specific establishments. 

The primary data used by OSHA for targeting establishments for
inspection are industrywide injury and illness incidence rates and
lost workday injury and illness (LWDII) rates provided by BLS.\3 For
its programmed inspections, OSHA uses the LWDII and incidence rates
to identify industries with the highest rates.  OSHA cannot
distinguish, however, between those establishments with injury and
illness rates above the industry average and those with rates below
the average because an industry's rates reflect the industrywide
average rates of all establishments in the industry.  As a result,
OSHA schedules inspections of establishments with more than 10
employees randomly selected from among the top 200 industries
regardless of the establishment's record of injuries and illnesses or
any previous safety and health violations.  In this way, OSHA treats
all establishments in a targeted industry alike.  Thus, OSHA cannot
target for inspection specific establishments most likely to be
hazardous; it may target others--less likely to be
hazardous--instead. 

We and others have said that OSHA's enforcement and outreach
activities, using industrywide data rather than
establishment-specific data, are ineffective in targeting the most
hazardous establishments for inspection and result in an inefficient
use of OSHA's resources.  Others have recommended, as we did in
1994,\4 that OSHA develop procedures to obtain establishment-specific
injury and illness data from employers to better target its
programmed inspections and education and training efforts and to use
limited resources more efficiently. 

Since 1996, OSHA has been collecting establishment-specific data so
that it could identify establishments with the highest injury and
illness rates and better target its inspections and other
interventions to these establishments.  Some employers and business
associations, however, have concerns about how OSHA planned and
implemented its data collection initiative.  Because of these
concerns, you asked us to answer the following questions: 

1.  What industries and establishments did OSHA include in its 1996,
1997, and 1998 data collection efforts and how did it select them? 
What changes, if any, does OSHA plan to make for subsequent data
collection efforts? 

2.  What data did OSHA collect from establishments in the selected
industries? 

3.  How did OSHA intend to use the data it collected, and how has it
actually used the data? 

4.  What specific assurance, if any, did OSHA give about
confidentiality and privacy rights associated with the data
collected?  What is OSHA's current policy on access to the data it
collected from these establishments? 

To do our work, we reviewed provisions of the Occupational Safety and
Health Act of 1970 and documents on OSHA's data collection
initiative.  We conducted interviews with OSHA, BLS, and other Labor
officials and business and worker representatives about this
initiative.  We determined whether OSHA pledged confidentiality of
the employer occupational injury and illness data that it collected. 
In addition, we reviewed the Freedom of Information Act (FOIA) and
discussed with Labor Department officials how Labor, including OSHA,
responds to FOIA requests for these data.  We conducted our study
from September 1997 to April 1998 in accordance with generally
accepted government auditing standards. 


--------------------
\1 BLS excludes farms with 10 or fewer employees and self-employed
individuals.  In addition, BLS reports but does not collect data for
employers in railroad transportation and for certain mining
operations; these data are provided to BLS by other federal agencies. 
Until 1995, BLS surveyed about 4 percent of all employers--about
280,000 employers--each year.  In 1996, BLS surveyed about 180,000
employers. 

\2 The SIC system classifies private-sector establishments by type of
primary economic activity.  It classifies private industries into
nine divisions, for example, construction, manufacturing, and
services.  Within the divisions are 74 major groups such as the
health services group in the services division.  Within the major
groups, industry groups are classified broadly by type of business
such as nursing and personal care facilities in the major health
services group.  Within the industry groups, industries are
classified by specific type of business such as skilled nursing care
facilities in the nursing and personal care facilities industry
group.  In this structure, major groups are designated by a two-digit
numerical code, industry groups are designated by a three-digit code,
and particular industries are designated by a four-digit code.  For
example, within the services division, 80 denotes the major health
services group, 805 denotes the nursing and personal care facilities
industry group, and 8051 denotes the skilled nursing care facilities
industry.  BLS compiles injury and illness data at the four-digit
level for manufacturing industries and at the three-digit level for
nonmanufacturing industries. 

\3 The LWDII rate is the number of injury and illness cases that were
serious enough for workers to lose work time or to be put on
restricted work activities per 100 full-time workers per year.  The
injury and illness incidence rate is the number of injury and illness
cases per 100 full-time workers per year. 

\4 Occupational Safety and Health:  Changes Needed in the Combined
Federal-State Approach (GAO/HEHS-94-10, Feb.  28, 1994). 


   RESULTS IN BRIEF
------------------------------------------------------------ Letter :1

With about $2.6 million available annually in 1996, 1997, and 1998
for its data collection surveys, OSHA determined it could survey
about 80,000 establishments each year.  Within that constraint, OSHA
used mainly BLS data to select industries with high rates of injuries
and illnesses.  Then it used size of establishment as a determining
factor for the number of establishments to survey.  For its 1996 and
1997 surveys, OSHA included all manufacturing industries and 14
selected industry groups and specific industries in nonmanufacturing
SIC codes.  The nonmanufacturing industries were among those with the
highest LWDII and injury and illness incidence rates, according to
data reported by BLS on occupational injuries and illnesses for
calendar year 1993.  In addition, OSHA knew some of the industries
had high numbers of work-related fatalities.  OSHA surveyed all
establishments in these industries with 60 or more employees in both
years.  About 88 percent of the 80,000 establishments surveyed
responded in both years. 

For its 1998 survey, OSHA again included all manufacturing industries
and, using calendar year 1995 data reported by BLS, chose the same 14
nonmanufacturing industry groups and specific industries included in
the two previous surveys.  In addition, OSHA replaced four specific
industries included in 1996 and 1997 with the three larger industry
groups of which they are a part and added three new industry groups. 
OSHA sent all establishments in the newly selected industries with 50
or more employees survey forms in March 1998 and sent forms to some
of the establishments that it surveyed in 1997. 

Employers surveyed were not required to develop new sets of injury
and illness data to respond to the OSHA surveys.  Instead, these
employers were already required by OSHA to keep at their
establishments records of specific information on work-related
injuries and illnesses.  For its surveys, OSHA instructed
participants to copy these data from their records onto the survey
forms.  OSHA also required surveyed establishments to provide
information on employees' total hours worked and on the average
number of employees who worked during the year.  The data that OSHA
collects in its survey are a part of the data that BLS collects in
its annual occupational injury and illness surveys. 

OSHA planned to use the data collected to better identify
establishments with the highest injury and illness rates so that it
could more accurately target on-site compliance inspections to
establishments with safety and health problems.  In addition, OSHA
planned to use the data to better target its technical assistance and
consultation efforts and to measure its performance under the
Government Performance and Results Act of 1993 (GPRA) in meeting its
goals of reducing establishment injuries and illnesses. 

As of April 1998, however, OSHA had made only limited use of the data
it collected in its 1996 and 1997 surveys mainly because of two
lawsuits.  To settle one lawsuit, in February 1997, OSHA agreed not
to use the data from the 1996 survey for enforcement activities.  In
another lawsuit, in February 1998, a federal court ordered OSHA to
halt implementation of a new program that, using the 1997 survey
data, identified specific establishments with the highest LWDII
rates.  At that time, OSHA had already scheduled for inspection 500
establishments that reported the highest LWDII rates for inspection
and had completed 89 inspections.  OSHA had also already invited
about 12,000 employers who reported LWDII rates of at least
7.0--almost double the national average--to participate in this new
program, which would involve working cooperatively with the agency to
eliminate hazardous conditions in return for a reduced probability of
being inspected.  Employers who declined to participate in this new
program would remain on OSHA's list of employers most likely to be
inspected.  The program has been suspended until the court issues a
decision.  An oral argument is scheduled for December 1998.  In April
1998, OSHA began an interim inspection plan using OSHA survey data to
select establishments with high LWDII rates for inspection. 

As part of its data collection effort, OSHA gave no assurances about
privacy or confidentiality when it requested establishment
information from employers.  OSHA did not compromise the privacy
rights of individual employees because it collected only summary
information on injuries and illnesses; the data do not identify
individual workers.  OSHA said it took steps to protect employers'
privacy rights and to maintain confidentiality of the information. 
The agency did not pledge confidentiality, however, because the data
collected could be subject to disclosure under FOIA. 

OSHA has received many FOIA requests for the names and addresses of
the 12,000 establishments with high injury and illness rates that it
invited to participate in the new program.  OSHA has provided this
information to the requesters.  OSHA is permitted under FOIA to
withhold from disclosure information that might reveal its inspection
scheduling criteria.  It does not disclose such information while it
is being used for inspection purposes, but the information may be
subject to disclosure after OSHA completes its enforcement
activities. 


   BACKGROUND
------------------------------------------------------------ Letter :2

The Occupational Safety and Health Act of 1970 covers more than 100
million working men and women and about 6.5 million employers.\5
Excluded from coverage are the self-employed; state and local
government employees in some states; and some transportation workers,
miners, and others covered by other federal laws.\6

OSHA regulations require most employers covered by the act to keep
records at each establishment, including a log and summary of
occupational injuries and illnesses (OSHA form 200 or an equivalent
form) and a supplementary record of occupational injuries and
illnesses (OSHA form 101 or an equivalent form).\7 On the log,
employers must briefly describe all occupational injuries and
illnesses that occur at the establishment and summarize that
information yearly.  Employers must make the log accessible to
authorized federal and state officials and to employees upon request
and post an annual summary of occupational injuries and illnesses for
the previous calendar year at each establishment.  The supplementary
record is to provide information about each injury and illness on the
log, such as the affected employee's name and the circumstances of
the injury or illness.  Authorized government officials must have
access to these records also. 

The records employers must keep provide useful information for (1)
employers and employees, raising their awareness of injuries and
illnesses and helping them in their efforts to address establishment
hazards; (2) OSHA staff for carrying out enforcement and outreach
programs; and (3) statistical purposes, by measuring the magnitude of
injury and illness problems nationwide.  The information also helps
OSHA develop safety and health standards and conduct research on the
causes and prevention of such injuries and illnesses.  In addition,
BLS collects injury and illness data from employers for its annual
survey of occupational injuries and illnesses.  OSHA's compliance
officers review and collect data from the records during on-site
inspections. 

In February 1996, as part of its initiative to enhance safety, reduce
paperwork, and reinvent the agency, OSHA proposed comprehensively
revising the current rule for record-keeping requirements.  The
overall rule addressed certifying the records' accuracy and
completeness, requiring employers to provide increased access to the
records, defining key terms, and updating records to reflect changes
in previously recorded data. 

The proposed rule would also create a system for OSHA to collect
injury and illness data at the establishment level.  Under the
proposed data collection system, OSHA would (1) identify industries
among the most hazardous based on their injury and illness rates as
reported by the BLS annual survey of occupational injuries and
illnesses and (2) survey establishments in these industries to
collect establishment-specific injury and illness data. 
Establishment-specific data would help identify individual
establishments with high rates of occupational injuries and
illnesses. 

OSHA said it would focus its enforcement and outreach efforts on
establishments with the highest injury and illness rates.  Its
inspection priority system would remain unchanged--highest priority
would still be given to unscheduled inspections--but its process for
scheduling programmed inspections would be based on these
establishment-specific data.  In addition, OSHA said the data
collection system would enhance the agency's ability to measure its
performance in achieving established goals for reducing injuries,
illnesses, and fatalities. 

Although the proposed overall rule to revise the record-keeping and
reporting requirements had not been completed, in February 1996, OSHA
initiated its survey of employers to collect injury and illness data
for calendar year 1995.  In March 1996, the American Trucking
Associations and others filed a lawsuit challenging OSHA's authority
to compel employers to participate in this survey in the absence of a
final rule.  A federal district court ruled that OSHA did not have
the authority to issue citations to employers who did not complete
and return the survey.  In February 1997, OSHA issued a final rule
implementing its authority to survey employers and cite them for
failing to respond.  Also in February 1997, OSHA and the parties
involved in the lawsuit agreed that OSHA would not use the survey
data collected in 1996 for enforcement purposes, but it could use the
data for other purposes. 


--------------------
\5 A state may develop and operate its own safety and health program
in place of the federal program if OSHA has approved a state plan and
confirmed through monitoring state performance that the program
remains "at least as effective as" OSHA.  Twenty-three states and two
territories operate such programs; two of these states cover the
public sector only. 

\6 Since 1976, an appropriations rider has exempted certain small
farms with 10 or fewer workers. 

\7 Employers exempt from keeping these records include employers with
10 or fewer employees and employers who conduct business in specific
SIC codes, including some retail trade businesses; service
businesses; and finance, insurance, and real estate businesses. 


   OSHA CONSIDERED INDUSTRYWIDE
   INJURY AND ILLNESS DATA AND
   ESTABLISHMENT SIZE IN DESIGNING
   ITS DATA COLLECTION SURVEYS
------------------------------------------------------------ Letter :3

OSHA used a two-stage process for selecting industries and
establishments to include in its data collection surveys.  First,
OSHA selected the industries for its surveys using mainly
industrywide data on injuries and illnesses.  Second, within the
industries selected, OSHA chose individual establishments to survey
on the basis of establishment size.  OSHA's objective was to survey
all establishments of a specific size in the industries with the
highest injury and illness rates.  OSHA said a major determinant of
the number of establishments it could survey in a year, however, was
the amount of funds available for conducting the survey.  With about
$2.6 million available annually to fund the surveys in 1996, 1997,
and 1998, OSHA determined it could survey up to 80,000 establishments
each year. 


      SELECTION OF INDUSTRIES
---------------------------------------------------------- Letter :3.1

To select the industries among the most hazardous, OSHA used data
obtained from BLS' annual surveys together with other factors such as
work-related fatalities and the number of establishments most likely
to be included in the survey in each industry.  Because data and
documentation supporting these decisions were not available, however,
we could not assess the extent to which each factor contributed to
OSHA's selecting--or excluding--an industry. 

According to OSHA officials, the agency decided to include in its
first three surveys all manufacturing and nonmanufacturing industries
considered most likely to be hazardous for which it has
responsibility.\8 OSHA officials said it included all manufacturing
industries because (1) manufacturing industries are required to
maintain OSHA injury and illness records, (2) OSHA compliance
standards to a large extent focus on manufacturing industries, (3)
some manufacturing industries have injury and illness rates that are
among the highest in all the SIC codes, and (4) the large number of
manufacturing SIC codes (and the large number of establishments in
each manufacturing group) made it impractical to separate high-hazard
manufacturing industries from low-hazard manufacturing industries for
the surveys. 

According to OSHA, the main factor it used to identify
nonmanufacturing industries most likely to be hazardous was the
industry three-digit SIC code LWDII rate as reported by BLS.  OSHA
also considered work-related fatalities when selecting industries to
survey.  Other factors OSHA considered in its selection process for
those years included whether the injuries reported by establishments
occurred at a fixed facility or at an off-site location and the
number of establishments in each industry that would most likely be
included in the surveys. 


--------------------
\8 OSHA excluded the construction industry from the surveys for
several reasons, according to OSHA officials.  For example, a
construction business may have many contractors and subcontractors. 
Although each contractor and subcontractor at a construction site is
required to keep injury and illness records, some of them (for
example, plumbers and electricians) may have a primary SIC code in an
industry other than construction.  For them, establishment injuries
and illnesses would be recorded in their primary SIC codes, not in a
construction code.  OSHA officials said construction industries may
be included in future surveys. 


      1996 AND 1997 SURVEYS
---------------------------------------------------------- Letter :3.2

For the 1996 and 1997 surveys, OSHA chose all manufacturing
industries, six industry groups (three-digit SIC codes), and eight
specific industries (four-digit SIC codes) in nonmanufacturing SIC
codes.  The nonmanufacturing industries selected had high LWDII and
injury and illness incidence rates at the three-digit SIC code level,
according to calendar year 1993 BLS data.  Four of the specific
industries and one industry group that OSHA selected, according to
the agency, also had high numbers of fatalities during the previous
10 years.  (See fig.  1.)

   Figure 1:  Industries Selected
   for OSHA Surveys, 1996, 1997,
   and 1998

   (See figure in printed
   edition.)

\a LWDII and incident rates for 1996-97 are BLS data reported for
calendar year 1993.  Rates for 1998 are BLS data reported for
calendar year 1995.  Also, rates for four-digit codes are the same as
those provided by BLS at the three-digit level of which they are a
part.  For example, rates for codes 4214 and 4215 are those reported
by BLS for SIC code 421. 

\b Selected SIC code also has a high fatality rate. 

As shown in figure 1, all 14 nonmanufacturing industry groups and
specific industries OSHA selected in 1996 and 1997 had LWDII rates
and injury and illness incidence rates that exceeded the national
averages for all industries.  Furthermore, most industries selected
were among those with the highest LWDII rates.  Because data were not
available, however, we could not assess the extent to which each
factor contributed to OSHA's selecting or excluding industries. 


      1998 SURVEY
---------------------------------------------------------- Letter :3.3

OSHA expanded the 1998 survey to include additional industries in the
survey database.  According to OSHA, it included all manufacturing
industries and the 14 nonmanufacturing industry groups or specific
industries in the previous surveys again.  It replaced four specific
industries on the survey list with the three industry groups of which
the specific industries are a part.  OSHA also included three
industry groups in 1998 that were not in the previous surveys.  As a
result, for 1998, OSHA selected a total of 16 nonmanufacturing
industry groups or specific industries for the survey.  According to
OSHA, it based selections for the 1998 survey exclusively on calendar
year 1995 injury and illness data obtained from BLS; fatalities and
other safety and health factors were not considered.  Each of the
newly selected industry groups included in OSHA's 1998 survey had
LWDII and injury and illness incidence rates that exceeded the
national averages for all industries.  (See fig.  1.)


      SELECTION OF ESTABLISHMENTS
---------------------------------------------------------- Letter :3.4

OSHA selected establishments to survey for all 3 years from within
each of the chosen industries on the basis of establishment size. 
OSHA officials said they did this to include in the surveys all
establishments of certain sizes in each of the industries selected,
rather than survey a sample of establishments in these industries. 
For the 1996 and 1997 surveys, OSHA mailed surveys to all
establishments with 60 or more employees in manufacturing industries
and in each of the 14 selected nonmanufacturing industries.  For the
1998 survey, OSHA mailed survey forms in March 1998 to all
establishments in the newly selected industries with 50 or more
employees.  It also mailed forms to some of the establishments
included in the 1997 survey:  (1) those that did not return their
forms, (2) the largest establishments in each state, and (3) those
that had reported an LWDII rate of 7.0 or higher. 


   ESTABLISHMENTS ALREADY COMPILE
   INJURY AND ILLNESS DATA THAT
   OSHA COLLECTS
------------------------------------------------------------ Letter :4

For its 1996, 1997, and 1998 surveys, OSHA asked employers to provide
summary information on their employees' injuries and illnesses during
the previous calendar year.  Because OSHA already requires the
establishments to compile this information, employers are not
required to develop new data sets.  OSHA also asked employers to
provide certain employment information for the establishments. 

Although OSHA and BLS collect the same injury and illness and
employment data from establishments that participate in their
respective programs, BLS collects the data from a small sample (less
than 3 percent) of all private-
sector industry establishments and uses the information to generate
aggregate statistics on occupational injuries and illnesses at the
state and national levels.  Because BLS pledges confidentiality of
the data to employers, it does not share these data with OSHA.  OSHA,
on the other hand, needs establishment-specific data to identify
individual establishments' LWDII rates and injury and illness
incidence rates to more effectively and efficiently carry out its
regulatory and enforcement activities.  Because it cannot obtain
these data from BLS and they are otherwise unavailable, OSHA collects
injury and illness data from all establishments of a certain size
within selected industries. 

The OSHA data collection survey form is identical to a portion of the
BLS annual survey of occupational injuries and illnesses form.  The
wording of the instructions, examples, and questions on the OSHA
survey form is identical to that on the BLS survey form.  In addition
to the injury and illness and employment data, both data collection
forms ask for the name, telephone number, date, and signature of the
person to contact if any questions arise about the information
provided.  This contact information also allows OSHA and BLS to
verify the data provided.  BLS also collects information that OSHA
does not on the demographics of injured and ill workers and the
circumstances of the injuries and illnesses for a sample of cases
that required recuperation away from work. 

To minimize employers' burden, OSHA and BLS instruct employers
responding to their surveys to copy on their survey forms the
requested injury and illness data from the log and summary of
occupational injuries and illnesses they are required to maintain. 
In addition, because some establishments from which OSHA collects
data may be included in the BLS sample in a given year, OSHA has
coordinated its data collection effort with BLS'.  (BLS estimated
that less than 10 percent of the establishments selected for the OSHA
data collection effort would be included in the BLS sample in any
year.) Establishments required to report to OSHA and BLS may use a
single form and send a copy to each agency. 

The data collection form includes a section in which the respondent
can provide summary information specific to the selected
establishment.  The first part of the summary section requests the
average annual number of employees and the total number of hours that
employees worked during the previous calendar year.  It also requests
information on conditions during the year, such as a strike or a
shutdown, that might have affected the number of employees or the
hours they worked. 

The second part of the form requests the following information from
the total line of the log and summary of occupational injuries and
illnesses maintained by each establishment: 

  -- total injuries, including the number of deaths as a result of
     injury, injuries with days away from work or restricted workdays
     or both, total days away from work, total days of restricted
     work activity, and injuries without lost workdays;

  -- total illnesses, including deaths as a result of illness,
     illnesses with days away from work or restricted workdays or
     both, total days away from work, total days of restricted work
     activity, and illnesses without lost workdays; and

  -- types of illnesses experienced by the workers, including skin
     diseases or disorders, diseases of the lungs due to dust,
     respiratory conditions due to toxic agents, poisonings,
     disorders due to physical agents, disorders associated with
     repeated trauma, and other occupational illnesses. 

The information collected enables OSHA to compute each
establishment's LWDII rate and injury and illness incidence rate. 
See the appendix for a copy of the OSHA data collection form. 


   OSHA PLANS TO USE
   ESTABLISHMENT-SPECIFIC DATA IN
   VARIOUS WAYS, BUT SO FAR USE
   HAS BEEN LIMITED
------------------------------------------------------------ Letter :5

OSHA, in announcing its plans to collect establishment-specific
injury and illness data by mail, indicated that such information
would be used in a variety of ways to help OSHA carry out its
responsibilities more efficiently and effectively.  The intended uses
were (1) directing OSHA's program activities, including the
scheduling of establishment inspections under its enforcement program
and the targeting of mailings of safety and health information to
employers under its nonenforcement programs; (2) monitoring and
tracking injury and illness incidents; (3) developing information for
promulgating, revising, and evaluating OSHA's safety and health
standards; (4) evaluating the effectiveness of OSHA's enforcement,
training, and voluntary programs; and (5) providing pertinent
information to the public.  In addition, OSHA stated that the
establishment-specific data were necessary for it to meet GPRA
requirements, which direct federal agencies to implement a program of
strategic planning, develop systematic measures of performance to
assess the impact of individual government programs, and produce
annual performance reports. 

Although OSHA collected establishment-specific injury and illness
data during 1996 and 1997, as of April 1998, it had made only limited
use of the data.  None of the intended purposes has been fully
implemented, and the data have not been used for other purposes. 
About 70,000 establishments responded to both the 1996 and 1997
surveys--about 88 percent of the establishments surveyed. 

According to OSHA officials, firm plans for using the data involve
enforcement activities and meeting performance goals it established
under GPRA.  The data will be used as part of Labor's performance
measurement system to track the impact of OSHA's enforcement and
compliance assistance interventions.  For example, to measure the
extent to which OSHA achieves its goal of reducing injuries and
illnesses by 15 percent in high-hazard industries, such as food
processing and logging, the agency will track survey data from
employers in these industries. 

According to OSHA's directive (CPL 2-0.119), the agency also plans to
use survey data to schedule enforcement activities for establishments
with the highest LWDII rates.  OSHA will use the data to identify the
500 establishments with the highest rates and schedule them for
on-site inspections.  In addition, OSHA wants to use the
establishment-specific injury and illness data to identify employers
for participation in its new Cooperative Compliance Program (CCP). 
Under this program, OSHA would invite employers who report high LWDII
rates on the survey to work cooperatively with OSHA to eliminate the
hazardous working conditions.  These employers would be put on a list
of those most likely to be inspected; however, if these employers
agree to participate, they must agree to establish an effective
safety and health program.  They must also agree to (1) find and
remove hazards, (2) work toward reducing injuries and illnesses, (3)
fully involve employees in their safety and health program, (4) share
injury and illness data, and (5) provide OSHA with information from
their annual injury and illness records.  Under CCP, employers with
100 or fewer employees who choose to participate and agree to seek
free assistance from their state OSHA consultation program to
establish effective safety and health programs reduce their
likelihood of being inspected by OSHA to 10 percent.  CCP
participants with more than 100 employees and smaller employers not
using consultation services face a 30-percent chance of being
inspected.  If identified employers do not agree to participate in
the program, they will remain on OSHA's list for on-site inspection. 

According to OSHA officials, inspections of CCP participants will
most likely be shorter than regular inspections and result in lower
penalties than normal because of these employers' commitment to
finding and eliminating hazardous working conditions in their
establishments.  OSHA believes that those who successfully fulfill
the requirements of the program should reduce injuries, illnesses,
and fatalities, leading to lower workers' compensation costs and
reduced insurance costs.  In addition, workers whose employers join
the program will be more involved in establishment safety and health
issues and should experience fewer injuries and illnesses and have an
improved quality of work life.  OSHA will also benefit by extending
its resources and expanding the base of employers with safety and
health programs, which OSHA believes is a major difference between
employers with low injury rates and those with high rates. 

OSHA used the information it collected in 1997 to develop a list of
about 12,500 establishments with the highest LWDII rates--that is,
LWDII rates of 7.0 or higher.  OSHA scheduled the 500 establishments
with the highest LWDII rates for inspection and began these
inspections in December 1997.  In November 1997, OSHA invited about
12,000 of these establishments--less than 20 percent of those that
responded--to participate in the CCP.  According to OSHA officials,
more than 89 percent of the employers invited by OSHA agreed to
participate in the program.  In response to a lawsuit filed by the
U.S.  Chamber of Commerce and others claiming that OSHA had not
followed proper procedures in implementing the CCP, however, a
federal court of appeals ordered OSHA in February 1998 to halt the
enforcement program that includes CCP until the court decides whether
the program is valid.  According to OSHA officials, oral argument is
scheduled for December 1998, and a decision is unlikely to be issued
until some time in 1999.  The order also required OSHA to stop
conducting its inspections of the 500 establishments with the highest
LWDII rates; 89 of these inspections had been completed when OSHA was
told to stop conducting them.  OSHA officials stated that this delay
in implementing the CCP will adversely affect many of its enforcement
and nonenforcement activities. 

In April 1998, OSHA began implementing an interim inspection
scheduling plan.  Under the plan, OSHA will schedule for inspection
establishments in 99 industries with LWDII rates of 6.4 or higher,
according to calendar year 1996 BLS data.  Establishments in these
industries for which OSHA collected data in 1997 with LWDII rates at
or above the national average of the industry of which they are a
part will be randomly selected for inspection.  The interim
inspection plan has no CCP component. 

In our 1994 report,\9 we noted that one problem with relying on
employer-
provided data is the risk that employers may underreport injuries and
illnesses if they know OSHA is collecting data about their
establishments that could be used to target them for on-site
compliance inspections.  To reduce the risk of employers
underreporting injury and illness data, OSHA needs to have a
successful combination of enforcement and education.  Therefore, we
recommended in that report that OSHA implement procedures for
ensuring that employers accurately record occupational injuries and
illnesses.  Because of its concerns about the quality of the data
provided by employers responding to its surveys, OSHA is conducting
on-site audits of employers' injury and illness records to assess
these records' accuracy.  OSHA has completed all of the 250 records
audits it had planned to conduct. 


--------------------
\9 GAO/HEHS-94-10, Feb.  28, 1994. 


   OSHA GAVE EMPLOYERS NO
   ASSURANCE OF CONFIDENTIALITY AS
   PART OF THE DATA COLLECTION
   INITIATIVE
------------------------------------------------------------ Letter :6

OSHA gave no assurances about privacy rights or confidentiality
associated with the data collected from employers selected to
participate in the data collection survey.  Privacy rights of
individual employees did not present a problem because the only
information about workers that OSHA collected was summary information
on injuries and illnesses, which does not identify individual
workers.  OSHA said it took steps to protect employers' privacy
rights and to maintain confidentiality of the information.  OSHA did
not pledge to employers that the data it collected in its surveys
would be kept confidential, however, because the data could be
subject to disclosure under FOIA.  According to OSHA, this
information would be made available to the public only in response to
specific FOIA requests.  The injury and illness data OSHA collected,
however, are the same data that employers are required to post in
their establishments each year. 

In contrast with OSHA, BLS pledges confidentiality to the full extent
permitted by law to all participating establishments and informs the
respondents that the data will be used for statistical purposes only. 
BLS tabulates and publishes data aggregated at the national and state
levels by various characteristics, such as industry group,
occupation, and age.  BLS does not tabulate or publish injury and
illness data on individual establishments.  Over the years, BLS has
received FOIA requests for the data, including requests for
establishment-specific injury and illness data but has refused to
disclose the data, relying on the FOIA exemption for confidential
commercial or financial information.  A federal district court has
upheld BLS' right to withhold from disclosure commercial or financial
information that has been voluntarily provided to BLS under a pledge
of confidentiality in large part because disclosure would impair the
government's ability to obtain the data in the future.\10 Whether
OSHA would have a valid basis to rely on the same exemption has not
been determined. 


--------------------
\10 See Husted v.  Norwood, 529 F.  Supp.  323 (S.D.  Fla.  1981). 


      FOIA REQUIRES DISCLOSURE OF
      AGENCY RECORDS WITH SOME
      EXCEPTIONS
---------------------------------------------------------- Letter :6.1

Under FOIA, a federal department or agency is required to disclose
information to anyone who requests it, unless the information is
covered by one of the law's exemptions.  Examples of such exemptions
include trade secrets and individuals' medical files.  The medical
files exemption excludes from disclosure any data from establishments
that identify individual employees' injuries and illnesses.  Another
exemption excludes information compiled for law enforcement purposes
that would disclose techniques, procedures, or guidelines for law
enforcement investigations.  This exemption excludes from mandatory
disclosure any data that might provide advance notice of an
inspection.  According to OSHA, it does not disclose collected
establishment-specific data while such data are being used for
scheduling inspections that might disclose the scheduling criteria. 
After the inspection is completed, however, the exemption no longer
applies and the data may be subject to disclosure, OSHA officials
said. 


      OSHA HAS RECEIVED AND
      RESPONDED TO MANY FOIA
      REQUESTS
---------------------------------------------------------- Letter :6.2

From 1996 to 1998, OSHA received many FOIA requests about the data
collection initiative.  Many of these requests specifically related
to requests about the CCP.  The only establishments asked to
participate in the CCP were those that responded to the data
collection initiative, but, as already noted, implementation of the
CCP has been postponed because of a lawsuit. 

Labor agencies handle all FOIA requests on a case-by-case basis. 
Most of the requests OSHA has received and responded to about the
data initiative asked for the names and addresses of establishments
identified as having LWDII rates high enough to be invited to
participate in the CCP.  OSHA has provided these requesters with the
names and addresses of the establishments only.  OSHA has also
provided its field offices with the names and addresses of
establishments in their regions to enable staff there to respond to
similar requests.  OSHA received one FOIA request for injury and
illness data collected in the 1996 survey.  As of April 1998,
however, the agency had not released the requested information. 

OSHA does not know the number of FOIA requests received by its
headquarters and field offices about the data collection initiative. 
Labor is not required to and does not collect data on the specific
subjects of FOIA requests.  In addition, although federal agencies
and departments must annually report to the Department of Justice on
the number and cost of FOIA requests and responses, detailed
information on the subjects of FOIA requests is not required. 
Moreover, although Labor has a national FOIA coordinator, it does not
centrally track all FOIA requests received.  FOIA requests concerning
establishments identified by the data collection initiative are
decentralized:  they may be responded to by OSHA headquarters,
regional, or area office staff.  OSHA headquarters officials told us
that generally they neither oversee nor approve FOIA responses
handled by OSHA regional and area staff; nor are they informed of all
FOIA requests received in the field.  According to OSHA officials,
however, area, regional, and national staff responsible for FOIA
activities may coordinate efforts when preparing FOIA responses. 


   AGENCY COMMENTS
------------------------------------------------------------ Letter :7

We provided a draft of this report to the Department of Labor for its
review and comment.  Although Labor did not provide written comments
on the draft report, officials from OSHA and other offices provided
technical comments, which we have incorporated as appropriate. 


---------------------------------------------------------- Letter :7.1

As arranged with your office, unless you publicly announce the
contents of this report earlier, we plan no further distribution
until 7 days after its issue date.  At that time, we will send copies
of this report to the Secretary of Labor and other interested
parties.  We will make copies available to others upon request. 

If you or your staff have any questions about this report, please
call me at (202) 512-7014 or Larry Horinko, Assistant Director, at
(202) 512-7001.  Other major contributors to this report are John T. 
Carney, Evaluator-
in-Charge; Ronni Schwartz, Senior Evaluator; and Robert G.  Crystal,
Assistant General Counsel. 

Sincerely yours,

Carlotta C.  Joyner
Director, Education and
 Education Issues




(See figure in printed edition.)Appendix
OSHA'S OCCUPATIONAL INJURY AND
ILLNESS DATA COLLECTION FORM
============================================================== Letter 



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)


*** End of document. ***