Every workday, about 25 American workers are killed on the job, and between 10,000 and 45,000 are injured seriously enough to require medical care or time off, according to a congressional study highly critical of the Occupational Safety and Health Administration (OSHA).

The 425-page study by the Office of Technology Assessment (OTA) said that contrary to Reagan administration statements, the decline in work-place deaths and injuries between 1979 and 1983 has been caused mostly by economic recession and unemployment rather than OSHA's efforts.

The study is one of the most comprehensive examinations of data on injury and illness in the work place. It recommends that Congress consider increased funding for research, data collection, training and enforcement of health and safety laws.

The study faults OSHA for issuing few new regulations to control hazards and for failing to update its health standards as medical evidence shows that certain substances are more dangerous than previously believed.

"Most OSHA standards are the same as they were 14 years ago" when the agency was created, the study said, adding that those rules "seriously lag behind" industry standards.

The OTA, a nonpartisan arm of Congress, noted that the typical factory will be inspected once every six years and will face fines of less than $200 for "serious" violations. "OSHA's generally infrequent inspections and comparatively low penalty levels have provided limited incentives for companies to comply with the standards," OTA said.

OSHA is facing growing criticism from organized labor. Labor Secretary-designate William E. Brock has said he plans to scrutinize the agency. Brock said in an interview that OSHA has been the leading area of concern among union officials with whom he has discussed Labor Department issues.

Rep. John D. Dingell (D-Mich.), chairman of the House Energy and Commerce Committee, who requested the OTA report, said it "seems to indicate that employers have the ability to do far more to protect workers on the job . . . . The report also points out that effective enforcement of existing OSHA regulations and a more active effort to provide information to both employers and workers could result in major reductions in occupational disease.

"The new secretary of labor and other policy-makers would do well to read this report," Dingell said.

Brock, who faces confirmation hearings later this month, is expected to replace OSHA chief Robert A. Rowland, who has not been confirmed by the Senate, according to department sources. Rowland has been criticized for his agency's failure to move promptly to develop new standards.

A Texas attorney and former Reagan campaign official, Rowland has also been criticized because he retains substantial investments in chemical and drug firms that OSHA regulates. Those investments are in a blind trust that has been approved by the Office of Government Ethics.

In November Rowland said Bureau of Labor Statistics (BLS) data on work-place injuries and illnesses showed the problems at the "lowest level ever recorded." He said that the findings were evidence of the effectiveness of OSHA's "nonadversary" approach to regulation, which emphasizes counseling employers rather than penalizing them for safety violations.

But the OTA study said the BLS results were incomplete and that "most studies on injury rates have found that OSHA, overall, has had either no effect or only a limited effect . . . . The decline in injury and fatality rates between 1979 and 1993 resulted mainly from the economic recession," which spurred unemployment and closed many manufacturing plants that had had high injury rates. The study also said the recession caused layoffs of mostly younger workers, who have higher injury rates than their older, more experienced colleagues.

While minimizing OSHA's overall impact, the study said that the agency had been effective in reducing illnesses when it moved in the past to establish stricter exposure standards for such substances as vinyl chloride, cotton dust and lead. But OSHA issued only 11 standards in its first 13 years, covering 24 substances, the study said.

OSHA officials had no immediate comment on the study.

The OTA said that data on work-place deaths and illnesses were sketchy and often inaccurate, and that "employer records underestimate the magnitude of the occupational disease problem." The study estimates that 6,000 workers are killed annually on the job, a figure that is considerably higher than the Bureau of Labor Statistics' estimate of 3,000 to 5,000 fatalities. The National Safety Council's estimate is 11,000 to 12,000.