Medical specialists urged the government yesterday to ban or sharply curb use of a popular prescription pain killer, saying that it is killing three to 10 Americans a day while nonprescription aspirin and acetaminophen are far safer, work better, and cost about one-tenth as much.

The drug is sold mainly as Darvon, the Eli Lilly & Company trade name for propoxyphene. Lilly introduced it in 1957 and has grossed from it, and drugs containing an estimated $500 million. Doctors have written 575 million new and refill prescriptions for Darvon products, and, in some years, more for them than for any other drug.

The experts told a hearing of the Senate Committee on Small Business that no group of patients has been isolated in whom nonprescription pain killers or codeine wouldn't be more effective and/or safer.

Mayo Clinic cancer specialist Charles G. Moertel, who has reviewed all comparative studies in the medical literature on Darvon's effectiveness as a pain killer, was asked if there is a medical justification for the narcotictype chemical. "I know of none," he replied.

He said that for mild pain, over-the-counter drugs do "just as good a job, if not better." For moderate pain, aspirin combined with other analgesics achieves "significantly superior" relief over Darvon combinations. For Severe pain, "Darvon either alone or in combination... is really inhumane to the patient."

Lilly, which will testify Monday, released a national survey showing that of 514 physicians polled, "a vast majority" rate Darvon as "a safe and effective analgesic when used as directed." Moertel commented that doctors "are no less vulnerable than the public at large through the persuasive influence of Madison Avenue."

In regard to the recommended doses, Dr. Sidney M. Wolfe of the Health Research Group emphasized that the margin for error is slim. This group, which is financed by Ralpa Nader's Public Citizen Inc., has petitioned Joseph A. Califano, secretary of Health, Education, and Welfare, either to ban Darvon or to endorse classifying it so that prescription refills can be prohibited.

The witnesses differed sharply on the number of Darvon deaths but agreed that almost always they involve doses higher than recommended.

North Carolina's chief toxicologist, Dr. Arthur J. McBay, attributed "at least 1,200 deaths a year" to propoxyphene, with the majority resulting from "suicidal overdoses." The same state's chief medical examiner, Dr. Page Hunter Hudson, calculated about 1,500 deaths, more than from "any drug, licit or illicit" are caused by the substance.

But Dr. Larry V. Lewman, medical examiner for Multnomah County (Portland) Ore., made a much higher estimate: 3,000 to 4,000 Darvon-related deaths annually. His state and North Carolina are among the few in which physicians, rather than coroners, investigate all violent or unexplained deaths.

Unlike McBay and Hudson, Lewman joined Wolfe in ascribing Darvon deaths less to suicides than to accidents, mainly among cronic users.

Wolfe testified that he had obtained an 11-page Lilly "progress report" on an animal study showing that it knew in February 1976 that nor-propoxyphene (NPX) could cause a serious heart condition.

Even though Lilly told the FDA, in a half-page abstract in March 1977, that it would provide the critical NPX information in a manuscript "now being prepared," it still hasn't done so, Wolfe charged.

In a letter to committee chairman Gaylord Nelson (D-Wis.), Lilly termed the charge "reckless and irresponsible" but Wolfe said the company was being "extremely misleading."

He said that the FDA is "considering a reprimand" to Lilly. An agency spokesman said, however, that the FDA is investigating whether Lilly's failure to supply the manuscript was attributable to a "breakdown in communications" or to conduct requiring a reprimand.