WASHINGTON (AP) _ Serious medical problems at veterans hospitals, including unsupervised interns and inexperienced medical personnel, have caused some patients to die, congressional investigators charge.

''I examined chart after chart where attending physicians rarely examined patients or did not examine them at all,'' said Mary Ann Curran, a congressional health care investigator. ''Nurses allowed patients with life- threatening illnesses to languish for hours, even days, without monitoring.''

Curran told a House subcommittee Wednesday that she found serious medical problems at all six Department of Veterans Affairs hospitals she visited. And she said a broader examination of records found 30 VA hospitals had high numbers of patient complications and other indicators of substandard care.

VA officials disputed the findings.

''At this moment, we really do provide first-rate care,'' Dr. James Holsinger Jr., chief medical director of the department, said in anticipation of the hearing.

Curran, a health care specialist for the congressional General Accounting Office, conducted her investigation for the House Government Operations subcommittee on human resources and intergovernmental relations.

She told the panel that the most serious problem found at the six medical centers was the lack of supervision of residents and interns, a problem she said had ''severe consequences for patients.''

''We discovered several cases of patients who had died because of errors made by unsupervised interns or residents,'' she testified.

Curran said the investigation also found VA doctors who were untrained and inexperienced in emergency responses, such as resuscitation, and that basic infection control was inadequate. She said many problems were caused by ''the lack of continuity of care to patients.''

''In one case, a patient lost a leg because he was not checked regularly. In another case, a patient's body was cold upon examination by a doctor, indicating that the patient had been dead for several hours before anyone discovered him,'' she said.

Curran said the worst case she found was at Cheyenne, Wyo., where a patient was diagnosed for bladder cancer and scheduled for surgery. While the patient was awaiting surgery the staff urologist resigned in a contract dispute, and the patient was allowed to go untreated for the cancer for 45 days.

When his condition was recognized by a surgical officer, the patient had lost 30 pounds and his cancer had spread to other parts of his body. He was transferred then to the Denver VA hospital, where he soon died, Curran said.

''The patient's demise was expedited by the lack of treatment,'' Curran said. ''The medical care provided to this patient, who was unnecessarily restrained in bed against his will, whose suffering was allowed to continue, amounted to negligence.''

Curran added that this ''is not an isolated case, but an example of uncorrected systemic problems.''

Rep. Ted Weiss, D-N.Y., the subcommittee chairman, said, ''There is a pervasive attitude among the hospitals and the central office to cover up findings of poor patient care.''

The six medical centers studied were at Cheyenne; Bay Pines, Fla.; Chillicothe, Ohio; Gainesville, Fla.; Milwaukee, Wis.; and Tampa, Fla. The VA has more than 170 hospitals.

Holsinger told the subcommittee that he has begun management reforms aimed at improving quality since taking the job last year.

''Our system is obviously not perfect - no health care system is,'' he said. ''Our patients are older, sicker and more complex than the average patient.''

One of the hospitals listed as having a high number of poor quality-of-care indicators, at Richmond, Va., was run previously by Holsinger.

Donna St. John, a spokeswoman for VA medical system, said in a telephone interview that the department took issue with Curran's methodology in her comparison of VA hospital records. She said the investigator had not compared the hospitals with non-VA hospitals and did not take into account the special nature of VA hospitals, some of which are referral centers with heavy caseloads of the sickest patients.

''We think there are limitations,'' St. John said. ''The analysis is not weighted to adjust for age, morbidity or volume'' of cases.