ORDER
Vietnam veteran Richard Espinosa appeals from the district court’s judgment, issued after a bench trial on his claim under the Federal Tort Claims Act (“FTCA”), 28 U.S.C. § 2671-80, that doctors at the Veteran’s Administration (“VA”) hospital committed in malpractice while treating him. He argues that the district court clearly erred in finding that the VA doctor’s negligence did not cause his injuries. Because the court’s decision rests on a reasonable credibility finding, we AFFIRM.
Espinosa served the United States Army in Viet Nam and Cambodia from 1969 to 1971. He received an honorable discharge, but upon his return to civilian life began suffering psychiatric problems. In the 1980s he was diagnosed with post-traumatic stress disorder. His tours of duty and subsequent psychiatric problems seem to have exacerbated his alcoholism, which he has struggled with since the early 1960s. Although sober now, Espinosa drank heavily for over twenty years. Since losing an eye and suffering a skull fracture in a 1975 car accident, Espinosa has also shot heroin intravenously, sometimes sharing needles.
*403In September 1994, Espinosa vomited blood. He was admitted to a VA hospital, where an endoscopy (an internal examination performed with a light and camera attached to a tube) revealed esophageal varices. Esophageal varices are dilated veins found at the bottom of the esophagus and the top of the stomach caused by increased pressure in the veins passing through the fiver. This increased pressure sometimes causes the veins to burst and bleed. In Espinosa’s case, the varices were traceable to his cirrhotic fiver, which constricted the veins passing through it, increasing the pressure blood exerted on their walls.
The VA doctors treated Espinosa’s var-ices through sclerotherapy, i.e., administering a scarring agent that toughens the veins and hopefully prevents re-bleeding. He was discharged in October 1994, with the specific instruction not to take the over-the-counter medication Motrin. It is unclear from the record why exactly Espinosa was told not to take Motrin, but Motrin’s active ingredient, ibuprofen, can cause fiver damage and ulcers.
About two weeks after his release, Espinosa was admitted to a different VA hospital; this time he was experiencing flashbacks and nightmares. About a month after his admission, on November 30, 1994, he told a nurse on the Posh-Traumatic Stress Unit that his back was bothering him. The nurse called a doctor, who, despite knowing that Espinosa had been told not to take Motrin, gave Espinosa a prescription for 600 mg ibuprofen pills. A VA pharmacist incorrectly filled this prescription, giving Espinosa 800 mg pills. Espinosa took one of these pills and several hours later began vomiting blood again. He was taken to the emergency room, where another surgical endoscopy revealed that his varices were more inflamed than during his previous endoscopy and that one of them had welts and a bright, cherry-red spot on it. According to the attending physician, the spot and welts suggested recent bleeding. VA doctors treated Espinosa with additional sclerotherapy and a blood transfusion, and again ordered that Espinosa not take non-steroidal anti-inflammatory drugs like ibuprofen.
Espinosa later brought this suit under the FTCA, charging that the VA doctor negligently prescribed ibuprofen and that the VA pharmacist negligently filled that prescription. He alleged that this negligence caused his second spate of internal bleeding. He also charged that the blood transfusion he received during the second endoscopy infected him with Hepatitis C, further damaging his already cirrhotic fiver.
During a bench trial, Espinosa presented the testimony of Dr. William Wehrmacher, a cardiologist who opined that the VA doctor breached the prevailing standard of care by prescribing ibuprofen to Espinosa when he knew or should have known that ibuprofen was contraindicated for someone with Espinosa’s condition. Dr. Wehrmacher further opined that the VA pharmacist negligently filled the prescription. Dr. Wehrmacher concluded that the combined negligence of the VA doctor and pharmacist led to Espinosa’s internal bleeding because, in his opinion, ingestion of the 800 mg ibuprofen tablet “irritated” or “inflamed” the wall of the spotted varix, precipitating the bleed. Dr. Wehrmacher based this conclusion partially on the emergency-room physician’s pre-endoscopy diagnosis, which states “Patient with upper GI bleed ... with recent use of nonsteroidals [i.e. ibuprofen], with previous history of peptic ulcer disease.” He also based it on three notes from Espinosa’s medical records that had been written some time after the November bleed, which all recite that the Motrin induced *404Espinosa’s bleeding. Finally, Dr. Wehrmacher stated that Espinosa likely contracted Hepatitis C from the blood transfusion.
The government countered with the testimony of Dr. Stephen Hanauer, a gastroenterologist at the University of Chicago. According to Dr. Hanauer, the VA doctor did not breach the standard of care by prescribing ibuprofen: Espinosa at the time was also taking Tagamet, a drug that reduces stomach acid and the- risk of gastrointestinal bleeding and that should have acted as a prophylactic against the effects of the ibuprofen. Dr. Hanauer also concluded that the ibuprofen did not cause Espinosa’s November 30 bleed. Ibuprofen’s impact, Dr. Hanhauer stated, is “systemic,” meaning that its damaging effects are more noticeable in patients who have taken it continuously for extended periods. It is therefore unlike aspirin, which can cause “topical” damage simply through contact with stomach lining. Although long-term ibuprofen use can cause bleeding ulcers, according to Dr. Hanhauer, a single dose would not cause varices to bleed. That Espinosa had received sclerotherapy did not sway Dr. Hanhauer from this conclusion; sclerotherapy is an ongoing process, not a one-time cure, and even treated varices sometimes simply start bleeding when pressure from the cirrhotic liver becomes too great.
Furthermore, neither the pre-operative diagnosis nor the medical histories relied upon by Dr. Wehrmaeher caused Dr. Hanhauer to doubt his conclusions. The emergency room doctor’s pre-operative diagnosis did not state that ibuprofen actually caused the bleed, and ibuprofen was not mentioned as a potential cause in the postoperative diagnosis, which focused on the esophageal varices. And the medical histories, continued Dr. Hanhauer, were likely the product of medical personnel recording what Espinosa told them, not actual diagnoses. Lastly, Dr. Hanhauer stated that Espinosa’s intravenous heroin use, and not the blood transfusion, likely led to his Hepatitis C.
The district court found that the VA doctor and pharmacist had breached a duty of care owed to Espinosa by writing and incorrectly filling the ibuprofen prescription. But the district court further concluded that Espinosa had not established by a preponderance of the evidence that the ibuprofen caused his internal bleeding. Espinosa’s expert witness was simply not credible on this point, explained the court, which instead credited Dr. Hanauer’s testimony regarding the negligible effects of a single ibuprofen dose. The court further found that Espinosa likely contracted Hepatitis C from sharing infected needles. The court therefore returned a judgment for the government, which Espinosa has appealed.
On appeal, Espinosa challenges only the district court’s finding that the ibuprofen did not cause his internal bleeding. He argues that the court clearly erred by discounting the emergency room physician’s pre-operative diagnosis and the three medical history entries that attribute the bleeding to Motrin. “It is simply not credible to believe,” continues Espinosa, “that Government physicians would consistently repeat an allegedly incorrect cause for Mr. Espinosa’s 1994 bleeding if they did not believe that the Motrin intake was the cause of the bleeding.”
Under the FTCA, the United States is liable for government employees’ negligent acts to the same extent that a private person would be under the law of the state where the cause of action arose. See 28 U.S.C. § 2674. Espinosa alleged malpractice in Illinois, which required that he show (1) the proper standard of care by which a physician’s conduct may be measured, (2) a negligent failure to comply *405with the applicable standard, and (3) a resulting injury proximately caused by the physician’s lack of skill or care. Donais v. United States, 232 F.3d 595, 598 (7th Cir.2000) (applying Illinois law). This court reviews a district court’s findings of fact after a bench trial for clear error. Id.
Espinosa’s argument that the district court clearly erred by crediting Dr. Hanauer’s testimony regarding the ibuprofen’s effects is unpersuasive. As Dr. Hanauer explained, ibuprofen has only an indirect effect on esophageal varices: it can further damage the already damaged liver that causes them. Furthermore, the medical records relied upon by Espinosa to show that the ibuprofen caused his bleed are not reliable diagnoses. The emergency room physician’s pre-operative diagnosis does not actually say that the ibuprofen caused the bleeding, and the absence of any mention of ibuprofen in the post-operative diagnosis — created after the physician had examined Espinosa internally and formed some opinion regarding the source of the bleed — seems to undercut a finding that ibuprofen caused it. And the other three notes from the medical records relied upon by Dr. Wehrmacher appear to reflect Espinosa’s own self-diagnoses, dictated by Espinosa himself and dated either months or years after the November 30 bleed. Espinosa did not produce any evidence — such as testimony from the medical personnel who wrote the histories— that calls Dr. Hanauer’s characterization of the histories into question.
In short, the district court made a credibility finding, as it was entitled to do, see, e.g., Keller v. United States, 58 F.3d 1194, 1199 (7th Cir.1995), and Espinosa has not produced any evidence undermining that finding. We therefore AFFIRM.