On Sept. 28, an ambulance carrying Duncan pulled into the hospital's emergency bay. It was just after 10 a.m.
Not 55 hours earlier, Duncan had come to the same emergency room complaining of a headache and abdominal pain. His temperature spiked to 103 at one point, and on a scale of one to 10, he rated his pain as an eight.
Doctors ran CT scans of his head and abdomen and did extensive blood tests before deciding it must be sinusitis. They sent him home with a course of antibiotics and told him to follow up with a doctor the next day.
A nurse's note said Duncan told her he recently had arrived from Africa. Somehow, that information did not make it to the attending physician.
Now, Duncan was back, only this time his symptoms included vomiting and diarrhea. His temperature was 103.1 degrees.
This time, the nurse's notes made it clear that Duncan had "just moved here from Liberia." This time, the doctor got the message.
"I followed strict CDC protocol," wrote Dr. Otto Javier Marquez-Kerguelen, referring to the U.S. Centers for Disease Control and Prevention guidelines for treating potentially infectious patients. Wearing a mask, gloves and full gown, Marquez-Kerguelen began his examination and took a history.
"Pt states he has not been to any rural areas or funerals recently," he noted. "Pt denies any sick contacts. Pt denies chills. The pt does not do (sic) any other associated signs of sx (symptoms) at this time."
Among the possible diagnoses: malaria, gastroenteritis, influenza and Ebola.
Duncan was put into isolation. The nurses notified county officials; the doctor called the CDC.
By evening, Duncan was suffering from explosive diarrhea, abdominal pain, nausea and projectile vomiting. Efforts to bring down his fever failed.
Three more doctors were put on the case. Duncan received intravenous fluids to counter the dehydration, but there was still no firm diagnosis.
Because Duncan had recently traveled from Liberia, Dr. Gebre Kidan Tseggay noted, of differential diagnoses.
Shortly before noon on Monday, Sept. 29, Duncan asked the nurse to put him in a diaper, "because he feels too tired to keep getting up to the bedside commode." His fever spiked again to 103, and Duncan was wracked with chills.
"Pt said he just doesn't feel good and doesn't want to stay in the hospital and expressed concern that the doctor had not been here to tell him what was going on," a nurse wrote.
Blood tests showing damage to the liver and kidneys, and fluctuating blood sugar levels kept doctors scrambling. Tests ruled out influenza, hepatitis, parasites and C-diff, the germ notorious for spreading diarrhea in hospitals and nursing homes.
"Feels miserable. Says he is suffering," Dr. Oghenetega Abraham Badidi wrote in the chart. "The patient seems to be deteriorating."
Finally, at 2 p.m. on Sept. 30, doctors received the confirmation that all had been dreading: "Patient has tested positive for Ebola ..." The staff attending to Duncan traded their gowns and scrubs for hazmat suits and attendants would scrub the room with bleach.
By Oct. 1, sepsis had set in. Doctors ordered aggressive IV care in hopes of preventing kidney and liver failure.
Duncan tried to remain upbeat.
"Pt requested to watch an action movie," the doctor noted. "States he is feeling better."
Duncan told his nurse he wanted to try solid food, then refused his lunch tray.
The next morning, Duncan told his attendants that his abdominal pain had lessened. He said he was "trying to keep up good spirits and 'stay strong.'"
But there was now blood in his urine. Worried about his lung function, doctors added advanced antibiotics to Duncan's medication.
That afternoon, he ate his crackers and drank less than 2 ounces of Sprite.
In the morning note on Friday, Oct. 3, a nurse-practitioner wondered, "concern for liver failure?"
Nurses were urged to tempt Duncan with applesauce, bananas and other bland foods; at one point they offer ice cream. A nutritionist wanted to start total parenteral nutrition, a kind of IV feeding used in the severely ill, and the doctors sought advice from the CDC.
That afternoon, a doctor's note declared Duncan's kidney function "much worse."
Stymied, physicians contacted Chimerix, a small pharmaceutical research firm based in Durham, North Carolina. They wanted to try the firm's experimental antiviral drug, brincidofovir.
The Food and Drug Administration gave its blessing.
"They were not holding anything back," said Dr. Amesh Adalja, an infection disease and critical care specialist at the University of Pittsburgh Medical Center who reviewed the records for the AP. "They were trying to support every organ system."
Just after midnight Oct. 4, nurses noted that "patient is restless. Coughing." Duncan's oxygen levels were dropping, and he went into multiple organ failure.
"Patient's condition is life-threatening and without immediate intervention would deteriorate," a doctor wrote. They placed a tube in Duncan's airway to help him breathe.
The body can compensate for a while, Adalja said. "But then "you go over this cliff."
"He's not slowly coming to a wreck, veering off the road. It's crashing," Adalja said.
Later that morning, a shipment of brincidofovir arrived and Duncan got the first dose.
The following day, Sunday, Oct. 5, the Texas doctors consulted with their colleagues at Emory University in Atlanta about additional options. Emory had cared for three Ebola-stricken aid workers airlifted there from West Africa; two had recovered and the other was stable.
One of those survivors, Dr. Kent Brantly, had donated blood for other patients, in hopes that his antibodies might help them fend off the virus. But Brantly's blood was not a match for Duncan's. There would be no transfusion for the Liberian.