The prep bag is navy blue with a gray zipper and a small UPMC logo embossed near the handle — she did not notice the logo when she bought it at the medical supply station on Craig Street, and she still does not think about it now. She packed it Sunday afternoon: four days before she needed to. A change of clothes she would not need because she is not staying. The consent forms printed twice — the first set from the standard-pathway printer in the lobby had a gray streak across the signature line, so she reprinted from the Adamski-protocol terminal on the second floor, which outputs on heavier stock and includes the protocol identifier watermark in the upper right corner. Her insurance card in the inner pocket. Her CAS-card: a thin plastic rectangle, cream-colored, her patient identifier printed above her monitoring status in gray type — THRESHOLD-STABLE, with the assessment date. She touches the inner pocket each morning to confirm both are still there even though she has not touched either since packing.
The bag is by the front door. It has been there for five days.
She woke at 4:13 AM for no particular reason — or rather, for the reason she has been waking at this hour all week, which is not a noise or a dream but a body that has learned the shape of what is coming and will not wait for the alarm. She lay still for eight minutes. The ceiling was the same off-white it always is. The room was warm. Then she got up.
The kitchen was dark except for the appliance lights: the refrigerator indicator, the CAS-monitor terminal's idle amber glow in the corner, the microwave clock showing 4:21 by the time she got the kettle on. The Adamski notepad is on the counter to the left of the stove — yellow cover, Adamski in green letters across the top, her patient identifier printed below in smaller type because the Adamski protocol requires documentation linking the notepad to the specific patient file. Dr. Adamski's nurse pressed it into her hands at the August consultation, which was held in the IRB-designated room on the third floor rather than a standard exam room, and where the wall-mounted recording system had been switched off first. For any questions that come up before the September imaging window, the nurse said. Write them down.
She has passed the notepad seven times since Sunday. The first pass was the evening she came home with it, when she set it on the counter and then went to make dinner without deciding anything. The second and third were Monday morning, back and forth to the coffee station. By Wednesday she had established a pattern: not slowing, not looking at the cover, not picking it up. The decision not to open it has not required active resistance. It requires only continuation — the same choice, each time, not to stop.
She passed it now without slowing.
The three questions are more accurate held in her head than written down. She has thought about whether to write them and concluded: not yet. Writing makes questions into tasks. A task requires action; a question can wait. These particular questions are waiting for something specific — the imaging results, the post-scan appointment, whatever Dr. Adamski decides to say when she sees the read-out — and writing them down would make them want answering now.
The first question is about trajectory. 84 IU/mL has been her biomarker for two years. The CAS-monitor, which runs her threshold assessments automatically every seventy-two hours, classifies her as threshold-stable: no deterioration, no escalation trigger, monitoring continues on current schedule. The question is whether threshold-stable is a description or a verdict. Whether 84 IU/mL, unchanged across eight consecutive assessments, means nothing is moving or means she is watching a thing that has not yet decided what to do. The CAS-monitor produces a classification. It does not produce an interpretation. She has learned the difference.
The second question is about the Adamski protocol itself. Standard-pathway patients at UPMC are assessed against the CAS-monitor's predictive model, which was trained on a cohort of 2.3 million cases and has a stated sensitivity of 94.7% for her category. She is not a standard-pathway patient. The Adamski protocol covers cases where CAS-monitor's predictive confidence falls below a threshold the IRB documentation describes, carefully, as insufficient for standard-pathway management. What she wants to know is what specific pattern in her case caused that designation. The nurse told her only that her profile matched three criteria and that Dr. Adamski had reviewed the flag personally. She has not been told what the three criteria are.
The third question is about radiological language. She has had four imaging scans before — all at Building C, all read by the same practice. She knows the corridor, the machine's resonance frequency, the copper weight of contrast medium moving through the scan. She has learned to read some of the report language: no significant interval change is different from within expected variance is different from threshold-stable imaging. She does not know all of the distinctions. She wants to ask Dr. Adamski to explain, in advance, what results language the report will use and what each category means for her specific case. She does not want to be reading the glossary function in the patient portal at eleven PM the night the results post.
She did not write any of this down.
The kettle finished. She poured the water over the grounds slowly and watched the bloom — the dark grounds rising with the CO2 and then sinking slowly as it escaped. Same ratio, same method, for eleven years. The stainless filter came from Gdansk in 2015. She was carrying it before the CAS-monitor existed in its current form, before the Adamski protocol had its eight-year IRB extension, before 84 IU/mL became a number she thought about every morning. The coffee is a continuity she maintains without thinking about maintaining it. Water, bloom, four minutes, the same filter. The result is the same cup it has always been.
The CAS-monitor terminal chimed once at 4:27 AM.
She looked at it before she had decided to. The terminal displays routine CAS-monitor outputs on a 72-hour cycle: threshold readings, interval summaries, automated flag notifications. She has read enough of these in the last two years to know the structure before she reads the content — header, current reading, interval comparison, classification. This one had a different header. It read: IMAGING-PROXIMATE REASSESSMENT — AUTOMATED NOTICE.
She read the full text. It was 47 words. The CAS-monitor had generated an automated pre-scan notice indicating that her 84 IU/mL reading, in combination with the upcoming imaging window, had triggered a protocol-standard imaging-proximate reassessment eligibility flag. The notice said this was automatic. It said the flag would be reviewed by her assigned protocol specialist prior to her September 22 appointment. It said no action was required on her part.
She read it again.
The three questions had been precise and she had been holding them cleanly for two weeks. The phrase imaging-proximate reassessment eligibility was not one she knew. It was distinct from threshold-stable — that classification she understood: 84 IU/mL, no movement, monitoring continues. This was something the CAS-monitor generated specifically in the period before imaging in cases meeting certain criteria. She did not know what those criteria were. She did not know whether imaging-proximate reassessment eligible was a standard designation for all Adamski-protocol patients in the week before their scan, or whether it was generated by her specific reading, her specific case profile, the combination of this particular 84 IU/mL and this particular imaging window.
The Adamski notepad was eight inches from her left hand.
She did not reach for it. She picked up the mug instead — still too hot — and held it without drinking. On the terminal, the notice waited in its gray-bordered box. The timestamp said 4:27. The CAS-monitor ran its automated routines on a schedule she had never inquired about: 72-hour threshold cycles, event-triggered flag notifications, imaging-proximate protocols she was now learning were a category. It did not require a response. It had said so: no action required on your part. She had been classifying its outputs for two years — threshold-stable, interval unchanged, monitoring current — and had developed, without exactly deciding to, a read on what the language meant when it appeared. This phrase she could not place.
She dimmed the terminal. The notice stayed in background state. The kitchen returned to appliance-amber and refrigerator hum.
The thing about the Adamski notepad is that it contains information she elected not to read first. Dr. Adamski uses it systematically with her protocol patients — she knows this because the nurse explained that the notepad is calibrated to the pre-scan period: what the imaging is designed to detect, how to read the results language when it posts, what questions are most productive to bring to the post-scan appointment. She had decided, specifically, that she wanted to arrive at the September 22 appointment with her own questions before being shaped by the framing Dr. Adamski had anticipated. It was a decision about epistemic order. She would see the imaging results and then, with those results in front of her, she would ask what she had prepared to ask.
The CAS-monitor notice had not been part of that calculation. It had arrived at 4:27 AM, four days before the imaging window, and introduced a term she did not have. Her three questions remained intact: trajectory, protocol criteria, radiological language. The fourth question was new: what imaging-proximate reassessment eligible meant, why the CAS-monitor generated the flag automatically, and whether her assigned protocol specialist had already seen it or would see it between now and the 22nd.
She stood at the counter for another few minutes. Outside the kitchen window, Pittsburgh was still entirely dark — the hillside across the Cut invisible, the bridge traffic just faint lights moving in the direction of the bridge. She could hear the automated transit line, the electric hum of the early morning run, nearly empty at this hour. The cars run the route regardless: pre-dawn, the hill, the bridge, the flatlands. They do not wait for passengers before departing.
She washed the filter. Dried it. Set it on the rack.
She did not open the Adamski notepad.
The fourth question is held with the first three now. She has it specific: what the phrase means in the CAS-monitor's classification logic, why the system generates it for imaging-proximate periods and whether it applies to all Adamski-protocol patients or to patients with her profile specifically, and what the protocol specialist's review before the 22nd is intended to determine. These are separate questions with different answers. She is holding all of them cleanly.
She went back to bed at 4:43 AM. The kitchen was dark except for the terminal's idle amber. The Adamski notepad was on the counter where she had not touched it.
Four questions. Held and waiting.
She did not slow down when she passed it.