5. Clara
CLARA
The facility runs a skeleton crew on Tuesday nights.
I know this facility from the inside out.
Not because someone told me, but because I took it apart myself until I understood how it worked.
Shift schedules cross-referenced against payroll cycles.
Security rotation mapped against camera coverage gaps.
Waste processing intake logged against the hospital's Tuesday overflow volume. Fourteen months of observation.
Nora handles the perimeter, I handle the interior. We have done this eleven times. We do not speak unless something changes.
Tonight, something changes.
The access corridor on the east side, the one that runs parallel to the loading bay and feeds directly into the overflow processing room, has been reorganized.
Not closed. The door is still there, the lock is still the same, the camera blind spot I have been using for four months still exists.
But the routing on the other side has shifted.
The overflow inventory that was staged in processing room three has been moved to processing room one, through a pathway that adds two minutes to my extraction window and passes within eight meters of the facility's active monitoring station.
I stand in the corridor for four seconds, then I adjust.
Room one is accessible from the north stairwell.
The stairwell camera has a seven-second sweep cycle I have never needed to use before because I have never needed the north side of this building.
I use it now. I count the seconds, move on the blind spot, and reach the room without incident.
The KD-9 inventory is there, a smaller cache than expected, but enough.
I take what I came for, reseal everything to its prior state, and leave through the north stairwell.
Nora is in the car when I reach it. She reads my face before I say anything.
"Problem?"
"Adjustment." I set the bag in the back and close the door quietly. "The east corridor routing changed. Someone moved the overflow staging from room three to one."
She is quiet for a moment as she pulls out of the lot. "Maintenance?"
"Maintenance doesn't touch inventory staging. That's a logistics decision." I watch the facility recede behind us. "Someone who understands how the overflow system works reorganized a single access point without disrupting the rest of the operation. And they left the corridor open."
Nora says nothing, she understands what I am not saying.
This is not the move of someone running a standard security review.
A standard review would have closed the corridor, added a camera, flagged the blind spot.
This left the blind spot intact and simply moved the target.
Whoever did it knows I am here, and they chose not to stop me.
I file this on the drive back and do not look at it again until I am ready.
Renata is worse.
Nora reaches her room first while I'm logging the supply run, and by the time I get through the door Nora has already taken her temperature and is reading the joint swelling with the practiced efficiency of someone who has done this assessment a hundred times.
She doesn't need to say anything. I can see it from the doorway.
The flare has accelerated, Renata's joints swollen past the point of functional movement, her temperature running two degrees above where it was three days ago.
I set up the line without discussion and draw the KD-9 while Nora adjusts the monitors.
The drug performs as it always has, not dramatically, not immediately, but in steady increments, a tide rather than a wave.
Within two hours, Renata's temperature drops.
Within three, she can move her fingers without wincing.
She sleeps. I stay until I am sure.
At some point Nora leans in the doorway with her arms crossed and says, "You know what happens if someone figures out you're the only one keeping her alive."
I don't answer. She isn't asking a question. She is naming a risk we both carry and have chosen not to put down. Answering it would only confirm that I've thought about it, which she already knows.
"Go home," she says. "You have a shift in six hours."
Hargrove finds me before my second patient the next morning. He falls into step beside me in the corridor with an unhurried ease, as though the approach has been precisely timed and does not intend for it to show.
"The Delgado file," he says. "The discharge notes reference a medication administered during the second overnight. There's no corresponding pharmacy order."
I keep walking. "I cross-referenced the dosage with his allergies before administering. The order was verbal, confirmed with the attending nurse. I logged it at sign-off."
"There's no sign-off notation in the system."
"Then there's a documentation gap." I stop at the door to exam room two and face him. "Is there a clinical concern about Mr. Delgado's recovery?"
Hargrove holds my gaze. He is a precise man, measured in his confrontations. Someone who has worked long enough to know which battles cost too much and has never once confused caution for weakness. He is also, I have always known, more perceptive than he lets on in formal settings.
"No," he says. "He's recovering well."
"Then I'll review the documentation and correct the gap." I push the door open. "Thank you for flagging it."
He nods once and moves on. Neither of us says what is sitting in the corridor between us. That this is the third documentation anomaly in two months, that he has flagged each one with the same careful neutrality, and that we are both performing a conversation neither of us is actually having.
What Hargrove is doing, I understand, is giving me rope.
Not to hang myself, he is not that kind of man.
He is giving me the length of it and watching what I do, and the fact that he has not taken this to the department head means something I do not take lightly.
He is extending a professional trust that I am spending on people whose names are not in any of his files.
I am aware of this every time he flags something and every time I give him nothing back.
I do not know how long his patience will hold, but it will not be indefinitely.
The meeting request from Kade Biologics arrives at 4:17 p.m.
It is framed as a follow-up on the trauma research initiative.
A second consultation on the intake methodology, with a proposed date two weeks out.
His assistant's name is on the message, but the framing is his.
I recognize his construction in it. Direct, no social cushioning, the question embedded in the structure of the meeting rather than stated outright.
I read it twice. I think about the access point that was reorganized but not closed.
I think about a mind that understands institutional architecture well enough to name its failures from first principles, not from a position paper.
I think about how he registered my analysis in that office.
Not with professional courtesy, but with the quality of attention that someone pays when they have encountered something they were not expecting to find.
I open the reply window.
There are four reasons to decline this meeting. Operational security. Professional distance. The fact that his company's supply chain is the source of half my clinic's pharmaceutical inventory. The fact that I have spent fourteen months building a case against him.
There is also the reorganized access point.
Someone with executive-level authority over that facility's logistics made a quiet adjustment that left my operation intact.
I do not know if it was him, I have no way to know.
But I am aware that I am not declining this meeting, in part, because I want to find out.
I accept before I have finished listing the reasons not to.
The laptop stays open after I send the reply. I navigate — not deliberately, without quite deciding to — to a search that has no connection to the supply chain or the clinical trial records or the distribution architecture.
An interview, archived, from twelve years ago.
Pre-empire, pre-Kade Biologics as it currently exists.
A photograph of him at a conference table, younger, without the controlled stillness he carries now.
He is looking at something outside the frame.
His expression is unguarded, nothing like any photograph taken after a certain point in his career.
I read the interview. It is about pharmaceutical access reform, the argument that treatment discontinuation decisions should require independent ethical review. His position is articulate, sourced, uncompromising in its conclusions.
His company buried KD-9 years after this interview was published.
I close the tab, then I close the laptop.
I press my thumb to the inside of my wrist, hold it there, release. A small, deliberate physical act. Not comfort, just the body confirming its own weight.
I do it twice more.
The laptop sits closed on the desk. The meeting is in two weeks. The interview is archived somewhere I have already decided not to revisit.
I leave it all exactly where it is.