Good Health Hospital: the nosocomial infection report

Read the scenario for what it actually establishes, because it is written to look more settled than it is. It reports six cases, then supplies a table listing four; it says the outbreak *was in fact* caused by spoiled cafeteria food while also saying the litigation question is whether one patient was already symptomatic before admission; and it cites the recommendation to change vendors on the grounds that the present suppliers cannot verify sanitation, which is a procurement failure rather than a demonstrated source. Those are three different epistemic states — a count, a conclusion, and an unresolved question — presented in one paragraph. Noticing the gap between them is the first analytic move, and it is what an evaluation report is for. Opening the report by stating which of the three you are treating as established, and which you are treating as open, also tells the reader what kind of document they are about to read.

The parameters are named for you and they are the standard epidemiological descriptors: *person, time, place, ethnicity, and gender*. Categorising the data into those is not a formatting exercise. Person and time together give you the epidemic curve, which is what distinguishes a common-source outbreak from person-to-person spread; place localises exposure and, in a hospital, separates ward acquisition from community acquisition; and the demographic parameters matter mainly for whether the cases share anything beyond being in the building. With four ages and two sexes and nothing else, the honest finding is that the supplied data cannot discriminate between hypotheses — which is a result, and a defensible one to report. Saying what each parameter would show if the data supported it is worth doing even where it does not, since the report is assessing a compilation and not only reporting one.

The six questions you are asked to propose for the health department are the part where the report either shows method or does not. Good questions here are ones whose answers would change the conclusion: when did each patient's symptoms begin relative to admission, which is the litigation question stated as an epidemiological one; what did the four cases eat and where; were any staff affected; is there a shared ward, shift or procedure; was the organism typed and did the isolates match; and are there comparable cases in other Tampa Bay facilities. Each of those is answerable and each discriminates. A question that cannot change what you would conclude is decoration. A finding that the available records cannot settle the question is also the finding that most directly supports the request for a health department investigation.

Onset relative to admission is the hinge of the whole scenario and deserves its own section. A hospital-acquired infection is conventionally defined by onset after a defined interval from admission, and that interval is what separates an infection the hospital is responsible for from one the patient brought with them. The litigation issue the administrator raised is exactly this question, and the incubation period of the organism is what answers it — which is why the report cannot resolve liability without dates. Say so plainly. An evaluation report that adopts the scenario's own conclusion without noticing that the decisive datum is missing has evaluated nothing. Setting out the definitional interval explicitly also gives the six questions their most important member, which is simply when each patient's symptoms began.

The second deliverable is a contingency plan built from state and federal surveillance data, and it should be structured rather than listed. A workable plan states what triggers activation, who leads, what is done in the first twenty-four hours, how cases are found and defined, what control measures are applied and in what order, what is reported and to whom, and how the plan is stood down. National surveillance data earns its place by telling you which infections are actually common in facilities like yours, so the plan is proportionate to the real distribution rather than to the outbreak that happens to be in front of you. Three academic sources minimum, and five to six pages. Naming who owns each step is what turns a plan from a description of good practice into something an administrator could actually put into effect on the day.