Health authorities are usually interested how many patients are newly diagnosed in a calendar year. Or perhaps a quarter or a single month depending on the actual question. The time period doesn't matter as long as it is clearly stated, and there cannot be guidelines on that. What you possibly mean is not recently diagnosed, but recently infected (incident infection). As incidence is usually given as new infections per investigated population per year it makes sense to also define the recent infection period as one year. Standardization on an annual basis is needed if you want to compare recent infection rates derived from different tests.