1/6 My SEIR model follows the daily data for patients in hospital and ICU. Like ALL other epidemiological models, it is sensitive to many parameters. Nevertheless, it gives trends and the opportunity to test different strategies for leaving the confinement. 
2/6 [ACTUAL SCENARIO = CONFINEMENT] Assuming that Re is around 0.8 (as claimed by @sciensano and neglecting the current fluctuation around 1.2) and remains unchanged after May 4, the situation in ICU will improve slowly. Still above 40% of total ICU capacity in June.
3/6 [UNREALISTIC SCENARIO] In an extreme situation : Re takes the value before #lockdown after May 4, the ICU are saturated within 12 days. Not a surprise since it is the average time to develop the disease and to be transferred to hospitals.
4/6 [REALISTIC SCENARIO] If Re increases by 50% after May 4, a plateau is obtained for May-June before growing again during the summer.
5/6 [REALISTIC SCENARIO] If Re increases by 100% after May 4 (doubling contacts), we recover the critical situation in hospitals by the end of June.
6/6 CONCLUSION : if we assume that Re will be affected after May 4 by leaving confinement, additional periods of confinements later in the year are necessary to avoid ICU saturation.
