Estimands and estimators of two-level methods using return to baseline strategy for longitudinal clinical trials with incomplete daily patient reported outcomes