An exemplary healthcare services recommender implementation may be configured to recommend selectively redeemable bundled services comprising care for conditions predicted as likely by a machine learning model based on patient data input. The likely conditions may be predicted for a patient or a patient population comprising a plurality of patients. The machine learning model may comprise a neural network trained to determine a probability distribution of patient condition trends for a future time period based on patient data input. The patient data input may comprise data from an Electronic Medical Record (EMR), diagnosis, lab test, patient-provided information such as survey/symptom data, and/or wearable device data. The machine learning model may be trained using a training data set including at least healthcare outcomes sampled during a past time period from a plurality of patient healthcare episodes each comprising a plurality of procedures performed for the patient in at least one encounter.