Identification of asthma triggering conditions based on medicament device monitoring for a patient

A non-transitory computer readable storage medium comprising computer program instructions that when executed by a computer processor cause the processor to:

• receive, from a medicament device sensor removably attached to a rescue inhaler unit and configured to monitor medicament usage of the rescue inhaler unit, a plurality of signals, wherein each signal of the plurality of signals describes a rescue inhaler usage event recorded by the medicament device sensor when the rescue inhaler unit dispense a rescue medication to the patient;
• access, from one or more remote servers, measurements of one or more environmental triggers of a rescue medication event recorded during each rescue inhaler usage event described in the plurality of signals;
• encode a plurality of feature vectors, wherein each feature vector of the plurality of feature vectors corresponds to a rescue inhaler usage event described in the plurality of signals and is encoded with patient data for the patient and a value assigned to each of the one or more environmental triggers based on the measurement of the environmental trigger recorded during the rescue inhaler usage event, wherein the value for each of the one or more environmental triggers describes whether a trigger condition for the environmental trigger was present when the rescue inhaler usage event was detected, the trigger condition representing a subset of measurements for the environmental trigger;
• input the plurality of feature vectors encoded with patient data for the patient into a function trained, for each trigger condition, to:
• categorize the plurality of feature vectors into a first subset of feature vectors where the trigger condition is present and a second subset where the trigger condition is not present;
• and
• determine, for the patient, a relative risk score for the trigger condition based on a count of feature vectors in the first subset relative to a count of feature vectors in the second subset;
• for each trigger condition,
• determine a sensitivity of the patient to the trigger, wherein instructions for determining the sensitivity of the patient to the trigger condition cause the processor to:
• determine, by a statistical model, a weighted normalization term based on a similarity between the patient and each secondary user of a plurality of secondary users in an aggregate dataset, the relative risk score for the trigger condition, and a number of days during which the trigger event was present and;
• compare the weighted normalization term against a set of confidence intervals to determine whether the relative risk score for the trigger indicates that the patient is sensitive to the trigger;
• and
• generate, for presentation to the patient, a notification describing the sensitivity of the patient to the trigger condition and suggestions for preventing rescue inhaler usage events when the trigger condition is satisfied.