The volume and diversity of healthcare data available using modern technology offers great potential for improving health services delivery. Giving clinicians and health system administrators the ability to easily investigate and analyze data from various perspectives can promote evidence-based decision-making. Current analysis approaches often result in process models where essential relations are difficult to depict and/or discern. Moreover, it is not easy to change the level of detail in order to accommodate user requirements by allowing them to analyze data from various perspectives or capture temporal aspects of the data. Further, inherent differences between patients and the vast variety of healthcare settings, even for one patient, make process mining extremely difficult. In this paper, we first discuss community detection methods which, together with filtering techniques based on dimensional modeling and ontologies, allow us to obtain “contextual” insights from event log data using what we call “contextual process mining”. Then, to capture time-dependent relations in patient data, we propose a linear temporal logic-based language, LTLEOT$$LTL_{EOT}$$ (LTLEOT$$LTL_{EOT}$$: Event-Ontology-Time Linear Temporal Logic), which can be used to express both time and order-dependent conditions. The LTLEOT$$LTL_{EOT}$$ formulas are used to filter the event log data to find those patients who satisfy the conditions, thus capturing “temporal” insights in event log data. Both the investigations using community detection methods and those using the temporal logic methods are validated learning approaches: the first allows the user to experiment with the level of abstraction, while the second allows the user to experiment with the temporal logic formulas. This approach can give healthcare professionals insights into patterns of admission, diagnosis, and treatment among patients, which, in turn, lead to the overall goal of this research, which is improved resource management, scheduling, and other aspects of health services delivery. In a simple case study, we show this approach leads to an improved understanding of healthcare processes among comorbid patients.