A process mining approach for clinical guidelines compliance: real-world application in rectal cancer

Abstract

In the era of evidence-based medicine, several clinical guidelines were developed, supporting cancer management from diagnosis to treatment and aiming to optimize patient care and hospital resources. Nevertheless, individual patient characteristics and organizational factors may lead to deviations from these standard recommendations during clinical practice. In this context, process mining in healthcare constitutes a valid tool to evaluate conformance of real treatment pathways, extracted from hospital data warehouses as event log, to standard clinical guidelines, translated into computer-interpretable formats. In this study we translate the European Society of Medical Oncology guidelines for rectal cancer treatment into a computer-interpretable format using Pseudo-Workflow formalism (PWF), a language already employed in pMineR software library for Process Mining in Healthcare. We investigate the adherence of a real-world cohort of rectal cancer patients treated at Fondazione Policlinico Universitario A. Gemelli IRCCS, data associated with cancer diagnosis and treatment are extracted from hospital databases in 453 patients diagnosed with rectal cancer. PWF enables the easy implementation of guidelines in a computer-interpretable format and visualizations that can improve understandability and interpretability of physicians. Results of the conformance checking analysis on our cohort identify a subgroup of patients receiving a long course treatment that deviates from guidelines due to a moderate increase in radiotherapy dose and an addition of oxaliplatin during chemotherapy treatment. This study demonstrates the importance of PWF to evaluate clinical guidelines adherence and to identify reasons of deviations during a treatment process in a real-world and multidisciplinary setting.

Keywords: clinical guidelines; computer-interpretable clinical guidelines; conformance checking; evidence - based medicine; process discovery; process mining; rectal cancer; treatment pathways.

Figure 1

Graphical representation of ESMO recommendations for rectal cancer treatment in early, intermediate, locally advanced and advanced risk categories. CRT, chemoradiotherapy; SCRT, short-course radiotherapy; TME, total mesorectal excision; TEM, transanal endoscopic microsurgery; Folfox, leucovorin/fluorouracil/oxaliplatin.

Figure 2

Example of rule, which is called trigger in PWF. Each rule is identified with a name (“is Nad RT dose between 45 and 54 Gray A?”) and includes a condition (e.g., if the patient belongs to the early risk group and the event is neoadjuvant radiotherapy with a total dose between 45 and 54 Gray and a fraction dose between 1.8 and 2.0). If the condition is met the patient transitions from one state (“Early”) to another (“Nad RT dose between 45 and 54 Gray A”).

Figure 3

Flow chart showing study inclusion and exclusion criteria. 483 adult patients were diagnosed with non-metastatic rectal cancer. We excluded 30 patients who either did not undergo treatment or had missing information regarding staging. The final dataset includes 453 patients.

Figure 4

Distribution of patients in the four TNM risk categories at the moment of diagnosis: early, intermediate, locally advanced and advanced.

Figure 5

Conformance checking results for Early (A) and Intermediate (B) risk groups. The number of patients who activated each rule, i.e. each transition, is displayed on the Pseudo-Workflow diagrams representing ESMO guidelines’ recommendations: boxes are rules, circles are statuses. Only patients that reach final nodes are patients fully compliant with ESMO guidelines, all others deviate from the guidelines at some point in their treatment pathway.

Figure 6

Conformance checking results for Locally Advanced (A) and Advanced (B) risk groups. The number of patients who activated each rule, i.e. each transition, is displayed on the Pseudo-Workflow diagrams representing ESMO guidelines’ recommendations: boxes are rules, circles are status. Only patients that reach final nodes are patients fully compliant with ESMO guidelines, all others deviate from the guidelines at some point in their treatment pathway.

Figure 7

CareFlow miner graph produced with pMineR software to represent the most common processes of patients not compliant to clinical guidelines. Numbers indicate patients passing through each node. The direct acyclic graph displays only transitions involving a number of patients greater than the threshold, set to 4.