PeriPlex

CSC Lead: Haleema

---

This project has been withdrawn. Along with stakeholder resource constraints, an exploratory data analysis demonstrated that the data for this project was too complex and too varied in missingness and meaningfulness to proceed at this point in time. There are currently no plans to revisit this project.

---

The process for determining patient suitability for elective surgery is currently fragmented and occurs close to the procedure date, which leads to cancellations and delays for high-risk patients.

Perioperative patient complexity assessment aims to develop a Machine Learning algorithm that categorises patients into three groups based on their perioperative risk. The algorithm will gather patient data from various sources, enabling appropriate pathway planning and optimisation. Ultimately, it will reduce cancellations and delays for high-risk patients. Additionally, low-risk patients will experience reduced waiting times as they can be fast-tracked to a non-admitted model of care, such as a high volume low complexity hub or day case setting.

The initial platform will cover an initial pilot at GSTT in Gastrointestinal Medicine and Surgery, and Orthopaedics, with a view to scaling in other specialties and Trusts if successful. The algorithm will be tested on approximately 200 patients across the two surgical specialties. The number of patients will be defined based on feasibility and validity of a three-month pilot study.

Clinical lead: Ramai Santhirpala

Rationale

The current process for determining patient suitability for elective surgery is fragmented and occurs close to the procedure date, leading to cancellations and delays for high-risk patients.

Patient pathway

Currently, suitability for surgery is determined at the pre-assessment stage, which occurs 1-2 days before the scheduled procedure. This process is not optimised and lacks information from various sources.

Training data

EHR records, workflow instructions, structured data contained in EDW.

Risks

Delayed optimisation of surgery planning for high-risk patients can lead to increased morbidity, mortality, and negative patient outcomes. Inappropriate stratification can also lead to increased waiting times for both low and high-risk patients, and inefficient allocation of clinical resources.

Goals

Develop an algorithm that categorises patients into three groups based on their perioperative risk, enabling appropriate pathway planning and optimisation.

Success criteria

Reduction in cancellations, reduced waiting times for low-risk patients, and optimised management for high-risk patients.

alternatives

Currently no commercial products identified.