JMIR Publications

ABSTRACT

Background:

Adults with haematological malignancies who develop febrile neutropenia are at high risk of sepsis, intensive care admission, and treatment disruption. Early recognition is difficult because classical clinical signs may be muted or delayed. Commonly used bedside tools such as MASCC, CISNE, NEWS 2, and qSOFA were not developed specifically for febrile neutropenia and show limited sensitivity in this immunocompromised group. There is a need for a nurse led sepsis risk score tailored to febrile neutropenia that combines clinical, laboratory, and continuously monitored physiological data.

Objective:

This protocol describes the development and temporal validation of a nurse led scoring system for early identification of Sepsis. 3 sepsis among adult patients with febrile neutropenia, using routinely available clinical and laboratory variables plus physiological data from an AI enabled remote patient monitoring system, and comparing its diagnostic accuracy with MASCC, CISNE, NEWS 2, and qSOFA.

Methods:

This two phase prediction model study will be conducted in the haematology and bone marrow transplantation department of a tertiary hospital in Bengaluru, India. Phase I is a nested case–control study within a retrospective cohort of about 140 febrile neutropenia episodes. In this phase, demographic, clinical, laboratory, and remote monitoring variables will be analysed with multivariable conditional logistic regression to develop a bedside sepsis score, and internal validation will use 1,000 iteration bootstrap resampling to obtain optimism corrected AUC estimates. Phase II is a prospective temporal validation cohort in which consecutive adult febrile neutropenia episodes will be enrolled at the same centre. At the onset of febrile neutropenia, nurses will calculate the new score and comparator scores (MASCC, CISNE, NEWS 2, qSOFA), and Sepsis 3 sepsis within 14 days will be adjudicated by blinded clinicians. Diagnostic accuracy will be estimated, and comparative performance versus existing tools will be examined.

Results:

Ethics approval was obtained in April 2025. The study was prospectively registered with the Clinical Trials Registry of India (CTRI/2025/06/088290) in June 2025. Phase I retrospective data extraction for model development and internal validation began in May 2025 and is projected to be completed by October 2025. Phase II temporal validation is planned from December 2025 to September 2026, with main results expected to be submitted for publication in 2027.

Conclusions:

This protocol describes a nurse led, febrile neutropenia–specific sepsis score that combines clinical and laboratory predictors with continuous remote patient monitoring in haematology and bone marrow transplantation care. The score is designed as a practical bedside tool for early sepsis risk stratification and will be compared with MASCC, CISNE, NEWS 2, and qSOFA. If it shows good discrimination and calibration, it could support earlier sepsis recognition, more consistent escalation of care, fewer interruptions to chemotherapy, and serve as a basis for later implementation and external validation studies. Clinical Trial: Clinical Trials Registry-India (CTRI), CTRI/2025/06/088290. Registered on 5th June 2025.