Development of a Fever Detection Algorithm Based on Non-invasive Skin-based Sensor Values in Infants up to 18 Months of Age

NCT ID: NCT04654559

Last Updated: 2024-03-15

Basic Information

• Recruitment Status: SUSPENDED
• Clinical Phase: NA
• Total Enrollment: 50 participants
• Study Classification: INTERVENTIONAL
• Study Start Date: 2020-04-07
• Study Completion Date: 2025-03-30

Brief Summary

Fever of infants up to 18 months of age will be monitored in the hospital using the standard clinical methods and wearable research prototypes. These research prototypes will be measuring continuously and non-invasively skin based parameters, with which the feasibility of developing a fever detection algorithm will be investigated.

Detailed Description

Background and Rationale:

In daily clinical practice, core body temperature (CBT) of infants with fever symptoms is monitored using sporadic rectal measurements. Because these sporadic invasive measurements are time consuming for the medical staff and displeasing for patients, an alternative method to assess CBT/fever is needed. The company greenTEG is developing a CBT algorithm which calculates CBT continuously form skin temperature (ST), corresponding heat flux (HF) and other skin-based parameters. The algorithm development will be achieved by collecting skin-based parameters and reference CBT values from infants having fever in a clinical setting.

Objective(s):

Develop and validate an algorithm that allows the detection of fever in infants through a non-invasive sensor system, which calculates CBT from ST, HF and other skin-based data streams, allowing a more effective patient management.

Statistical Considerations:

The measures of quality will be: 1) The mean absolute difference (MAD) between the CBT prediction and the reference signal where the mean is taken over the whole measurement of a single patient. An aggregate performance measure over a group of patients is defined by averaging the MAD values of each patient in the group. 2) the 2σ (standard deviation) range of the Bland-Altmann-Plot between the CBT prediction and the reference signal. This is calculated either for individual patient data or for the combined data of all patients together. As we have defined a group of patients for algorithm validation, the total improvement will be defined by comparing the above aggregate performance measures of old and new algorithm for the validation group. We expect the factors age, sex to influence the algorithmic prediction. Balancing the probability of occurrence of the factors in the population and the overall size of the study, a final size of 50 patients is reasonable.

Study procedures:

Infants will be recruited and screened 1 days before the measurements starts. Two research prototypes will be applied to the patient on the left side of the body (lateral ribcage and foot), after being admitted to the hospital and parents having signed the informed consent. For all infants rectal measurement will be asses every 8 hours, as reference temperature. For infants older than 6 months, in addition to rectal temperature, ear temperature will be assessed every 4 hours with an infrared ear thermometer. The whole measurement procedure will last 18-72 hours, depending on the stationary stay of the individual patient.

Conditions

Fever

Study Design

• Allocation Method: NA
• Intervention Model: SINGLE_GROUP
• Primary Study Purpose: OTHER
• Blinding Strategy: NONE

Study Groups

Unobtrusive data collection

Group Type: OTHER

Unobtrusive data collection using wearable non-invasive sensors

Intervention Type: OTHER

Two wearable non-invasive sensors will be mounted to the patients skin (lateral rib cage and foot). These wearable non-invasive sensors are measuring unobtrusively skin based parameters (no diagnosis will be made), which will be used for fever detection algorithm development..

Interventions

Unobtrusive data collection using wearable non-invasive sensors

Two wearable non-invasive sensors will be mounted to the patients skin (lateral rib cage and foot). These wearable non-invasive sensors are measuring unobtrusively skin based parameters (no diagnosis will be made), which will be used for fever detection algorithm development..

Intervention Type: OTHER

Eligibility Criteria

Inclusion Criteria

• Infants from two weeks after birth and due date up to 18 months of age
• Infants which are staying at the Nursery and the University Children's Hospital Zurich or Children's Hospital of Eastern Switzerland, St. Gallen for more than 1 day
• Informed consent signed by parents or by legal representatives

Exclusion Criteria

• Acute medical contradiction against medical plaster and adhesives (e.g. skin diseases or allergies).
• Irritated or damaged skin (e.g. burn, rush, eczema)
• Infant with congenital anomalies (e.g. malformations of the anus)
• Disease or brain injury which cause a disturbance of the thermoregulation
• Comatose infant's
• Infants with implants (e.g. pacemaker, cochlear implants)

• Minimum Eligible Age: 2 Weeks
• Maximum Eligible Age: 18 Months
• Eligible Sex: ALL
• Accepts Healthy Volunteers: No

Sponsors

University Children's Hospital, Zurich

OTHER

Sponsor Role: collaborator

Children's Hospital of Eastern Switzerland, St. Gallen

UNKNOWN

Sponsor Role: collaborator

greenTEG AG

INDUSTRY

Sponsor Role: lead

Responsible Party

Responsibility Role: SPONSOR

Principal Investigators

Christoph Berger, Prof. Dr.

Role: PRINCIPAL_INVESTIGATOR

University Children's Hospital, Zurich

Locations

Children's Hospital of Eastern Switzerland, St. Gallen

Sankt Gallen, , Switzerland

University Children's Hospital Zurich

Zurich, , Switzerland

Countries

Switzerland

Other Identifiers

6058_Baby_study_1

Identifier Type: -

Identifier Source: org_study_id