A pressure sore risk early warning device based on internet of things near-infrared tissue oxygenation sensing and body position recognition
Patent Information
- Application Number
- CN202610998233.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-06
- Publication Date
- 2026-08-28
AI Technical Summary
[0007]本发明的目的在于提供一种基于物联网的近红外组织氧合感知与体位识别的压疮风险预警装置及方法,用于解决现有单一体位监测、单一组织氧合监测和人工巡视难以及时、连续、共定位解释局部受压暴露与组织缺氧关系的问题
[0024] 1. IoT patch deployment: The patch itself has sensing, computing, wireless transmission and device identification capabilities, and can be connected to the information platform as an independent IoT device.
Smart Images

Figure FT_1 
Figure FT_2 
Figure FT_3
Abstract
Description
Technical Field
[0001] This invention belongs to the technical fields of IoT medical devices, wearable physiological monitoring, near-infrared spectroscopy tissue oxygenation monitoring, and pressure ulcer prevention. Specifically, it relates to a waterproof, skin-friendly flexible patch, an information platform, and a pressure ulcer risk warning device and method that work in conjunction with an IoT wristband or watch terminal. The patch, acting as an IoT sensing terminal, is attached to bony prominences such as the sacrum, greater trochanter of the femur, heel, ischial tuberosity, scapular region, or lateral malleolus. It collects local tissue oxygenation and positional data in situ and automatically uploads it to the information platform via a built-in wireless transmission module. After receiving, storing, and processing the data, the information platform synchronously transmits tissue oxygenation values, positional status, alarm information, and treatment suggestions to the IoT wristband or watch terminal, which then receives, prompts, and confirms the data on-site. Background Technology
[0002] Pressure ulcers, also known as pressure injuries, are localized tissue injuries resulting from the combined effects of prolonged pressure, shear force, impaired local perfusion, tissue hypoxia, abnormal microclimate, and decreased individual tissue tolerance. They commonly occur at bony prominences such as the sacrum, coccyx, greater trochanter of the femur, heel, and ischial tuberosity. Early pressure ulcers do not necessarily present as skin breaks; in some cases, the skin surface remains relatively intact, but subcutaneous tissues or deeper soft tissues have already experienced decreased perfusion, insufficient oxygen supply, or slow recovery after decompression. Therefore, pressure ulcer risk warning requires not only determining whether the body is in a pressure position but also assessing whether specific bony prominences exhibit abnormal local tissue oxygenation under those positions.
[0003] Current pressure ulcer risk assessment methods largely rely on manual rounds, turning records, skin observation, rating scales, movement records, or single physiological parameters. Manual observation struggles to detect subcutaneous tissue hypoxia in a timely manner and is affected by the experience of nursing staff, the frequency of rounds, and differences in patient body size. While single-tissue oxygenation monitoring can reflect changes in local perfusion and oxygen supply, without the positional semantics of the attachment site, the system struggles to determine whether a decrease in oxygenation occurs at a corresponding bony prominence under pressure exposure, and it also struggles to distinguish its relationship with turning, movement artifacts, changes in attachment quality, or systemic circulatory fluctuations. Single-position monitoring can only identify supine, lateral, semi-recumbent, or turning events, and cannot directly indicate whether tissue oxygenation abnormalities have occurred at a specific site.
[0004] For the reasons mentioned above, this invention employs a co-location monitoring method of "tissue oxygenation + body position recognition" at the patch end. The patch acquires changes in tissue oxygenation at the application site through near-infrared monitoring, and obtains posture angle, turning events, and body position duration through the body position recognition module. The patch has a site identifier, and the information platform can map the body position category to the pressure exposure state of the corresponding bony prominence. For example, the risk weight of the sacrococcygeal patch increases in the supine or semi-recumbent position, the risk weight of the greater trochanter patch increases in the ipsilateral supine position, and the heel patch indicates increased risk when the heel is under continuous pressure and tissue oxygenation decreases.
[0005] To improve comfort and deployability during long-term wear, this invention does not superimpose a pressure-sensitive array on the pressure-bearing surface of the patch. Instead, it integrates a near-infrared emitting and detecting unit, a body position recognition unit, a wireless communication unit, a low-power processing unit, and a power supply unit into a low-profile flexible IoT patch. This patch uses waterproof encapsulation and skin-friendly contact materials, making it suitable for use in daily care, sweat, wiping, and short-term damp environments.
[0006] Furthermore, pressure ulcer early warning requires a closed-loop information system encompassing "patch sensing—platform reception—wristband or watch notification—nursing care confirmation." Tissue oxygenation and positional data collected by the patch are automatically uploaded to the information platform. The platform performs threshold checks, data quality verification, personnel-patch site matching, and event recording, then simultaneously sends alarm information to the IoT wristband or watch. Nursing staff, caregivers, or authorized receivers can receive vibration, sound, or screen notifications on the wristband or watch and complete actions such as confirmation, turning the patient, applying local pressure relief, or verifying patch status. Summary of the Invention
[0007] The purpose of this invention is to provide a pressure ulcer risk warning device and method based on near-infrared tissue oxygenation sensing and body position recognition using the Internet of Things, which solves the problem that existing single body position monitoring, single tissue oxygenation monitoring and manual inspection are difficult to timely, continuously and co-locate to interpret the relationship between local pressure exposure and tissue hypoxia.
[0008] I. The core technical solution of this invention includes the following components and processing steps:
[0009] 1. Waterproof, skin-friendly, flexible near-infrared monitoring patches are installed on bony prominences such as the sacrum, coccyx, greater trochanter of the femur, heel, ischial tuberosity, scapular region, or lateral malleolus; the patches acquire multi-wavelength, multi-source near-infrared signals and output tissue oxygenation parameters such as HbO2, HHb, StO2, or their relative changes, change rates, and recovery capabilities.
[0010] 2. A removable skin-friendly contact layer or a removable adhesive component is provided on the side of the patch that contacts the skin. The contact layer can be removed and replaced from the patch body when used once, when contaminated, when adhesion decreases, or when skin care is required. The sensing, processing, and wireless transmission units of the patch body can be reused or maintained according to the product design.
[0011] 3. Set up a body position recognition module at the same patch end to collect triaxial acceleration, angular velocity, posture angle or rolling events, and map the body position information to the pressure exposure state of the corresponding bony prominence through patch site identification.
[0012] 4. The patch integrates a low-power processor, storage unit, wireless transmission module, antenna, and power supply unit, supporting Bluetooth, Wi-Fi, cellular IoT, LoRa, or equivalent wireless communication methods. Tissue blood oxygenation and body position data collected by the patch are automatically transmitted to the information platform.
[0013] 5. The information platform is used to receive patch data, establish the association between patients, patch numbers, application sites, body position categories, tissue blood oxygen values, signal quality, and alarm events, and synchronously push the processing results to IoT wristbands or watch terminals.
[0014] 6. IoT wristbands or watches are used to receive tissue blood oxygenation data, body position data, alarm status and nursing care suggestions from the platform, and provide prompts through vibration, sound, icons, text or a combination thereof.
[0015] 7. The alarm parameters adopt a graded triggering method applicable to pressure ulcer risk warning scenarios for the elderly in nursing homes, retirement homes, long-term care facilities, etc.: When there is no stable individualized baseline or when rapid review is required, the effective tissue oxygenation channel StO2≤70% is set as the early warning triggering parameter, StO2≤65% is set as the alarm processing parameter, and StO2≤60% is set as the emergency review parameter; at the same time, it is combined with continuous data confirmation, signal quality, duration of pressure position, downward trend, hypoxia load, and recovery after decompression for joint judgment.
[0016] II. Device Composition
[0017] 1. Flexible near-infrared monitoring patch: including multi-wavelength near-infrared emitting unit, short / medium / long source detection distance unit, light-shielding isolation structure, flexible circuit, waterproof encapsulation layer, low profile shell, and a removable skin-friendly contact layer or removable adhesive component located on the skin-contact side of the patch.
[0018] 2. Body position recognition module: including a three-axis accelerometer, gyroscope or combination thereof, used to recognize supine, left lateral, right lateral, semi-recumbent, prone, rolling events and body position duration.
[0019] 3. Wireless transmission module: including Bluetooth, Wi-Fi, cellular IoT, LoRa or equivalent communication modules, as well as communication protocol units for data encryption, retransmission, caching and device registration.
[0020] 4. Data processing module: including multi-wavelength tissue oxygenation calculation, superficial signal correction, subcutaneous fat thickness influence correction, body position-site mapping, signal quality evaluation, individualized baseline establishment, 70% warning threshold judgment, 65% alarm threshold judgment, 60% emergency review threshold judgment, and alarm event generation.
[0021] 5. Information Platform: Includes functions such as device management, patient and patch binding, data reception, alarm judgment, log storage, alarm push, nursing treatment record, and system permission management.
[0022] 6. IoT wristband or watch terminal: including alarm reception, vibration or sound alerts, screen display, event confirmation, reminders for turning over or reviewing, data summary display, and communication functions for synchronization with the platform.
[0023] III. Beneficial Effects and Innovations
[0024] 1. IoT patch deployment: The patch itself has sensing, computing, wireless transmission and device identification capabilities, and can be connected to the information platform as an independent IoT device.
[0025] 2. Co-localization of tissue oxygenation and body position: The same patch simultaneously obtains StO2 trend, posture angle, turning events and motion artifact markers, so that local oxygenation changes have clear attachment site and body position semantics.
[0026] 3. Waterproof, skin-friendly and low profile design: It adopts a skin-friendly medical adhesive layer, flexible base and waterproof encapsulation to reduce long-term adhesion irritation and help improve the reliability of use in daily care scenarios.
[0027] 4. Platform-Wristband / Watch Closed Loop: Data collected by the patch automatically enters the information platform and is then synchronously pushed to the IoT wristband or watch terminal, which helps to make the terminal receive the data more intuitive and improve the timeliness of care response.
[0028] 5. Simplified alarm parameters: StO2 drops to 70% or below as the early warning trigger parameter, 65% or below as the alarm processing parameter, and 60% or below as the emergency verification parameter, combined with body position exposure, downward trend, hypoxia load, continuous data confirmation, and signal quality evaluation.
[0029] 6. Traceable nursing records: The platform records alarm time, patch site, StO2 value, body position, wristband or watch confirmation time and nursing treatment results, providing data support for pressure ulcer risk management. IV. Description of the attached drawings
[0031] Figure 1 This is a schematic diagram of the overall structure of a system based on IoT patches, an information platform, and IoT wristbands or watches (Schematic diagram of the overall structure of an IoT pressure injury risk warning system).
[0032] Figure 2 A schematic diagram of the structure of a waterproof, skin-friendly, flexible near-infrared monitoring patch.
[0033] Figure 3 This is a schematic diagram of multi-source near-infrared tissue oxygenation acquisition and body position recognition co-localization (pattern-type near-infrared multi-source tissue oxygenation acquisition and body position semantic mapping).
[0034] Figure 4 A schematic diagram of the data flow for automatically uploading patch data to the platform and synchronizing it to the wristband or watch terminal (data transmission and closed-loop feedback process diagram).
[0035] Figure 5 A flowchart (alarm and handling process diagram) for graded alarms based on tissue blood oxygenation threshold, body position exposure time, continuous confirmation, and recovery status after decompression.
[0036] Figure 6 A schematic diagram of tissue oxygenation changes before and after turning over or local decompression (a schematic diagram of tissue oxygen saturation changes over time and recovery). Detailed Implementation
[0037] I. Monitoring Sites and Patch Placement
[0038] The system can apply patches to the sacrum, coccyx, left / right greater trochanter of the femur, both heels, ischial tuberosities, scapular region, lateral malleolus, or other bony prominences. Each patch has a unique device number and site identifier. Upon initial application, the platform records patient identity, patch number, application site, left / right side information, application time, initial posture angle, and initial tissue oxygenation level. Additional patch sites can be added as needed for patients in prone positions, postoperative immobilization, long-term immobilization, or those who are extremely emaciated.
[0039] II. Principle of Multi-wavelength Near-infrared Monitoring
[0040] The near-infrared emitting unit selects at least two or three wavelengths within the commonly used near-infrared tissue oxygenation window of 650 nm to 950 nm. In this embodiment, combinations of 730 nm to 760 nm, around 805 nm, 850 nm to 880 nm, and 930 nm to 940 nm can be selected. The data processing module calculates the relative changes in HbO2, HHb, total hemoglobin, and StO2-related parameters based on the improved Beer-Lambert model, spatial resolution model, or empirical correction model. The system prioritizes using the changes, rates of change, and recovery capacity of the same patient and the same site relative to the individualized baseline as auxiliary judgments, while retaining the absolute value of StO2 for the 70% warning, 65% alarm, and 60% emergency review threshold judgments.
[0041] III. Multi-source range-depth resolution structure
[0042] Short source-probe distances can be set from 5 mm to 13 mm, preferably 5 mm to 10 mm, for superficial tissue, ambient light, adhesion status, and motion artifact correction; medium source-probe distances can be set from 15 mm to 25 mm for subcutaneous and superficial soft tissue trend identification; long source-probe distances can be set from 30 mm to 45 mm to improve sensitivity to oxygenation changes in deeper soft tissues. Increasing the source-probe distance generally improves sensitivity to deeper tissues, but it also reduces the signal-to-noise ratio and increases the influence of individual variability. Therefore, the system does not interpret long-distance channels as absolute measurements of a specific anatomical layer, but rather as deep-sensitivity signals, and makes a comprehensive judgment based on short-distance channels, signal quality, fat thickness influence, and individualized baseline.
[0043] Table 1 Recommended Hardware Parameter Range (Example)
[0044] Wavelength combination At least 2-3 wavelengths within the 650-950 nm near-infrared window; optional wavelengths include 730-760 nm, around 805 nm, 850-880 nm, and 930-940 nm. Used to distinguish the absorption differences between HbO2 and HHb and improve the stability of tissue oxygenation calculations; the specific wavelength should be determined in conjunction with device calibration. Short source detection range 5-13 mm; preferably 5-10 mm Used for correction of superficial tissue, ambient light, attachment status, and motion artifacts. Mid-source distance 15-25 mm Used for trend identification in subcutaneous and superficial soft tissues; the specific range is adjusted based on patch size and calibration results. Long Source Distance Exploration 30-45 mm It is used to improve sensitivity to changes in oxygenation in deeper soft tissues; it needs to be combined with signal-to-noise ratio and fat thickness correction. NIRS sampling frequency 0.25-10 Hz Low-power continuous monitoring can use a lower frequency; the sampling frequency can be increased when verifying alarms, motion events, or signal quality. IMU sampling frequency 10-50 Hz Used for attitude angle calculation, rollover event recognition, and motion artifact labeling, it can be adjusted according to power consumption and algorithm requirements. Waterproof and skin-friendly structure Waterproof encapsulation layer, removable skin-friendly contact layer, removable adhesive components, flexible substrate, low-profile housing Medical-grade silicone, medical pressure-sensitive adhesive, hydrogel, polyurethane film, or equivalent skin-friendly materials can be used; the layer in contact with the skin is removable and easy to replace, and the main body of the patch remains waterproof and sealed. Communication methods Bluetooth, Wi-Fi, cellular IoT, LoRa or equivalent wireless communication The patch is automatically uploaded to the information platform; the platform then pushes the information to the IoT wristband or watch terminal.
[0045] IV. Waterproof and Skin-Friendly IoT Patch Structure
[0046] The patch may include, from the outside in, a waterproof encapsulation layer, a flexible circuit layer, a light-shielding structure, a near-infrared emitting and detecting unit, a body position recognition module, a battery or energy management unit, a wireless communication unit, a removable skin-friendly contact layer, and a removable adhesive component. The waterproof encapsulation layer protects the electronic components from the effects of sweat, cleaning solutions, and everyday humid environments. The removable skin-friendly contact layer, located on the side of the patch that contacts the skin, can be connected to the patch body via clips, inserts, release films, peelable medical adhesive, positioning and fastening structures, or equivalent removable structures. After use, it can be removed from the patch body for replacement, reducing the risk of cross-contamination, minimizing skin irritation, and facilitating daily care. The light-shielding structure reduces ambient light interference and emitter-detector crosstalk.
[0047] As an IoT device, the patch has a unique device identification and can be bound to patient identity, nursing location number, application site, and wristband or watch terminal. The patch can perform data acquisition, preliminary filtering, signal quality judgment, short-term buffering, and abnormal retransmission locally.
[0048] V. The Influence of Subcutaneous Fat Thickness on Signal Quality Correction
[0049] For obese patients, patients with thick subcutaneous fat in the sacrococcygeal region, or patients with thick soft tissue around the greater trochanter of the femur, the system estimates the equivalent optical path factor and layered channel confidence based on the multi-source probe distance attenuation slope, short / medium / long channel signal ratio, channel SNR, location identification, BMI, and skinfold or ultrasound thickness. When the signal quality is low, the platform prompts for verification of patch attachment, light-blocking status, adhesion status, or motion artifacts, and generates corresponding alarm events based on the results confirmed by continuous data.
[0050] VI. Body Position Recognition and Semantic Mapping of Compression Locations
[0051] The body position recognition module uses accelerometer gravity vectors to estimate local attitude angles and gyroscopes to identify rolling events. The data processing module maps the patch site to body position categories and assigns them to a risk model: when supine, the pressure semantics of the sacrococcygeal region and heel are enhanced; when lateral, the pressure semantics of the ipsilateral greater trochanter and lateral malleolus are enhanced; and when semi-recumbent, the sacrococcygeal region shows enhanced shear risk. Body position recognition provides pressure semantics and exposure time information for changes in tissue oxygenation at the patch site.
[0052] VII. Data Upload, Platform Reception, and Synchronization with Bracelet / Watch Terminal
[0053] Data collected by the patch, including StO2, HbO2, HHb, posture angle, body position type, turning events, patch location, signal quality, battery level, and device status, are automatically uploaded to the information platform via a wireless transmission module. Uploading can be done through periodic reporting, event-triggered reporting, alarm-priority reporting, or resume interrupted transmission.
[0054] After receiving the patch data, the information platform completes data parsing, patient-patch-site binding verification, timestamp unification, signal quality review, individualized baseline comparison, 70% warning threshold judgment, 65% alarm threshold judgment, 60% emergency review threshold judgment, alarm event generation, and push management. The platform will synchronously send blood oxygen data, body position data, alarm status, patch site, confirmation button, and nursing treatment suggestions to the IoT wristband or watch terminal.
[0055] Upon receiving a push notification from the platform, the IoT wristband or watch can display the patient's identification, patch location, current StO2 value, body position, alarm duration, and treatment suggestions, and provide alerts via vibration, sound, screen flashing, or a combination thereof. Nursing staff, caregivers, or authorized recipients can confirm alarms, record actions such as turning over, local decompression, skin examination, or patch verification on the wristband or watch, with the treatment records synchronously transmitted back to the platform.
[0056] Table 2. Data Fields of Platform and IoT Bracelet / Watch Terminal (Example)
[0057] Patch identity and body part Device serial number, patient identification, application site, left or right side, application time Used to establish a unique correspondence between data sources and alarm locations. Tissue oxygenation data StO2, HbO2, HHb, total hemoglobin, rate of change, signal quality It is used to reflect the local tissue oxygenation status and data reliability. Body position data Posture angle, position category, rolling event, position duration, motion artifact markers This is used to explain whether the attachment site is under pressure and exposed. Alarm data Alarm trigger time, trigger channel, StO2 value, duration, alarm status Used for platform alarm recording and notifications from wristbands or watches. Processing feedback Confirmation of time, turning record, local decompression, skin reassessment, patch reapplication, and battery management. Used to create traceable nursing care records.
[0058] VIII. Alarm Parameters and Processing Procedures
[0059] In this embodiment, the tissue oxygenation alarm parameters adopt a hierarchical triggering method that combines "absolute value assistance, body position semantic participation, continuous confirmation, and recovery evaluation". For pressure ulcer risk warning scenarios for the elderly, such as nursing homes, retirement homes, and long-term care facilities, the platform first judges the signal quality and continuous confirmation results, and then combines the absolute value of tissue oxygenation, the downward trend, the hypoxia load, the exposure time of the pressure position, and the recovery status after decompression for judgment. Generally, when the effective tissue oxygenation channel StO2 at the patch site is ≤70% and confirmed by continuous data, and the site is under the corresponding pressure position semantics or the StO2 shows a continuous downward trend, the system outputs an early warning. When the effective tissue oxygenation channel StO2 is ≤65% and confirmed by continuous data, or when the condition does not improve under pressure position after the early warning and is accompanied by a downward trend or an increase in hypoxia load, the system outputs an alarm and prompts the patient to turn over, relieve local pressure, or check the patch status. When the effective tissue oxygenation channel StO2 is ≤60%, or when the recovery within the preset observation window after pressure relief is insufficient, the system outputs an emergency check prompt and prompts the nursing staff to check the skin, patch, position, and systemic circulation status. The above 70%, 65%, and 60% are initial engineering parameters for the example and are not used as diagnostic criteria for pressure ulcers. In actual applications, they can be adjusted according to the patch site, equipment calibration, signal quality, duration of pressure position, and clinical validation results.
[0060] To facilitate the implementation of the platform algorithm and the understanding of nursing staff, the meaning and calculation method of the main alarm parameters in this embodiment are as follows:
[0061] (1) StO2(t): Represents the current effective tissue oxygenation value at time t at the attachment site. This parameter is obtained from near-infrared multi-wavelength signals through an improved Beer-Lambert model, a spatially resolved model, or a calibrated equivalent algorithm; in terms of algorithmic expression, it can be expressed as HbO2 / (HbO2+HHb)×100%, or directly output by the device calibration model. StO2(t) is used for the 70%, 65%, and 60% absolute threshold judgment.
[0062] (2) B_StO2: Represents the individualized tissue oxygenation baseline for the same patient and the same patch application site. This parameter can be determined by taking the mean or median of StO2 within a preset time window when the patch is initially stably applied, the signal quality is acceptable, and there are no significant motion artifacts. It can also be updated daily, by nursing cycle, or by platform strategy. B_StO2 is used to reduce the absolute differences caused by different patients, different sites, and different subcutaneous tissue thicknesses.
[0063] (3) Slope_StO2: Represents the rate of change in tissue oxygenation, used to reflect whether StO2 shows a downward trend over a period of time. This parameter can be calculated as the difference between adjacent time windows: Slope=[StO2(t)-StO2(t-Δt)] / Δt, or it can be calculated using the slope of a linear regression over a 5-15 minute sliding window. In the embodiment, when Slope_StO2 is negative within the preset sliding time window and confirmed by stable data for at least two consecutive sliding windows or three consecutive minutes, and Q_signal is qualified, it can be determined as a downward trend in tissue oxygenation. Slope_StO2 is used as a trend auxiliary parameter and is not used alone as an alarm basis. It should be used in combination with parameters such as StO2 absolute threshold, AUC_low, Exposure_posture, or Recovery_20min.
[0064] (4) AUC_low: Indicates hypoxia load, used to reflect the degree and duration of tissue oxygenation below a preset threshold. The calculation formula is AUC_low=Σmax(0,Th-StO2_i)×Δt, with units of %·min, where Th can be 70%, 65% or an individualized threshold, StO2_i represents the tissue oxygenation value at the i-th sampling point or within the statistical window, and Δt represents the corresponding time interval. The larger the AUC_low value, the deeper the degree of hypoxia or the longer the duration; this parameter is used to distinguish between transient fluctuations and persistent hypoxia, and its specific trigger threshold can be determined based on the patch site, sampling frequency, equipment calibration and clinical validation results, and is not used as a sole basis for the diagnosis of pressure ulcers.
[0065] (5) Exposure_posture: Indicates the duration of the current patch site under the corresponding pressure semantic. The platform accumulates the exposure time of the pressure position based on the patch site identifier and IMU position recognition results; for example, the exposure time of the sacral and coccygeal patch is accumulated when the patient is supine or semi-recumbent, the exposure time of the left / right femoral greater trochanter patch is accumulated when the patient is in the corresponding lateral decubitus position, and the exposure time of the heel patch is accumulated when the heel is under continuous pressure and the pressure is not relieved. In the embodiment, Exposure_posture ≥ 60 min can be used as the parallel condition for early observation, Exposure_posture ≥ 120 min can be used as the parallel condition for alarm processing, and Exposure_posture ≥ 180 min can be used as the parallel condition for emergency review; the timer can be reset or the weight can be reduced after turning over, local decompression, or change of position.
[0066] (6) Confirm_3min: This parameter represents the continuous data confirmation parameter, used to reduce false alarms caused by motion artifacts or instantaneous attachment disturbances. In the embodiment, a rule that at least three consecutive 1-minute sliding windows meet the corresponding threshold conditions and the signal quality is qualified can be used, or an equivalent rule of three consecutive minutes of stable data confirmation can be used.
[0067] (7) Recovery_20min: This parameter represents the recovery after decompression and is used to evaluate the recovery of tissue oxygenation after turning over, local decompression, or relief of pressure. In the example, it can be determined whether StO2 has recovered to above 65% within 20 minutes after decompression, or whether it has reached the recovery criterion set by the platform based on individualized baseline and clinical validation results; when B_StO2-StO2_min is greater than the preset minimum effective difference, the recovery rate R=[StO2_post-StO2_min] / [B_StO2-StO2_min]×100% can also be calculated. When the denominator is too small, the signal quality is insufficient, or the baseline is not yet stable, the judgment should be made based on whether it has recovered to above 65%.
[0068] (8) Q_signal: Represents the signal quality parameter, used to determine whether the current near-infrared data can be used for threshold alarm. This parameter can be combined with light intensity, SNR, ambient light interference, loose attachment, motion artifacts, signal saturation, power and channel consistency to form a qualified, pending review or unqualified state; low quality data will be given priority to trigger patch review and will not be directly used as the basis for tissue hypoxia alarm.
[0069] The above parameters are combined according to preset rules to trigger alarm events. The platform first determines whether Q_signal is qualified and Confirm_3min is met, and then determines parameters such as StO2(t), Slope_StO2, AUC_low, Exposure_posture, and Recovery_20min. Except for cases requiring immediate verification, such as StO2≤60% or insufficient recovery after decompression, the platform generally does not directly generate nursing alarms based on a single instantaneous StO2 value; Slope_StO2 and AUC_low are used as dynamic auxiliary parameters, mainly to distinguish between transient fluctuations, continuous declines, and continuous hypoxia. When absolute thresholds, continuous downward trends, hypoxia load, exposure to pressure positions, or insufficient recovery after decompression are mutually corroborated, the alarm event level can be increased. If low tissue oxygenation occurs under non-pressure position semantics, or the mapping relationship with the patch site is inconsistent, the platform can first output a verification prompt, observation prompt, or systemic circulatory status verification prompt.
[0070] Continuous data confirmation can be achieved by ensuring that at least three consecutive 1-minute sliding windows meet the corresponding trigger conditions, or by equivalent confirmation of three consecutive minutes of stable data. Alternatively, equivalent confirmation rules can be set based on the patch sampling frequency, filtering algorithm, and clinical validation results. If the patch detects significant motion artifacts, loose adhesion, ambient light interference, signal saturation, insufficient signal-to-noise ratio, or low battery, the platform will prioritize outputting a "Check Patch Status" prompt and requesting re-collection or re-application. Early warning events are used to prompt observation, verification, or early decompression, while alarm processing events are used to prompt for turning over, local decompression, or skin verification.
[0071] Upon triggering an alarm or review prompt, the wristband or watch displays the alarm location, current tissue oxygenation level, body position, duration of exposure to pressure, and suggested interventions. Suggested interventions include observing the local skin, assisting with turning, applying local pressure relief, adjusting limb position, reviewing patch application, and recording the results. After intervention, the patch continuously monitors tissue oxygenation recovery. If StO2 does not rise above 65% within 20 minutes after pressure relief or turning, or if the recovery criteria set by the platform based on individualized baselines and clinical validation results are not met, the platform maintains the alarm or generates an emergency review event and records the recovery curve.
[0072] Table 3 Simplified Alarm Rules (Example)
[0073] Q_signal is valid and continuously confirmed by Confirm_3min; at the same time, at least one of the following conditions must be met: 1) StO2 ≤ 70%; 2) Slope_StO2 is negative within a 5-15 minute sliding window and continuously confirmed; 3) AUC_low increases compared to the previous observation window. The above conditions occur under the corresponding pressure position semantics, or Exposure_posture ≥ 60 min, or insufficient recovery after recent decompression. Generate early warning events and record patch location, StO2 value, body position type, Exposure_posture, trend parameters, hypoxia load parameters, and timestamp. Vibration / sound / screen notification; displays "Tissue blood oxygen is low, please observe, reduce pressure or recheck". Used for observing and verifying patch status and arranging nursing interventions in advance; if not in the corresponding pressure position semantics, prompts can be made to verify patch status, position recognition, or systemic circulation status first. Q_signal is qualified and continuously confirmed by Confirm_3min; at the same time, at least one of the following must be met: 1) StO2≤65%; 2) AUC_low reaches the preset hypoxic load; 3) after the early warning, there is no improvement under the pressure position. The above conditions are accompanied by Exposure_posture≥120min, the pressure position is present continuously, or Slope_StO2 shows a decreasing trend after continuous confirmation. An alarm event is generated, prompting a check of the skin, body position, and patch status, and prompting the patient to turn over, relieve local pressure, or adjust limb position. Enhance vibration / sound alerts; display "Tissue blood oxygenation is low, please turn over, relieve pressure and record." Used to trigger nursing interventions; determined by a combination of absolute tissue oxygenation, exposure to pressure positions, hypoxia load, and trend parameters. Q_signal is qualified and meets at least one of the following: 1) StO2 ≤ 60% and continuously confirmed; 2) Exposure_posture ≥ 180 min and Slope_StO2 continuously confirmed to be decreasing or AUC_low continuously increasing; 3) Failure to recover to 65% or more within 20 minutes after decompression / turning over, or failure to meet the recovery criteria set by the platform based on individualized baseline and clinical validation results. An emergency review event is generated, prompting the user to check the skin, body position, patch status, systemic circulation status, and nursing treatment results; if necessary, the event is continuously pushed to the terminal. The message continues to display "Hypoxia persists, please review and record immediately". Used for situations of persistent hypoxia, prolonged exposure to pressure, or insufficient recovery; the recovery observation window is used to evaluate the effectiveness of turning over or local decompression. Numerical conditions are met, but motion artifacts, loose attachment, ambient light interference, signal saturation, low battery, or inconsistency between body position and location mapping exist. Output a verification prompt and request re-acquisition, re-attachment, check for light shading, check battery level, or verify body position recognition results. The message reads, "Please check patch attachment / light shielding / battery level / body position." After continuous confirmation that the signal quality is qualified and the body position semantics are reasonable, the corresponding early warning or alarm event is generated. After the alarm is triggered, the patient should turn over or undergo local decompression, and StO2 should recover to more than 70%, or reach the recovery criteria set by the platform based on individualized baselines and clinical validation results, and be confirmed by continuous stable data; at the same time, Exposure_posture should decrease or the compressed position should be relieved. Record alarm cancellation, changes in body position, and recovery curves, and retain the event log. The message reads "Alarm deactivated / Continue monitoring". The release criteria can be determined based on clinical validation, patch site, and nursing procedures.
[0074] The alarm parameters mentioned above were set with reference to factors such as the sacral StO2 observation range under bedridden pressure scenarios, the slower recovery of elderly people after pressure, and the influence of equipment and location on the absolute value of NIRS. These are the initial engineering parameters of the embodiment. Because near-infrared tissue oxygenation measurement is affected by equipment model, source-detection distance, attachment site, subcutaneous fat thickness, motion artifacts, and systemic circulatory status, the somatic StO2 measurement verification study indicates that a single measurement has a bias and precision range of several percentage points. Therefore, this embodiment uses continuous data confirmation, signal quality judgment, and positional semantics to participate in the judgment, and forms a progressive event level of 70%, 65%, and 60%: an early warning is generated when StO2 ≤ 70% combined with pressure position exposure or a downward trend; an alarm processing event is generated when StO2 ≤ 65% or when there is no improvement under pressure position after the early warning and accompanied by hypoxia or a continuous downward trend; and an emergency review event is generated when StO2 ≤ 60% or when recovery is insufficient after decompression / turning. When implementing the final product, the final alarm threshold, body exposure time window, continuous confirmation duration, and deactivation rules should be determined through optical phantom experiments, human body adhesion experiments, patch site calibration, preclinical validation, and clinical data review and analysis.
Claims
1. A pressure ulcer risk early warning device based on near-infrared tissue oxygenation sensing and body position recognition using the Internet of Things, characterized in that, The system includes a waterproof, skin-friendly, flexible near-infrared monitoring patch, an information platform, and an IoT wristband or watch terminal. The waterproof, skin-friendly, flexible near-infrared monitoring patch is attached to at least one bony prominence among the sacrum, coccyx, greater trochanter of the femur, heel, ischial tuberosity, scapular region, or lateral malleolus. It includes a near-infrared tissue oxygenation acquisition unit, a body position recognition module, a data processing module, a wireless transmission module, a power supply unit, a waterproof encapsulation structure, and a detachable skin-friendly contact layer or detachable adhesive component located on the skin-contacting side of the patch. The near-infrared tissue oxygenation acquisition unit is used to acquire near-infrared signals from the attachment site and obtain changes in tissue oxygenation. The body position recognition module is used to acquire posture angles, body position types, turning events, or duration of body position. The wireless transmission module is used to upload tissue oxygenation data and body position data to the information platform; the information platform is used to establish the association between patient, patch number, patch application site, body position category, tissue oxygenation data, signal quality, and alarm events, and to generate graded alarm events based on at least two of the following: absolute tissue oxygenation value, body position semantics, continuous confirmation results, signal quality, downward trend, hypoxia load, and recovery status after decompression; the IoT wristband or watch terminal is used to receive tissue oxygenation data, body position status, alarm information, and processing suggestions from the information platform, and to transmit back confirmation, turning over, local decompression, or patch review processing records.
2. The pressure ulcer risk warning device according to claim 1, characterized in that, The waterproof, skin-friendly, flexible near-infrared monitoring patch comprises, from the outside to the inside, a low-profile shell, a waterproof encapsulation layer, a flexible circuit layer, a light-shielding or isolation structure, a near-infrared emission and detection structure, a body position recognition module, a data processing module, a wireless transmission module, a power or energy management unit, a removable skin-friendly contact layer, and a removable adhesive component; wherein, the removable skin-friendly contact layer or the removable adhesive component can be detached and replaced relative to the patch body.
3. The pressure ulcer risk warning device according to claim 1, characterized in that, The near-infrared tissue oxygenation acquisition unit includes at least two or three working wavelengths, which are located within the 650 nm to 950 nm near-infrared tissue oxygenation monitoring window; the near-infrared tissue oxygenation acquisition unit also includes at least two of a short-source detection distance channel, a medium-source detection distance channel, and a long-source detection distance channel, wherein the short-source detection distance is 5 mm to 13 mm, the medium-source detection distance is 15 mm to 25 mm, and the long-source detection distance is 30 mm to 45 mm.
4. The pressure ulcer risk warning device according to claim 1, characterized in that, The body position recognition module includes a three-axis accelerometer, a gyroscope, or a combination thereof, used to identify supine, left lateral, right lateral, semi-recumbent, prone, or rolling events; the information platform establishes a semantic mapping of pressure-affected body positions based on the patch location identifier and the body position category, so that the sacral and coccygeal patch corresponds to the pressure-affected state in supine or semi-recumbent positions, the femoral greater trochanter patch corresponds to the pressure-affected state in ipsilateral lateral position, and the heel patch corresponds to the pressure-affected state under continuous heel pressure.
5. The pressure ulcer risk warning device according to claim 1, characterized in that, The removable skin-friendly contact layer or removable adhesive component is connected to the patch body via a snap, insert, release film, peelable medical adhesive, positioning and fastening structure or equivalent removable structure; the removable skin-friendly contact layer or removable adhesive component can be removed from the patch body and replaced when used once, when contaminated, when adhesion decreases, or when skin care is required.
6. The pressure ulcer risk warning device according to claim 1, characterized in that, The wireless transmission module includes Bluetooth, Wi-Fi, cellular IoT, LoRa or equivalent wireless communication modules; the wireless transmission module supports at least one of the following communication methods: periodic reporting, event-triggered reporting, alarm-priority reporting, breakpoint resumption, abnormal retransmission or short-term buffering.
7. The pressure ulcer risk warning device according to claim 1, characterized in that, The information platform is configured to first determine the Q_signal signal quality parameter and the Confirm_3min continuous confirmation parameter, and then perform a combined judgment on StO2(t), Slope_StO2, AUC_low, Exposure_posture, and Recovery_20min; where Q_signal represents the signal quality state formed by light intensity, signal-to-noise ratio, ambient light interference, loose attachment, motion artifacts, signal saturation, power or channel consistency, and Confirm_3min represents the continuous data confirmation state.
8. The pressure ulcer risk warning device according to claim 7, characterized in that, When Q_signal is qualified and continuously confirmed, the information platform will use an effective tissue oxygenation channel StO2 ≤ 70% as one of the early warning trigger conditions, an effective tissue oxygenation channel StO2 ≤ 65% as one of the alarm processing trigger conditions, and an effective tissue oxygenation channel StO2 ≤ 60% as one of the emergency review trigger conditions; and generate alarm events of corresponding levels by combining the semantics of the pressure position, Exposure_posture, Slope_StO2, AUC_low, or Recovery_20min.
9. The pressure ulcer risk warning device according to claim 8, characterized in that, Slope_StO2 represents the rate of change in tissue oxygenation, which is obtained by the difference between adjacent time windows or the slope of linear regression in a sliding window of 5 to 15 minutes; AUC_low represents the hypoxic load, which is calculated as AUC_low=Σmax(0,Th-StO2_i)×Δt, where Th is the preset threshold, StO2_i is the tissue oxygenation value at the i-th sampling point or within the statistical window, and Δt is the corresponding time interval; Exposure_posture represents the duration of the current patch site under the corresponding pressure semantics; Recovery_20min represents the tissue oxygenation recovery status within the preset recovery observation window after turning over, local decompression, or release of pressure.
10. The pressure ulcer risk warning device according to claim 8, characterized in that, The information platform generates an early warning event when Exposure_posture ≥ 60 min and tissue oxygenation data meets the warning conditions; generates an alarm processing event when Exposure_posture ≥ 120 min and tissue oxygenation data continues to fail to improve or hypoxia load increases; and generates an emergency review event when Exposure_posture ≥ 180 min and tissue oxygenation data continues to deteriorate or recovery after decompression is insufficient. Among these events, Exposure_posture is reset, its accumulation is paused, or its weight is reduced after turning over, local decompression, or a change in body position.
11. The pressure ulcer risk warning device according to claim 1, characterized in that, The IoT wristband or watch terminal is used to display patient identification, patch site, current StO2 value, body position, alarm duration and treatment suggestions, and to provide prompts through at least one of vibration, sound, screen flashing, icons or text; the IoT wristband or watch terminal is also used to receive alarm confirmations, turning records, local decompression records, skin review records, patch reapplication records or treatment results input by nursing staff, caregivers or authorized personnel.
12. A method for pressure ulcer risk early warning based on near-infrared tissue oxygenation sensing and body position recognition using the Internet of Things, characterized in that, The pressure ulcer risk warning device according to any one of claims 1 to 11, the method comprising: attaching a waterproof, skin-friendly, flexible near-infrared monitoring patch to a target bony prominence; collecting near-infrared tissue oxygenation data and body position data by the patch; uploading the near-infrared tissue oxygenation data and body position data to an information platform; the information platform completing the binding and verification of patient, patch number, attachment site, and body position semantics, and judging the signal quality and continuous data confirmation results; the information platform generating an early warning, alarm processing, or emergency review event by combining the StO2 absolute threshold, the exposure time of the pressure position, the trend of tissue oxygenation decline, the hypoxia load, and the recovery status after decompression; synchronously pushing the event to an IoT wristband or watch terminal; the IoT wristband or watch terminal receiving the alarm information and transmitting back the processing records of confirmation, turning over, local decompression, skin review, or patch review.
13. The pressure ulcer risk warning method according to claim 12, characterized in that, After an alarm event is generated, the information platform continues to receive tissue oxygenation data collected by the patch and determines whether StO2 has risen to above 65% or has reached the recovery criteria set by the platform based on individualized baselines and clinical validation results within a 20-minute recovery observation window after turning over or local decompression. If the recovery criteria are not met, the alarm is maintained or an emergency review event is generated.
