Spine postoperative care quality sensitive index evaluation method

Abstract

The invention discloses a spine postoperative care quality sensitive index evaluation method, and relates to the technical field of rehabilitation medicine. Comprising the following steps: establishing a mapping relation table between to-be-evaluated nursing quality sensitive indexes and postoperative nursing record entries; postoperative patient recovery data are collected, dynamic space-time association with the mapping relation table is established, and recovery data changes and nursing operation records are matched through a time window synchronization algorithm; and calculating a standardized deviation degree delta between actual data of each index and a nursing record expected value based on a correlation result, dynamically adjusting an index weight according to a formula, visually outputting an evaluation result on the three-dimensional spine model, and highlighting a deviation degree over-limit region and a corresponding nursing record abstract. According to the invention, through a multi-source data real-time acquisition and intelligent analysis technology, a spinal post-operation nursing quality evaluation system with dynamic feedback is constructed, so that the defects of poor timeliness, complex calculation, insufficient dynamic evaluation capability and the like of a traditional method are overcome.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of rehabilitation medicine, in particular to a method for evaluating sensitive indicators of postoperative nursing quality of the spine. BACKGROUND

[0002] With the continuous development of medical technology, as a high-risk surgical procedure, spinal surgery has become an important means of treating spinal diseases. However, the quality of postoperative nursing of the spine directly affects the recovery speed, complication rate and quality of life of patients. The paper "Construction of perioperative nursing quality sensitive indicator system for scoliosis" (Wu QF et al. "Construction of perioperative nursing quality sensitive indicator system for scoliosis." Chinese Journal of Spine and Spinal Cord 31.7 (2021): 8.) discloses the construction process of the nursing quality sensitive indicator system as follows: first, a research team composed of 10 experts from clinical nursing management, research and other fields is established; the team preliminarily formulates the indicators according to the evidence-based method, searches multiple Chinese and English databases, and constructs a preliminary item pool after screening and evidence evaluation, combined with diagnosis and treatment guidelines, nursing standards, etc.; then, 20 experts in the field of spinal surgery from hospitals in Beijing, Shanghai, Guangdong and other places are selected for two rounds of Delphi expert consultation; the "critical value method" is used to screen the indicators, and the analytic hierarchy process and weight method are used to determine the weight of each indicator; after two rounds of consultation, the indicators are continuously adjusted, and finally a sensitive indicator system for perioperative nursing quality of scoliosis is constructed, which includes 3 first-level indicators, 6 second-level indicators and 14 third-level indicators.

[0003] However, the existing construction method of the nursing quality sensitive indicator system can screen out effective evaluation indicators, but the indicator effectiveness evaluation process has obvious deficiencies, mainly reflected in the following aspects: (1) evaluation relies on statistical verification, poor timeliness: the actual effect evaluation of sensitive indicators mainly relies on clinical trials and statistical analysis, and a large amount of long-term data (such as complication rate, nursing compliance rate, etc.) needs to be collected to draw a conclusion on effectiveness, resulting in a long evaluation period and difficulty in timely optimizing and improving nursing programs. (2) Complex calculation, high application threshold: the construction of the indicator system needs to go through expert consultation, weight analysis and other processes, and involves complex parameter calculation, such as analytic hierarchy process weight calculation and critical value index screening, which requires high data analysis ability, increasing the application difficulty in actual nursing management. (3) Lack of real-time dynamic evaluation capability: the existing method lacks real-time effect monitoring and feedback mechanism for nursing indicators, and cannot dynamically adjust the evaluation system during postoperative nursing, which may affect the precision and adaptability of nursing measures.

[0004] Especially for such high-risk scenarios as postoperative care of the spine, traditional static statistical analysis methods are difficult to quickly evaluate the effectiveness of sensitive indicators, resulting in the lag of nursing quality improvement. Therefore, a method and system capable of real-time evaluation and dynamic optimization of nursing quality sensitive indicators are needed to improve the accuracy and efficiency of postoperative care. SUMMARY

[0005] In view of the deficiencies of the prior art, the present application provides a method for evaluating sensitive indicators of postoperative care of the spine, which uses multi-source data real-time acquisition and intelligent analysis technology to construct a dynamic feedback postoperative care quality evaluation system for the spine to solve the problems of poor timeliness, complex calculation and insufficient dynamic evaluation capability of traditional methods.

[0006] The specific technical solutions are as follows: A method for evaluating sensitive indicators of postoperative care of the spine, comprising: establishing a mapping relationship table between the sensitive indicators to be evaluated and the postoperative care record items, each indicator in the mapping relationship table being associated with a nursing record data field and an expected threshold, and the anatomical site matching relationship being labeled based on the spinal segment code; collecting recovery data of postoperative patients and establishing a dynamic spatio-temporal association with the mapping relationship table, the recovery data including real-time electromyographic signals of L4-L5 segment paraspinal muscle groups monitored by surface electromyographic sensors, three-dimensional mechanical parameters of spinal flexion / rotation obtained by an inertial measurement unit, and wound microenvironment data dynamically collected by a temperature-sensitive sensor array implanted around the incision; using a time window synchronization algorithm to match the recovery data changes and the nursing operation records, including the correspondence between the electromyographic signal abnormal period and the turning over nursing record, the association between the spinal motion data mutation point and the body position adjustment, and the control of the wound parameter fluctuation and the dressing change operation, if data abnormalities are detected but the nursing record has no response, a warning is triggered; based on the association results, calculating the standardized deviation δ of the actual data of each indicator and the expected value of the nursing record according to the formula: dynamically adjusting the indicator weight, and upgrading the indicator to a nursing quality control alarm if δ is greater than 0.7 for three consecutive times and no effective nursing intervention is matched; finally, the evaluation results are visualized and output on a three-dimensional spinal model, and the deviation exceeding the limit area and the corresponding nursing record summary are highlighted.

[0007] The above scheme further comprises: the establishment of the dynamic spatio-temporal association specifically includes: the L4-L5 segment paraspinal muscle group signals collected in real time by the surface electromyographic sensors, the establishment of event sequence matching according to the time stamp and the turning over operation in the nursing record, and when the electromyographic signal fluctuation amplitude exceeds 30% of the baseline and no turning over operation is recorded for 5 minutes, a first-level nursing omission warning is automatically generated.

[0008] Further, in the time window synchronization algorithm, an adaptive sliding window mechanism is adopted for matching the spinal motion data, and the window width W is dynamically adjusted according to the motion speed: when the inertial measurement unit detects that the angular velocity ω xyz > 15° / s, the matching window is compressed from ±5 minutes to ±2 minutes, and the corresponding time period of the body position adjustment missing record is highlighted on the nursing record interface.

[0009] Further, in the index weight formula, the decay coefficient α is dynamically configured according to the wound type: α = 0.25 for patients undergoing posterior lumbar disc resection, and α = 0.4 for patients undergoing anterior vertebral fusion, and the wound exudate pH value fed back in real time by the implanted temperature-sensitive sensor is fine-tuned by ±0.1.

[0010] Further, in the three-dimensional spinal model visualization, the vertebral region with a deviation δ > 0.7 is marked with a pulsed red light, and the red light flickering frequency f is positively correlated with the deviation: f = 2δ Hz, and a text scrolling reminder box containing the last 3 times of unexecuted nursing measures is generated on the right side of the model.

[0011] Further, the standardization processing of the nursing record expected value includes: Z-score normalization for electromyographic signals, Min-Max standardization to the [0, 1] interval for spinal motion parameters, and relative temperature difference ΔT / T0 calculation based on the basic body temperature on the operation day for wound temperature data.

[0012] Further, in the anatomical site matching relationship, for multi-segment surgical patients, the correlation rules of intervertebral fusion device position and adjacent upper and lower vertebral segments are established in the mapping relationship table, and when T12-L1 segment anomaly is detected, L2-L3 segment data is checked synchronously.

[0013] Further, the deployment density of the temperature-sensitive sensor array is determined according to the length of the surgical incision L, with (L / 2+1) sensors deployed per centimeter, the sensor spacing error is controlled within ±1mm, and the data acquisition is transmitted to the central processor through LoRa wireless transmission with a delay of no more than 200ms.

[0014] Further, the nursing quality control alarm escalation logic contains a double verification mechanism: when δ > 0.7 is detected for the first time, the nursing group leader review process is started, and if no review confirmation is received within 12 hours and the δ value continues to rise, the alarm level is automatically upgraded and the event record is marked as red emergency state.

[0015] Further, the mapping relationship table contains a pain control index item, and the pain control index item configuration requires: VAS score value is collected through electronic pain assessment scale, and a dynamic correlation matrix is established between the VAS score value and the L4-L5 segment paraspinal muscle group electromyographic signals and spinal motion data.

[0016] Compared with the prior art, the application has the beneficial effects that: The real-time dynamic evaluation system for postoperative nursing quality of the spine is constructed, and breakthrough improvement is made on the traditional nursing evaluation method by means of multi-source data fusion, so that the evaluation quality and efficiency of the nursing sensitive index can be significantly improved.

[0017] The application effectively solves the problem of poor timeliness of the traditional method, adopts a state-time-space correlation mechanism, can match the electromyographic signal, the spinal motion data, the wound parameter and the nursing record in real time, the traditional evaluation may need weeks or more for statistical analysis, and the application realizes real-time abnormal detection response and timely feedback of the evaluation situation, so that nursing adjustment is facilitated.

[0018] The application converts the complex weight calculation which depends on expert consultation and hierarchical analysis into standardized deviation δ calculation by presetting the threshold of the mapping relationship table and using a time window synchronization algorithm for automatic matching, so that the nursing staff can obtain the evaluation result without professional statistical knowledge, and the application threshold is reduced.

[0019] The application breaks through the technical bottleneck, adopts a heavy dynamic adjustment mechanism, can automatically correct the index sensitivity according to real-time monitoring data such as wound temperature mutation, and combines three-dimensional visualization output to optimize the adjustment cycle of the nursing scheme from monthly to real-time update as needed. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 It is the evaluation method flowchart of the application. DETAILED DESCRIPTION

[0021] The application embodiments are further described in detail below in combination with the drawings of the specification, so that the purposes, technical solutions and technical effects of the application are more clearly presented.

[0022] As shown in the drawings, Figure 1 The application discloses a postoperative nursing quality sensitive index evaluation method for the spine, which comprises: A mapping relationship table of the nursing quality sensitive index to be evaluated and the postoperative nursing record item is established, each index in the mapping relationship table is associated with the nursing record data field and the expected threshold, and the anatomical site matching relationship is labeled based on the spinal segment code; The recovery data of the postoperative patient is collected and a dynamic space-time correlation with the mapping relationship table is established, the recovery data comprises real-time electromyographic signals of the L4-L5 segment paraspinal muscle group monitored by a body surface electromyographic sensor, three-dimensional mechanical parameters of spinal flexion / rotation obtained by an inertial measurement unit, and wound microenvironment data dynamically collected by a temperature-sensitive sensor array implanted around the incision; The time window synchronization algorithm is used to match the recovered data changes and nursing operation records, including the correspondence between the abnormal period of electromyographic signal and the turning-over nursing record, the correlation between the mutation point of spinal motion data and the body position adjustment, and the control between the wound parameter fluctuation and the dressing change operation. If data abnormalities are detected but the nursing record does not respond, a warning is triggered. Based on the correlation results, the standardized deviation δ of the actual data of each index and the expected value of the nursing record is calculated according to the formula: The index weight is dynamically adjusted. If δ is greater than 0.7 for three consecutive times and there is no matched effective nursing intervention, the index is upgraded to a nursing quality control alarm. Finally, the evaluation results are visualized and output on the three-dimensional spinal model, and the deviation exceeding the limit area and the corresponding nursing record summary are highlighted.

[0023] The above scheme forms a complete closed-loop processing flow: first, a mapping relationship table of indexes and nursing records is established. The body surface electromyographic sensor, inertial measurement unit and temperature sensitive sensor array are used to collect paraspinal muscle electromyographic signals to monitor muscle spasm, collect three-dimensional mechanical parameters of the spine to evaluate body position stability, and collect wound microenvironment data to detect infection risk. These physiological parameters and nursing operation records are correlated through a time window synchronization algorithm, and the standardized deviation δ is calculated. The index weight is dynamically adjusted, and finally the deviation area is presented in three dimensions. Each step forms a synergistic mechanism of data acquisition, time sequence matching, deviation analysis and dynamic optimization.

[0024] Here, the establishment of dynamic spatio-temporal correlation specifically includes: the L4-L5 segment paraspinal muscle group signal collected by the body surface electromyographic sensor is matched with the turning-over operation in the nursing record according to the time stamp, and when the electromyographic signal fluctuation amplitude exceeds the baseline by 30% and lasts for 5 minutes without recording the turning-over operation, a first-level nursing omission warning is automatically generated. By setting the matching mechanism of body surface electromyographic signal fluctuation and turning-over operation record, the typical characteristics of abnormal activity of paraspinal muscle group in patients after spinal surgery are targeted. When the electromyographic signal fluctuation amplitude exceeds the baseline by 30% and lasts for 5 minutes without recording the turning-over operation, a first-level nursing omission warning is forced to intervene, which can effectively prevent pressure ulcers or nerve compression caused by long-term body position fixation. The threshold is determined through clinical pre-test: statistics of 100 cases of postoperative spinal patients show that when the electromyographic signal lasts for more than 30% of the baseline, the probability of skin ischemic injury within 2 hours significantly increases (P<0.01).

[0025] In the time window synchronization algorithm, an adaptive sliding window mechanism is adopted for the matching of spinal motion data. The window width W is dynamically adjusted according to the motion speed. When the inertial measurement unit detects an angular velocity ωxyz>15° / s, the matching window is compressed from ±5 minutes to ±2 minutes, and the corresponding period of body position adjustment is highlighted in the nursing record interface. The setting of the adaptive sliding window mechanism is based on the nonlinear characteristics of spinal motion data. When the patient suddenly stands up or turns over violently, ωxyz>15° / s, the traditional fixed time window will increase the mismatch rate of body position adjustment records. By compressing the window to ±2 minutes, the matching accuracy can be improved to more than 92% while ensuring timeliness. The highlight function in the nursing interface solves the problem that nurses easily ignore rapid body position changes when multitasking.

[0026] For the index weight formula , the decay coefficient α is dynamically configured according to the wound type. For posterior lumbar disc removal patients, α=0.25, and for anterior vertebral fusion patients, α=0.4. The wound exudate pH value is fine-tuned by ±0.1 through the real-time feedback of the implanted temperature-sensitive sensor. The setting of the weight decay coefficient for different surgical types is based on the biological differences in wound healing of the two types of surgery: anterior surgery involves exposure of large blood vessels, and the pH value of the wound exudate fluctuates 40% more than posterior surgery. Therefore, a higher α value is needed to prevent false positives. The real-time fine-tuning mechanism monitors changes in the wound microenvironment through temperature-sensitive sensors: clinical data shows that when the pH value exceeds the range of 7.25-7.35, the risk of wound infection increases by 3 times. At this time, a coefficient correction of ±0.1 can provide 12-24 hours of early warning of infection signs.

[0027] In three-dimensional spinal model visualization, the pulse red light marker is used in the vertebral region with a deviation δ>0.7, and the red light flicker frequency f is positively correlated with the deviation: f=2δHz. At the same time, a text scrolling reminder box containing the last 3 times of unexecuted nursing measures is generated on the right side of the model. The pulse red light marker design conforms to human visual perception characteristics. The setting of the flicker frequency f=2δHz makes the high-risk area with δ>0.7 correspond to the "urgent intervention" state in clinical classification; this can attract the attention of medical staff first. Eye tracking experiments show that the red flashing marker at this frequency shortens the gaze dwell time by 0.8 seconds compared to the conventional fixed direction display, and the emergency response speed is improved by 22%. The text scrolling reminder box is limited to the last 3 records, which not only avoids information overload, but also conforms to the "recent events are given priority" principle in nursing operations.

[0028] Here, the standardization of nursing record expectation value includes: Z-score normalization for electromyographic signals, Min-Max standardization to the [0, 1] interval for spinal motion parameters, and relative temperature difference ΔT / T0 calculation based on the basic body temperature on the operation day for wound temperature data. The differential standardization processing method is aimed at the physical meaning of different data types: Z-score is suitable for electromyographic signals that conform to normal distribution; Min-Max standardization can preserve the original proportional relationship of spinal motion parameters, such as the percentage of flexion angle and physiological limit; the relative temperature difference calculation eliminates the interference of individual basic body temperature difference. Comparative tests show that this scheme can increase the sensitivity of indicators by 18%-25% compared with single standardization method.

[0029] In the anatomical site matching relationship, for multi-segment operation patients, the correlation rules of interbody fusion cage position and adjacent upper and lower vertebral segments are established in the mapping relationship table, and when T12-L1 segment abnormalities are detected, L2-L3 segment data are checked synchronously. Multi-segment correlation rules are derived from the biomechanical properties of the spine, and the incidence of abnormal activity of the fusion cage adjacent segment is many times that of the distal segment. After setting the synchronous checking mechanism of T12-L1 and L2-L3, the missed detection rate of adjacent segment lesions is reduced from 14.7% to 3.2% in the clinical application of 256 multi-segment operation patients. This rule is particularly suitable for early dynamic stability evaluation after interbody fusion cage implantation.

[0030] Here, the deployment density of the temperature-sensitive sensor array is determined according to the length L of the surgical incision, with L / 2+1 sensors deployed per centimeter, the sensor spacing error is controlled within ±1mm, and the data acquisition is transmitted to the central processor through LoRa wireless transmission with a delay of no more than 200ms. The setting of sensor density L / 2+1 per centimeter ensures that there are at least two sensors covering the edge area of the wound. The 1mm spacing error control ensures that the temperature gradient monitoring resolution reaches 0.1℃ / mm: experiments have proved that this accuracy can detect early infection foci with a diameter of 5mm. The 200ms delay limit meets the real-time requirements of nursing: when the threshold is exceeded, the warning delay of postoperative acute bleeding and other events will exceed the clinical golden disposal window.

[0031] A double verification mechanism is provided for the quality control alarm escalation logic. When the first detection of delta > 0.7, the nurse-in-charge review process is started, if the review confirmation is not received within 12 hours and the delta value continues to rise, the alarm level is automatically upgraded and the event record is marked as a red emergency state. The double verification mechanism balances the sensitivity and specificity of the alarm, and the first delta > 0.7 can filter out false positives through the nurse-in-charge review; the 12-hour observation period is set according to the evolution rule of postoperative complications: 80% of nursing-related complications will have a continuous increase in delta value within this period. The red emergency state mark makes the display priority of this event in the electronic medical record system the highest, ensuring that the forced processing process starts.

[0032] The implementation process includes the following steps: (1) Establish a mapping relationship table between the sensitive indicators to be evaluated for nursing quality and the postoperative nursing record items. In the mapping relationship table, each indicator is associated with a nursing record data field and an expected threshold, and is annotated with an anatomical site matching relationship based on spinal segment coding. For patients with multiple segment surgery, establish the association rule between the intervertebral fusion cage position and the adjacent upper and lower vertebrae in the mapping relationship table. When detecting T12-L1 segment abnormalities, check the L2-L3 segment data simultaneously.

[0033] (2) The specific implementation of collecting postoperative patient recovery data is to collect L4-L5 segment paraspinal muscle group signals in real time through surface electromyography sensors, and establish event sequence matching according to the time stamp and the turning operation in the nursing record. At the same time, obtain the three-dimensional mechanics parameters of spinal flexion / rotation through the inertial measurement unit, and dynamically collect the wound microenvironment data through the temperature-sensitive sensor array deployed around the incision. The deployment density of the temperature-sensitive sensor array is determined according to the length L of the surgical incision, with L / 2+1 sensors per centimeter, and the sensor spacing error is controlled within ±1 mm. The collected data is transmitted to the central processor through LoRa wireless transmission with a delay of no more than 200 ms.

[0034] (3) When matching the recovery data changes and nursing operation records using the time window synchronization algorithm, for spinal motion data matching, an adaptive sliding window mechanism is used, and the window width W is dynamically adjusted according to the motion speed: when the inertial measurement unit detects that the angular velocity ωxyz > 15° / s, the matching window is compressed from ±5 minutes to ±2 minutes. The algorithm realizes the correspondence of electromyographic signal abnormal period and turning nursing record, the correlation of spinal motion data mutation point and body position adjustment, and the control of wound parameter fluctuation and dressing operation. When the electromyographic signal fluctuation amplitude exceeds 30% of the baseline and no turning operation is recorded for 5 minutes, a first-level nursing omission warning is automatically generated.

[0035] (4) The process of calculating the actual data of each index and the expected value of the nursing record based on the correlation results is as follows: the electromyographic signal is normalized by Z-score, the spinal motion parameters are standardized to the [0, 1] interval by Min-Max, and the wound temperature data are calculated based on the relative temperature difference AT / T0 of the basic body temperature on the day of the operation. After calculating the standardized deviation δ, the index weight is dynamically adjusted according to the formula, and the decay coefficient α is dynamically configured according to the wound type: α = 0.25 for posterior lumbar disc resection patients and α = 0.4 for anterior vertebral fusion patients, and the wound exudate pH value is adjusted by ±0.1 through the real-time feedback of the implantable temperature-sensitive sensor. The index is upgraded to a nursing quality control alarm for three consecutive times δ > 0.7 and no matching effective nursing intervention. The nursing quality control alarm upgrade logic includes a double verification mechanism: when δ > 0.7 is detected for the first time, the nursing group leader review process is started, and if no review confirmation is received within 12 hours and the δ value continues to rise, the alarm level is automatically upgraded and the event record is marked as a red emergency state.

[0036] (5) When the evaluation results are finally visualized and output on the three-dimensional spinal model, the vertebral segment area with a deviation δ > 0.7 is marked with a pulsatile red light, and the red light flickering frequency f is positively correlated with the deviation: f = 2δ Hz, and a text scrolling reminder box containing the last 3 nursing measures not performed is generated on the right side of the model. The entire implementation process realizes closed-loop management from data collection, analysis to result display, forming a complete technical solution for dynamic optimization of nursing quality.

[0037] Here, the mapping relationship table also includes pain control index items, and the VAS score is collected through the electronic pain assessment scale, and a dynamic correlation matrix is established with the L4-L5 segment paraspinal muscle group electromyographic signal and spinal motion data. When the VAS score exceeds the threshold value of the corresponding activity state, a three-level verification mechanism is started: check whether the electromyographic signal frequency spectrum power in the 50-150 Hz frequency band increases by ≥20%, verify whether the sagittal plane offset angle in the three-dimensional mechanics parameters of the spine breaks through the expected value of the rehabilitation process, and compare whether the Pearson correlation coefficient between the wound temperature gradient and the pain score is >0.4. If at least two verifications are true, a nursing measure missing alarm is generated and pushed to the pain management priority queue, and the deviation calculation of other sensitive indexes in this period is frozen until the alarm is lifted. The three-level verification mechanism for pain control aims to overcome the subjectivity of traditional VAS score, with electromyographic spectrum power growth verifying neuropathic pain, sagittal plane offset angle verifying mechanical pain, and wound temperature gradient verifying inflammatory pain. Only when at least two verifications are true can the alarm be triggered, with a false positive rate controlled below 7%. The design of freezing other index calculations avoids multiple alarm interference and ensures the priority allocation of pain management resources.

[0038] The above merely describes preferred and feasible embodiments of the present application, and is not intended to limit the patent application scope of the present application. Any equivalent changes, equivalent replacements or modified changes within the technical spirit and principles suggested by the present application should be included in the patent protection scope of the present application.

Claims

1. A method for evaluating a quality of postoperative care of a spine, characterized by, The method comprises the following steps: Establish a mapping table of sensitive indicators for evaluating nursing quality and postoperative nursing record items, each indicator in the mapping table is associated with a nursing record data field and an expected threshold, and the mapping table is annotated with an anatomic site matching relationship based on a spinal segment code; Collect postoperative patient recovery data and establish a dynamic spatiotemporal association with the mapping table, the recovery data includes real-time electromyography signals of L4-L5 segment paraspinal muscle groups monitored by a body surface electromyography sensor, three-dimensional mechanics parameters of spinal flexion / rotation obtained by an inertial measurement unit, and wound microenvironment data dynamically collected by a temperature-sensitive sensor array implanted around the incision; Use a time window synchronization algorithm to match the changes in the recovery data and the nursing operation records, including the correspondence between abnormal periods of electromyography signals and turning over nursing records, the correlation between mutation points of spinal movement data and body position adjustment, and the comparison between wound parameter fluctuations and dressing change operations, if data abnormalities are detected but there is no response in the nursing records, a warning is triggered; Based on the results of the correlation calculation of each index actual data and the expected value of nursing records standard deviation δ, according to the index weight formula: Dynamic adjustment of index weight, for three times δ> 0.7 and no matching effective nursing intervention of index upgrade for nursing quality control alarm; Finally, the evaluation results are visualized and output on a three-dimensional spinal model, and the areas with deviation exceeding the limit and the corresponding nursing record summaries are highlighted.

2. The method of claim 1, wherein: The establishment of the dynamic spatiotemporal association specifically includes: the L4-L5 segment paraspinal muscle group signals collected in real time by the body surface electromyography sensor are matched with the turning over operation in the nursing records according to the time stamp, when the fluctuation amplitude of the electromyography signals exceeds 30% of the baseline and no turning over operation is recorded for 5 minutes, a first-level nursing omission warning is automatically generated.

3. The method of claim 1, wherein the method further comprises: determining a quality of care for the patient based on the one or more post-operative indicators. In the time window synchronization algorithm, an adaptive sliding window mechanism is used for matching the spinal movement data, the window width W is dynamically adjusted according to the movement speed: when the inertial measurement unit detects that the angular velocity ωxyz is greater than 15° / s, the matching window is compressed from ±5 minutes to ±2 minutes, and the body position adjustment omission record in the corresponding time period is highlighted on the nursing record interface.

4. The method of claim 1, wherein: The decay coefficient α in the index weight formula is dynamically configured according to the wound type: for patients undergoing posterior lumbar disc resection, α=0.25, for patients undergoing anterior vertebral fusion, α=0.4, and the wound exudate pH value fed back in real time by the implanted temperature-sensitive sensor is adjusted by ±0.

1.

5. The method of claim 1, wherein: In the three-dimensional spinal model visualization, the regions of the vertebrae with a deviation δ>0.7 are marked with a pulsatile red light, the red light flickering frequency f is positively correlated with the deviation: f=2δHz, and a text scrolling reminder box containing the last 3 nursing measures not performed is generated on the right side of the model.

6. The method of claim 1, wherein: The standardization processing of the expected value of the nursing records includes: Z-score normalization for electromyography signals, Min-Max standardization to the [0,1] interval for spinal movement parameters, and relative temperature difference ΔT / T0 calculation based on the basal body temperature on the day of the operation for wound temperature data.

7. The method of claim 1, wherein the method is a method of evaluating postoperative care quality sensitive indicators for a spine. In the anatomic site matching relationship, for patients with multiple segment surgery, an association rule between the intervertebral fusion cage position and the adjacent upper and lower vertebrae is established in the mapping table, and when T12-L1 segment abnormalities are detected, L2-L3 segment data are checked synchronously.

8. The method of claim 1, wherein: The deployment density of the temperature-sensitive sensor array is determined according to the length L of the surgical incision, (L / 2+1) sensors per centimeter are deployed, the sensor spacing error is controlled within ±1 mm, and the data acquisition is transmitted to the central processor through LoRa wireless transmission with a delay of no more than 200 ms.

9. The method of claim 1, wherein: The nursing quality control alarm escalation logic includes a double verification mechanism: when δ>0.7 is detected for the first time, the nursing group leader review process is started, and if no review confirmation is received within 12 hours and the δ value continues to rise, the alarm level is automatically upgraded and the event record is marked as a red emergency state.

10. The method of claim 1, wherein: The mapping relationship table includes pain control index items, and the pain control index items are configured to require VAS score values to be collected through an electronic pain assessment scale, and to establish a dynamic correlation matrix with L4-L5 segment paraspinal muscle group electromyographic signals and spinal motion data.

Images