A medical institution patient satisfaction evaluation method

Abstract

This invention specifically relates to a method for evaluating patient satisfaction in medical institutions, falling under the technical field of medical service quality assessment. The method includes: performing a logarithmic ratio calculation and normalization on the waiting time of each stage in the actual waiting time sequence with the historical median waiting time of the same department during the same period, generating a waiting deviation index sequence. In this invention, objective timestamp data and interaction behavior logs are directly obtained from the medical institution's information system. This automatic data collection replaces patients' subjective recollection afterward, solving the technical problems of distorted assessment results caused by recall bias, selective responses, and low questionnaire return rates. This achieves the technical effect of objective assessment data sources and complete sample coverage. By converting absolute waiting time into a waiting deviation index based on the historical distribution of the same department during the same period, the assessment index reflects the relative degree of deviation in the patient's experience under specific circumstances.

Description

Technical Field

[0001] This invention relates to the field of medical service quality assessment technology, and in particular to a method for assessing patient satisfaction in medical institutions. Background Technology

[0002] When conducting patient satisfaction assessments, medical institutions typically distribute paper or electronic questionnaires after the patient's visit to collect the patient's subjective evaluation of each service aspect and calculate scores for each dimension using statistical methods.

[0003] The above methods rely on patients' active participation in completing the questionnaires and use their post-hoc recollections as the source of information. In practice, the questionnaire return rate is generally low, resulting in insufficient sample representativeness and systematic bias in the assessment results.

[0004] Furthermore, static, single-response ratings cannot reflect the changes in a patient's experience throughout the entire medical process. The assessment results deviate significantly from the actual patient satisfaction, making it impossible to accurately pinpoint specific service weaknesses. Summary of the Invention

[0005] The purpose of this invention is to provide a method for evaluating patient satisfaction in medical institutions in order to solve the above-mentioned problems.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: A method for assessing patient satisfaction in a medical institution includes: The system obtains the full-process timestamp data of a single patient's visit from the medical institution's information system, calculates the time difference between each adjacent step, and generates an actual waiting time sequence. The waiting time of each stage in the actual waiting time sequence is logarithmically compared with the historical median waiting time of the same department during the same period and then normalized to generate a waiting deviation index sequence. Acquire interactive behavior data of target patients during their medical visits, and sum the behavioral data in each time period by type to generate a sequence of behavioral anxiety scores; The waiting deviation index sequence and the behavioral anxiety score sequence are input into the satisfaction evolution prediction model based on long short-term memory network, and the step-by-step satisfaction prediction curve is output. The difference between adjacent stages is calculated on the step-by-step satisfaction prediction curve to identify key turning points where the decline in satisfaction exceeds a preset threshold, and a patient satisfaction assessment report is generated.

[0007] Preferably, the full-process timestamp data includes the registration time, the start time of waiting, the time of being called, the time of entering the consultation room, the time of ordering examinations and tests, the time of waiting for examinations, the time of completion of examinations, the time of picking up medication, and the time of leaving the hospital.

[0008] Preferably, the step of performing a logarithmic ratio calculation and normalization of the waiting time at each stage of the actual waiting time sequence with the historical median waiting time of the same department during the same period to generate a waiting deviation index sequence includes: Extract a sample set of patients from historical medical records who have visited the same department and during the same time period as the target patient, statistically analyze the distribution of waiting time at each stage, and calculate the historical median waiting time at each stage; For each stage, the natural logarithm of the ratio of the actual waiting time of the target patient to the historical median waiting time of that stage is taken, and then divided by the standard deviation of the logarithm of the historical waiting time of that stage to obtain the waiting deviation index of that stage. The waiting deviation indexes of each stage are arranged in order of consultation time to generate a waiting deviation index sequence.

[0009] Preferably, the same period refers to the time period corresponding to the historical records that are on the same workday and within the same hour period as the time of the target patient's visit.

[0010] Preferably, the step of acquiring interactive behavior data during the target patient's medical visit, and weighting and summing the behavioral data by type for each time period to generate a behavioral anxiety score sequence includes: Acquire interactive behavior data generated by the target patient at each stage of the medical visit, including the number of self-service machine operations, the number of triage consultations, the number of complaint window access records, and the frequency of online platform message sending; According to the preset time period division rules, the interactive behavior data is allocated to each time period, and the behavior data in each time period is counted according to the behavior type. Based on the negative emotion correlation weights of each behavior type, the behavior counts in each time period are weighted and summed to obtain the behavior anxiety score for each time period, thus generating a behavior anxiety score sequence.

[0011] Preferably, in the negative emotion correlation weighting, the weight of accessing the complaint window is higher than the weight of operating the self-service machine, and the weight of sending high-frequency messages on the online platform is higher than the weight of triage consultation.

[0012] Preferably, before inputting the waiting deviation index sequence and the behavioral anxiety score sequence into the satisfaction evolution prediction model, the two are standardized to make the two types of features have the same numerical scale. The input to the satisfaction evolution prediction model is a two-dimensional feature vector sequence consisting of the waiting deviation index of each stage and the behavioral anxiety score of the corresponding time period. At each time step, the long short-term memory network unit receives the two-dimensional feature vector of the current stage and updates the hidden state. The hidden state is passed sequentially between stages and projected by the linear fully connected layer to become the satisfaction prediction value of that stage.

[0013] Preferably, the satisfaction evolution prediction model uses historical patient data that simultaneously possesses complete timestamp records of the medical treatment process, interaction behavior logs, and valid satisfaction questionnaire scores as training samples, uses questionnaire scores as supervision labels, and employs a mean squared error loss function for training; during the inference stage, it relies solely on objectively collected timestamp data and interaction behavior logs to generate prediction results.

[0014] Preferably, the step of identifying key turning points where the decline in satisfaction exceeds a preset threshold and generating a patient satisfaction assessment report includes: Calculate the change in satisfaction between adjacent stages in the satisfaction prediction curve to obtain the satisfaction gradient sequence. The links in the satisfaction gradient sequence with negative values ​​and absolute values ​​exceeding a preset threshold are marked as key turning points; For each key turning point, the corresponding waiting deviation index and behavioral anxiety score are standardized and compared. When the standardized value of the waiting deviation index is higher than the standardized value of the behavioral anxiety score, the decrease in satisfaction is attributed to abnormal waiting time. When the standardized value of the behavioral anxiety score is higher than the standardized value of the waiting deviation index, the decrease in satisfaction is attributed to abnormal service interaction. The system integrates and outputs key turning point markers, the degree of decline in satisfaction, and attribution information to generate a patient satisfaction assessment report.

[0015] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: 1. This invention solves the technical problems of distorted assessment results caused by recall bias, selective responses, and low questionnaire return rates by directly acquiring objective timestamp data and interaction behavior logs from medical institution information systems to automatically collect data and replace patients' subjective recollections afterward. This achieves the technical effect of objective assessment data sources and complete sample coverage. By converting absolute waiting time into a waiting deviation index based on the historical distribution of the same department during the same period, the assessment index reflects the relative degree of deviation in the patient's experience in a specific situation.

[0016] 2. This invention models two types of temporal features using a Long Short-Term Memory (LSTM) network, capturing the cumulative dependencies between each stage of the patient experience. This enables dynamic satisfaction prediction at each stage, solving the technical problem that a single static rating cannot reflect changes in the overall experience. Based on gradient analysis and attribution of the satisfaction prediction curve, it can pinpoint the specific stages leading to a significant decrease in satisfaction and their main related factors, achieving the technical effect of accurately identifying weak service links. Attached Figure Description

[0017] Further details, features, and advantages of this application are disclosed in the following description of exemplary embodiments in conjunction with the accompanying drawings, in which: Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0018] Several embodiments of this application will now be described in more detail with reference to the accompanying drawings to enable those skilled in the art to implement this application. This application may be embodied in many different forms and for various purposes and should not be limited to the embodiments set forth herein. These embodiments are provided to make this application thorough and complete, and to fully convey the scope of this application to those skilled in the art. The embodiments described do not limit this application.

[0019] Unless otherwise defined, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. It will be further understood that terms such as those defined in commonly used dictionaries shall be interpreted as having a meaning consistent with their meaning in the relevant field and / or the context of this specification, and shall not be interpreted in an idealized or overly formal sense unless expressly defined herein.

[0020] Example 1 Its specific implementation method is combined with the appendix Figure 1 Please provide a detailed explanation.

[0021] In this embodiment, it includes: On a weekday morning in 20XX, the Department of Gastroenterology at a general hospital conducted a full-process satisfaction assessment on a patient (patient number P0047). The patient presented with abdominal discomfort, and the entire process included registration, waiting, consultation, examination, and medication dispensing. The computing device, through its data interface with the hospital's HIS system, read P0047's visit timestamp record and terminal interaction logs in real time, sequentially performing calculations and analyses for each step.

[0022] Step S1 extracts the full-process timestamp of P0047's current visit from the HIS system, calculates the time difference between adjacent steps, and generates the actual waiting time sequence.

[0023] Table 1. Timestamps and actual waiting times of patient P0047's entire medical process. Step S2 involves selecting historical samples from the medical records that are in the same department (gastroenterology) and during the same time period (8:00 AM to 11:00 AM on weekdays) as P0047, calculating the historical median waiting time and logarithmic standard deviation for each step, and then applying the formula... Calculate the waiting deviation index for each stage. Among them, For the first The actual waiting time for each step Waiting time to reach the historical median The standard deviation is the logarithm. Taking step 5 as an example, the actual waiting time is... minutes, historical median duration minutes, log standard deviation Substituting into the formula, we get This indicates that the waiting time in this step is significantly higher than expected. The calculation results for each step are as follows.

[0024] Table 2 Historical benchmark parameters and waiting deviation index for each stage Step S3 extracts the interactive behavior data generated by P0047 at each stage of the medical visit from the HIS system and terminal device logs, distributes and counts them according to the time period corresponding to each stage, and then performs a weighted sum based on the negative emotion correlation weight of each behavior type to generate a behavioral anxiety score sequence. In the weight setting, the complaint window access... Online platform message sending Guidance and consultation Self-service machine operation Taking stage 5 as an example, during this period, the patient initiated 2 triage consultations, accessed the complaint window 1 time, and sent 3 online messages; all other actions were 0. Substituting these into the formula... have to .in, For the first Weights of class behaviors For the first In the first stage The number of times a class of behaviors occurs.

[0025] Table 3. Interactive behavior data and behavioral anxiety scores for each time period. Step S4 involves Z-score standardization of the waiting deviation index sequence and behavioral anxiety score sequence for each stage, followed by concatenation into a two-dimensional feature vector sequence. This sequence is then input into the trained LSTM satisfaction evolution prediction model in chronological order. At each time step, the model receives the feature vector of the current stage, updates the hidden state, and maps the hidden state to the predicted satisfaction value for that stage via a fully connected linear layer. Since the hidden state is sequentially passed between stages, the predicted value for stage 5 incorporates the accumulated experience information from previous stages. The predicted satisfaction values ​​output by each stage constitute a stage-by-stage satisfaction prediction curve, with satisfaction scores ranging from 1 to 5.

[0026] Table 4. Input features of the LSTM model (after standardization) and predicted satisfaction values Step S5 calculates the difference between adjacent stages on the above satisfaction prediction curve to obtain the satisfaction gradient sequence. The preset threshold is set to 0.8 points. The change in satisfaction from stage 4 to stage 5 is... When the absolute value exceeds the threshold, it is marked as a critical turning point. The standardized value of the waiting deviation index (1.53) and the standardized value of the behavioral anxiety score (1.89) for this point are extracted. Both are significantly high, but the standardized value of behavioral anxiety is higher than that of the standardized value of the waiting deviation index. Therefore, the decrease in satisfaction at this point is attributed to both abnormal waiting time and abnormal service interaction, with abnormal service interaction being the primary contributing factor. A patient satisfaction assessment report is then generated.

[0027] Table 5. Satisfaction Gradient Analysis and Attribution of Key Turning Points Throughout the evaluation process, data starts with raw timestamps from the HIS system. Step S1 transforms this data into an actual waiting time sequence, which is then compared with historical distributions in Step S2 to create a waiting deviation index sequence comparable across departments. Step S3 converts terminal interaction behavior logs into a quantified behavioral anxiety score sequence. In Step S4, both sequences are input as time-series features into an LSTM model, outputting a satisfaction prediction value that integrates accumulated experience information at each stage. Step S5 performs gradient analysis on the prediction curve, accurately pinpointing the examination waiting stage as a critical turning point, attributing it primarily to abnormal service interaction. This results in a satisfaction evaluation report with clear locational evidence, all without relying on patients actively completing questionnaires.

[0028] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

[0029] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0030] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for evaluating patient satisfaction in medical institutions, characterized in that, include: The system obtains the full-process timestamp data of a single patient's visit from the medical institution's information system, calculates the time difference between each adjacent step, and generates an actual waiting time sequence. The waiting time of each stage in the actual waiting time sequence is logarithmically compared with the historical median waiting time of the same department during the same period and then normalized to generate a waiting deviation index sequence. Acquire interactive behavior data of target patients during their medical visits, and sum the behavioral data in each time period by type to generate a sequence of behavioral anxiety scores; The waiting deviation index sequence and the behavioral anxiety score sequence are input into the satisfaction evolution prediction model based on long short-term memory network, and the step-by-step satisfaction prediction curve is output. The difference between adjacent stages is calculated on the step-by-step satisfaction prediction curve to identify key turning points where the decline in satisfaction exceeds a preset threshold, and a patient satisfaction assessment report is generated.

2. The method for evaluating patient satisfaction in a medical institution according to claim 1, characterized in that, The full-process timestamp data includes the registration time, the start time of waiting, the time of being called, the time of entering the consultation room, the time of ordering examinations and tests, the time of waiting for examinations, the time of completing examinations, the time of picking up medication, and the time of leaving the hospital.

3. The method for evaluating patient satisfaction in a medical institution according to claim 1, characterized in that, The waiting time at each stage of the actual waiting time sequence is logarithmically compared to the historical median waiting time for the same period in the same department, and then normalized to generate a waiting deviation index sequence, including: Extract a sample set of patients from historical medical records who have visited the same department and during the same time period as the target patient, statistically analyze the distribution of waiting time at each stage, and calculate the historical median waiting time at each stage; For each stage, the natural logarithm of the ratio of the actual waiting time of the target patient to the historical median waiting time of that stage is taken, and then divided by the standard deviation of the logarithm of the historical waiting time of that stage to obtain the waiting deviation index of that stage. The waiting deviation indexes of each stage are arranged in order of consultation time to generate a waiting deviation index sequence.

4. The method for evaluating patient satisfaction in a medical institution according to claim 3, characterized in that, The same time period refers to the time period corresponding to the historical records that are on the same workday and within the same hour period as the target patient's visit time.

5. The method for evaluating patient satisfaction in a medical institution according to claim 1, characterized in that, Acquire interactive behavior data of target patients during their medical visits, weight and sum the behavioral data for each time period according to type, and generate a sequence of behavioral anxiety scores, including: Acquire interactive behavior data generated by the target patient at each stage of the medical visit, including the number of self-service machine operations, the number of triage consultations, the number of complaint window access records, and the frequency of online platform message sending; According to the preset time period division rules, the interactive behavior data is allocated to each time period, and the behavior data in each time period is counted according to the behavior type. Based on the negative emotion correlation weights of each behavior type, the behavior counts in each time period are weighted and summed to obtain the behavior anxiety score for each time period, thus generating a behavior anxiety score sequence.

6. The method for evaluating patient satisfaction in a medical institution according to claim 1, characterized in that, In the weighting of negative emotion correlation, the weight of accessing the complaint window is higher than that of operating the self-service machine, and the weight of sending high-frequency messages on the online platform is higher than that of triage consultation.

7. The method for evaluating patient satisfaction in a medical institution according to claim 1, characterized in that, Before inputting the waiting deviation index sequence and the behavioral anxiety score sequence into the satisfaction evolution prediction model, both are standardized to make the two types of features have the same numerical scale. The input to the satisfaction evolution prediction model is a two-dimensional feature vector sequence consisting of the waiting deviation index of each stage and the behavioral anxiety score of the corresponding time period. At each time step, the long short-term memory network unit receives the two-dimensional feature vector of the current stage and updates the hidden state. The hidden state is passed sequentially between stages and projected by the linear fully connected layer to become the satisfaction prediction value of that stage.

8. The method for evaluating patient satisfaction in a medical institution according to claim 1, characterized in that, The satisfaction evolution prediction model uses historical patient data with complete timestamp records of the medical process, interaction behavior logs, and valid satisfaction questionnaire scores as training samples, with questionnaire scores as supervision labels, and is trained using the mean squared error loss function; the inference stage relies only on objectively collected timestamp data and interaction behavior logs to generate prediction results.

9. A method for evaluating patient satisfaction in a medical institution according to claim 1, characterized in that, Identify key turning points where satisfaction declines beyond a preset threshold, and generate a patient satisfaction assessment report, including: Calculate the change in satisfaction between adjacent stages in the satisfaction prediction curve to obtain the satisfaction gradient sequence. The links in the satisfaction gradient sequence with negative values ​​and absolute values ​​exceeding a preset threshold are marked as key turning points; For each key turning point, the corresponding waiting deviation index and behavioral anxiety score are standardized and compared. When the standardized value of the waiting deviation index is higher than the standardized value of the behavioral anxiety score, the decrease in satisfaction is attributed to abnormal waiting time. When the standardized value of the behavioral anxiety score is higher than the standardized value of the waiting deviation index, the decrease in satisfaction is attributed to abnormal service interaction. The system integrates and outputs key turning point markers, the degree of decline in satisfaction, and attribution information to generate a patient satisfaction assessment report.

Images