In the complex and ever-changing medical business environment, the result reporting function of the LIS system not only carries the mission of information transmission, but also is a key link to improve diagnosis and treatment efficiency and optimize service processes. This article will discuss the design logic and optimization strategies of the reporting function in multiple business scenarios from the practical application, and provide a strong reference for the product design of the LIS system through case analysis and experience sharing.
1. The evolution from data carrier to value bridge
As the final outlet for LIS to interact directly with users, the design quality of the result report directly determines the conversion efficiency of test data into clinical value.
Clinical scenarios require immediacy and accuracy of reporting: in the emergency room, troponin results in patients with myocardial infarction need to be captured by doctors within 3 minutes, otherwise the best intervention window may be missed; In the outpatient clinic, diabetic patients hold a biochemical report full of professional terms, confused about the meaning of 6.8% glycated hemoglobin; At the laboratory quality control meeting, the director needs to quickly extract the error trend of different instruments from the report to evaluate the quality of this month’s test – these real-world scenarios point to a core problem: LIS result reports need to have the ability to adapt to the scenario.
The needs of different roles vary significantly:
- Clinicians: need to be efficiently located and quickly correlate abnormal indicators with clinical diagnosis (e.g., directly linking blood potassium critical values to arrhythmia risk);
- Patients: It is necessary to understand the meaning of the results and health action suggestions (such as positive urine protein is recommended to reduce a high-protein diet);
- Laboratory managers: data must be traceable to cover the quality and efficiency indicators of the entire inspection process (such as the distribution of time from sampling to reporting).
Therefore, the design and optimization of the LIS result reporting function need to start from the whole chain of data generation, format presentation to delivery channels, so that a report can be accurately used in multiple scenarios.
2. Result report generation function
2.1 Data integration and verification
The collection scenarios of test data are complex and diverse: the biochemical analyzer pushes liver function data in batches every 15 minutes, the hemocytometer transmits blood routine results in real time, and the pathological biopsy diagnosis requires technicians to manually enter descriptive text…… These data often have format dialects: for the same red blood cell count, some instruments output × 10¹²/L, and some abbreviations are 10^12/L; Manually entered Rh blood type positive, may be written as Rh (+) or Rh positive. If standardized processing is not carried out and a source fidelity mechanism is established, subsequent reports will be reduced to garbled data.
How can product managers do a good job in B-end digitalization?
All walks of life have taken advantage of the ride-hail of digital transformation and achieved the rapid development of the industry. Since B-end products are products that provide services for enterprises, how should enterprises ride the digital ride?
View details >
2.1.1 Design details of the multi-level verification system
1）Basic check-up layer: Adopt the double threshold + dictionary mapping combination strategy.
- Numerical data sets dual thresholds of clinical reference range and instrument theoretical range: for example, blood glucose detection, the clinical reference range is 9-6.1mmol/L (for indicating abnormalities), while the theoretical range of the instrument is 0.5-33.3mmol/L (used to intercept obvious errors, such as 50mmol/L, which is directly judged as instrument failure).
- Character data is standardized through multi-level dictionary mapping: taking blood type as an example, the system establishes a three-level conversion table of original input-intermediate mapping-standard encoding (such as AB blood → AB→003, AB→AB→003), while retaining the special value processing mechanism – for non-standard inputs such as unknown, the system automatically marks to be verified and triggers a technician review reminder.
2) Associate verification layer: Error interception through patient identity binding + logic rules. With the patient’s unique identification (medical record number) as the core, the test data is bound to the basic information (age, gender, clinical diagnosis), and cross-field logical verification rules are set:
- Age correlation: In the blood routine report of newborns (< 28 days), if the age field is filled in with 3 years old, the system pop-up window indicates that the age does not match the test type (the newborn test items need to match the age of < 28 days);
- Gender association: prostate-specific antigen (PSA) test report of female patients, triggering gender-item mismatch reminder (PSA is a male-specific indicator);
- Clinical diagnosis association: If the results of glycated hemoglobin test in diabetic patients are normal but diabetic ketoacidosis is diagnosed, the system suggests that the results are contradictory to the diagnosis, and it is recommended to review the specimen.
Practical cases: A tertiary hospital once lost one decimal place in the total cholesterol data of a batch of blood lipid testing due to instrument interface failure (the actual 5.2mmol/L was shown as 52mmol/L). Because the system presets the calibration rule of the upper limit of the total cholesterol instrument range of 8.0mmol/L, the data is intercepted in real time and triggers the engineer’s maintenance reminder, avoiding clinical misjudgment.
2.2 Intelligent audit and exception identification
The limitations of manual review are significant: one technician reviews an average of 200 reports per day (each containing 20-50 items), and the sensitivity to identifying hidden anomalies decreases after working for a long time. The intelligent audit system can not only share repetitive work through the rule engine + machine learning combination strategy, but also improve the comprehensiveness of anomaly identification.
2.2.1 Rule Engine
The core of the rule engine is to encode clinical consensus and transform medical consensus into executable rules, which needs to be designed to meet dynamic updates and scenario segmentation:
- Base rule base: Covers core scenarios such as critical values, project associations, and reference ranges. For example: potassium ion <2.8mmol/L or >6.2mmol/L→ critical value, immediately trigger a clinical alarm (phone + system pop-up); If urine protein is positive in the urine routine, urine microalbumin must be checked simultaneously (if not tested, mark ‘recommended supplementation’).
- Dynamic update mechanism: The rules are revised quarterly by the hospital inspection expert committee. Taking the reference range for children as an example, when the “Consensus on Laboratory Medicine for Children (2023 Edition)” is released, the system needs to complete the adjustment of the reference range for age groups such as 0-1 years old, 1-3 years old, and 3-6 years old within one week, and simultaneously update the judgment logic of the rule engine.
2.2.2 Machine learning
The machine learning model is used to supplement the blind spots of the rule engine and identify hidden anomalies and is designed to be trained on real clinical data:
- Data preparation: The model training of a tertiary hospital contains 5 years of historical data (1 million reports + corresponding clinical diagnosis), and invalid data such as specimen contamination detection errors need to be eliminated in the pretreatment stage, and finally 850,000 valid samples are retained.
- Model features: Identify patterns where a single metric is normal but the combination is unusual. For example, the white blood cell count is normal (4-10×10⁹/L), but the proportion of neutrophils is increased (>70%) + C-reactive protein is slightly increased (5-10mg/L), which is easy to be ignored in manual review, but the model can correlate the probability of early bacterial infection (based on 78% of similar combinations in historical data).
- Effect evaluation: After the model is launched, the identification rate of hidden anomalies has increased from 62% to 91% of manual review, but the manual review link is still retained (anomalies marked by the model need to be confirmed by technicians before being included in the report).
2.2.3 Scenario-based presentation of abnormal identification
Abnormal identification should be designed differently according to the degree of urgency and user role:
- Critical value: red bold font + exclamation mark icon, and a separate critical value reminder column at the top of the report (such as troponin T: 5ng/mL (reference value< 0.014ng/mL)→ indicates the risk of acute myocardial infarction, and Dr. Zhang has been notified to the emergency department;
- Mild abnormalities: orange italics with a brief description of clinical significance (eg, uric acid: 450 μmol/L (reference value 208-428 μmol/L)→ mildly elevated, possibly related to a high-purine diet);
- Dynamic changes: Items that require long-term monitoring, such as tumor markers, are marked with ↑↑ (significantly higher than the last time) and ↓ (slightly decreased), and a trend chart (such as the numerical change curve of CEA in the past 6 months) is attached.
2.3 Report content generation
Report content should be generated to avoid data piling and instead provide in-depth interpretation based on the type of inspection:
2.3.1 General Items
Taking the blood routine report as an example, the content design follows clinical priorities, strives to be concise and clear, and focuses on the core information:
- Core indicators (leukocytes, erythrocytes, platelets) are in the forefront, with reference range + result interpretation (e.g., leukocytes: 12×10⁹/L (4-10×10⁹/L) → results are abnormally elevated, suggesting possible infection);
- Segmentation indicators (such as neutrophil proportion and lymphocyte ratio) were sorted according to their correlation with core indicators.
- Add a comprehensive tip at the end (if the overall result suggests a high possibility of bacterial infection, it is recommended to combine body temperature and symptom assessment).
2.3.2 Special Projects
Taking the tumor genetic test report as an example, the content needs to cover the whole chain of detection method, result, significance, recommendation, and limitations, so as to extend from data to clinical decision-making suggestions:
- Detection method: ARMS-PCR technology is used, and the lower limit of detection is 5% (i.e., mutation abundance ≥ 5% can be detected);
- Interpretation of results: EGFRexon19 deletion mutation (positive) → this mutation is seen in 40% of patients with non-small cell lung cancer;
- Medication recommendations: It is recommended to give priority to gefitinib (250mg per day), and osimertinib can be considered if resistance occurs;
- Limitations: This test only covers 10 common mutations, and if the treatment effect is not good, NGS panoramic testing (covering 500+ genes) is recommended.
According to a cancer hospital, the optimized genetic test report shortened the interpretation time of doctors from 15 minutes to 5 minutes, and the adoption rate of drug recommendations increased by 32%.
3. Result report format design
3.1 General format specifications
The core principle of format design is to make information transmission more efficient, and it is necessary to balance standardization and practicality to reduce the reading burden:
3.1.1 Pages and fonts
- Margin: Uniformly set to 5cm (slightly wider to 3cm on the left side, reserve binding space). The comparative test shows that 2cm margins will cause overcrowding of content (especially multi-item reports), and 3cm will increase the number of pages (emergency reports need to be presented on a single page);
- Font: The main text is in Song (No. 5), and the title is in bold (No. 12). According to the geriatric doctor’s survey, the recognition efficiency of Song font is 20% higher than that of italic font, and the positioning speed of bold font titles is 1.5 seconds faster than that of Song font titles.
3.1.2 Information partitioning
The report page is divided into 4 core areas:
- Patient information area(upper left corner): left-aligned table, containing name, gender, age, medical record number, sampling time, and reporting time 6 items, of which the sampling-reporting time difference (TAT) is marked separately (e.g., TAT: 2 hours and 15 minutes) – emergency items require TAT to be < 1 hour, and the overtime will be marked in orange;
- Inspection project area(Core area): The order of the table columns is the name of the project→ the result → unit→ reference range→ abnormal prompt, which is in line with the doctor’s reading habit of looking at the item first and then whether the results are normal. Multi-set items (such as all biochemistry items) are grouped according to the clinical logic of liver function→ kidney function→ electrolytes → blood lipids, and each group is distinguished by a light gray background (gray value 240 to avoid visual fatigue);
- Notes area(below the project area): indicate the type of specimen (e.g., serum/urine) and the instrument model (e.g., biochemical items tested on Beckmann AU5800);
- Area of responsibility(footer): Contains the electronic signature of the inspector and reviewer (with time stamp), disclaimer (this report is only responsible for this specimen, please consult the clinician for the interpretation of the results) and laboratory contact number (to facilitate the communication of clinical questions).
3.2 Personalize the template
3.2.1 Doctor templates
Taking the myocardial enzyme profile report as an example, a clinical hint column is added to the routine results to strengthen the clinical association:
- When the creatine kinase isoenzyme (CK-MB) > 25U/L and troponin > 0.04ng/mL, automatic insertion: it is consistent with the changes in biochemical indicators of acute myocardial infarction, and 85% of the same cases in the past 3 months were finally diagnosed, and it is recommended to combine electrocardiogram (ST-segment elevation) and chest pain symptom assessment;
- Attached is a reference to recent similar cases: showing the final diagnosis of patients with the same index combination in the past 3 months (such as acute myocardial infarction 85%, myocarditis 10%, other 5%), providing doctors with a decision-making anchor.
3.2.2 Patient template
The design of patient templates should be popularized + graphical to lower the threshold of understanding:
- Project name: Total cholesterol labeling (a type of blood lipid, too high may increase the vascular burden);
- Interpretation of results: It is recommended to reduce the intake of animal offal and fried foods with ↑ corresponding to the high result, and the positive correspondence is detected to detect the substance, and it is recommended to review it after 3 months;
- Graphical presentation: Blood glucose results are displayed in a bar chart of green normal interval (9-6.1 mmol/L), yellow warning interval (6.1-7.0 mmol/L), and red abnormal interval (>7.0 mmol/L), and patients can intuitively locate their result position.
A community hospital pilot showed that patient templates increased the comprehension rate of reports from 32% to 78%, and the proportion of active consultations with doctors increased by 45%.
3.2.3 Administrator template
The laboratory manager template needs to cover the two dimensions of testing quality + process efficiency:
- Quality indicators: intra-batch error of the batch (e.g., CV=2.1% in the batch of liver function items, meeting ISO15189 standards <3%), daytime precision (detection deviation of the same quality control product in the past 7 days);
- Efficiency indicators: the distribution of time from sampling to receiving (average 30 minutes) and detection to review (average 45 minutes), compared with the average value of the previous month (for example, the time spent in the detection process increases by 5 minutes compared with the previous month, and the instrument load needs to be checked);
- Abnormal statistics: The proportion of abnormal results of this month (such as 12% abnormal rate of blood routine, up 3% from the previous month, and the specimen collection specifications need to be checked).
4. Distribution channels for result reports
The delivery of reports should take into account both convenience and security, and cover different user needs through offline + online + system integration through multiple channels.
4.1 Self-service printing
4.1.1 Equipment layout and operation process
- Layout: There are 3-5 units in the outpatient hall (the service radius of each unit is < 5 meters during peak hours), and 1 unit on each floor of the inpatient department (close to the nurse station, which is convenient for assisting elderly patients);
- Operation: Adopt a three-step process – scan the barcode (or ID card) of the receipt form→ select the report (support multiple selections), confirm → printing, voice prompt throughout the process (if please align the barcode with the scanning area), and enlarge the screen font to 14 (suitable for users with poor eyesight).
4.1.2 Privacy and consumables management
1) Privacy protection:
- The report will be taken out of the paper, and if it is not taken within 30 seconds, it will be automatically retracted and destroyed (to avoid others taking it by mistake);
- After the operation is completed, the screen will be automatically cleared, and no input information will be retained;
- Upgrade secondary verification: In view of the risk of the barcode being scanned by others, the last 4 digits of the patient’s mobile phone number need to be entered when printing (matching the registration information).
2) Consumables management: built-in low margin warning (when paper/toner < 20%), automatically send SMS to the consumables administrator (including equipment number + missing material type) to ensure that more than 98% of the equipment does not lack consumables during working hours.
4.2 Online push
4.2.1 WeChat push
After the patient binds the information to the hospital’s official account, the system sets the push strategy according to the test type:
- Push timing: Push immediately after the review of the emergency report is completed (with emergency sign), and the general outpatient report will be pushed before 17 o’clock on weekdays (to avoid disturbance at night);
- Push content: The copywriting takes into account both conciseness and focus, such as your blood routine report has been issued: white blood cell results are abnormal (↑), it is recommended to consult a doctor within 24 hours→ click to view details;
- Function support: The report page supports long-pressing to save pictures (presented during offline medical treatment), adding to health records (associated with WeChat health cards), and one-click consultation of abnormal indicators (jumping to the online consultation entrance of the corresponding department).
4.2.2 SMS push
For WeChat blind spot users such as elderly patients, the content of the text message needs to be short and fast:
- The number of words is controlled within 50 words: [XX Hospital] Your biochemical report has been released, and the results are normal. Check details: http://xxx log in with ID number + verification code;
- Security design: The link is valid for 24 hours, and the web page requires ID number + SMS verification code to avoid information leakage.
4.2.3 APP push
Hospital APP push is not only a result notification, but also an entrance to the connection between diagnosis and treatment, realizing a closed loop of in-depth services:
- Intelligent interpretation: generate trend charts for multiple testing items (such as blood potassium testing of hypertensive patients, mark that the result is 3mmol/L lower than the last time, and it is recommended to monitor the post-medication reaction);
- Service linkage: Provide online consultation entrance (directly hang the specialist number corresponding to the test item), review appointment (such as glycated hemoglobin needs to be reviewed after 3 months, click to make an appointment for the next time), to realize the closed loop of reporting→ consultation→ diagnosis and treatment.
4.3 Electronic medical record integration
The integration of LIS with electronic medical records (EMRs) was a pain point: the early use of document import mode required manual report uploads, which was time-consuming and error-prone. Real-time two-way synchronization is now achieved through HL7FHIR standard interface, breaking down system barriers:
- Data synchronization: The test results module of the EMR is automatically written to the LIS report within 30 seconds after the review is completed, and the doctor can directly view it in the EMR (no need to switch systems);
- Information linkage: Modify the patient’s basic information (such as correcting age) in the EMR, and LIS will synchronously update the association report (to avoid interpretation errors caused by inconsistent information);
- Scientific research empowermentA teaching hospital used integrated data to carry out a study on the association between glycated hemoglobin and complications in diabetic patients, and completed the screening and analysis of 10-year cases within 1 day by extracting the association data of diabetes diagnosis + glycated hemoglobin detection in EMR (traditional manual collection takes 2 weeks).
5. Direction of technological evolution
LIS results reporting is designed with a user-centric approach to technology—from data verification to formatting, every detail should revolve around the goal of getting the right information to the right person at the right time.
When emergency doctors can lock in troponin critical values within 30 seconds, when patients can understand the intervention direction of dyslipidemia through graphical reports, and when laboratory managers can optimize the testing process through data traceability – these subtle improvements will eventually converge into a joint force to improve medical quality.
The future optimization direction is gradually becoming clear:
- Natural language processing technology: Automatically generate report summaries (such as core conclusions: high probability of acute myocardial infarction; Key indicators: troponin 5ng/mL; Recommendation: immediate cardiology consultation);
- Blockchain technology: realize mutual recognition of cross-hospital reports (solve the problem of duplicate inspection), and ensure that data cannot be tampered with through distributed bookkeeping;
- AR visualization technology: allow patients to intuitively understand abnormal indicators (such as using 3D models to show the crystallization damage of high uric acid to joints).
Technology is changing, but making test data better serve the original intention of diagnosis and treatment has always been the core driving force for the evolution of the LIS system.
