2.1 Study Design
This prospective, multi-center, large sample observational cohort study was conducted in three comprehensive traditional Chinese medicine hospitals in Henan Province, China, and one tertiary-level and 10 primary-level medical and health institutions in Shanghai. The selection of monitoring hospitals was as follows: first, we selected hospitals with relatively high sales of NXT in various regions of China, and then finally determined the monitoring hospitals based on the scientific research capabilities of each hospital and the willingness to participate in project cooperation. Based on the Hospital Information System (HIS) data, the active monitoring research model of adverse reactions after the listing of drugs in the real world was adopted.
This study was funded by the Ministry of Science and Technology of the People's Republic of China (Ministry of Science and Technology of the People's Republic of China, No. 2015ZX0950104-001-007) and approved by the China Clinical Trials Registration Center through international registration (No. ChiCTR-OPC-17013912).
2.2 Study Population
All inpatients and outpatients who were orally administrated with NXT in 14 surveillance hospitals, including 3 comprehensive traditional Chinese medicine hospitals in Henan Province, and 11 tertiary and primary medical and health institutions in Shanghai, between January 2018 and December 2018 were included in the study (see Fig. 1 and Table 1 for details).
Following the sample size requirements defined by the "Guidelines for Key Drug Monitoring of Manufacturers", the number of cases included in statistical analysis should generally be ≥ 3,000 . A total of 7345 cases were included in this study, and 6399 cases were finally used for statistical analysis.
2.3 Inclusion and Exclusion Criteria
Inclusion criteria were the following: all cases that were followed up for 3 times with a maximum monitoring time of 90 days. Cases with adverse reactions/events that occurred after NXT use were considered valid cases.
Dropout criteria were the following: patients who visited once but did not return for follow-up monitoring.
Exclusion criteria were the following: cases in the patient enrollment system, for which the information was not consistent with the data extracted by the HIS.
2.4 Database Information
The database covers two aspects of hospital centralized monitoring data and hospital HIS data. The centralized monitoring data of the hospital include the report data of the Monitoring Information Form A and B. Form A mainly involves patient demographic information, personal allergy history, allergic disease history, family allergy history, indication information, NXT usage information, and administration status, etc. (see Appendix 1). Table B mainly relates to NXT ADRs/ADEs category, past and family ADRs/ADEs status, ADRs/ADEs process description and treatment status, NXT usage conditions, ADRs/ADEs results, and conversion Return (see Appendix 2). The hospital HIS data accurately store the clinical actual electronic medical record information of the patients in the clinic, and include medical record home page information in the medical record management system (patient clinic/hospital number, gender, age, ethnicity, height, weight, working conditions, living area, personal history, family history, allergy history, diagnosis, treatment category, treatment course, treatment results, hospitalization days and frequency, etc.), information related to hospitalization course in the electronic medical record system (admission record, discharge summary, information related to vital signs in the medical workstation (body temperature, pulse, fluid intake and output, blood pressure, etc.), hospitalization course record), inpatient medical order information in the drug management system (drug name/specification/manufacturer, usage and dosage, frequency of administration, route of administration, start and end dates of medication, information on the execution of medical orders and information on patient medication, etc.), inspection related information in inspection system (names and results of inspection items, normal reference values of biochemical indicators, inspection report results, etc.).
2.5 Data source, Collection, and Quality Control
The data collection method is shown in Fig. 2. In this study, the active follow-up monitoring by third-party pharmacists was conducted three times to obtain information on ADRs/ADEs, and the monitoring form was filled out. All patients who used NXT were actively monitored and/or were followed up by the pharmacist via telephone for 3 times. The follow up data were reported to the Patient Enrollment Registration System on the Web network, including patient medical record number, name, gender, age, visit time, outpatient or inpatient service, usage and dosage of NXT, as well as the number of prescriptions, visit times, and the occurrence of ADRs/ADEs, and other information. Monitoring Information Table A was also filled out (see Appendix 1). Pharmacists registered patients by using the Patient Enrollment Registration System at each visit and/or telephone follow-up, eliminated duplications, and kept track of patient identification. At the first visit, the pharmacist filled in the Monitoring Information Form A, which they actively tracked and monitored. They also cooperated with the follow-up three times in the monitoring period, and completed the telephone follow-up on the 15th − 18th day, 45th − 48th day and 90th − 93rd day after patients took the medicine, respectively. Among those who were discharged from the hospital with NCT, the two-way contact information was kept between pharmacists and patients, and the patient was kept informed of the course of treatment and the time of next visit. If there were any medication related problems, pharmacists were contacted or the patient was sent directly to outpatient clinic, and was then followed up by pharmacists.
The monitoring period was terminated for the patients without ADRs/ADEs. If ADRs/ADEs occurred during the monitoring period, due to which monitoring needed to be stopped, the monitoring of the patient was finished by tracking the patient to recovery or improvement. At the same time, the Monitoring Information Form B (see Appendix 2) and ADRs/ADEs Report Form were filed and reported to the National Adverse Drug Reaction Monitoring System (http://www.adrs.org.cn/). If ADRs/ADEs occurred during the monitoring period, and mild symptoms did not affect the follow-up of continuous use of the medicine until the end of 3 months, the monitoring could be finished. At the same time, Monitoring Information Form B and ADRs/ADEs Report Form were filled and reported.
Within one month after monitoring, the HIS data were extracted, and valid fields were set for information collection and mining, and the centralized monitoring ADRs/ADEs database of NXT clinical safety hospital was established with the integration of monitoring table data and other information. A scientific and standardized quality control system was established to ensure the controllability of the whole research and the quality of monitoring data, which mainly included data quality control during and after monitoring. At the same time, the dynamic quality control method was used in the monitoring research process. Three-level quality control was established in this study, including first-level quality control in each monitoring hospital, second-level quality control in sub-centers, and third-level quality control in the project monitoring unit, thus forming the quality control system that could be used to ensure the quality of centralized monitoring research in hospitals. A database was established by using EpiData data management software. Two persons recorded the monitoring information table separately and used the computer to carry out synchronous entry consistency check and logical error detection. Errors and inconsistencies were immediately corrected according to the monitoring table. After the data were verified and corrected by computer, 2.5% of the sample size monitoring table was randomly selected, and all the fields in the monitoring table were manually checked with the database. The main field error item is required to be 0, and the secondary field error is within 0.3%. Otherwise, it is necessary to re-enter all data. Outpatient and inpatient medical record data of the electronic information system of each research hospital were directly imported, and the data were reviewed in detail. The questionable parts were confirmed by the data export summary and were then reviewed again. After the data were checked and standardized again, the database was locked under the supervision of statisticians and researchers.
2.6 Statistical Analysis
Descriptive statistics on the types and outcomes of adverse reactions related to NXT, the demographic characteristics of the monitored objects, count data such as baseline variables, and categorical variables are described by frequency (%). Logistic regression analysis was used to explore the risk factors of ADRs, including variables such as BMI, personal drug/food allergy history, disease allergy history, family allergy history, whether this was the first medication that was used, and the first medication dose. SPSS 23.0 software package was used for data analysis. A p value < 0.05 was considered to be statistically significant.