2.1. Data Sources

In this study, all data originated from the Hospital Prescription Analysis Cooperation Project database, administered by the Hospital Pharmacy Professional Committee of the Chinese Pharmaceutical Association. This database encompassed information gathered from nine cities, including Beijing, Shanghai, Guangzhou, Chengdu, Hangzhou, Tianjin, Zhengzhou, Shenyang, and Harbin. From each quarter, a random selection of outpatient prescriptions and inpatient medical orders spanning 10 working days was extracted from the respective hospital information systems and subsequently incorporated into the project’s prescription database. The information of inpatient medical orders from 86 hospitals (Table S1) in nine cities was extracted from this database for a total of 240 working days from 2016 to 2021. The extracted information included region, time, patient code, prescription number, name of the department, reimbursement, source of prescription, drug code, generic name, trade name, drug strength, route of administration, number of prescriptions, number of drugs taken, prescription cost, drug price, drug usage (frequency of medication), single dosage, patient age, and primary diagnosis.

2.2. Inclusion and Exclusion Criteria

The extracted prescription data were cleaned. The inclusion criteria were (1) women aged 18 and above; (2) women at 20 to 36⁺⁶ weeks of gestation; (3) women with “premature delivery” or “abortion” or “premature rupture of membranes” or “pregnancy” as the primary diagnosis; and (4) prescriptions with “magnesium sulfate” or “nifedipine” or “indomethacin” or “ritodrine” or “atosiban” as the primary generic drug name. The exclusion criteria were (1) women who terminated the pregnancy (including stillbirth, fetal arrest, spontaneous abortion, medical abortion, and ectopic pregnancy); (2) prescriptions with “magnesium sulfate” as the primary generic drug name and women with “eclampsia” or “hypertension” or “high blood pressure” as the primary diagnosis; (3) prescriptions with “nifedipine” as the primary generic drug name and women with “eclampsia” or “hypertension” or “high blood pressure” as the primary diagnosis; and (4) prescriptions with “indomethacin” as the primary generic drug name and women with “inflammation” or “fever” as the primary diagnosis.

2.3. Analysis of Rationality

Prescriptions with complete information on frequency of usage, single dosage, or total dosage were evaluated according to the drug instructions and off-label uses. Magnesium sulfate is administered as an intravenous drip, and the total dose in 24 hours cannot exceed 30 g. Nifedipine (immediate-release) is given orally at 10–20 mg every 6 hours [10, 13]. Initially, indomethacin is administered as a loading dose, rectally or orally between 50 and 100 mg. Subsequently, a maintenance dose of 25 mg is given every 6 hours [10, 11]. However, its usage is strictly limited to no more than 3 days and should be prescribed before the 32nd week of gestation. In contrast, ritodrine is administered initially as an intravenous drip, followed by oral maintenance. The initial oral dosage of ritodrine is 10 mg every 2 hours for the first 24 hours, and then 10 to 20 mg every 4 to 6 hours, with a total daily dosage not exceeding 120 mg. Atosiban is administered as an intravenous drip, and the total dose in 24 hours cannot exceed 330 mg.

Inappropriate prescriptions include inappropriate selection of drug dosage forms, overdosing, underdosing, low-frequency, and high-frequency. Inappropriate selection of drug dosage forms refers to the situation where nifedipine intended for the treatment of preterm labor should be in its immediate-release form [13]. Using the sustained-release or controlled-release form instead would constitute inappropriate selection of drug dosage forms. Overdosing refers to a dose higher than the recommended dosage, while underdosing refers to a dose lower than the recommended dosage. High-frequency medication refers to a situation where the daily frequency of dosing exceeds the recommended number, whereas low-frequency medication occurs when the daily frequency of dosing is less than the recommended.

All prescription combinations and clinical diagnoses were checked to determine whether the prescribing of medications was appropriate. The combined use of two or more tocolytic agents is considered inappropriate. Prescriptions with gestational weeks exceeding 34 weeks in the “original diagnosis” category are considered to have inappropriate clinical diagnoses.

2.4. Statistical Methods

Data classification and data analysis were performed using Excel 2021 (Microsoft, Redmond, WA, USA) and SPSS 25 (IBM, Armonk, NY, USA) software packages. The defined daily dose (DDD) refers to the average daily dose established for a specific drug to treat its primary indication in adults, and these DDD values are derived from data formulated by the World Health Organization (WHO) [14]. The defined daily dose (DDD) was defined as the consumption of a drug in DDD and calculated as the total dosage of a drug/DDD [15]. The larger the DDDs are, the greater the tendency they have to reflect the clinical selection of the drug. The defined daily cost (DDC) was defined as the average daily expense of a specific drug used by a patient and calculated as the annual sales amount of a drug/DDDs. The larger the DDC, the heavier the economic burden caused to the patient.

2.5. Ethics Statement

This study protocol was approved by the Ethics Committee of the Obstetrics and Gynecology Hospital affiliated with Fudan University (approval number: 2023-01). This was a retrospective study with an informed consent exemption. Patient consent was not needed.

3.1. Patient Characteristics

The data of 54,497 tocolytic agents were extracted from the database, and the data of 39,842 tocolytic agents remained after excluding the ineligible data. Data with the same prescription number (6,604) were combined into one prescription, and finally 33,129 prescriptions were included in this study. In brief, 95.75% of patients were aged 18 to 39 years (mostly 30 to 39 years old, followed by 18 to 29 years) (Table 1). More than half of prescriptions were distributed in non-first-tier cities, but the mean number of prescriptions in hospitals in first-tier cities (446 prescriptions/hospital) was higher than that in non-first-tier cities (351 prescriptions/hospital) (Tables 1 and S2).

3.2. Number of Prescriptions for Tocolytics

The total number of prescriptions for tocolytics was the highest in 2019, with an increase of 20.44% from 2016 to 2019, a slight decrease after 2019, and a decrease of 11.89% from 2019 to 2021 (Table S3). From 2016 to 2021, the number of tocolytic prescriptions was ranked as magnesium sulfate > ritodrine > nifedipine >indomethacin and atosiban (Figure 1). From 2018 to 2019, the number of prescriptions for atosiban underwent a significant surge, exhibiting a substantial growth of 185.71% in 2018 and a remarkable jump of 871.43% in 2019, when compared to 2017 (Figure 1, Table S3).

3.3. Prescription Cost of Tocolytics

The prescription costs of nifedipine and indomethacin were lowest from 2016 to 2021. The prescription cost of tocolytic agents was ranked as ritodrine > atosiban > magnesium sulfate in 2016. The number of magnesium sulfate prescriptions increased by 4.08% in 2017, whereas the cost of magnesium sulfate increased by 691.11%. From 2018 to 2019, there was a substantial leap in the prescription cost of atosiban, experiencing a rise of 104.21% in 2018 and a steep jump of 558.23% in 2019 compared to 2017 (Table S4). The prescription cost of tocolytic agents was ranked as atosiban > magnesium sulfate > ritodrine from 2018 to 2021 (Figure 2).

3.4. Daily Defined Doses for Tocolytics

From 2016 to 2021, the ranking of DDDs was magnesium sulfate > ritodrine >nifedipine > indomethacin > atosiban (Figure 3).

3.5. Daily Defined Cost of Tocolytics

From 2016 to 2021, the ranking of DDC was atosiban > ritodrine > magnesium sulfate > nifedipine = indomethacin (Figure 4).

3.6. Relationship between Daily Defined Doses and Daily Defined Cost of Atosiban

From 2016 to 2021, the DDC of atosiban gradually decreased, while the DDDs continued to rise. The natural logarithm (ln) value of DDC for atosiban exhibits a negative correlation with the ln value of DDDs (R² = 0.9149, P < 0.01) (Figures 5 and S1).

3.7. Appropriate Evaluation of the Use of Tocolytic Agents

4.1. Demographic Characteristics of Patients

Studies have shown that the relationship between the incidence of PTB and maternal age follows a “U-shaped” pattern, with a higher occurrence rate of PTB in younger and older pregnant women [20–22], the lowest occurrence rate of PTB in pregnant women aged 30–34 years (5.7%), and the highest occurrence rate of PTB in pregnant women aged 40 years or older (7.8%) [21]. In China, there is an increasing trend in the proportion of older maternal age (≥35 years), rising from 10.6% in 2016 to 16.7% in 2017 and 15.9% in 2018 [3]. In the United States, there has been a steady increase in pregnant women aged 45 years or older, rising from 0.19% in 2009 to 0.26% in 2020 [23]. The likelihood of PTB among pregnant women aged 45 years or older is significantly higher compared to those under 45 years (16.5% vs 8.3%) [23]. In summary, the annual increase in the proportion of older pregnant women is accompanied by a growing potential for preterm delivery. Therefore, it is inevitable that the proportion of older pregnant women experiencing PTB will also increase. This is also confirmed by the findings of this study, which show an increasing trend in the use of tocolytic therapy among older pregnant women (for example, those aged 50 or above) (Table 1). In this study, the average age of PTB patients was 30.85 years, with 77.66% of patients aged between 20 and 34 years old (Table S2). This may be due to the larger number of pregnant women between 20 and 34 years old, so even if the incidence of PTB is lower in this age group, there will still be a large number of patients who experience PTB. Systematic reviews have reported that PTB is related to environmental factors, such as fine particulate matter, ozone, and high temperature [24–26]. Exposure to ozone and high fine particulate matter levels may have significant impacts on the cardiovascular system through oxidative stress and inflammatory responses, leading to endothelial dysfunction and arterial vasoconstriction, which in turn increases placental vascular resistance, leading to an increased risk of PTB [27]. The relationship between temperature and PTB is U-shaped, with both high and extremely low temperatures increasing the risk of PTB [28]. The impact of low temperatures is primarily explained by changes in hemodynamics, while the effects of high temperatures are mainly attributed to alterations in immunity and inflammation. Low temperatures are associated with peripheral vasoconstriction during pregnancy, which may alter uteroplacental perfusion and have adverse effects on the developing fetus [28]. Exposure to excessive heat can trigger the secretion of cytokines, including prostaglandins, oxytocin, and heat shock proteins. This can potentially upset the intricate cytokine equilibrium at the interface between mother and fetus, ultimately precipitating the activation of the delivery mechanism prematurely [28]. In this study, the average number of prescriptions for PTB patients in first-tier cities was higher than that in second-tier cities, which may be associated with the different environmental characteristics of the cities [29]. Another study has demonstrated that the rate of intrauterine transfer of preterm infants differs greatly across different geographical regions, and the intrauterine transfer of preterm infants to well-equipped medical institutions can significantly improve survival outcomes [30]. Therefore, it cannot be excluded that PTB patients in second-tier cities were transferred to medical institutions in first-tier cities.

4.2. Pharmacoeconomic Evaluation

DDC can guide the market price of drugs, which is an important determinant of drug consumption [31]. In order to improve the affordability of medicines, the Chinese government has implemented various measures, including establishing a basic medical insurance system, conducting price negotiations, and executing centralized drug procurement policies, to reduce drug prices. Previous studies have shown a negative correlation between DDC and DDDs during the process of price reduction for drugs [15, 32, 33]. Although atosiban in this study is the safest tocolytic agent approved so far, its sales are very low due to its far higher price than other tocolytic agents. The DDC of atosiban decreased significantly in 2018, which may be related to the donation of atosiban by social welfare projects and the market entry of domestic generic drugs. From January 2018 to December 2019, the Beijing Renze Foundation launched the “Yibao Patient Assistance Program,” which provided atosiban injections free-of-charge to financially challenged women who were experiencing a threatened PTB. In March 2018, generic atosiban was approved for marketing in China. Compared with patent-protected drugs, the prices of domestic generic drugs are significantly lower. In 2019, the atosiban injectable was listed in the National Medical Insurance Drug List (implemented on January 1, 2020), and the DDC of atosiban further decreased in 2020. During the period from 2018 to 2020, social welfare funds, generic drugs, and the national medical insurance policy contributed to a notable decline in the cost of atosiban. Consequently, the percentage of prescriptions rose from 0.21% in 2016 to 4.76% in 2021. As expected, the DDC and DDDs of atosiban in this study were negatively correlated.

4.3. Drug Characteristics