The World Health Organization announced that the coronavirus disease 2019 (COVID-19) outbreak had become an international public health emergency on 30 January 2020; on 11 March 2020, it declared that the outbreak had become a pandemic.1 Governments and public health authorities worldwide implemented public health policies to reduce the risk of viral transmission, including strict physical distancing, severe travel restrictions, and the closure of many businesses and schools. On 25 January 2020, China’s Central Government announced a nationwide travel ban and quarantine policy2; it initiated nationwide school closures as an emergency measure to prevent the spread of COVID-19.3 Thus, >220 million school-aged children and adolescents were confined to their homes; online classes were offered and delivered via the internet.4

Vision problems are public health challenges; among school-aged children, these problems often involve asthenopia and vision impairment. Asthenopia is defined as a subjective sensation of visual fatigue, eye weakness, or eyestrain; it can manifest through various symptoms, including epiphora, ocular pruritis, diplopia, eye pain, and dry eye.5 Vision impairment is defined as visual acuity (VA) of 6/12 or worse in either eye6; it is often caused by uncorrected refractive errors, and its estimated prevalence is 43%.7 Although both asthenopia and vision impairment have negative effects on students, the effects of vision impairment are greater. A previous global analysis revealed that vision impairment was present in 12.8 million children aged 5 to 15 years, half of whom lived in China.8 Moreover, students with vision impairment have lower scores on various motor and cognitive tests.9 10

Excessive use of digital devices contributes to increases in asthenopia prevalence and vision impairment among school-aged children.4 11 12 13 14 15 The COVID-19 pandemic has led to increased use of digital device–supported online classes,16 17 18 which require extended exposure to those devices.19 20 Importantly, long durations of exposure to digital devices can contribute to many vision problems in children.14

Asthenopia and vision impairment related to the excessive use of digital devices during the COVID-19 pandemic have been investigated in developed countries and urban China.4 11 12 To our knowledge, no similar studies have been conducted in western rural China. Additionally, online classes are increasingly implemented in rural areas, and the use of digital devices is becoming more prevalent11; thus, there is a need for research that focus on vision health in students.

The primary purpose of this study was to assess screen time, asthenopia prevalence, and vision impairment progression during the COVID-19 pandemic among students in western rural China. To achieve this goal, we first conducted a general descriptive analysis of student characteristics and screen time trends before and during the pandemic. We then investigated the prevalence of asthenopia and progression of vision impairment. Finally, we explored factors influencing the prevalence of asthenopia and progression of vision impairment before and during the pandemic.

This study focused on areas that were broadly representative of rural western China because of limited resources. Thus, the study was conducted in Shaanxi and Ningxia regions in western China. In 2019, the per capita gross domestic product in Shaanxi Province was US$10 167; this is similar to that in Ningxia Autonomous Region (US$8236).21

Vision data were acquired from local vision care centres (VCs), which had been established by the Center for Experimental Economics in Education at Shaanxi Normal University, in cooperation with county-level organisations such as the local education ministries and hospitals.

Before the pandemic, VC screenings were performed in each county, except during summer and winter vacations. Staff conducted one to two screenings per week (covering 2 to 4 schools); they completed one round of screening in one town each month. In practice, approximately 1 year is needed to complete one round of vision screening for all eligible children in a particular county. The second round and subsequent rounds of vision screening were performed using a similar workflow. After the completion of vision screening, students who required further assessment were referred to the VC for full eye and refractive examinations. This study included students who had visited the VC 3 months before the beginning of the COVID-19 pandemic.

During the pandemic, VC staff could not attend schools to perform vision screenings. To maintain vision screening services for students, we telephoned all students who had visited the VC before the pandemic. Participants in this panel study were students who participated in data collection before and during the COVID-19 pandemic.

We conducted two cycles of surveys in the VC. The first survey cycle was conducted from October to December 2019 (before the pandemic); the second survey cycle was conducted among a group of students who visited the VC for follow-up from July to December 2020 (during the pandemic), based on their enrolment in the study before the pandemic. The same information was collected during the two survey cycles. During the vision screening process, VC staff administered questionnaires to students for collection of the following information: sex (male=1), age, ethnicity (Han=1), residence (non-rural=1), only-child status (yes=1), parental education (parents with ≥12 years of education=1), and parental migration status (one or both out-migrated=1; defined as one or both parents worked away from home during the semester). Household assets were calculated by summing the values of 13 items owned by the family, in accordance with the China Rural Household Survey Yearbook.22

The survey also included the collection of information regarding screen time and asthenopia. Students completed a previously described, self-administered questionnaire concerning mean time spent throughout the day on near activities (including computer and smartphone use, television viewing, and studying/homework after school). Reports of time spent on near activities during different parts of the day were categorised as screen time while learning and screen time while playing. Information regarding asthenopia was collected via three questions focused on ocular discomfort: whether the student had experienced dry eyes (yes=1), eye pain and swelling (yes=1), and eye fatigue and watery eyes (yes=1). Asthenopia was defined as the presence of at least one of these three types of vision health problems (yes=1).23 Furthermore, information regarding VA was collected when students visited the VC. The optometrist in the VC conducted a VA test to measure the clarity of each student’s vision. All students completed VA tests without refractive correction; students with spectacles completed VA tests with their routine method of vision correction.

The questionnaire regarding asthenopia was developed and reviewed by a group of health experts from Shaanxi Normal University and Zhongshan Ophthalmic Center, a well-known ophthalmology institution in China. The included questions were constructed to ensure that they could be clearly understood by students aged 9 to 17 years with the aid of trained VC staff. These three questions can serve as good indicators of symptoms representing different degrees of asthenopia in students, and they have been used in previous research.23

Visual acuity was assessed using Early Treatment Diabetic Retinopathy Study tumbling-E charts (Precision Vision, La Salle [IL], United States). In an indoor area with sufficient light, VA was separately assessed for each eye without refraction at a distance of 4 m. Students were first examined using a 6/60 line; if they correctly identified the orientation of at least four of five optotypes, they were examined using a 6/30 line, followed by a 6/15 line and a 6/3 line. In this manner, the VA for an eye was defined as the lowest line on which four of five optotypes were correctly identified. If the participant could not read the top line at a distance of 4 m, they were tested at a distance of 1 m, and the VA result was divided by 4.

In this study, VA levels were calculated and compared using the logarithm of the minimum angle of resolution (logMAR) scale, which is a linear scale with regular increments that offers a reasonably intuitive interpretation of VA measurement.24 In this study, vision impairment was defined as logMAR ≥0.3 (ie, VA of 6/12 or worse) in either eye.

This balanced panel study compared student data between two periods (before and during the COVID-19 pandemic). Mean screen time, asthenopia prevalence, and vision impairment progression were compared among students using t tests, after stratification according to various demographic and behavioural factors. Fixed-effects logistic and regression models were used to explore factors influencing the prevalence of asthenopia and progression of vision impairment before and during the pandemic. Fixed-effects models were adjusted for sex, age, ethnicity, rural or non-rural residence, only-child status, parental migration status, parental education level, household assets, screen time while learning, and screen time while playing. All analyses were performed using Stata Statistical Software, version 14.1 (StataCorp, College Station [TX], United States). All tests were two-sided, and P values <0.05 were considered statistically significant.

This study included 128 students from western rural China (mean age before pandemic, 11.82 ± 1.46 years; mean age during pandemic, 12.32 ± 1.54 years; 80 girls [62.5%] and 48 boys [37.5%]). All participants had vision impairment and were attending online classes (Table 1).

During the pandemic, screen time significantly increased because of enrolment in online classes. The mean total screen time during the pandemic was 3.22 hours per day, compared with 1.97 hours during the pre-pandemic period (P<0.001). The mean screen time while learning during the pandemic was 1.70 hours per day, compared with 0.90 hours during the pre-pandemic period (P<0.001); the mean screen time while playing during the pandemic was 1.52 hours per day, compared with 1.33 hours during the pre-pandemic period (P=0.019). Additionally, rural students had significantly greater screen time while learning during the pandemic, compared with the pre-pandemic period (P<0.001); there was no such difference among non-rural students (Table 1).

The prevalence of asthenopia and progression of vision impairment significantly differed between the pandemic and pre-pandemic periods. The prevalence of asthenopia during the pandemic was 55% (71/128), whereas it was 27% (35/128) during the pre-pandemic period (P<0.001). The mean logMAR VA was worse during the pandemic compared with the pre-pandemic period (0.81 vs 0.65; P<0.001). The prevalence of asthenopia was higher during the pandemic than during the pre-pandemic period, regardless of the characteristics used to stratify participants. The mean logMAR VA was worse during the pandemic than during the pre-pandemic period, although the difference being insignificant among participants with non-Han ethnicity and participants in the top quartile of household assets (Table 2).

Fixed-effects logistic models for asthenopia revealed that screen time while learning was associated with asthenopia prevalence, and the probability of asthenopia increased by 24.6% for each 1-hour increase in screen time while learning (95% confidence interval [CI]=1.02-1.53; P=0.034). Additionally, older age (odds ratio [OR]=2.073, 95% CI=1.13-3.81, P=0.019), Han ethnicity (OR=2.405, 95% CI=1.22-4.74; P=0.011), and only-child status (OR=0.488, 95% CI=0.21-1.13; P=0.095) were factors associated with asthenopia; screen time while playing was not (Table 3).

Fixed-effects regression models showed that residence in a non-rural area (OR=-0.200, 95% CI=-0.355 to -0.046; P=0.011) and only-child status (OR=-0.099, 95% CI=-0.197 to 0.000; P=0.049) were factors associated with logMAR VA. The probability of worse logMAR VA increased by 0.200 in non-rural areas, compared with rural areas. However, screen time while learning and screen time while playing were not associated with vision impairment (Table 4).

The global spread of the COVID-19 pandemic has affected the education of >1.5 billion children and adolescents worldwide.25 The participants in our study were representative of this important population. They demonstrated declines in VA and vision health during the pandemic, in relation to the excessive use of digital devices; these findings were consistent with the results of previous studies.19 26

All students in our study were attending online classes during the pandemic. We observed an increase in the mean daily time spent on digital devices between the pre-pandemic and pandemic periods; these results are consistent with international findings that screen time was greater during the pandemic than before the pandemic.19 Notably, we found that total screen time and screen time while learning significantly changed among rural students but not among non-rural students; these results are also consistent with previous findings.19 This difference presumably occurred because, compared with rural students, non-rural students were more likely to use digital devices and online classes before the pandemic.

We observed a significant difference in asthenopia prevalence among students in low-income areas of western China before and during the pandemic; this finding supports the results of previous studies.26 27 Although the risk of asthenopia reportedly increases with screen time,28 there is no published literature concerning changes in asthenopia among students in relation to the COVID-19 pandemic. Similar to previous studies,14 we found that the prevalence of asthenopia was approximately twofold greater among students aged 13 to 17 years than among those aged 9 to 12 years. Furthermore, Moon et al26 reported that symptoms of dry eye diseases were more common among older children than among younger children. Older children spend more time using digital devices, leading to a higher prevalence of asthenopia.29

This study showed significant progression of vision impairment in relation to the pandemic; similarly, a study in eastern China revealed that students had worse vision during the pandemic, compared with their vision at pre-pandemic examinations.4 However, screen time has not been associated with vision impairment among students. Furthermore, evidence regarding the impact of digital devices use on vision impairment has been inconsistent,30 31 with computer screen time made students’ vision worse while television viewing had no effect. We speculate that the association will become clearer as school-aged children spend increasing amounts of time using these devices.

This study had three important limitations. First, the screen time data were retrospectively collected through a self-reporting mechanism, which may have led to recall bias. However, considering the resource and measurement limitations that researchers encountered during the pandemic, self-reported recall was regarded as the optimal method for collection of screen time data in the present study. Second, the selection of students with poor vision may lead to underestimation of screen time effects on the general population, and the results should be generalised with caution. Third, the study was not designed to accurately distinguish between vision impairment caused by intrinsic factors and vision impairment caused by pandemic-related eye strain.

Our findings provide new evidence regarding the effects of increased screen time on asthenopia and vision impairment among students in western rural China during the pandemic; they can also serve as a basis for future research. Although pandemic-related school closures are temporary, the increasing popularity of online classes may accelerate the overall acceptance of digital devices. The use of online learning approaches is associated with multiple vision problems, which merit attention in future studies.

The present study demonstrated that asthenopia and vision impairment among students in western rural China were also affected by the pandemic; these findings provide critical insights regarding the effects of the pandemic on vision health in rural students. Moreover, the findings highlight important issues related to childhood vision health during the pandemic; parents, teachers, and eye care providers should consider evidence-based measures to avoid asthenopia and vision impairment in children. The current pace of economic and technological development is leading to increased use of digital devices and online learning approaches, but vision problems in rural China have not received sufficient consideration. Thus, there is a critical need for greater efforts to monitor VA and vision health among students in this region.

Concept or design: All authors.
Acquisition of data: Y Ding, H Guan, K Du.
Analysis or interpretation of data: Y Ding, H Guan, K Du, Y Shi.
Drafting of the manuscript: Y Ding, Y Zhang, Z Wang.
Critical revision of the manuscript for important intellectual content: H Guan, Y Shi.

All authors contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

As an International Editorial Advisory Board member of the journal, Y Shi was not involved in the peer review process. Other authors have disclosed no conflicts of interest.

We thank Dr Wenting Liu, Dr Jiaqi Zhu, and staff from the Center for Experimental Economics in Education of Shaanxi Normal University, China for their valuable contributions.

H Guan received funding for this study from the National Natural Science Foundation of China (Grant No.: 7180310) and Soft Science Project of Shaanxi Province (Grant No.: 2023-CX-RKX-127). Y Ding received funding for this study from the Fundamental Research Funds for the Central Universities (Grant No.: 2020CSWY018). This study was supported by the 111 Project (Grant No.: B16031). The funders had no role in designing the study, collecting, analysing or interpreting the data, or in drafting this manuscript.

This study protocol was approved by Sun Yat-sen University, China (Registration No.: 2013MEKY018) and all procedures followed the principles of the Declaration of Helsinki. Permission was obtained from the local boards of education in the study area, as well as the principals of all participating schools. All participating children provided oral assent before baseline data collection, and legal guardians provided written informed consent for their children to be enrolled in the study.

1. World Health Organization. WHO timeline—COVID-19. 2020. Available from: https://www.who.int/news-room/detail/27-04-2020-who-timeline---covid-19. Accessed 13 Sep 2021.

2. Li D, Liu Z, Liu Q, et al. Estimating the efficacy of quarantine and traffic blockage for the epidemic caused by 2019-nCoV (COVID-19): a simulation analysis. medRxiv [Preprint]. 25 Feb 2020. Available from: https://doi.org/10.1101/2020.02.14.20022913. Accessed 13 Sep 2021. Crossref

3. Wang G, Zhang Y, Zhao J, Zhang J, Jiang F. Mitigate the effects of home confinement on children during the COVID-19 outbreak. Lancet 2020;395:945-7. Crossref

4. Wang J, Li Y, Musch DC, et al. Progression of myopia in school-aged children after COVID-19 home confinement. JAMA Ophthalmol 2021;139:293-300. Crossref

5. Kowalska M, Zejda JE, Bugajska J, Braczkowska B, Brozek G, Malińska M. Eye symptoms in office employees working at computer stations [in Polish]. Med Pr 2011;62:1-8.

6. Cumberland PM, Peckham CS, Rahi JS. Inferring myopia over the lifecourse from uncorrected distance visual acuity in childhood. Br J Ophthalmol 2007;91:151-3. Crossref

7. Pascolini D, Mariotti SP. Global estimates of visual impairment: 2010. Br J Ophthalmol 2012;96:614-8. Crossref

8. Resnikoff S, Pascolini D, Mariotti SP, Pokharel GP. Global magnitude of visual impairment caused by uncorrected refractive errors in 2004. Bull World Health Organ 2008;86:63-70. Crossref

9. Jan C, Li SM, Kang MT, et al. Association of visual acuity with educational outcomes: a prospective cohort study. Br J Ophthalmol 2019;103:1666-71. Crossref

10. Roch-Levecq AC, Brody BL, Thomas RG, Brown SI. Ametropia, preschoolers’ cognitive abilities, and effects of spectacle correction. Arch Ophthalmol 2008;126:252-8. Crossref

11. Zhang Z, Xu G, Gao J, et al. Effects of e-learning environment use on visual function of elementary and middle school students: a two-year assessment—experience from China. Int J Environ Res Public Health 2020;17:1560. Crossref

12. Wong CW, Tsai A, Jonas JB, et al. Digital screen time during COVID-19 pandemic: risk for a further myopia boom? Am J Ophthalmol 2021;223:333-7. Crossref

13. Kim J, Hwang Y, Kang S, et al. Association between sexposure to smartphones and ocular health in adolescents. Ophthalmic Epidemiol 2016;23:269-76. Crossref

14. Mohan A, Sen P, Shah C, Jain E, Jain S. Prevalence and risk factor assessment of digital eye strain among children using online e-learning during the COVID-19 pandemic: digital eye strain among kids (DESK study-1). Indian J Ophthalmol 2021;69:140-4. Crossref

15. Guan H, Yu NN, Wang H, et al. Impact of various types of near work and time spent outdoors at different times of day on visual acuity and refractive error among Chinese school-going children. PLoS One 2019;14:e0215827.Crossref

16. Sultana A, Tasnim S, Hossain MM, Bhattacharya S, Purohit N. Digital screen time during the COVID-19 pandemic: a public health concern. Available from: https://f1000research.com/articles/10-81. Accessed 13 Sep 2021. Crossref

17. Nigg CR, Wunsch K, Nigg C, et al. Are physical activity, screen time, and mental health related during childhood, preadolescence, and adolescence? 11-year results from the German Montorik–Modul Longitudinal Study. Am J Epidemiol 2021;190:220-9. Crossref

18. Schmidt SC, Anedda B, Burchartz A, et al. Physical activity and screen time of children and adolescents before and during the COVID-19 lockdown in Germany: a natural experiment. Sci Rep 2020;10:21780. Crossref

19. Aguilar-Farias N, Toledo-Vargas M, Miranda-Marquez S, et al. Sociodemographic predictors of changes in physical activity, screen time, and sleep among toddlers and preschoolers in Chile during the COVID-19 pandemic. Int J Environ Res Public Health 2020;18:176. Crossref

20. Bates LC, Zieff G, Stanford K, et al. COVID-19 impact on behaviors across the 24-hour day in children and adolescents: physical activity, sedentary behavior, and sleep. Children (Basel) 2020;7:138. Crossref

21. National Bureau of Statistics of China, PRC Government. China Statistical Yearbook 2020. Available from: http://www.stats.gov.cn/tjsj/ndsj/2020/indexch.htm. Accessed 14 Sep 2021.

22. National Bureau of Statistics of China, PRC Government. China Statistical Yearbook 2013. Beijing, China: China State Statistical Press; 2013.

23. Seguí Mdel M, Cabrero García J, Crespo A, Verdú J, Ronda E. A reliable and valid questionnaire was developed to measure computer vision syndrome at the workplace. J Clin Epidemiol 2015;68:662-73. Crossref

24. Yi H, Zhang L, Ma X, et al. Poor vision among China’s rural primary school students: prevalence, correlates and consequences. China Econ Rev 2015;33:247-62. Crossref

25. United Nations International Children’s Emergency Fund. Don’t let children be the hidden victims of COVID-19 pandemic. Available from: https://www.unicef.org/press-releases/dont-let-children-be-hidden-victims-covid-19-pandemic. Accessed 6 Oct 2020.

26. Moon JH, Kim KW, Moon NJ. Smartphone use is a risk factor for pediatric dry eye disease according to region and age: a case control study. BMC Ophthalmol 2016;16:188. Crossref

27. Moon JH, Lee MY, Moon NJ. Association between video display terminal use and dry eye disease in school children. J Pediatr Ophthalmol Strabismus 2014;51:87-92. Crossref

28. Rechichi C, De Mojà G, Aragona P. Video game vision syndrome: a new clinical picture in children? J Pediatr Ophthalmol Strabismus 2017;54:346-55.Crossref

29. Mowatt L, Gordon C, Santosh AB, Jones T. Computer vision syndrome and ergonomic practices among undergraduate university students. Int J Clin Pract 2018;72:e13035. Crossref

30. Terasaki H, Yamashita T, Yoshihara N, Kii Y, Sakamoto T. Association of lifestyle and body structure to ocular axial length in Japanese elementary school children. BMC Ophthalmol 2017;17:123. Crossref

31. Fernández-Montero A, Olmo-Jimenez JM, Olmo N, et al. The impact of computer use in myopia progression: a cohort study in Spain. Prev Med 2015;71:67-71. Crossref

Patients with breast cancer receiving (neo)adjuvant treatment exhibit improved prognoses.1 However, chemotherapy regimens for breast cancer are associated with various degrees of chemotherapy-induced nausea and vomiting (CINV). The doxorubicin/cyclophosphamide (AC) regimen is one of the most frequently prescribed regimens for patients with breast cancer who are receiving (neo) adjuvant chemotherapy; AC is among the highly emetogenic chemotherapies with ≥90% risk of nausea and vomiting.

In situations where a neurokinin-1 receptor antagonist (NK1RA) is accessible, most current guidelines for AC(-like) chemotherapy recommend the use of a prophylactic triplet antiemetic regimen that consists of an NK1RA, a 5-hydroxytryptamine type-3 receptor antagonist (5HT3RA), and a corticosteroid, with or without olanzapine.2 3 4 In addition to earlier NK1RAs (eg, aprepitant, fosaprepitant, and rolapitant), netupitant/palonosetron (NEPA) [Akynzeo], which is a combination of an NK1RA (netupitant 100 mg) and a second-generation 5HT3RA (palonosetron 0.5 mg), has been available in the past decade. Although palonosetron constitutes a more potent 5HT3RA,5 it also has synergistic interactions with netupitant that include interference with 5HT3 receptor cross-talk and enhancement of the netupitant-mediated effect on NK1 receptor internalisation.6 7

In a recent systematic review and meta-analysis, Yokoe et al8 compared different antiemetic regimens to assess their control of CINV in patients receiving highly emetogenic chemotherapy regimens. The authors arbitrarily defined the ‘conventional’ regimen as a three-drug regimen that contained dexamethasone, a first- or second-generation secondgeneration 5HT3RA, and an earlier NK1RA compound (aprepitant, fosaprepitant, or rolapitant); they defined ‘new’ regimens as regimens that contained NEPA or olanzapine. The results indicated that, compared with conventional regimens, new regimens containing NEPA were more effective in terms of producing a complete response (ie, absence of vomiting and no use of rescue therapy). Additionally, Yokoe et al8 showed that olanzapine-containing regimens were most effective in terms of producing a complete response, particularly when olanzapine was added to a triplet regimen of an NK1RA, a 5HT3RA, and dexamethasone. These findings were supported by the results of a prospective randomised study published in 2020, which directly compared an olanzapine-containing four-drug regimen with a standard triplet antiemetic regimen (consisting of aprepitant, ondansetron, and dexamethasone) for the prevention of CINV in patients receiving AC chemotherapy.9

Here, we conducted a post-hoc analysis through retrospective assessment of individual patient data from two previously reported prospective antiemetic studies that involved Chinese patients with breast cancer.9 10 We hypothesised that a four-drug antiemetic regimen (consisting of an NK1RA, a 5HT3RA, dexamethasone, and olanzapine) would remain superior to a three-drug regimen (consisting of an NK1RA, a 5HT3RA, and dexamethasone) that included NEPA as a combination NK1RA and 5HT3RA agent. The primary objective was to compare the efficacies of olanzapine- and NEPA-containing antiemetic regimens in terms of controlling CINV during the first cycle of AC. The secondary objectives were: (1) to assess quality of life (QOL) outcomes in patients receiving these treatments during the first cycle of AC, and (2) to assess emesis control outcomes in patients receiving these treatments over multiple cycles of AC.

This study constituted a post-hoc analysis of data from two recently reported prospective studies. The first prospective study investigated emesis outcomes in patients with breast cancer who received a standard triplet antiemetic regimen (ie, aprepitant, ondansetron, and dexamethasone) with or without olanzapine9; after the first study, a second prospective study was conducted to assess the antiemetic efficacy of NEPA and dexamethasone.10 These studies were conducted with institutional ethics approval and were registered at ClinicalTrials.gov (NCT03386617 and NCT03079219, respectively). For the post-hoc analysis, data were extracted from the first study regarding patients who received an olanzapine plus aprepitant-containing four-drug antiemetic regimen; data were extracted from the second study regarding patients who received NEPA and dexamethasone. These patients were categorised into the ‘olanzapine’ and ‘NEPA’ groups, respectively.

Inclusion criteria were similar for the two studies. Specifically, patients were eligible if they were women of Chinese ethnicity, were aged >18 years, had early-stage breast cancer, and planned to receive a regimen of (neo)adjuvant AC. All study participants were required to read, understand, and complete study questionnaires and diaries in Chinese. Exclusion criteria included abnormal bone marrow, renal, or hepatic functions; receipt or planned receipt of radiation therapy to the abdomen or pelvis within 7 days prior to initial administration of study treatment; presence of grade 2 to 3 nausea, as defined by the National Cancer Institute Common Terminology Criteria for Adverse Events version 4.0,11 or vomiting within 24 hours prior to initial administration of the study treatment; presence of an active infection or any uncontrolled disease; a history of illicit drugs, including marijuana or alcohol abuse; mental incapacitation; and/or presence of a clinically significant emotional or psychiatric disorder. Written consent was provided by eligible patients prior to enrolment in the studies.

Patients in the olanzapine group received olanzapine 10 mg, aprepitant 125 mg, dexamethasone 12 mg, and ondansetron 8 mg before chemotherapy on day 1; they also received ondansetron 8 mg 8 hours after chemotherapy. Subsequently, they received aprepitant 80 mg daily on days 2-3 and olanzapine 10 mg daily on days 2-5.

Patients in the NEPA group received one capsule of NEPA (netupitant 300 mg/palonosetron 0.50 mg) with dexamethasone 12 mg before chemotherapy on day 1. Subsequently, they received dexamethasone 4 mg twice per day on days 2-3.

At the initiation of chemotherapy on day 1, individual patients were provided a diary to record the date and time of their symptoms of vomiting and nausea for 120 hours after the AC infusion; the use of any rescue medication was also recorded. On days 2-6, patients rated their symptoms of nausea for the previous 24 hours using a visual analogue scale (in which 0 mm implied no nausea, whereas 100 mm implied nausea that was ‘as bad as it could be’). Additionally, on day 1 (before infusion of AC) and day 6 (after completion of the diary), patients completed the Functional Living Index-Emesis (FLIE) questionnaire. A research nurse/assistant called individual patients on days 2-6 to remind them to take the study medications, complete the patient diary, and complete the FLIE questionnaire.

Antiemetic efficacy was measured across three overlapping time periods. The ‘acute’ phase comprised 0 to 24 hours from the infusion of AC; the ‘delayed’ phase comprised 24 to 120 hours from the infusion of AC; the ‘overall’ phase comprised 0 to 120 hours from the infusion of AC.

Variables used to assess antiemetic efficacy were ‘complete response’, ‘no vomiting’, ‘no significant nausea’, ‘no nausea’, ‘no use of rescue therapy’, ‘complete protection’, and ‘total control’; definitions of these variables are provided in Table 1. The proportions of patients who exhibited these variables were recorded separately. Additionally, the ‘time to first vomiting’ in cycle 1 was determined using information recorded in patients’ diaries.

Quality of life was evaluated using the Chinese version of self-reported FLIE questionnaires from individual patients.12 The FLIE questionnaire consists of a nausea domain (9 items) and a vomiting domain (9 items). All scores were transformed to ensure that higher scores indicated worse impact on QOL.

A modified intention-to-treat approach was used for all efficacy analyses; specifically, analyses included patients who had received chemotherapy, had completed the study procedures from 0 to 120 hours in cycle 1 of AC, and had no major protocol violations.

To achieve the primary objective of this study, the efficacies of the two antiemetic regimens were based on the proportions (including 95% confidence intervals) of patients who achieved complete response during the acute, delayed, and overall phases after AC infusion in cycle 1. Other parameters compared in cycle 1 of AC were ‘time to first vomiting’, ‘no vomiting’, ‘no significant nausea’, no nausea’, ‘no use of rescue therapy’, ‘complete protection’, and ‘total control’.

To achieve the secondary objectives, QOL was compared between the two antiemetic regimens based on assessments of the nausea domain, vomiting domain, and total score (sum of nausea and vomiting domains) of the FLIE questionnaire during cycle 1 of AC. Emesis control over multiple cycles was compared between the two antiemetic regimens by assessing the proportions (including 95% confidence intervals) of patients who achieved ‘complete response’, ‘complete protection’, and ‘total control’ in the acute, delayed, and overall phases.

Comparisons between the two antiemetic regimens were made using the Wilcoxon rank-sum test for continuous data and Pearson’s Chi squared test for dichotomous data. Two-sided P values <0.05 were considered statistically significant. The SAS Software version 9.4 (SAS Institute, Cary [NC], United States) was used for analyses.

Data from 120 patients were included in this study; 60 patients each were enrolled in the NEPA and olanzapine groups. Fifty-six patients (93.3%) in the olanzapine group completed all four cycles of AC, whereas 60 patients (100%) in the NEPA group completed all four cycles of AC.

Patient characteristics, including characteristics that could potentially affect CINV, are shown in Table 2. The olanzapine and NEPA groups had very similar patient characteristics, with median ages of 54.5 and 56 years, respectively. Nearly two-thirds of patients in each group had Stage II breast cancer (63.3% and 66.7%, respectively). The percentage of patients with a history of motion sickness was higher in the NEPA group (35%) than in the olanzapine group (16.7%). Furthermore, 30% of patients in the NEPA group and 20% of patients in the olanzapine group received AC as neoadjuvant treatment.

Antiemetic efficacies during cycle 1 of AC in the olanzapine and NEPA groups are shown in Table 3. Complete response rates in acute, delayed, and overall phases in cycle 1 did not differ between groups. In the acute phase, the olanzapine group exhibited a higher rate of ‘no use of rescue therapy’ (olanzapine vs NEPA: 96.7% vs 85.0%, P=0.0225). No parameters differed between groups in the delayed phase. In the overall phase, the olanzapine group exhibited significantly higher rates of ‘no use of rescue therapy’ (91.7% vs 76.7%, P=0.0244) and ‘no significant nausea’ (91.7% vs 78.3%, P=0.0408).

The median time to first vomiting was not reached in either group (P=0.3902). Quality of life results during cycle 1 of AC in the olanzapine and NEPA groups, determined using the FLIE questionnaire, are shown in the Figure. There were no significant differences in the nausea domain, vomiting domain, or total score of the FLIE questionnaire between the two groups.

Antiemetic efficacies over multiple cycles of AC in the olanzapine and NEPA groups are shown in Table 4. In the acute phase, the NEPA group exhibited significantly higher rates of total control in cycle 2 (olanzapine vs NEPA: 59.6% vs 81.7%, P=0.0087) and cycle 4 (63.2% vs 86.7%, P=0.0032). No parameters differed between groups in the delayed phase. In the overall phase, the NEPA group exhibited significantly higher rates of total control in cycle 3 (55.4% vs 73.3%, P=0.0430) and cycle 4 (54.4% vs 75.0%, P=0.0195).

Chemotherapy-induced nausea and vomiting is a frustrating adverse effect for patients receiving anticancer treatment.13 The administration of optimal antiemetic prophylaxis can help to maintain QOL, while potentially improving patient compliance in terms of completing planned therapies. In current antiemetic prophylaxis guidelines, the European Society of Medical Oncology/Multinational Association of Supportive Care in Cancer, the American Society of Clinical Oncology, and the United States National Comprehensive Cancer Network offer several options regarding antiemetic regimens for patients receiving AC(-like) chemotherapy. These options mainly involve the combination of a 5HT3RA and corticosteroids, with or without an NK1RA and olanzapine.2 3 4 In particular, the incorporation of olanzapine, an antipsychotic drug with antagonistic effects on various receptors (eg, dopamine and serotonin receptors),14 is increasingly regarded as a component of antiemetic prophylaxis for patients receiving anticancer treatment.

In an attempt to identify the best antiemetic regimen, Yokoe et al8 conducted a meta-analysis of randomised trials that tested various antiemetic regimens. The results indicated that olanzapine-based regimens demonstrated the best efficacy. Specifically, olanzapine in combination with an NK1RA, a 5HT3RA, and dexamethasone exhibited the greatest efficacy; other olanzapine-containing regimens (consisting of a 5HT3RA and dexamethasone) were also superior to regimens that lacked olanzapine. Moreover, even in the presence of earlier NK1RAs (eg, aprepitant, fosaprepitant, or rolapitant), regimens lacking olanzapine remained inferior.

Similar to the findings with olanzapine, Yokoe et al8 reported that triplet antiemetics involving NEPA were superior to conventional NK1RAs (eg, aprepitant, fosaprepitant, or rolapitant). Furthermore, Zhang et al15 directly compared NEPA-based antiemetic regimens with aprepitant-based triplet regimens in a randomised study that involved 800 patients who underwent administration of a cisplatin-containing regimen. Their results revealed that patients receiving NEPA and dexamethasone exhibited similar control of CINV, compared with patients receiving aprepitant, granisetron, and dexamethasone; however, NEPA-treated patients had a significantly lower requirement for rescue therapy. Additionally, in a recent study focused on patients with breast cancer who were undergoing AC chemotherapy, patients who received NEPA and dexamethasone demonstrated significantly higher rates of complete response, complete protection, and total control with enhanced QOL, compared to historical controls who received aprepitant, ondansetron, and dexamethasone; these benefits persisted over multiple cycles of chemotherapy.10

To our knowledge, no study has directly compared olanzapine- and NEPA-containing regimens. Using an indirect comparison approach, the present study showed that the olanzapine-based regimen had higher rates of ‘no use of rescue therapy’ and ‘no significant nausea’ in cycle 1 of AC, compared to the NEPA-based regimen. In contrast, assessments in subsequent cycles revealed that the NEPA-based regimen led to higher rates of total control in the acute phase (cycles 2 and 4) and the overall phase (cycles 3 and 4). The lack of difference in QOL between the two groups of patients may be related to the difference in adverse-effect profiles of the antiemetics used. For instance, the continued use of dexamethasone on days 2-3 in the NEPA group may have affected QOL among those patients because of its effects on mood, insomnia, gastrointestinal symptoms, and metabolic profiles.16 Indeed, a recent meta-analysis showed that, among patients receiving AC or moderately emetogenic chemotherapy, 3 days of dexamethasone did not provide additional benefit compared to 1 day of the agent.17 However, olanzapine has been associated with sedation and somnolence.18 Thus, after the completion of a phase 2 trial in Japan that suggested olanzapine was more effective at 5 mg than at 10 mg,19 the same group of investigators conducted a phase 3 study in which they tested the addition of daily olanzapine 5 mg to an aprepitant-based three-drug regimen; the results showed that, even at a lower dose of olanzapine, the olanzapine-containing regimen remained more efficacious than the olanzapine-free regimen for patients receiving cisplatin.20 Other adverse effects have been reported. Our analysis of olanzapine in combination with aprepitant, ondansetron and dexamethasone revealed a significantly higher incidence of grade ≥2 neutropenia in the olanzapine arm than in the standard arm, although this altered incidence was not associated with a significant difference in neutropenic fever.9 A few cases of olanzapine-induced neutropenia have been reported21; additionally, a recent randomised antiemetic study showed that patients who received an olanzapine-containing regimen had a higher frequency of severe neutropenia (without an increased incidence of neutropenic fever).22 Although the underlying mechanism remains unknown, the results of the aforementioned Japanese study20 suggest that olanzapine 5 mg could reduce the incidence of neutropenia. In contrast, in our previous trial regarding a NEPA-based regimen, we found that patients in the NEPA arm had significantly lower incidences of grade ≥2 neutropenia and neutropenic fever, compared to historical controls who received an aprepitant-based regimen.9 10

This study had some potential limitations. First, dexamethasone was only used for 1 day in the olanzapine-based regimen, whereas it was administered for 3 days in the NEPA-based regimen; this difference may have influenced the findings. Second, the use of data from two separate studies may have affected the generalisability of the findings because of slight variations in patient characteristics; the lack of blinding in both studies also increased the potential for patient-related reporting biases. Nonetheless, the original studies were consecutively conducted during the period from 2017 to 2019; both the data from Chinese patients enrolled in a homogenous group with early-stage breast cancer who were receiving (neo)adjuvant AC chemotherapy and the present analysis were analysed based on individual patient data. These factors support the validity of our comparison approach.

In conclusion, the present findings do not conclusively support the superiority of either the olanzapine-based regimen or the NEPA-based regimen in terms of antiemetic efficacy or QOL among patients with breast cancer who are receiving AC. Our previous study demonstrated that aprepitant has a limited effect when used with a 5HT3RA and dexamethasone23; we also found that NEPA was superior to aprepitant.10 Overall, the available data suggest that olanzapine-containing antiemetic regimens can be used without aprepitant, particularly when seeking to reduce medical expenses. Moreover, the available data support the previous conclusion that, in parts of the world where socio-economic limitations restrict the availability of NK1RAs, the use of olanzapine combined with a 5HT3RA and dexamethasone may be an effective low-cost alternative antiemetic regimen.8 24 Antiemetic efficacy may be enhanced if NEPA is administered in combination with dexamethasone and olanzapine as a four-drug antiemetic regimen; however, the efficacy of an olanzapine plus NEPA regimen in terms of controlling CINV should be confirmed in a trial setting.

Concept or design: W Yeo.
Acquisition of data: FKF Mo, W Yeo.
Analysis or interpretation of data: W Yeo, CCH Yip, FKF Mo.
Drafting of the manuscript: W Yeo, CCH Yip.
Critical revision of the manuscript for important intellectual content: L Li, TKH Lau, VTC Chan, CCH Kwok, JJS Suen, FKF Mo.

All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

W Yeo has been involved in the Chemotherapy-Induced Nausea and Vomiting (CINV) Network in Asia and has provided lectures on CINV at events organised by Mundipharma International Limited, which supported the design of the NEPA study10 analysed in this post-hoc analysis but had no role in the present comparative analysis, data collection and analysis, decision to publish, or preparation of the manuscript.

We thank Ms Dong KT Lai, Ms Elizabeth Pang, Ms Vivian Chan, and Ms Maggie Cheung of the Department of Clinical Oncology, The Chinese University of Hong Kong for their support in contributing to patient enrolment and study monitoring.

Data from this study were presented at the European Society of Medical Oncology Asia Virtual Congress 2020 on 27 November 2020.

This study was supported by an education grant from Madam Diana Hon Fun Kong Donation for Cancer Research (Grant No.: CUHK Project Code 7104870). The Donation had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

The studies examined in this post-hoc analysis were approved by The Joint Chinese University of Hong Kong–New Territories East Cluster Institution Review Board of The Chinese University of Hong Kong and the Hong Kong Hospital Authority, and the Kowloon West Cluster Research Ethics Committee of the Hong Kong Hospital Authority (Ref No.: CREC 2016.013, CREC 2017.1609 and KW/FR-18-019[119-19]). All patient data in this study were anonymous and were based on the abovementioned reported studies. There was no additional work on retrieving patient records in this study.

1. Early Breast Cancer Trialists’ Collaborative Group (EBCTCG); Peto R, Davies C, et al. Comparisons between different polychemotherapy regimens for early breast cancer: meta-analyses of long-term outcome among 100,000 women in 123 randomised trials. Lancet 2012;379:432-44. Crossref

2. National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology, Antiemesis Version 1; 2019.

3. Herrstedt J, Roila F, Warr D, et al. 2016 updated MASCC/ESMO consensus recommendations: prevention of nausea and vomiting following high emetic risk chemotherapy. Support Care Cancer 2017;25:277-88. Crossref

4. Hesketh PJ, Kris MG, Basch E, et al. Antiemetics: American Society of Clinical Oncology Clinical Practice Guideline update. J Clin Oncol 2017;35:3240-61. Crossref

5. Popovic M, Warr DG, Deangelis C, et al. Efficacy and safety of palonosetron for the prophylaxis of chemotherapy-induced nausea and vomiting (CINV): a systematic review and meta-analysis of randomized controlled trials. Support Care Cancer 2014;22:1685-97. Crossref

6. Thomas AG, Stathis M, Rojas C, Slusher BS. Netupitant and palonosetron trigger NK1 receptor internalization in NG108-15 cells. Exp Brain Res 2014;232:2637-44. Crossref

7. Stathis M, Pietra C, Rojas C, Slusher BS. Inhibition of substance P-mediated responses in NG108-15 cells by netupitant and palonosetron exhibit synergistic effects. Eur J Pharmacol 2012;689:25-30. Crossref

8. Yokoe T, Hayashida T, Nagayama A, et al. Effectiveness of antiemetic regimens for highly emetogenic chemotherapy-induced nausea and vomiting: a systematic review and network meta-analysis. Oncologist 2019;24:e347-57. Crossref

9. Yeo W, Lau TK, Li L, et al. A randomized study of olanzapine-containing versus standard antiemetic regimens for the prevention of chemotherapy-induced nausea and vomiting in Chinese breast cancer patients. Breast 2020;50:30-8. Crossref

10. Yeo W, Lau TK, Kwok CC, et al. NEPA efficacy and tolerability during (neo)adjuvant breast cancer chemotherapy with cyclophosphamide and doxorubicin. BMJ Support Palliat Care 2022;12:e264-70.

11. National Cancer Institute. Cancer Therapy Evaluation Program. 2021. Available from: https://ctep.cancer.gov/protocoldevelopment/electronic_applications/ctc.htm#ctc_40. Accessed 3 Feb 2023.

12. Martin AR, Pearson JD, Cai B, Elmer M, Horgan K, Lindley C. Assessing the impact of chemotherapy-induced nausea and vomiting on patients’ daily lives: a modified version of the Functional Living Index-Emesis (FLIE) with 5-day recall. Support Care Cancer 2003;11:522-7. Crossref

13. Griffin AM, Butow PN, Coates AS, et al. On the receiving end. V: patient perceptions of the side effects of cancer chemotherapy in 1993. Ann Oncol 1996;7:189-95. Crossref

14. Bymaster FP, Nelson DL, DeLapp NW, et al. Antagonism by olanzapine of dopamine D1, serotonin2, muscarinic, histamine H1 and alpha 1-adrenergic receptors in vitro. Schizophr Res 1999;37:107-22. Crossref

15. Zhang L, Lu S, Feng J, et al. A randomized phase III study evaluating the efficacy of single-dose NEPA, a fixed antiemetic combination of netupitant and palonosetron, versus an aprepitant regimen for prevention of chemotherapy-induced nausea and vomiting (CINV) in patients receiving highly emetogenic chemotherapy (HEC). Ann Oncol 2018;29:452-8. Crossref

16. Roila F, Ruggeri B, Ballatori E, et al. Aprepitant versus metoclopramide, both combined with dexamethasone, for the prevention of cisplatin-induced delayed emesis: a randomized, double-blind study. Ann Oncol 2015;26:1248-53. Crossref

17. Okada Y, Oba K, Furukawa N, et al. One-day versus three-day dexamethasone in combination with palonosetron for the prevention of chemotherapy-induced nausea and vomiting: a systematic review and individual patient data-based meta-analysis. Oncologist 2019;24:1593-600. Crossref

18. Navari RM, Qin R, Ruddy KJ, et al. Olanzapine for the prevention of chemotherapy-induced nausea and vomiting. N Engl J Med 2016;375:134-42. Crossref

19. Abe M, Hirashima Y, Kasamatsu Y, et al. Efficacy and safety of olanzapine combined with aprepitant, palonosetron, and dexamethasone for preventing nausea and vomiting induced by cisplatin-based chemotherapy in gynecological cancer: KCOG-G1301 phase II trial. Support Care Cancer 2016;24:675-82. Crossref

20. Hashimoto H, Abe M, Tokuyama O, et al. Olanzapine 5 mg plus standard antiemetic therapy for the prevention of chemotherapy-induced nausea and vomiting (J-FORCE): a multicentre, randomised, double-blind, placebo-controlled, phase 3 trial. Lancet Oncol 2020;21:242-9. Crossref

21. Malhotra K, Vu P, Wang DH, Lai H, Faziola LR. Olanzapine-induced neutropenia. Ment Illn 2015;7:5871. Crossref

22. Gjafa E, Ng K, Grunewald T, et al. Neutropenic sepsis rates in patients receiving bleomycin, etoposide and cisplatin chemotherapy using olanzapine and reduced doses of dexamethasone compared to a standard antiemetic regimen. BJU Int 2021;127:205-11. Crossref

23. Yeo W, Mo FK, Suen JJ, et al. A randomized study of aprepitant, ondansetron and dexamethasone for chemotherapy-induced nausea and vomiting in Chinese breast cancer patients receiving moderately emetogenic chemotherapy. Breast Cancer Res Treat 2009;113:529-35. Crossref

24. Babu G, Saldanha SC, Kuntegowdanahalli Chinnagiriyappa L, et al. The efficacy, safety, and cost benefit of olanzapine versus aprepitant in highly emetogenic chemotherapy: a pilot study from South India. Chemother Res Pract 2016;2016:3439707. Crossref

The coronavirus disease 2019 (COVID-19) pandemic, which began in early 2020, has resulted in unprecedented large-scale public health responses. Stringent regional social distancing measures (eg, quarantine, school closures, and restrictions at work and recreation destinations) were rapidly implemented during the early days of the pandemic as forms of non-pharmacological intervention.1 Although there is evidence that such measures can temporarily contain the spread of severe acute respiratory syndrome coronavirus 2,2 collateral effects among non–COVID-19–related conditions have also been reported.3 Trauma is the leading cause of death and disability among young adults worldwide,4 but the effects of the COVID-19 pandemic on injuries and fracture incidence within Hong Kong have not been fully elucidated. This uncertainty has hindered healthcare resource deployment and clinical service demand estimation in times of stringency. We sought to address this problem using ‘big data’ sources and regional clinical data repositories, which allow researchers to rapidly delineate epidemiological associations with potential applications in forecasting models, while avoiding resource-intensive collection of conventional epidemiological information and protecting patient anonymity.

We presumed that restrictions on citizen mobility, in concert with social distancing, were associated with reductions in musculoskeletal injuries during the early days of the COVID-19 pandemic. Specifically, we hypothesised that reduced population mobility was associated with reductions in fracture incidence and fracture-related healthcare needs during the early days of the pandemic. We investigated these relationships by analysing daily multicentre hospital data registries in Hong Kong, along with digital population mobility datasets published by a technology company. Our main outcome measurement was skeletal fractures, which served as a specific surrogate for musculoskeletal trauma.

This study was conducted in Hong Kong, a highincome region (with gross domestic product per capita of HK$357 667 in 20205) that was among the first areas affected by COVID-19; social distancing measures were implemented during the early days of the pandemic.

Using the Clinical Data Analysis and Reporting System of the Hospital Authority, anonymised patient records were retrieved from all 43 public hospitals in Hong Kong for the period from 22 November 2016 to 20 May 2020. In Hong Kong, up to 90% of hospital bed-days occur in public hospitals, which manage nearly all critical emergencies in the region.6 Anonymised clinical data were retrieved, including time of initial injury presentation, emergency department triage, trauma category, hospital admission, diagnosis, and surgical procedures. Diagnoses and procedures were encoded in accordance with the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) by treating physicians based on clinical and radiological investigations, intraoperative findings, and date of hospital discharge. The ICD-9-CM codes that met the inclusion criteria (which included fractures under the purview or commonly admitted under the care of orthopaedic and traumatology service) were all codes from 805 to 829 (inclusive). Duplicate records from fracture reassessment related to follow-up attendances, hospitalisation after emergency department attendance, and elective hospital re-admissions (ie, episodes assigned to the same patient unique identifier with identical diagnostic codes, which occurred within 30 days of the index episode) were regarded as a single event to avoid double counting. Pathological fractures and records with missing diagnosis codes or admission times were excluded from the analysis.

The ‘COVID-19 epoch’ was defined as 25 January 2020 (activation of the government’s ‘emergency’ response and commencement of social distancing policies7) to 26 March 2020; this arbitrarily chosen 62-day period included all patients with fractures who presented during that period. This epoch was compared with the 9 weeks preceding the onset of the COVID-19 pandemic (ie, 22 November 2019 to 24 January 2020), as well as the same period over the past 3 years to adjust for seasonality-related variations8 (ie, 25 January to 26 March in 2017, 2018, and 2019). Differences between actual and projected daily fracture incidences were calculated based on mean values at the same time of year over the past 3 years. Fracture-related mortality rates, defined as the numbers of deaths within 30 days after initial fracture presentation per 100 000 person-years, were compared. The Chi squared test was used to detect differences in fracture incidence and fracture-related mortality during the COVID-19 pandemic and pre-pandemic epochs.

Surrogate data concerning population mobility were retrieved from Mobility Trends Reports9—an aggregate daily measure of geographical direction requests on Apple Maps, a service established by Apple Inc, which holds the largest market share of electronic mobile devices (including smartphones and tablets) in Hong Kong.10 Walking index was regarded as an index of population mobility, considering the smartphone penetration of 91.5%11 among the 7.50 million residents of Hong Kong.12

Associations between daily fracture incidence and population mobility were determined by incidence rate ratios (IRRs) using quasi-Poisson regression. Secondary analysis involved associations between mobility index and fracture repair surgeries, all types of orthopaedic emergency department attendances, orthopaedic hospital admissions, and emergency orthopaedic surgeries.

Because medical records were timestamped in Hong Kong time (8 hours ahead of Greenwich Mean Time), they were converted to Pacific Time to match the time intervals listed in Mobility Trends Reports; this conversion ensured that data were temporally matched for analysis.

To determine whether mobility associations simply reflected health-seeking behaviour, we included analyses of diseases which lacked a physiological basis and were not associated with population mobility—these ‘controls’ included appendicitis, cellulitis, and abscess (ICD-9-CM diagnosis codes 540 and 682). Statistical analysis was performed using R software, version 3.6.2 (R Foundation for Statistical Computing, Vienna, Austria). Quasi-Poisson regression was used to model the relationship between the population mobility index and the daily incidences of fractures and fracture-related events; the population mobility index was the explanatory variable, whereas the

daily incidences of various events were response variables. A quasi-Poisson distribution was preferred over a Poisson distribution, considering the presence of significant overdispersion among some response variables (in the form of count data) when a dispersion parameter was included. In accordance with standard statistical methods, the natural logarithm was utilised as the link function. Incidence rate ratios were reported and represented by the following formula:

Estimated incidence = IRR^PMI × BIR

where IRR represents the incidence rate ratio, PMI represents the population mobility index, and BIR represents the baseline incidence rate. The IRR, which quantifies the relationship between the mobility index and fracture incidence, is multiplicative in nature—for every unit increase in the mobility index, there is a corresponding multiplicative increase in the IRR. If the IRR is <1, it is expected to decrease in a multiplicative manner for every unit decrease in the mobility index. Multiple comparisons were adjusted by Bonferroni correction, and the threshold for statistical significance was regarded as P<0.00227 (0.05/22).

In total, 59 931 fracture-related medical records from orthopaedic emergency department attendances, hospital admissions, and surgeries were reviewed. After exclusion of 11 498 linked episodes, 284 pathological fractures, 786 follow-up attendances, 175 hospital re-admissions, and two episodes with missing admission times, 47 186 fractures were included in the analysis. Descriptive statistics regarding daily fracture incidences, controls, and fracture-related surgeries during COVID-19 social distancing are shown in Table 1. Intra-year and inter-year comparison cohorts are presented in Table 2.

A reduction of 1748 fractures in the actual versus projected incidence (321.9 vs 459.1 per 100 000 person-years, P<0.001) was observed during the COVID-19 epoch; the relative risk was 0.690 (95% confidence interval [CI]=0.678-0.702), compared with mean incidences in the previous 3 years (ie, inter-year cohort) [Table 2]. Differences in fracture incidence between the pandemic and pre-pandemic epochs are shown in Figure 1.

Fracture incidences, population mobilities, and controls are depicted in Figure 2. The first two COVID-19 cases in Hong Kong were reported on 23 January 202013; three additional cases were reported on 24 January 2020. Social distancing measures were implemented on 25 January 2020; these included suspension of schools, initiation of ‘work from home’ measures among civil servants, and suspension of hospital visitations. Mandatory border quarantine was enforced on 8 February 2020. The sharpest decrease in mobility was observed on 24 January 2020; population mobility subsequently remained at low levels, in conjunction with cancellations of large-scale social and sporting events, as well as the imposition of travel restrictions with quarantine measures for returning travellers.7

Fracture incidence was positively associated with the population mobility index (IRR=1.0055, 95% CI=1.0044-1.0066, P<0.001). Analyses of fracture incidence according to anatomical location revealed associations of the population mobility index with upper limb fractures (IRR=1.0073, 95% CI=1.0057-1.0088, P<0.001) and lower limb fractures (IRR=1.0045, 95% CI=1.0030-1.0060, P<0.001) [Fig 3].

The population mobility index was associated with the incidences of fractures involving the radius and ulna (IRR=1.0079, 95% CI=1.0057-1.0101, P<0.001), hand and fingers (IRR=1.0069, 95% CI=1.0039-1.0098, P<0.001), femoral neck (IRR=1.0065, 95% CI=1.0035-1.0095, P<0.001), and tibia and fibula (IRR=1.0097, 95% CI=1.0044-1.0151, P<0.001) [Fig 4]. However, after Bonferroni correction, the population mobility index did not exhibit statistically significant associations with trochanteric hip fractures (IRR=1.0008, P=0.683), spine fractures (IRR=0.996, P=0.183), or pelvic fractures (IRR=1.0064, P=0.00799).

Stronger associations were observed among fractures, such that some patients presented at a younger age (eg, patients with tibia, fibula, hand, and finger fractures), whereas other patients presented at an older age (eg, patients with femoral neck fractures). Digital literacy, manual dexterity and visual acuity, and higher internet and smartphone usage among younger residents11 are among the factors that cause the population mobility index to have increased sensitivity for analysis in such age-groups.

The incidences of cellulitis, abscesses, and appendicitis were not associated with the population mobility index (P>0.00227). These findings support the hypothesis that changes in associations between fracture incidence and population mobility were not solely caused by changes in health-seeking behaviour; if they had been caused by changes in such behaviour, corresponding reductions in those conditions would have been observed.

The daily population mobility index was associated with the number of patients admitted on a particular day who subsequently underwent fracture repair surgeries (IRR=1.0041, 95% CI=1.0020-1.0062, P<0.001). The population mobility index was also associated with all types of emergency orthopaedic surgeries (IRR=1.0040, 95% CI=1.0021-1.0058, P<0.001), attendances at orthopaedic emergency departments (IRR=1.0076, 95% CI 1.0064-1.0087, P<0.001), and emergency orthopaedic hospital admissions (IRR=1.0054, 95% CI=1.0043-1.0064, P<0.001). Additionally, the numbers of orthopaedic patients triaged as critical, emergent, and urgent (ie, patients who require physician attention within 30 minutes of attendance) were also associated with the population mobility index (IRR=1.0063, 95% CI=1.0054-1.0073, P<0.001). Whereas the numbers of traffic-related and sports-related trauma cases were associated with the population mobility index (IRR=1.008, 95% CI=1.0063-1.0097 and IRR=1.013, 95% CI=1.0092-1.0158, respectively, both P<0.001), the number of assault-related trauma cases was not (P=0.238).

Forty-nine patients with fractures died within 30 days of presentation during the COVID-19 epoch. This constituted a mortality rate of 3.22 deaths per 100 000 person-years, which was lower than the rate of 4.70 deaths per 100 000 person-years during the period before the pandemic (P<0.001); thus, there were around 19 fewer fracture-related deaths in the Hong Kong population during the 62-day study period. Four patients with fractures had COVID-19 (ie, they had positive results in nasopharyngeal swab reverse transcriptase-polymerase chain reaction tests for severe acute respiratory syndrome coronavirus 2) and survived beyond 30 days after initial fracture presentation. The change in mortality was presumably explained by reduced fracture incidence: 30-day mortality among patients with fractures did not significantly differ between the COVID-19 epoch (1.2%, 49 deaths in 4101 patients) and the preceding period (1.0%, 175 deaths in 17 198 patients) [P=0.305].

This study analysed 47 186 fractures in Hong Kong, prior to and during the early days of the COVID-19 pandemic. Population mobility was assessed through aggregate digital footprints using the volume of location service requests as a surrogate marker, considering the high smartphone and internet penetration in Hong Kong11; importantly, datasets of aggregate digital footprints have been published to facilitate efforts to control COVID-19.9 The findings support our hypothesis in terms of the relationship between fracture incidence and population mobility.

Fractures incur substantial healthcare costs; for example, fragility fracture-related costs incurred costs of 37.5 billion euros, along with the loss of 1.0 million quality-adjusted life years, among the six largest European countries in 2017.14 Some fractures (eg, hip fractures) warrant early surgical management to mitigate the morbidity and mortality associated with surgical delays.15 Guidance regarding early surgical management remained in effect, even during the early days of the COVID-19 pandemic.16 Despite the best available tools, fracture prediction remains difficult; there are additional challenges associated with epidemiological projections of specific time points when such fractures occur. Accordingly, hospitals and public health entities experience difficulties in terms of estimating emergency trauma service load and allocating limited healthcare resources. Our findings suggest that population mobility indices, which are freely and publicly accessible, can provide insights regarding fracture incidence. Population mobility may be useful in quantitative modelling of fracture-related inpatient and surgical theatre service demand, using the IRRs described in this study.

Although there is evidence to support the efficacy of social distancing measures with respect to COVID-19 transmission,2 our findings emphasise the collateral impacts of pandemic-related interventions on non-communicable diseases. We found that fracture incidence decreased when population mobility was hindered by social distancing measures; the relative reduction in overall fractures appeared to be similar to the effect of established pharmacological interventions on fragility fractures.17 Although this relationship appears to contradict the common notion that physical activity confers a protective effect against fractures in both young and old age-groups, 18 19 associations of increased fracture risk with specific types of exercises (eg, bicycling), or regular participation in other exercise and sports activities, have been described.20 Thus, long-term benefits (eg, increased bone mineral density) may be accrued at the expense of increased exposure to fracture risk when engaging in physical activity. Although the long-term impact of reduced population mobility on fracture incidence remains unclear, vitamin D deficiency caused by prolonged time indoors (ie, without sunlight exposure) is an established risk factor for future fractures.21

The strengths of our study include its inclusion of data from all public hospitals in Hong Kong, which allowed extensive analysis of rare events such as fractures. Our database has a high (>96%) positive predictive value for fractures,22 presumably because data entry is conducted by impartial registered medical practitioners. Furthermore, high internet and smartphone penetration increased the sensitivity of the population mobility analysis, such that the mobility index was geographically specific to the study population. Pedestrian and road traffic densities, which are indirectly represented by the population mobility index, could also precipitate accidents, falls, and subsequent fracture risk. Additionally, potential confounding based on health-seeking behaviour was partially mitigated by the inclusion of ‘control’ groups. Fortunately, all hospitals involved in the study maintained full emergency service during the early days of the COVID-19 pandemic23; this maintenance of emergency service minimised potential confounding by hospitals that were unable to provide service to patients with fractures.

Limitations of the study involved deficiencies in the population mobility index. For example, travel between familiar places and travel where navigation guidance is unnecessary, as well as the usage of alternative electronic service providers, were not considered. Therefore, the population mobility index served as a more specific (rather than sensitive) tool for assessment of population mobility. Global positioning system (GPS)–based mobility tracking would theoretically allow more extensive data collection, thus providing greater detection sensitivity; however, such mobility tracking would cause substantial privacy issues, resulting in legal and ethical challenges.

Notably, older adults are less adept in smartphone usage (62.2% of residents aged ≥65 years reported internet usage in 202011), and the digital population mobility index does not adequately illustrate this division in the population. Furthermore, fractures in older adults are largely caused by osteoporosis, whereas high-energy injury mechanisms are observed in younger individuals.24 Therefore, social distancing may have a negligible effect on the incidences of osteoporotic fractures sustained indoors. We caution against using population mobility data as the sole source of estimates for health service planning because that approach could underestimate fragility fracture service demand.

Additionally, the use of fracture incidence data from a public healthcare database only included approximately 90% of the population health demand. During the early days of the COVID-19 pandemic, instances of diversion to the private sector, attendances in private clinics, and visits to alternative practitioners were not coded; the lack of these data may have led to underestimation of total fracture incidence. Finally, we caution against generalising these findings to regions with less internet and smartphone penetration.

During the early days of the COVID-19 pandemic, fracture incidence and fracture-related mortality considerably decreased with the implementation of government social distancing measures that targeted population mobility. This unique opportunity enabled the identification of collateral associations and revealed that population mobility could be used (along with many other factors) to estimate clinical service demand.

Concept or design: JSH Wong, DKH Yee.
Acquisition of data: JSH Wong.
Analysis or interpretation of data: JSH Wong, ALH Lee, DKH Yee, CX Fang.
Drafting of the manuscript: JSH Wong, ALH Lee, CX Fang.
Critical revision of the manuscript for important intellectual content: CX Fang, DKH Yee, FKL Leung, KMC Cheung.

All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

All authors have disclosed no conflicts of interest.

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Ethics approval was granted by the Institutional Review Board of The University of Hong Kong/ Hospital Authority Hong Kong West Cluster (HKU/HA HKW IRB Ref No.: UW 20-275), and investigations were carried out in accordance with the Declaration of Helsinki. The requirement for patient informed consent was waived by the Board because the study used anonymised data and the risk of identification was low.

1. Leung K, Wu JT, Liu D, Leung GM. First-wave COVID-19 transmissibility and severity in China outside Hubei after control measures, and second-wave scenario planning: a modelling impact assessment. Lancet 2020;395:1382-93. Crossref

2. Cousins S. New Zealand eliminates COVID-19. Lancet 2020;395:1474. Crossref

3. De Filippo O, D’Ascenzo F, Angelini F, et al. Reduced rate of hospital admissions for ACS during Covid-19 outbreak in Northern Italy. N Engl J Med 2020;383:88-9. Crossref

4. Krug EG, Sharma GK, Lozano R. The global burden of injuries. Am J Public Health 2000;90:523-6. Crossref

5. Census and Statistics Department, Hong Kong SAR Government. Table 31: Gross Domestic Product (GDP), implicit price deflator of GDP and per capita GDP. 2020. Available from: https://www.censtatd.gov.hk/en/web_table.html?id=31. Accessed 26 Apr 2020.

6. Food and Health Bureau, Hong Kong SAR Government. Report of the Strategic Review on Healthcare Manpower Planning and Professional Development. 2017. Available from: https://www.fhb.gov.hk/en/press_and_publications/otherinfo/180500_sr/srreport.html. Accessed 20 May 2020.

7. Leung GM, Cowling BJ, Wu JT. From a sprint to a marathon in Hong Kong. N Engl J Med 2020;382:e45. Crossref

8. Yee DK, Fang C, Lau TW, Pun T, Wong TM, Leung F. Seasonal variation in hip fracture mortality. Geriatr Orthop Surg Rehabil 2017;8:49-53. Crossref

9. Apple Inc. COVID-19–mobility trends reports. Available from: https://www.apple.com/covid19/mobility. Accessed 26 Apr 2020.

10. Statcounter GlobalStats. Mobile & tablet vendor market share Hong Kong. Jan – Mar 2020. Available from: https://gs.statcounter.com/vendor-market-share/mobile-tablet/hong-kong/#monthly-202001-202003. Accessed 26 Apr 2020.

11. Census and Statistics Department, Hong Kong SAR Government. Thematic Household Survey Report No. 69: Personal Computer and Internet Penetration. 2020. Available from: https://www.ogcio.gov.hk/en/about_us/facts/doc/householdreport2020_69.pdf. Accessed 5 May 2020.

12. Census and Statistics Department, Hong Kong SAR Government. Population– Overview. 2020. Available from: https://www.censtatd.gov.hk/hkstat/sub/so20.jsp. Accessed 12 Apr 2020.

13. Centre for Health Protection, Department of Health, Hong Kong SAR Government. Latest situation of novel coronavirus infection in Hong Kong. Available from: https://chp-dashboard.geodata.gov.hk/covid-19/en.html. Accessed 12 Apr 2022.

14. Borgström F, Karlsson L, Ortsäter G, et al. Fragility fractures in Europe: burden, management and opportunities. Arch Osteoporos 2020;15:59. Crossref

15. Leung F, Lau TW, Kwan K, Chow SP, Kung AW. Does timing of surgery matter in fragility hip fractures? Osteoporos Int 2010;21 Suppl 4:S529-34. Crossref

16. British Orthopaedic Association. COVID BOAST-Management of patients with urgent orthopaedic conditions and trauma during the coronavirus pandemic. Available from: https://www.boa.ac.uk/resources/covid-19-boasts-combined1.html. Accessed 13 Feb 2023.

17. Tsuda T, Hashimoto Y, Okamoto Y, Ando W, Ebina K. Meta-analysis for the efficacy of bisphosphonates on hip fracture prevention. J Bone Miner Metab 2020;38:678-86. Crossref

18. Fritz J, Cöster ME, Nilsson JA, Rosengren BE, Dencker M, Karlsson MK. The associations of physical activity with fracture risk—a 7-year prospective controlled intervention study in 3534 children. Osteoporos Int 2016;27:915-22. Crossref

19. Morseth B, Ahmed LA, Bjørnerem Å, et al. Leisure time physical activity and risk of non-vertebral fracture in men and women aged 55 years and older: the Tromsø study. Eur J Epidemiol 2012;27:463-71. Crossref

20. Appleby PN, Allen NE, Roddam AW, Key TJ. Physical activity and fracture risk: a prospective study of 1898 incident fractures among 34,696 British men and women. J Bone Miner Metab 2008;26:191-8. Crossref

21. Nilson F, Moniruzzaman S, Andersson R. A comparison of hip fracture incidence rates among elderly in Sweden by latitude and sunlight exposure. Scand J Public Health 2014;42:201-6. Crossref

22. Sing CW, Woo YC, Lee AC, et al. Validity of major osteoporotic fracture diagnosis codes in the Clinical Data Analysis and Reporting System in Hong Kong. Pharmacoepidemiol Drug Saf 2017;26:973-6. Crossref

23. Hospital Authority, Hong Kong SAR Government. HA adjusts service provision to focus on combatting epidemic. 2020. Press Release. Available from: https://www.info.gov.hk/gia/general/202002/10/P2020021000711.htm. Accessed 20 May 2020.

24. Bergh C, Wennergren D, Möller M, Brisby H. Fracture incidence in adults in relation to age and gender: a study of 27,169 fractures in the Swedish Fracture Register in a well-defined catchment area. PLoS One 2020;15:e0244291. Crossref

Village clinics, the first tier of rural health systems in China, are responsible for preventing and treating common diseases among rural residents.1 2 However, the quality of healthcare provided by village clinicians may be unsatisfactory in rural China.3 4 Village clinicians generally have a low level of education and limited medical qualifications.4 There is some evidence that, among village clinicians, the first records of formal schooling are primarily vocational school degrees; most (84.3%) of these clinicians only have the basic medical certification necessary to practise medicine in rural areas.5 Moreover, despite limited empirical evaluation, available data indicate that rural primary clinicians have low diagnostic quality and provide poor management of chronic diseases.6 A 2012 study in Shaanxi Province revealed that 41% of diagnoses were incorrect; treatments were considered correct or partially correct in 53% of clinician-patient interactions.5 A systematic review of 24 studies between 2000 and 2012 showed the rate of antibiotic use in rural clinics was much higher than the rate recommended by the World Health Organization.7 8

The Chinese Government has recognised the need to strengthen primary healthcare in rural areas. To improve health among rural residents, the government has recently issued multiple policies that are intended to improve service capacity within primary medical systems.9 10 For example, to improve clinical knowledge among village clinicians, several government departments jointly implemented a plan in 2013, which focused on the provision of continuing education for clinicians.11 In 2019, the Basic Medical and Health Promotion Law of the People’s Republic of China emphasised the need to support the development of primary medical institutions and implement various policies that would improve primary medical service capabilities.12

Although improvements in internal work motivation among village clinicians may help to enhance their medical performance, few empirical studies have examined the relationship between these two characteristics among village clinicians in rural China. Theory-focused researches indicate that internal work motivation is important for improvements to clinician performance.13 14 Other theory-based researches in China have suggested that clinicians with higher internal motivation are more likely to deliver higher-quality work.15 16 Quantitative analyses of clinician behaviour, primarily conducted in other countries, have also revealed positive effects of internal work motivation on healthcare quality and work performance of clinicians.17 18 19 To our knowledge, empirical studies of work motivation in China have primarily focused on individuals in business careers and similar occupations; few have considered groups of clinicians.20 21 22 Thus, there have been few empirical studies involving village clinicians in rural China.

This study explored the relationship between internal work motivation and healthcare quality among village clinicians in rural China. First, using a standardised patient method and questionnaire interviews, we evaluated healthcare quality and internal work motivation among village clinicians. Second, we examined the relationships between internal work motivation and healthcare quality among village clinicians. Third, we conducted heterogeneity analysis with a focus on clinician workload and financial incentives.

Our study sampling was conducted in the rural areas of three prefectures, each located in one of the following three provinces: Sichuan, Shaanxi, and Anhui. Representative samples were selected using a multi-level random method. First, 21 sample counties were randomly selected from the sample prefectures. Next, 10 townships from each sampled county were randomly chosen as sample townships; 209 sample townships were selected because one sample county contained only nine townships. Then, one village was randomly selected from each township. Finally, all village clinics in the sample village were included; one standardised patient interaction was completed in the sample village.

We conducted two sets of surveys to collect data regarding basic characteristics, internal work motivation, and healthcare quality among village clinicians in 2015. In the first set of surveys, we primarily gathered information regarding the characteristics of village clinics and village clinicians. Specifically, we used a facility structured questionnaire to enquire about the value of each sample clinic’s medical instruments and institutional net income in 2014 (both in Renminbi [RMB]), and length of daily lunch break (hours). We recorded the following characteristics of sample village clinicians: age, gender, level of education and clinical qualifications, duration of service, monthly salary (in RMB), number of training days in 2014, clinician workload (mean number of patients per month), mean duration of consultation per patient (minutes), and any financial incentives. Additionally, we asked village clinicians to respond to questions regarding internal work motivation.

In the second set of surveys, we used a standardised patient method to evaluate the quality of healthcare provided by sample village clinicians. This method avoids problems such as the Hawthorne effect and recall bias, accurately assesses healthcare quality among clinicians, and is widely used in other countries.23 24 We recruited 63 individuals (ie, standardised patients; 21 in each province) to present three predetermined disease cases of diarrhoea, tuberculosis, and unstable angina in a standardised manner. Generally, we randomly allocated one standardised patient to each sample clinic to report a case that had been randomly selected prior to allocation.

We evaluated the quality of healthcare provided by village clinicians using three indicators: process quality, diagnostic accuracy, and treatment accuracy. We assigned a process quality value of 1 to clinicians who completed more than the mean percentage of suggested items, indicating a high-quality enquiry process. Otherwise, the process quality value was 0. Regarding diagnostic and treatment accuracies, we assigned a value of 0 to an ‘incorrect’ result, based on predetermined criteria. Otherwise, ‘correct’ or ‘partly correct’ results were assigned an accuracy value of 1. The treatment was also considered correct if the clinician referred the patient to a higher-level hospital.

According to Amabile and Mueller,25 an individual’s work motivation is defined as internal work motivation if it originates from love and interest. The internal motivation instrument in our study included four items, such as ‘because I like what I do for a living'. The responses of four items were rated on a 7-point Likert-type scale, ranging from 1 = strongly disagree to 7 = strongly agree. In this study, we assigned a value of 0 to responses indicating disagreement or neutrality (with original score of 1-4) and a value of 1 to responses indicating agreement (with original score of 5-7). The total score of the four items on our instrument represented a clinician's level of internal work motivation. The total score ranges from 0 to 4; a higher score indicated a higher level of motivation.

The Cronbach’s α value of the internal work motivation questionnaire was 0.826, which indicated that the scale had good internal consistency. The Kaiser–Meyer–Olkin value of the questionnaire was 0.705, indicating that the scale had good structural validity. These results confirmed that the questionnaire was an acceptable measurement tool.

STATA15.0 software (Stata Corporation; College Station [TX], United States) was used to perform descriptive and regression analyses of the collected data. Logistic regression models with a significance threshold of P<0.1 were used to analyse relationships between internal work motivation and healthcare quality.26 27 28 Two items, clinician workload × internal motivation interaction and financial incentive × internal motivation interaction, were added to the model for analyses of heterogeneity. All regression analyses were adjusted for fixed effects of disease cases, standardised patients, and the coder.

In total, 225 village clinicians from 225 village clinics were included in this study. Table 1 describes the basic characteristics of sample village clinicians. The mean age of the clinicians was 49.20 years, and 196 clinicians (87.11%) were men. Among the 225 clinicians, 25 (11.11%) had attended college or above, whereas seven (3.11%) had a practising clinician qualification. Each clinician examined a mean of 171 patients per month. Mean salaries for village clinicians were particularly low (slightly >1900 RMB per month), and 103 clinicians (45.78%) had received financial incentives.

Table 1 also describes the characteristics of sample village clinics. The mean value of medical equipment was 920 RMB, and the mean institutional net income in 2014 was 25 500 RMB. However, only 86 clinics (38.22%) had a medical equipment value above the mean. This result indicates that the value of medical equipment considerably varied among sample clinics, and the value of medical equipment in most clinics was inadequate. Notably, clinics had a mean lunch break length of <1 hour.

The unannounced standardised patients completed 225 disease cases (57, 87, and 81 cases of diarrhoea, angina, and tuberculosis, respectively). Table 2 shows the healthcare quality among sample village clinicians determined via three disease cases. On average, the clinicians completed 17% of the recommended consultation and examination items. Furthermore, 129 clinicians (57.33%) completed fewer than the mean number of recommended consultation and examination items. Among all types of cases, 73 clinicians (32.44%) provided a completely or partially correct diagnosis. Furthermore, 94 clinicians (41.78%) provided correct or partly correct treatments across all types of cases. Although the results of these three indicators varied among diseases, the percentages of clinicians with number of recommended consultation and examination items above the mean, number of correct diagnoses, and number of treatments for each disease were generally low.

Table 3 shows the levels of internal work motivation among sample village clinicians. Overall, 213 clinicians (94.67%) believed that ‘I like what I do for a living’ or ‘I enjoy my job’ motivated their work in clinics. Furthermore, 187 (83.11%) and 206 (91.56%) clinicians indicated that their respective main work motivations were ‘because my job is interesting’ and ‘because my job is fun’. Integration of the scores for the four items revealed that the mean overall score for internal work motivation was 3.64 ± 0.85 (range, 0-4).

Table 4 presents the results of logistic regression analysis of the relationship between internal work motivation and healthcare quality among village clinicians. Internal work motivation had a positive effect on clinical performance among sample clinicians. Specifically, for each one-unit increase in internal work motivation, village clinicians were 42.17% (P<0.1) and 45.61% (P<0.1) more likely to provide a correct or partially correct diagnosis and treatment, respectively.

Table 5 shows the results of heterogeneity analysis from the perspective of clinician workload and financial incentives. The clinician workload × internal motivation interaction was significantly negatively correlated with diagnostic accuracy, whereas the financial incentive × internal motivation interaction was significantly positively correlated with treatment accuracy (P<0.1). These results indicate that a heavier workload could hinder the positive effect of internal motivation on diagnostic accuracy among village clinicians. Furthermore, among village clinicians who received financial incentives, the positive effect of their internal work motivation on their treatments was stronger than the corresponding effect among village clinicians who did not receive financial incentives.

This study evaluated the healthcare quality among village clinicians in rural China and its relationship with internal work motivation among these clinicians, through an analysis of 225 rural village clinicians from three provinces in 2015. There were three main findings. First, healthcare quality among village clinicians needed to be improved. Second, village clinicians with stronger internal work motivation were more likely to offer appropriate treatment. Third, village clinicians with a lighter workload (fewer patients per month) or financial incentives exhibited a stronger positive correlation between internal motivation and healthcare quality.

Generally, interactions between unannounced standardised patients and sample village clinicians showed that poor healthcare quality was provided by village clinics in rural China. On average, village clinicians completed only 17% of the recommended consultation and examination items. The rates of diagnostic accuracy and treatment accuracy (including correct or partly correct treatment) were 32.44% and 41.78%, respectively. Our findings of poor healthcare quality are comparable with the results of other studies performed at primary health centres in rural China. For example, a study based on the patient’s perspective, conducted in Guangdong Province, highlighted the difficulty in maintaining adequate coordination among primary medical services.29 A survey using a standardised patient method revealed that healthcare quality was worse in rural China than in primary care settings in Nairobi, Kenya.30 A systematic analysis of rural township health centres in Shandong Province also indicated a need for improved healthcare quality among primary care clinicians.16

We found that internal work motivation was generally high among village clinicians. The mean internal work motivation score was 3.64 ± 0.85, indicating that most village clinicians liked their jobs and were interested in their careers. Consistent with our findings, previous studies in other countries showed that most medical workers had high levels of internal work motivation.31 32 33 Although few empirical studies have evaluated internal work motivation among village clinicians, there is some evidence that rural primary care clinicians in China experience meaning and pleasure from engaging in medical work.13 Additionally, similar to results in other countries, we found that among the intrinsic factors, most village clinicians believed that a love for their career motivated them to work.33 34 35

Consistent with data from studies in other countries,17 19 our empirical analysis demonstrated significant positive correlations between internal work motivation and healthcare quality among village clinicians in rural China. According to affect heuristic theory, this relationship presumably arises because individuals rely on emotions to make behavioural decisions, and a positive attitude will lead to higher-quality behaviours.36 Empirical results from other countries support this assumption. A study in the United States demonstrated the importance of internal work motivation in medical behaviour decisions; clinicians with higher internal motivation were more willing to maintain higher quality in their work.17 The findings of studies in developing countries, such as Ghana and Indonesia, also indicated that work motivation can significantly improve the quality of medical services provided by clinicians.19 37 Thus, efforts to stimulate internal work motivation among village clinicians may help to improve their healthcare quality.

The results of heterogeneity analysis showed that the positive effect of internal work motivation on healthcare quality varied according to clinician workload and financial incentives. Specifically, internal work motivation had a stronger positive effect on clinical performance among village clinicians who had a lighter workload (fewer patients per month). This is presumably because clinicians with a heavier workload (more patients) are more likely to experience burnout38 and a decreased sense of autonomy,39 40 which could reduce internal motivation and ultimately lead to a decline in work performance.14 39 41 Additionally, compared with village clinicians who did not receive financial incentives, clinicians who received financial incentives experienced a stronger positive effect on healthcare quality because of their internal work motivation. Studies of clinicians, combined with the results of theoretical analyses in other fields (ie, motivational synergy theory and self-decision theory), suggest that the provision of financial incentives encourages a belief of greater competence among clinicians; this belief, in conjunction with internal work motivation, enables clinicians to maintain high quality in their work.13 14 41 42 Previous empirical studies have also demonstrated that performance-related financial incentives can improve internal work motivation among employees, leading to improvements in performance.43

The results of these heterogeneity analyses support efforts to enhance the positive effect of internal work motivation on healthcare quality by providing appropriate incentives for clinicians. Consistent with this perspective, the Chinese Government has been implementing incentive programmes during the past decade to improve healthcare quality among primary care clinicians in rural China.10 11 For example, the government is actively restructuring the salary and performance system, while asserting that healthcare systems at all levels should engage in combined efforts to provide additional financial incentives.44

To further promote internal work motivation among clinicians and improve their work performance, we recommend the revision of governmental incentives policies, based on existing policies. Specifically, medical institutions at all levels should establish performance accountability45; and emphasis should be placed on including physician performance in assessments to incentivise high-quality healthcare. Furthermore, medical institutions at all levels should provide additional financial incentives to clinicians based on assessments of patient experiences. These programmes could strengthen the positive effect of internal motivation on work performance among clinicians and improve their healthcare quality. Additionally, the workload of primary clinicians should be carefully managed to preserve the positive effect of their intrinsic motivation on job performance.

This study had a few limitations. First, it was a cross-sectional study, and the results represent correlations rather than causal relationships. Second, because we randomly selected samples from Sichuan, Shaanxi, and Anhui provinces, our results may not be fully representative of village clinicians and village clinics throughout rural China. Third, the reported level of internal work motivation may have been overestimated because this variable was self-reported by village clinicians.

Overall, healthcare quality was poor among village clinicians in rural China. Furthermore, there were positive correlations between internal work motivation and healthcare quality among rural village clinicians; these positive correlations were stronger among clinicians with financial incentives and lighter workload. Our findings suggest that the Chinese Government should implement policies to provide financial incentives for clinicians, with the goal of enhancing internal work motivation among village clinicians and improving their healthcare quality.

Concept or design: Q Gao, Y Shi, L Peng.
Acquisition of data: L Peng, S Song, Y Zhang.
Analysis or interpretation of data: L Peng, Q Gao, S Song.
Drafting of the manuscript: L Peng, Q Gao, Y Shi.
Critical revision of the manuscript for important intellectual content: All authors.

As an International Editorial Advisory Board member of the journal, Y Shi was not involved in the peer review process. Other authors have disclosed no conflicts of interest.

All authors thank the standardised patients and investigators for their contribution and hard work.

This work received funding from 111 Project (Grant No.: B16031), National Natural Science Foundation of China (Grant No.: 72203134) and Innovation Capability Support Program of Shaanxi, China (Grant No.: 2022KRM007).

Ethical approval was obtained from the Institutional Review Board of Sichuan University, China (Protocol No.: K2015025). The board approved the verbal consent procedure. Participants in this study were informed of the survey procedure and consented to publication.

1. Babiarz KS, Miller G, Yi H, Zhang L, Rozelle S. China’s new cooperative medical scheme improved finances of township health centers but not the number of patients served. Health Aff (Millwood) 2012;31:1065-74. Crossref

2. Shi Y, Xue H, Wang H, Sylvia S, Medina A, Rozelle S. Measuring the quality of doctors’ health care in rural China: an empirical research using standardized patients [in Chinese]. Stud Labour Econ 2016;4:48-71.

3. Guo W, Sylvia S, Umble K, Chen Y, Zhang X, Yi H. The competence of village clinicians in the diagnosis and treatment of heart disease in rural China: a nationally representative assessment. Lancet Reg Health West Pac 2020;2:100026. Crossref

4. Tan J, Yao Y, Wang Q. Comparative analysis of the development and service utilisation of health resources in Guangxi township hospitals and the whole country [in Chinese]. Soft Sci Health 2019;33:46-50.

5. Sylvia S, Shi Y, Xue H, et al. Survey using incognito standardized patients shows poor quality care in China’s rural clinics. Health Policy Plan 2015;30:322-33. Crossref

6. Li X, Lu J, Hu S, et al. The primary health-care system in China. Lancet 2017;390:2584-94. Crossref

7. Yin X, Song F, Gong Y, et al. A systematic review of antibiotic utilization in China. J Antimicrob Chemother 2013;68:2445-52. Crossref

8. Virtual Health Library. Using indicators to measure country pharmaceutical situations: fact book on WHO level I and level II monitoring indicators. 2006. Available from: https://pesquisa.bvsalud.org/portal/resource/pt/mis-19254. Accessed 30 Sep 2021.

9. Communist Party of China Central Committee, State Council, People’s Republic of China. “Healthy China 2030” Programme [in Chinese]. 2016. Available from: http://www.gov.cn/zhengce/2016-10/25/content_5124174.htm. Accessed 30 Sep 2021.

10. State Council, People’s Republic of China. “Thirteenth Five-Year” Sanitation and Health Plan [in Chinese]. 2017. Available from: http://www.gov.cn/zhengce/content/2017-01/10/content_5158488.htm. Accessed 30 Sep 2021.

11. National Health and Family Planning Commission, National Development and Reform Commission, Ministry of Education of the People’s Republic of China, Ministry of Finance of the People’s Republic of China, National Administration of Traditional Chinese Medicine. National Rural Doctor Education Plan (2011-2020) [in Chinese]. 2013. Available from: http://www.gov.cn/gzdt/2013-10/30/content_2518099.htm. Accessed 20 Jan 2023.

12. State Council, People’s Republic of China. The 15th Meeting of the Standing Committee of the 13th National People’s Congress. Basic Medical Hygiene and Health Promotion Law of the People’s Republic of China [in Chinese]. 2019. Available from: http://www.gov.cn/xinwen/2019-12/29/content_5464861.htm. Accessed 5 Oct 2021.

13. Yuan B, Meng Q. Behaviour determinants of rural health workers: based on work motivation theory analysis [in Chinese]. Chin Health Econ 2012;31:50-2.

14. Kao AC. Driven to care: aligning external motivators with intrinsic motivation. Health Serv Res 2015;50 Suppl 2:2216-22. Crossref

15. Yuan B, Meng Q, Hou Z, Sun X, Song K. Analysis on incentive mechanism and motivation of rural health providers [in Chinese]. Chin J Health Policy 2010;3:3-9.

16. Yuan S, Meng Q, Sun X. Studying on the quality of medical services in township health centres in view of structural quality [in Chinese]. Chin Health Serv Manage 2012;29:841-4.

17. Green EP. Payment systems in the healthcare industry: an experimental study of physician incentives. J Econ Behav Organ 2014;106:367-78. Crossref

18. Mangkunegara AP, Agustine R. Effect of training, motivation and work environment on physicians’ performance. Acad J Interdiscip Stud 2016;5:173-88. Crossref

19. Al Aluf W, Sudarsih S, Musemedi DP, Supriyadi S. Assessing the impact of motivation, job satisfaction, and work environment on the employee performance in healthcare services. Int J Sci Technol Res 2017;6:337-41.

20. Hou X, Lu F. Effects of work values of millennial employees, intrinsic motivation on job performance: the moderating effect of organizational culture. Manage Rev 2018;30:157-68.

21. Li W, Mei J. Intrinsic motivation and employee performance: based on the effects of work engagement as intermediary [in Chinese]. Manage Rev 2013;25:160-7.

22. Liao J, Jing Z, Liu W, Wang X. Promotion opportunity stagnation and job performance: the mediating—role of intrinsic motivation and perceived insider status. Ind Eng Manage 2015;20:15-21.

23. Das J, Holla A, Das V, Mohanan M, Tabak D, Chan B. In urban and rural India, a standardized patient study showed low levels of provider training and huge quality gaps. Health Aff (Millwood) 2012;31:2774-84. Crossref

24. Rethans JJ, Gorter S, Bokken L, Morrison L. Unannounced standardised patients in real practice: a systematic literature review. Med Educ 2007;41:537-49. Crossref

25. Amabile TM, Mueller J. Studying creativity, its processes, and its antecedents: an exploration of the componential theory of creativity. In: Zhou J, Shalley CE, editors. Handbook of Organizational Creativity. New York: Lawrence Erlbaum Associates; 2008: 33-64.

26. Lee KI, Koval JJ. Determination of the best significance level in forward step-wise logistic regression. Commun Stat Simul Comput 1997;2:559-75. Crossref

27. Maneejuk P, Yamaka W. Significance test for linear regression: how to test without P-values? J Appl Stat 2020;31:827-45. Crossref

28. Feise R. Do multiple outcome measures require p-value adjustment? BMC Med Res Methodol 2002;17:1471-2288-2-8. Crossref

29. Feng S. Analysis of the quality of primary care and the influencing factors in rural Guangdong Province—based on demander perspective [in Chinese]. Chin Health Serv Manage 2016;33:824-7.

30. Daniels B, Dolinger A, Bedoya G, et al. Use of standardised patients to assess quality of healthcare in Nairobi, Kenya: a pilot, cross-sectional study with international comparisons. BMJ Glob Health 2017;2:e000333.

31. Settineri S, Merlo EM, Frisone F, et al. The experience of health and suffering in the medical profession. Mediterr J Clin Psychol 2018;6:1-14.

32. Toode K, Routasalo P, Helminen M, Suominen T. Hospital nurses’ individual priorities, internal psychological states and work motivation. Int Nurs Rev 2014;61:361-70. Crossref

33. Tsounis A, Sarafis P, Bamidis PD. Motivation among physicians in Greek public health-care sector. Br J Med Med Res 2014;4:1094-105. Crossref

34. Ferraro T, dos Santos NR, Moreira JM, Pais L. Decent work, work motivation, work engagement and burnout in physicians. Int J Appl Posit Psychol 2020;5:13-35. Crossref

35. Lochner L, Wieser H, Mischo-Kelling M. A qualitative study of the intrinsic motivation of physicians and other health professionals to teach. Int J Med Educ 2012;3:209-15. Crossref

36. Slovic P, Finucane ML, Peters E, MacGregor DG. The affect heuristic. Eur J Oper Res 2007;177:1333-52. Crossref

37. Alhassan RK, Spieker N, van Ostenberg P, Ogink A, Nketiah-Amponsah E, de Wit TF. Association between health worker motivation and healthcare quality efforts in Ghana. Hum Resour Health 2013;11:37. Crossref

38. Ward ZD, Morgan ZJ, Peterson LE. Family physician burnout does not differ with rurality. J Rural Health 2021;37:755-61. Crossref

39. Wang Y, Hu XJ, Wang HH, et al. Follow-up care delivery in community-based hypertension and type 2 diabetes management: a multi-centre, survey study among rural primary care physicians in China. BMC Fam Pract 2021;22:224. Crossref

40. Shirom A, Nirel N, Vinokur AD. Work hours and caseload as predictors of physician burnout: the mediating effects by perceived workload and by autonomy. Appl Psychol 2010;59:539-65. Crossref

41. Deci EL, Ryan RM. Intrinsic Motivation and Self-Determination in Human Behavior. New York: Springer; 1985. Crossref

42. Amabile TM. Motivational synergy: toward new conceptualizations of intrinsic and extrinsic motivation in the workplace. Hum Resour Manage Rev 1993;3:185-201. Crossref

43. Eisenberger R, Aselage J. Incremental effects of reward on experienced performance pressure: positive outcomes for intrinsic interest and creativity. J Organiz Behav 2009;30:95-117. Crossref

44. Qin J, Li S, Lin C. Reform progress and development strategy of incentive mechanism for training and use of general practitioners in China [in Chinese]. Chin Gen Pract 2020;23:2351-8.

45. Li X, Krumholz HM, Yip W, et al. Quality of primary health care in China: challenges and recommendations. Lancet 2020;395:1802-12. Crossref

Maternal depression is a common mental health problem during the prenatal and postpartum periods. The World Health Organization estimates that approximately 10% of pregnant women and 13% of postpartum women worldwide have mental health problems, mainly depression.1 In China, the prevalence of maternal depression ranges from 8.2% to 28.5%.2 3 4 5 6 7 Women in urban areas have access to specialised maternity care services and mental health services that can help manage these mental health problems and difficulties.8 However, these commercialised services are usually expensive and distant from poor rural areas of China. Therefore, it is particularly important for pregnant and postpartum women in poor rural areas to rely on family and social relationships for reasonable care and support.

There is evidence that the level of perceived social support, particularly family support, is associated with a woman’s mental health status during pregnancy.9 10 China’s rapid societal and economic development have resulted in substantial changes to family structure in both urban and rural areas. For example, modern couples are more likely to live with only their children, rather than with family members from multiple generations.11 When grandparents are absent from a family’s daily life, the role of the husband becomes more important because he must be more engaged in housework12 and provide greater support.

The changes in primary family caregiver identity during prenatal and postpartum periods reflect this transformation of family structure.13 The results of multiple studies in developed countries and the urban areas of China have suggested that husbands are able to care for their wives and children during pregnancy and after delivery; moreover, a husband’s companionship has a positive impact on the mental health status of his pregnant wife.2 10 14 However, in poor rural areas, no consensus has been reached concerning whether a husband can provide effective family support for his pregnant wife.15 For example, husbands usually have lower awareness of maternity care because of limited education and limited housework experience. However, in a traditional Chinese family with patrilocal features, the husband is the main worker and is responsible for the economic well-being of the family,16 whereas the wife stays at home and cares for the family. This stereotype of traditional household arrangement prevents some men from providing maternal care, regardless of their presence at home. Accordingly, grandmother, the mother of the baby’s mother, becomes a possible caregiver for the mother and baby,17 although this may lead to mother-in-law conflict.18

Here, using data from a large-scale survey of pregnant and postpartum women in poor rural areas, we analysed the status of maternal mental health in poor rural areas, with family support as an intermediate variable, to understand the correlation between primary family caregiver identity and maternal depression risk.

The data analysed in this study were collected during a survey of maternal and neonatal health and nutrition statuses among residents of poor rural villages in the Qinba Mountains area; the survey was conducted by Shaanxi Normal University from March 2019 to April 2019. The Qinba Mountains area spans six provinces including Gansu, Sichuan, Shaanxi, Chongqing, Henan, and Hubei. Its primary portion is situated in Shaanxi’s southern region. In 2019, the per capita annual disposable income in the Qinba Mountains area was RMB 11 443, similar to that of rural residents in poverty-stricken counties (RMB 11 567).19 In 2018, the mean poverty rate in this area was 3.6%; for comparison, the national mean was 1.7% and the rate in poverty-stricken counties was 4.5%.20 This study included women aged ≥18 years who were either pregnant (≥4 weeks of gestation) or in the postpartum period (0-6 months after delivery).

The following multilevel cluster-based random sampling method was used in this study. First, 13 national-level poor counties in two prefectures in the Qinba Mountains area were selected. Then, a list of villages was obtained for each county, and the total numbers of pregnant women and households with babies aged ≤6 months in each village were counted with assistance from local government officials. Considering the financial limitations and overall feasibility of the study, villages with a small sample size (<3) or large sample size (>15) were excluded. Finally, we used Stata 15.0 (Stata Corp, College Station [TX], United States) to analyse the data. The sample size was estimated to achieve, for an average incidence of independent variables of 0.15 in consideration of our pilot study, a sampling standard error (SE) of 0.03 with a 95% confidence interval (CI). The final 131 villages were randomly selected as sample villages, and all households in the sample villages that met the above criteria were considered eligible for the study.

The data used in this study were collected through face-to-face interviews. To ensure accuracy and consistency during data collection, enumerators were selected from a group of interested university students in Xi’an. The enumerators underwent extensive training, then completed a pilot study with 20 participants prior to formal data collection. Each eligible participant received a consent form with information regarding programme objectives, procedures, potential risks, and benefits, as well as an explanation of privacy protection. Participants provided oral consent for inclusion in the study before engaging in a face-to-face interview with a single enumerator. Each interview only involved the participant, and interruptions from other family members were avoided.

A questionnaire was used to collect basic participant information, including their age, education level, and self-rated health status, along with whether the baby had been born and whether it was the firstborn child. The women were also asked whether they had access to any support groups where mothers could seek help and exchange information concerning parenting experiences. Furthermore, they were asked nine yes/no questions regarding family assets (eg, possession of a computer, an air conditioner, and a car). The above questions were also included in our questionnaire to better understand maternal social interactions and household assets in order to control for them in the regression analysis and thus produce more accurate regression results. Each participant’s decision-making power was measured using a scale of eight items compiled by Peterman et al.21 A higher score on the decision-making power scale was presumed to indicate greater autonomy concerning childcare and the management of other family issues.

A questionnaire was used to collect information about all family members living in the participant’s home for >3 months, who were more likely to be the primary caregivers and to have an impact on maternity. Each participant was asked to identify the family member who served as the primary family caregiver, providing the most care for the participant and her baby during the prenatal and postpartum periods. Considering the sample size and sample distribution, three primary family caregiver categories were used in this study: the husband, the baby’s grandmother (the mother of the baby's mother or the baby's father), and other family members or no caregivers.

The Edinburgh Postnatal Depression Scale (EPDS) is a 10-item scale used to identify women at risk of maternal depression.22 23 The total EPDS score ranges from 0 to 30, where a higher score indicates a greater risk of depression. Although the original cut-off value was an EPDS score of ≥13 points, we used the standard cut-off value in China (≥9.5 points24 25) as an indicator of sufficient depression risk to merit psychiatric examination and possible treatment. Previous research has demonstrated that the EPDS has satisfactory reliability and validity. Specifically, Wang et al26 reported that the EPDS had a content validity ratio of 0.93 and good internal consistency (Cronbach’s α coefficient of 0.76). The correlation coefficients between the 10 individual item scores and the total score ranged from 0.37 to 0.67, with P values <0.01.

The Perceived Social Support Scale, developed by Zimet et al27 and translated into Chinese by Jiang,28 is a 12-item self-assessment questionnaire that measures three sources of social support (ie, three subscales): family support, friends’ support, and other people’s support. Responses to questionnaire items are recorded using a seven-point Likert scale that ranges from ‘completely negative’ to ‘completely positive’ (1-7 points), indicating the respondent’s level of agreement with each item. The total score is 84 points (28 points per subscale), and a higher score indicates the receipt of greater social support. The Cronbach’s α coefficient of the scale is 0.88; the Cronbach’s α coefficients for family support, friends’ support, and other people’s support subscales are 0.81, 0.85, and 0.91, respectively.27 Because this study focused on family support, only the family support subscale was used as an intermediate variable to analyse the correlation between primary family caregiver identity and maternal depression risk.

STATA 15.1 software was used to clean the data and perform statistical analysis. Descriptive statistical analysis was performed and presented as mean ± standard deviation. F-test and t test were used to detect differences in depression scores among subgroups of women with different characteristics. Multiple linear regression was used to explore correlations between primary family caregiver identity and maternal depression risk or family support score. P values <0.05 were considered statistically significant. Additionally, we adjusted the SE at the village level and calculated coefficients with greater precision because individual values within the same village are correlated, which might result in biased SE in multiple linear regression.

In total, 715 women were interviewed, including 220 pregnant women and 495 new mothers. Twenty-two samples with missing values were excluded to ensure sample uniformity throughout the analysis procedure. Finally, analyses in this study were based on the data of 693 participants (220 pregnant women and 473 new mothers) and the questionnaire return efficiency was 96.9%, which is the percentage of survey responses that were valid.

Among the 220 pregnant women, 37 (16.8%), 66 (30.0%), and 117 (53.2%) were in the early, middle, and late stages of pregnancy, respectively (Table 1). In total, 226 of the 473 new mothers (47.8%) had babies aged 1 to 3 months, whereas 247 new mothers (52.2%) had babies aged 4 to 6 months.

The mean maternal EPDS score was 5.85 and the proportion of women at risk of depression was 18.9% (131/693). The proportion of women at risk of depression was generally stable regardless of pregnancy stage. Specifically, the proportion of women at risk of depression during early pregnancy was 16.2% (6/37); during middle and late pregnancy, the proportions of women at risk were slightly increased. The proportions of women at risk of depression were 16.8% (38/226) and 20.2% (50/247) at 1-3 months and 4-6 months after delivery, respectively. However, the maternal EPDS scores and proportions of women at risk of depression did not significantly differ according to pregnancy stage or time since delivery.

Overall, the mean participant age was 28.13 ± 4.70 years. In total, 239 women (34.5%; mean age, 25.52 ± 3.95 years) reported that the current pregnancy or ≤6-month-old baby was their firstborn child. The remaining 454 women (65.5%; mean age, 29.50 ± 4.49 years) were experienced mothers who have already had children and are familiar with caring for them. Overall, 116 women (16.7%) had an education level above junior high school. The self-rated health status was good in 89 women (12.8%), and 102 women (14.7%) were involved in a parenting support group. Table 2 summarises the participant characteristics.

As shown in Table 2, the participants were clustered into three groups according to primary caregiver identity: the husband for 151 women (21.8%), the baby’s grandmother for 452 women (65.2%), and other family members or no caregiver for 90 women (13.0%). The mean EPDS scores of women in the three groups were 6.23 ± 4.34, 5.56 ± 4.01, and 6.63 ± 4.84, respectively (P=0.039). Additionally, univariate analysis revealed statistically significant differences in depression scores according to education level, self-rated health status, and parenting support group involvement. There were no statistically significant differences in other variables.

As shown in Table 3, identification of the baby’s grandmother as the primary family caregiver was significantly negatively correlated with EPDS score (β=-0.979, 95% CI=-1.946 to -0.012, P=0.047). However, identification of the husband as the family caregiver was not significantly correlated with EPDS score (β=-0.499, 95% CI=-1.579 to 0.581, P=0.363).

As shown in Table 4, after adjustment for other variables, there was no significant correlation between identification of the husband as the primary family caregiver and the family support score (β=0.375, 95% CI=-0.704 to 1.455, P=0.493). However, identification of the baby’s grandmother as the primary family caregiver was significantly positively correlated with family support score (β=0.967, 95% CI=-0.062 to 1.996, P=0.065). Furthermore, identification of the baby’s grandmother as the primary family caregiver had the largest standardised regression coefficient among the three caregiver categories, indicating that pregnant and postpartum women felt the greatest family support when the baby’s grandmother was the primary family caregiver.

In this study, the overall proportion of women at risk of maternal depression was 18.9%, including a mean proportion of 19.5% among pregnant women and a mean proportion of 18.6% among women ≤6 months postpartum. This overall proportion of women at risk of maternal depression is much higher than the proportion in a western urban area of China (12.4%)29 and comparable with the proportions in low- and middle-income countries such as Ethiopia (19.9%)30—both previous studies also used the EPDS to identify women at risk of maternal depression. The high proportion in the present study may be related to the location (poor rural areas): compared with women in urban areas, women in poor rural areas are more likely to have a lower socio-economic status.31 The lack of knowledge regarding mental health and its services in rural areas also makes women in such areas more likely to become depressed if they do not receive timely treatment for mental health problems.32 Therefore, the mental health of rural mothers should receive greater attention from their family members and the relevant health departments.

This study also revealed a persistent risk of depression during the prenatal and postpartum periods (Table 1). Notably, the proportion did not substantially decrease by 6 months after delivery. Yue et al33 investigated the mental health of caregivers for babies aged 6 to 36 months in a rural area in western China. Their results showed that the proportion of caregivers at risk of depression was similar to the proportion in the present study. These findings suggest that maternal depression persists in the absence of external intervention. Thus, there is an urgent need for timely external mental health interventions among pregnant women and mothers of young children. The present study also showed that the maternal depression risk in poor rural areas is influenced by factors such as a woman’s education level, self-rated health status, and parenting support group involvement. These results are consistent with the findings by Zhou et al,7 Lancaster et al,10 and Lee et al.18

Our results showed that identification of the husband as the primary family caregiver was not significantly correlated with maternal depression risk in poor rural areas (Table 3). This finding was considerably different from the results of previous studies in urban areas. Xie et al34 found that insufficient or poor-quality emotional support from the husband was significantly associated with an increased risk of postpartum depression among mothers in Changsha, Hunan Province, China. In contrast, Wan et al2 found that the proportions of women at risk of maternal depression were 1.9- to 2.6-fold higher among women without support from the husband before and after delivery than among women with support from the husband, based on a study of mothers in Beijing, China. The results of these studies suggest that the husband’s involvement as the primary family caregiver can reduce the risk of maternal depression in urban areas, but this effect was not apparent in poor rural areas.

We also found that maternal depression risk was significantly lower when the baby’s grandmother was identified as the primary family caregiver (Table 3). Our results are consistent with the findings by Wan et al2 in a study of 342 pregnant women in Beijing, China: during the ‘confinement’ period, care and support from the baby’s grandmother(s) were important for relieving depression. However, Lee et al18 showed that mother-in-law conflict remains prominent in China, which may have negative emotional outcomes for grandmothers and new mothers. Although pregnant and postpartum women in poor rural areas may experience similar conflict, our findings suggest that support from the baby’s grandmother(s) remains predominantly positive.

We attempted to determine why support from the husband did not reduce maternal depression risk in poor rural areas through the analysis of an intermediary variable. Initially, we hypothesised that the positive effect of the husband acting as the primary family caregiver would be offset by the loss of income caused by the husband’s inability to seek work opportunities in other locations. However, data analysis revealed that the husband’s role as the primary caregiver had no impact on the family income and family asset index (online supplementary Table 1). Thus, we explored the effect of family support. Multiple previous studies demonstrated that family support influenced maternal depression risk14; consistent with those findings, our analysis showed that family support was significantly negatively correlated with maternal depression risk (online supplementary Table 2).

There may be two main reasons for this negative correlation. First, husbands in poor rural areas have insufficient knowledge and skills related to maternal care.16 Husbands do not have first-hand experience in childbirth and can only acquire it through education. However, compared with men in urban areas, men in poor rural areas have lower levels of education and may be less inclined to learn on their own, making it more difficult to acquire such knowledge and skills.35 In contrast, grandmothers are more experienced overall, which may enable them to provide more effective family support. For example, based on their own experience, grandmothers can help new mothers to prepare for and manage pain that sometimes occurs during breastfeeding, which can alleviate anxiety and provide a feeling of greater support.17 Second, in poor rural areas, husbands may lack sufficient time and energy to provide effective family care. Compared with families in urban areas, families in poor rural areas are more economically disadvantaged33; therefore, husbands in such families may prioritise financial stability and be unable to expend time or energy in support of maternal care, despite their physical presence in the home. In contrast, the baby’s grandmother(s) may have sufficient time and energy to provide effective maternal care (eg, by feeding the baby and changing its diapers), thus relieving the mother’s psychological stress.

The findings in this analysis of women in poor rural areas differ from the results of studies in urban areas, indicating important differences in family structure between urban and rural areas. There is evidence that a gradual transformation of the family is underway in urban areas, whereby husbands have begun to actively engage in caregiving. However, the transformation of family structure is much slower in poor rural areas,13 and husbands in those areas are not yet prepared for this new role. Because of constraints regarding their education level and skills, as well as family finances, husbands in poor rural areas continue to prioritise financial stability36; their support does not have a positive impact on the risk of maternal depression. Thus, women in poor rural areas must continue to rely on family members outside of the nuclear family, such as the baby’s grandmother(s), to assume some caregiving responsibilities.

Commercialised and specialised mental health counselling services in urban areas play important roles in improving maternal mental health.8 Xiao37 found that postnatal care through a menstrual club provided continuous physical, psychological, and emotional support that was sufficient to reduce the incidence of postpartum depression. However, such clubs are not available in poor rural areas. Therefore, it is important to promote better caregiving from family members, including husbands. For example, husbands could receive training that enables them to provide practical support, as well as guidance concerning the early identification of depressive tendencies and the development of communication skills for psychological adjustment.

This study had some limitations. First, its cross-sectional design prevented the assessment of maternal depression trends during pregnancy and after delivery, although such an assessment could have been conducted in a cohort study. Second, this study focused on primary family caregiver identity and did not explore the type or form of caregiving provided. Third, all participants were residents of rural northwest China, and thus the results may not be generalisable to other populations. These limitations should be addressed in future studies.

The prevalence of maternal depression is high in poor rural areas of Shaanxi Province. Identification of the husband as the family caregiver was not significantly correlated with maternal depression risk, whereas the involvement of the baby’s grandmother in that role was significantly negatively correlated with maternal depression risk. Based on our findings, we make the following suggestions. In rural areas, high-quality family support is necessary to ensure that pregnant women maintain good mental health. Compared with husbands, grandmothers may be better primary caregivers because they are more experienced in terms of parenting and housework. Husbands in poor rural China should receive training that enables them to provide effective maternal care.

Concept or design: N Wang, M Mu, J Yang, Y Shi, J Nie.
Acquisition of data: N Wang, M Mu, Z Liu, R Zulihumaer, W Nie.
Analysis or interpretation of data: N Wang, M Mu, J Yang, J Nie.
Drafting of the manuscript: N Wang, M Mu, J Yang, Y Shi, J Nie.
Critical revision of the manuscript for important intellectual content: All authors.

All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

As an International Editorial Advisory Board member of the journal, Y Shi was not involved in the peer review process. Other authors have disclosed no conflicts of interest.

The authors thank the study participants and the enumerators who conducted data collection.

The authors are supported by the 111 Project (Grant No. B16031), Soft Science Research Project of Xi’an Science and Technology Plan (Grant No. 2021-0059), the Fundamental Research Funds for the Central Universities (Grant No. 2021CSWY024) and the Fundamental Research Funds for the Central Universities (Grant No. 2021CSWY025) of China.

The study was approved by the Medical Ethics Committee of Shaanxi Normal University and Xi’an Jiaotong University of China (No: 2020-1240). Each eligible participant received a consent form with information regarding programme objectives, procedures, potential risks, and benefits, as well as an explanation of privacy protection. Participants provided oral consent for inclusion in the study before engaging in a face-to-face interview with a single enumerator.

1. Ceulemans M, Foulon V, Ngo E, et al. Mental health status of pregnant and breastfeeding women during the COVID- 19 pandemic—a multinational cross-sectional study. Acta Obstet Gynecol Scand 2021;100:1219-29. Crossref

2. Wan EY, Moyer CA, Harlow SD, Fan Z, Jie Y, Yang H. Postpartum depression and traditional postpartum care in China: role of Zuoyuezi. Int J Gynecol Obstet 2009;104:209-13. Crossref

3. Zeng Y, Cui Y, Li J. Prevalence and predictors of antenatal depressive symptoms among Chinese women in their third trimester: a cross-sectional survey. BMC Psychiatry 2015;15:66. Crossref

4. Zhang Y, Zou S, Cao Y, Zhang Y. Relationship between domestic violence and postnatal depression among pregnant Chinese women. Int J Gynecol Obstet 2012;116:26-30. Crossref

5. Song Y, Li W. The impact of social support and antepartum emotion on postpartum depression [in Chinese]. China J Health Psychol 2014;22:909-11.

6. Xu M, Li C, Zhang K, Sun S, Zhang H, Jin J. Comparison of prenatal depression and social support differences between Korean and Han pregnant women in Yanbian region [in Chinese]. Matern Child Health Care China 2015;30:524-7.

7. Zhou X, Liu H, Li X, Zhang S, Li F, Zhao Z. Depression and risk factors of women in their second-third trimester of pregnancy in Shaanxi province [in Chinese]. Chin Nurs Manage 2019;19:1005-11.

8. Liu H, Zhang C. Comparative study on service efficiency of China’s urban and rural health systems [in Chinese]. China Soft Sci 2011;10:102-13.

9. Leahy-Warren P, McCarthy G, Corcoran P. First-time mothers: social support, maternal parental self-efficacy and postnatal depression. J Clin Nurs 2012;21:388-97. Crossref

10. Lancaster CA, Gold KJ, Flynn HA, Yoo H, Marcus SM, Davis MM. Risk factors for depressive symptoms during pregnancy: a systematic review. Am J Obstet Gynecol 2010;202:5-14. Crossref

11. Tang C. A review of modernisation theory and its development on family [in Chinese]. Sociol Stud 2010;25:199-222,246.

12. Högberg U. The World Health Report 2005: “make every mother and child count”—including Africans. Scand J Public Health 2005;33:409-11. Crossref

13. Zhao F, Ji Y, Chen F. Conjugal or parent-child relationship? Factors influencing the main axis of Chinese family relations in transition period [in Chinese]. J Chin Women Stud 2021;4:97-112.

14. Li D, Xu X, Liu J, Wu P. The relationship between life event and pregnancy stress: the mediating effect of mental health and the moderating effect of husband support [in Chinese]. J Psychol Sci 2013;36:876-83.

15. Dennis CL, Ross L. Women’s perceptions of partner support and conflict in the development of postpartum depressive symptoms. J Adv Nurs 2006;56:588-99. Crossref

16. Yang H, Lu Y, Ren L. A study on the impact of bearing two children on urban youth’s work-family balance: the empirical analysis based on the third survey of Chinese women’s social status [in Chinese]. Popul Econ 2016;9:1-9.

17. Hu J, Wang Y. Study on related risk factors of patients with pre-and postnatal depression [in Chinese]. Chin Nurs Res 2010;24:765-7.

18. Lee DT, Yip AS, Chan SS, Lee FF, Leung TY, Chung TK. Determinants of postpartum depressive symptomatology—a prospective multivariate study among Hong Kong Chinese Women [in Chinese]. Chin Ment Health J 2005;9:54-9.

19. National Bureau of Statistics of China. Income of villagers in poor rural areas in 2019 [in Chinese]. Available from: http://www.stats.gov.cn/xxgk/sjfb/zxfb2020/202001/t20200123_1767753.html. Accessed 30 Nov 2022.

20. Sun J, Zhang J, Li C, Lu Y. Strategic judgements and development suggestions on the development of poverty-stricken areas in China [in Chinese]. Manage World 2019;35:150-9,185.

21. Peterman A, Schwab B, Roy S, Hidrobo M, Gilligan DO. Measuring women’s decision making: indicator choice and survey design experiments from cash and food transfer evaluations in Ecuador, Uganda, and Yemen. World Dev 2021;141:105387. Crossref

22. Adouard F, Glangeaud-Freudenthal NM, Golse B. Validation of the Edinburgh postnatal depression scale (EPDS) in a sample of women with high-risk pregnancies in France. Arch Womens Ment Health 2005;8:89-95. Crossref

23. Zhang H, Li L. Comparative analysis of 3 postpartum depression scales abroad [in Chinese]. Chin J Nurs 2007;2:186-8.

24. Lee DT, Yip SK, Chiu HF, et al. Detecting postnatal depression in Chinese women. Validation of the Chinese version of the Edinburgh Postnatal Depression Scale. Br J Psychiatry 1998;172:433-7. Crossref

25. Liu Y, Zhang L, Guo N, Li J, Jiang H. Research progress of Edinburgh Postnatal Depression Scale in screening of perinatal depression [in Chinese]. Chin J Mod Nurs 2021;27:5026-31.

26. Wang Y, Guo X, Lau Y, Chan KS, Yin L, Chen J. Psychometric evaluation of the Mainland Chinese version of the Edinburgh Postnatal Depression Scale. Int J Nurs Stud 2009;46:813-23. Crossref

27. Zimet GD, Dahlem NW, Zimet SG, Farley GK. The multidimensional scale of perceived social support. J Pers Assess 1988;52:30-41. Crossref

28. Yan R, Liu L, Zhang L. Analysis of the relationship between social support self-efficacy and health behaviours among college students [in Chinese]. Chin J Sch Health 2010;31:362-3.

29. Zhang X, Li X, Li Y, Quan X, Li X. Study on status and influencing factors of antenatal depression among urban women in western China [in Chinese]. Matern Child Health Care China 2019;34:5275-7.

30. Dibaba Y, Fantahun M, Hindin MJ. The association of unwanted pregnancy and social support with depressive symptoms in pregnancy: evidence from rural southwestern Ethiopia. BMC Pregnancy Childbirth 2013;13:135. Crossref

31. Wu H, Rao J. Research review on left-behind women [in Chinese]. J Chin Agric Univ Soc Sci 2009;26:18-23.

32. Hou F, Cerulli C, Wittink MN, Caine ED, Qiu P. Depression, social support and associated factors among women living in rural China: a cross-sectional study. BMC Womens Health. 2015;15:28. Crossref

33. Yue A, Gao J, Yang M, Swinnen L, Medina A, Rozelle S. Caregiver depression and early child development: a mixed-methods study from rural China. Front Psychol 2018;9:2500. Crossref

34. Xie RH, Yang J, Liao S, Xie H, Walker M, Wen SW. Prenatal family support, postnatal family support and postpartum depression. Aust N Z J Obstet Gynaecol 2010;50:340-5. Crossref

35. Wei Q, Zhang C, Hao B, Wang X. Study on parenting concepts and behaviours of caregivers with children under 3 years old in rural areas of China [in Chinese]. Matern Child Health Care China 2017;32:1759-61.

36. Hao J, Wang F, Huang J. Male labour migration, women empowerment and household protein intake: evidence from less developed rural areas in China [in Chinese]. China Rural Econ 2021;8:125-44.

37. Xiao G. A practical study of puerperal health care in a menstrual centre [in Chinese]. Chin J Mode Drug Appl 2016;10:271-2.