Large-scale survey of parental antibiotic use for paediatric upper respiratory tract infections in China: implications for stewardship programmes and national policy

BACKGROUND Inappropriate use of antibiotics for upper respiratory tract infections among Chinese children is rampant. To identify key constructs for effective interventions targeting the public, we investigated parents’ decision-making processes with respect to treatment choices and antibiotic use for paediatric URTIs. METHODS Data were collected between June 2017-April 2018 from a cluster random sample of 3,188 parents of children aged 0-13 across three Chinese provinces, representing different stages of economic development. Risk factors of parents’ treatment choices and antibiotic use for paediatric URTIs were assessed, using binary and multinomial logistic regressions, adjusting for socio-demographic characteristics.


HIGHLIGHTS
 Caregivers account for at least 40% of outpatient antibiotic use on Chinese children. Parents' perception of antibiotics as efficacious for treating URTIs and easy access to antibiotics, with or without a prescription, drives antibiotic misuse in China. Simultaneously enhancing both prescribing guidelines, doctor-patient communication skills, and patient education targeting the family as a unit is critical. Patient education interventions should prioritise urban parents with low socio-economic status in less developed regions and be disseminated via medical professionals or media.

RESULTS
Of the 3,188 parents who self-diagnosed their children with a URTI, 46.0% children were given antibiotics, with or without prescription (n=1465).Among them, 40.5% were self-medicated with antibiotics by parents and 56.1% obtained further antibiotic prescriptions at healthcare facilities.About 70% of children with URTI symptoms sought healthcare (n=2197); of them, 54.8% obtained antibiotic prescriptions and 7.7% asked for antibiotic prescriptions with a 79.4% success rate to obtain them.Those perceiving antibiotics as effective for treating common cold and fever (aOR=1.82[Acute, uncomplicated upper respiratory tract infections (URTIs)often benign, selflimiting, and untreatable by antibiotics -are diagnosed on symptomatology and treatments are mainly symptomatic rather than focusing on changes in viral titres in the airway or viral shedding. 1Considered the most common infectious disease among humans, URTIs are the most common cause of primary care visits and unnecessary use of antibiotics for children around the world, especially in China, [2][3][4] which has contributed to the rise of antimicrobial resistance (AMR), an imminent global health threat.Antibiotic-resistant pathogens, such as streptococcus pneumonia, have been reported in children across China. 5,6Antibiotic treatment changes gut microbiota and adversely impacts the development of the immune system, making it difficult for children to recover from repeated antibiotic exposure. 7,8Given the long-term consequences on human development and that children may experience URTIs seven to ten times on average annually, 1 misuse of antibiotics in children is particular harmful.
Nevertheless, 48.2% of urban parents 9 and 62% of rural parents 10 in China reported to have self-medicated children with antibiotics outside of clinical settings within the last six months.
Understanding the underlying reasons that drive the parental decision to use antibiotics for paediatric URTIs without professional guidance is important for developing strategies to reduce antibiotic misuse.To date, most public-targeted health behaviour research and interventions on antibiotic use have centred on knowledge-attitudes-practice (KAP), with the underlying assumption that individuals would make more risk-conscious choices if informed of the risks of AMRan approach has long been criticised for its overemphasis on personal responsibility. 11When faced with an acute infection in themselves or a loved one, individuals may disproportionately focus on the immediate outcome of curing the illness (i.e.perceived antibiotic efficacy), and discount long term risks such as AMR. 12 As such, parents' decision-making for treating URTIs might not be as rational or informed as a KAP approach would assume.

Study population
We used data from a survey of parents with children aged 0-13 years between June 2017 and April 2018.Three Chinese provinces, which represent different geographical areas and stages of economic development, 13 were chosen.These included Zhejiang (East, ranked 5 th out of 31 in the 2017 provincial GDP ranking of economic development), Shaanxi (Central-Northwest, ranked 12 th ), and Guangxi (Southwest, ranked 26 th ) provinces.A multistage stage random clustering sampling design was applied.The four-stage sampling units are provinces, prefecture-level cities, urban and rural areas, and local sites: primary schools (age 6-13), kindergartens (age 3-5) and community health centres (age 0-2), where most children received vaccination. 14Parents were identified and recruited through their children from all selected sites.They were asked to complete a structured questionnaire, which was tailored to the Chinese sociocultural context informed by literature review 10,15,16 and formative/qualitative interviews with stakeholders and experts.The questionnaire was comprised of four sections: 1) parental socio-demographic information, 2) healthcare-and antibiotic-related knowledge and perceptions, 3) last episode of URTI symptoms experienced by the child within the past month, and 4) treatment and parental care-seeking process and behaviours for the child's illness (i.e. the chemical or brand names of antibiotics obtained from clinics and retail pharmacies).To minimise the burden for the parents and ensure high quality of the response data, the survey was designed to take no more than 10 minutes and an IP address control was put in place to detect random responses or duplications.Parents could complete the questionnaire on a mobile device, online, or using a paper version and they were informed that participation was confidential, voluntary and could be terminated at any time.A consent form was presented in the first section of the questionnaire and was signed by the participants.To validate the questionnaire, we conducted a pilot study with 315 respondents to evaluate potential sources of response error and improve the instrument.The questionnaire was completed by 9,526 parents, with a response rate of 89%.Of those, 33.5% (n=3,188) reported that their children had experienced symptoms of a URTI within a month prior to the survey, including cold (cough, runny/stuffy nose), fever, sore throat, headache, and flu, either alone or in combination 1 .

Outcome variables
Participating parents reported whether they (1) self-treated children with antibiotics: did not use antibiotics, self-medication with antibiotics, and seeking formal care after self-medication with antibiotics at home; (2) sought care and/or requested antibiotics: did not seek care, sought care, and sought care and explicitly requested antibiotics for their children.In addition, parents also reported whether clinicians' prescribed antibiotics for their child: no antibiotic prescription, antibiotic prescriptions without being prompted, and inappropriate antibiotic prescriptions due to parental demands.

Exposure variables
Informed by the Health Belief Model 17,18 , we included the following potential risk factors in our analyses: (1) Whether parents had a medical background (yes/no), as it is relevant to parents' self-efficacy for making healthcare decisions for their children.
(2) Parents' ability to identify antibiotics, measured by number of commonly available drugs correctly identified by parents as antibiotics or non-antibiotics: low (0-1), medium (2-3), high (4 or higher); (3) Parents' perceptions: (a) perceived benefits of antibiotic use, measured by two factual statements about antibiotics' efficacy to treat the common cold or fever; and (b) perceived severity of the infection, measured by the number of self-diagnosed URTI symptoms the child experienced; Covariates: Socio-demographic characteristics were included as potential confounders for the association between each exposure and treatment decisions, including sex and age of the child, household income, parental education, urbanicity and province.

Statistical analysis
We first developed a flow diagram (Figure 1) to illustrate parental decision-making process of treatment and antibiotic use in their children for URTIs, from (non-clinical) household to (clinical) facility.We summarised the distributions of socio-demographic characteristics and factors by treatment decision/behavioural outcomes.
To examine the association between each factor and outcome, we applied logistic regressions to estimate the OR (95% CI) for (1) 'self-medication with antibiotics' (vs 'no self-medication with antibiotics') and ( 2) 'seeking healthcare' (vs 'without seeking healthcare').Factors considered include parental medical background, ability to identify antibiotics, perceived antibiotic efficacy for cold or fever, self-diagnosed severity, cues to action, and access to antibiotics.We explored the associations with subgroups of antibiotic and healthcare use, and applied multinomial logistic regressions to estimate the relative risk ratio, RRR (95% CI) for (1) 'self-medication with antibiotics without seeking healthcare' and 'self-medication with antibiotics then sought healthcare' (vs 'no self-medication with antibiotics') and ( 2) 'sought healthcare without requesting antibiotic prescriptions' and 'sought healthcare and requested prescriptions' (vs 'no seeking healthcare').For parents who sought healthcare for their children, we estimated RRR (95%CI) for "receiving prescriptions without patients' request' and 'receiving prescriptions due to patients' request' (vs 'without an antibiotic prescription').
For each outcome and risk factor, we first fitted an unadjusted model, and then adjusted for the potential confounders to establish whether the association was independent of these socio-demographic characteristics.Because different risk factors tend to co-occur, as sensitivity analyses, we mutually adjusted for all risk factors simultaneously.

Healthcare seeking and parents request for antibiotic prescription (Table 3)
Parents who perceived antibiotics as effective for the common cold and fever, who had high levels of perceived severity of infection, or presence of fever in children were more likely to seek healthcare and request antibiotic prescriptions, compared to their respective counterparts.
Parents who had a medical background, obtained health information from family, or kept antibiotics at home were less likely to seek healthcare for their children (aOR=0.65,0.81, and 0.84, respectively).Among parents who sought healthcare for their children, keeping antibiotics at home was associated with increased risk of requesting antibiotic prescriptions (aRRR=3.63[2.54-5.17]).

Antibiotic prescriptions for the treatment of URTIs (Table 4)
Children whose parents could identify most antibiotics, perceived antibiotics as efficacious for common cold or fever, perceived higher severity in their children, and kept antibiotics at home were more likely to receive antibiotic prescriptions, with a greater risk of receiving

Strengths and Limitations
This study is based on a large survey conducted in geographical areas representing various stages of economic development in China.This is the first study to comprehensively examine parental treatment decisions with respect to antibiotic use in children in both rural and urban settings across China.Though the cross-sectional study design limited us from drawing causal relationships, it helped generate causal hypotheses and offered several points for intervention.This study showed that the high childhood antibiotic consumption in China is largely driven by a combination of excessive use of formal care for URTIs, high prescription rates, and large population size.The actual antibiotic consumption in Chinese children is expected to be much more prevalent than what has been reported in this study, considering repeated infections throughout a year and non-prescription use at home. 19We found, before the parent sought formal care, 18% of children with URTIs had already received antibiotics, without prescription.

Interpretation of our findings
Evidence generated from this study will inform intervention design to reduce unnecessary antibiotic use for paediatric URTIs in China as well as other low-/middle-income countries that share similar challenges, including rising antibiotic consumption 20 and unsupervised (e.g.use of leftover antibiotics) or inappropriate use of antibiotics (e.g. for viral infections or prevention). 21,22First, our findings highlighted the continued need to tackle the non-clinical drivers of inappropriate prescribing behaviours (e.g.patients' or caregivers' profile and behaviours) which should be addressed along with other factors such as poor diagnostic capacity and financial incentives, 23,24 especially in primary care and rural settings 25 .More than half of paediatric patients with non-complicated URTI symptoms were prescribed with antibiotics while roughly 80% of those who demanded antibiotics were prescribed antibiotics, accounting for an estimate of 45% outpatient paediatric antibiotic use in the country.About 8% of Chinese parents admitted to having asked doctors for antibiotics for paediatric URTIs, which is similar to what has been reported in some European countries. 26Our study highlighted a preference for IV infusion for children that remains prevalent among Chinese parents.This phenomenon is a product of Chinese hospitals' financial incentives, as well as the expectations of consumers for rapid recovery, fueled by widespread accepting attitudes towards the use of needles in Chinese society 27 an attitude influenced by the concept of acupuncture, an ancient traditional Chinese medical treatment.Since 2012, many Chinese hospitals have made an effort to reduce outpatient infusion treatments, 28 yet these regulations have not been adopted by most lower level hospitals and have excluded paediatric patients.Furthermore, over-prescription in rural China may be due to the lack of diagnostic knowledge among providers 29 and therefore, improving their professional capacity is necessary.Further, the influence of doctor-patient encounters on antibiotic prescriptions might be more complex than verbal communication.Our data identified a surprisingly similar set of risk factors influencing antibiotic prescription outcomes for paediatrics URTIs between parents who explicitly demand antibiotics and those who did not.If Chinese doctors' prescribing behaviours for paediatrics URTIs are mainly driven by poor diagnostic capacity or financial incentives, as suggested by previous literature, 23,24 we would have expected no association between these risk factors of parents and doctors' prescription decisions.This phenomenon might be explained by possible non-verbal cues (whether true or not) that prescribers pick up from their interactions with parents who showed certain character traits or profiles during consultation that signalled to the prescribers that an antibiotic prescription was desired.This explanation is supported by a study that identified a misalignment between parents' reported expectations, their communication messages, and physicians' perceptions of parents' expectations and their reaction to those perceptions. 26,33These data pointed to an urgent need to enhance clinician training focusing on 1) clinical guidelines and appropriate prescribing for paediatric URTIs and 2) doctor-patient communication skills that aimed to help clinicians (a) neutralise the perceived expectation on/pressure from parents' demand for antibiotics and (b) inquire about possible parental self-medication with antibiotics on children before reaching the facility to avoid multiple doses.
Further, compared with the estimate regarding university students, 15 parents appeared to be more cautious, but still drove 40% of antibiotic misuse in children.Overuse of medical care for self-limiting illnesses combined with a high prescription rate and the population size of the country drove the overall high antibiotic consumption in China.In our data, approximately 70% of children with common cold symptoms in the past month sought care, which was about twice as many as those in UK (34-40%). 34The possibility of receiving an antibacterial prescription for such symptoms was around 33% in UK, 35,36 compared to 55% in our survey.As such, we estimated that an average Chinese child consumes more than three times the amount of antibiotics as their peers in UK or other European countries. 26,35,37,38The gap is even wider for Chinese children in infancy and early childhood, as they have higher usage of medical care than older children.This estimate is alarming considering it did not account for non-prescription use antibiotics in Chinese children.Our data indicates about 18.6% of children with the common cold within the month before the survey were self-medicated with non-prescription antibiotics by parents; additionally, a previous study 39 reported about 20% were given antibiotics for prophylaxis in the past year.Therefore, we estimate non-prescription use of antibiotics for paediatric URTIs among Chinese children is at least 4-6 times higher than that of some European countries. 26,37,38,40The true magnitude of this problem is underestimated because repeated use was not included in the calculation.This estimate is consistent with a survey conducted in 1995 and demonstrates that Chinese parental antibiotic misuse for their children has not improved over the past two decades. 41ntext-tailored patient/caregiver education interventions on appropriate home care for paediatric URTIs and prudent antibiotic use are needed.Content should prioritise correcting perceived antibiotic efficacy for relieving or eradicating URTI symptoms and appropriate care for self-diagnosed paediatric URTIs symptoms and fever, and be delivered by medical professionals or mass media -both were identified as effective channels for health information.
1][32] In China, we found antibiotic misuse in children was mainly associated with parents' access to antibiotics, within or outside of a clinical setting.Household antibiotic storage mainly came from leftover antibiotics from previous prescriptions (60.6%) and over-the-counter purchases (37.5%).Cephalosporines, Amoxicillins, and Azithromycins were the most commonly used antibiotics to treat paediatric URTIs, both with and without a prescription (data not shown).These antibiotics, Cephalosporines especially, are broad-spectrum antibiotics effective against a wide range of bacteria, which kill more normal microorganisms in children's body compared with narrow-spectrum antibiotics, and should only be used under professional supervision on patients who are sick on presentation.Furthermore, participants from all regions reported to have obtained antibiotics from retail pharmacies.Currently, antibiotic prescriptions are fulfilled and dispensed by packs, often more than the prescribed doses, leading to leftover antibiotics for unsupervised self-medication at home later on.Therefore, in addition to improving responsible prescribing practice, interventions should address the loopholes in current Chinese antibiotic dispensing system, including (1) strengthening the enforcement of Chinese government's AMR policies 42 that ban over-the-counter purchases and cap antibiotic prescriptions (e.g. at 20% for county hospitals), and (2) enabling responsible dispensing antibiotics according to prescribed doses.

Policy implications
Findings from this study suggest that context-appropriate multifaceted Antibiotic use for paediatric upper respiratory tract infections in China ABSTRACT BACKGROUND Inappropriate use of antibiotics for upper respiratory tract infections among Chinese children is rampant.To identify key constructs for effective interventions targeting the public, we investigated parents' decision-making processes with respect to treatment choices and antibiotic use for paediatric URTIs.METHODS Data were collected between June 2017-April 2018 from a cluster random sample of 3,188 parents of children aged 0-13 across three Chinese provinces, representing different stages of economic development.Risk factors of parents' treatment choices and antibiotic use for paediatric URTIs were assessed, using binary and multinomial logistic regressions, adjusting for socio-demographic characteristics.

( 4 )
Cues to action: included (a) presence of fever and (b) information sources for treatment decisions: medical advice, family, and media including social media.(5) Parents' access to antibiotics (with or without prescriptions), including: (a) non-prescription antibiotics: parents' habits of keeping antibiotics at homes for children in the past year; and (b) antibiotic prescriptions: when a child received formal care, point of care used for treatment was assessed, including hospitals above county level, county hospitals, township hospitals, and local clinics.

of 3 ,
188 parents whose children had URTI symptoms within the last month, 594 (18.6%) were self-medicated by parents without medical prescription -56% of these children further obtained antibiotic prescriptions at healthcare facilities.Approximately 70% of children with URTI symptoms (n=2197) sought healthcare; of them, 1204 (54.8%) obtained antibiotic prescriptionsa third of which (33.9%) contained intravenous antibiotics injected directly into the bloodstream, mostly combined with oral antibiotics.Patients or caregivers -the demand-side of the healthcare systemwho are engaged in self-medication and who have demanded antibiotic prescriptions were estimated to have contributed to 41% of antibiotic use for paediatric URTIs [(594+135)/(594+1204)]. (See Caregivers account for at least 40% of outpatient antibiotic use.Antibiotic misuse for paediatric URTIs can be summarised into three forms: (1) self-medication among children by caregivers in the community; and in clinical settings from either (2) unnecessary prescriptions by doctors, or (3) inappropriate prescriptions due to parental demand.Parents' perception of antibiotics as efficacious for treating URTIs and the nearly non-existent barriers to antibiotics are key risk factors in antibiotic misuse behaviours, including self-medication children with antibiotics and the demand and receipt of antibiotic prescriptions.Presence of fever leads to formal care seeking and the demand and receipt of antibiotics prescriptions.Those mainly taking advice from family members are more likely to self-medicate children with antibiotics and less likely to seek care; when they do seek care, they are more likely to receive antibiotic prescriptions.A majority of parents (n=1,728, 54.2%) reported having kept antibiotics at home for their children for the possibility of a future cold.Pressuring doctors for antibiotic prescriptions occurred at all levels of healthcare facilities with a high success rate (79.4%).
interventions are vital to untangle the perpetual problem of over-prescription and ill-informed demands for antibiotics.Simultaneously enhancing both prescribing guidelines, doctor-patient communication skills, and patient education targeting the family as a unit is critical.A blanket antibiotic awareness campaign in China and in other low-and middle-income countries will likely not be effective unless it is rigorously adapted to local context.Interventions enhancing parental self-efficacy of healthcare decision-making, especially regarding care management for paediatric URTIs, and correcting (mis-)perceptions around antibiotic efficacy for URTI symptoms, might reduce misuse.Education interventions should prioritise urban parents with low socio-economic status in less developed regions and be disseminated via medical professionals or media in order to effectively cue parents to a proper response.Enforcing regulations regarding the sale of antibiotics and pack-based antibiotic dispensing systems to reduce household antibiotic stockpiling could curb the main sources of non-prescription antibiotics for self-medication use in Chinese children.CONCLUSIONS Our data pointed to an urgent need for context-appropriate multifaceted interventions to untangle the perpetual problem of over-prescription and ill-informed demands for antibiotics.Having effective stewardship programmes that improve adherence to clinical practice guidelines for antibiotic prescribing and enhance doctor-patient communication over antibiotic use in China is vital.Risk factors influencing caregivers' antibiotic use identified in this study can inform much-needed interventions addressing the challenges posed by both the supply-and demand-side of healthcare system in China.Our findings emphasize the need to prioritise interventions enhancing clinical training, neutralising the pressure from patients for antibiotics, educating on appropriate home care, discouraging antibiotic self-medication, and improving antibiotic dispensing.

Figure 1 .
Figure 1.Antibiotic use for upper respiratory tract infections (URTIs) among Chinese children To develop effective interventions to reduce unnecessary or inappropriate use of antibiotics for paediatric URTIs in the Chinese community, evidence is needed on parents' decision-making for care and how these decisions influence antibiotic use within or outside

Table 3 . Estimated odds ratio (OR, 95% CI) of 'healthcare seeking' for URTIs among Chinese children and relative risk ratio (RRR, 95% CI) of 'seeking formal care without requesting for antibiotics' and 'Seeking antibiotic prescriptions' (vs 'no formal care') for factors affecting parental treatment decisions (N=3188)
OR, odds ratio; RRR, relative risk ratio; CI, confidence interval.*Reference group: Parents who did not seek formal care for their children (n=991, 31.1%) a Adjusted for sex, age, household income, parents' education, urbanicity and province.