Perceptions, challenges, and improvement strategies for reducing intravenous infusion use in hospitalised patients within Chinese medical institutions: a qualitative interview study Authors: Xiaoyan Zhong, Yuexi Huang, Qiming Wei, Kun Chen, Shurong Wang, Yilan Huang Published: 2026 ABSTRACT Background Intravenous infusions are widely used in inpatient care, yet their overuse may increase safety risks, waste healthcare resources, and impose unnecessary burdens on patients. In China, where infusion use remains high, understanding the factors influencing efforts to reduce unnecessary intravenous infusion use is of clear clinical relevance. This study explored stakeholder perceptions, perceived barriers, and potential strategies for reducing unnecessary intravenous infusion use within the broader framework of standardised and rational infusion use.Methods A qualitative study using semi-structured interviews was conducted with healthcare professionals and hospitalised patients. Purposive sampling ensured variation in participant roles, hospital levels, and demographic characteristics. Data were analysed using a hybrid qualitative approach, with inductive thematic analysis as the overarching framework and selected grounded theory coding procedures used to support code development, category construction, and theme refinement, rather than to generate a formal grounded theory.Results Stakeholders showed varying levels of awareness regarding national expectations for standardised and rational intravenous infusion use. References by participants to “reducing infusion use” were interpreted cautiously as indicating uneven understanding or dissemination of the broader rational-use policy framework, including the reduction of unnecessary infusion, rather than lack of awareness of a reduction-specific national target. Physicians’ prescribing decisions were influenced by clinical considerations, patient expectations, and medical insurance audit pressures. Nurses generally viewed reducing unnecessary infusion use as beneficial for improving workflow and reducing workload. Pharmacists highlighted both their important role in promoting rational drug use and the challenges they faced, including staffing shortages, limited training opportunities, and role ambiguity in clinical collaboration. Patients expressed mixed attitudes toward intravenous infusions, with some favouring them because of their perceived rapid effects and limited awareness of potential risks. Across stakeholder groups, reducing unnecessary infusion use emerged as a negotiation between clinical need, patient expectations, policy and insurance audit pressures, and the uneven authority and capacity of different professional groups. Strategy-related findings, derived mainly from healthcare professional accounts and contextualised by patient accounts of expectations, risk awareness, and acceptability, concerned strengthened education and training, better policy alignment, pharmacist capacity building, clinical pathway development, and interdepartmental collaboration.Conclusion This study identified key barriers to reducing unnecessary intravenous infusion use within the broader framework of standardised and rational infusion use in Chinese medical institutions and highlighted stakeholder-informed strategies for improvement. The findings may inform quality-improvement efforts to promote more rational intravenous infusion use. Further research is needed to evaluate the effectiveness of specific interventions in practice. Background Intravenous infusions play an important role in inpatient care by correcting electrolyte imbalances, replenishing nutrients, and enabling rapid medication administration. However, inappropriate intravenous infusion use may waste healthcare resources, increase patients’ financial burden, and expose patients to avoidable safety risks [1]. In China, safety concerns related to injectable and intravenous administration are reflected in national adverse drug reaction (ADR) monitoring data. According to the annual report [2] (Chap. 2, Sect. 2.1.8, Fig. 7), injection route accounted for 56.3% of all ADR/event reports, and intravenous injection accounted for 91.1% of injection-related ADR/event reports. In this ADR report, “intravenous injection” is used as a route category for adverse-event reporting and should not be interpreted as identical to the narrower inpatient intravenous infusion monitoring indicator used in this study. These ADR data are therefore cited to indicate safety concerns related to parenteral/intravenous medication administration rather than to estimate ADRs specifically attributable to intravenous infusions as defined below. From a utilisation perspective, a national survey of 3,964 secondary-level or higher general hospitals across 31 provincial-level administrative regions and the Xinjiang Production and Construction Corps reported an inpatient intravenous infusion use rate of 86.10% in 2020, with an average of 2.69 bags/bottles and 556.56 mL per bed-day [3]. In the present study setting in Luzhou, Sichuan Province, the inpatient intravenous infusion use rate was approximately 90% in 2023, slightly higher than the national benchmark but broadly consistent with previous evidence that inpatient intravenous infusion use remains high in China [3, 4]. Recent national medical quality and safety improvement goals issued by the National Health Commission of China have emphasised the standardised and rational use of intravenous infusions among hospitalised patients, rather than reduction alone [5–8]. Accordingly, in this study, reducing intravenous infusion use refers specifically to reducing unnecessary or irrational intravenous infusion use while preserving clinically indicated infusion therapy within the broader framework of standardised and rational medication use. “Intravenous infusion use” was defined according to the national monitoring indicators for inpatient intravenous infusion use. It referred to intravenous medication administration of ≥ 50 mL, including intravenous drip and pump infusion; for special populations, such as children, neonates, and patients with kidney disease, intravenous administration volumes of < 50 mL were also classified as intravenous infusion [9]. This operational definition of “intravenous infusion use” should therefore be distinguished from the broader ADR-reporting category of “intravenous injection” described above. After the initial qualitative data collection, subsequent institutional monitoring indicated that the local inpatient intravenous infusion rate had decreased to approximately 85% by late 2025. This later figure was not included in the qualitative dataset and should not be interpreted as an outcome of the present study; it is provided only as contextual evidence of continuing practice change in the study setting. Previous studies have mainly examined the prevalence, overuse, and clinical outcomes of intravenous infusion, with less attention to how policy expectations, professional roles, institutional pressures, and patient preferences shape infusion-related decision-making in routine practice. To address this gap, this study adopted a qualitative approach informed by a policy-to-practice and interprofessional implementation perspective. We examined how expectations regarding standardised and rational intravenous infusion use were interpreted, negotiated, and enacted by physicians, nurses, pharmacists, and hospitalised patients. The study addressed three research questions: (1) How do physicians, nurses, pharmacists, and hospitalised patients perceive efforts to reduce unnecessary intravenous infusion use within the broader framework of standardised and rational infusion use? (2) What barriers and contextual factors do stakeholders identify in implementing these efforts? and (3) What strategies suggested by participants may inform future quality-improvement initiatives? Methods Research design This study used an exploratory qualitative design based on semi-structured interviews to examine stakeholders’ perceptions and experiences of reducing unnecessary intravenous infusion use within clinical, institutional, and policy contexts. The study adopted a hybrid qualitative analytic approach: inductive thematic analysis was used as the overarching methodological framework, and selected coding procedures commonly associated with grounded theory were used pragmatically to strengthen the systematic development of codes, categories, and themes. We did not conduct a full grounded theory study or seek to generate a substantive theory. Instead, open, axial, and selective coding were used as structured analytic tools to identify patterns within and across stakeholder groups. This hybrid approach was appropriate because the research questions focused on exploring perceptions, barriers, and participant-suggested strategies, while also requiring an analytic process capable of examining relationships among clinical practice, patient expectations, professional roles, and institutional or policy pressures. The study received approval from the Ethics Review Committee of the Affiliated Hospital of Southwest Medical University (No. KY2024279). Participants Participants were recruited from 10 secondary and tertiary hospitals in Luzhou, Sichuan Province, China, and included pharmacists, physicians, nurses, and hospitalised patients. Recruitment took place during outreach visits supported by the regional pharmaceutical quality control centre affiliated with our hospital’s pharmacy department. Eligible individuals were invited to take part in on-site semi-structured interviews, and written informed consent was obtained before participation. Eligible hospitalised patients were approached on-site after the study had been introduced and after they were considered able to provide informed consent. Purposive, iterative sampling was used to achieve variation in professional role, hospital level, age, sex, and disease type. The number of participants recruited from each hospital varied according to eligibility, willingness to participate, and iterative sampling needs. The study sought to capture diverse perspectives rather than support formal between-hospital comparisons. Eligible healthcare professionals were pharmacists, physicians, and nurses with more than 3 years of hospital work experience. Eligible patients were adults who were currently hospitalised and able to provide informed consent. Recruitment continued until thematic saturation was reached across interviews and participant groups rather than within individual interviews. Subgroup sizes were not determined by prespecified quotas. Refusal rates were not systematically recorded and therefore were not analysed quantitatively. For patients, the current length of hospital stay at the time of interview was recorded as contextual information to help interpret their current inpatient experience; it was not an eligibility criterion and did not refer to years since first hospitalisation or lifetime exposure to hospital care. Data collection Data were collected through on-site semi-structured interviews conducted by trained members of the research team. Interviews explored participants’ perspectives on reducing unnecessary intravenous infusion use in hospitalised patients, associated practice challenges, and potential improvement strategies. Each interview lasted approximately 30–40 min and was audio-recorded with participants’ consent. The semi-structured interview guide was developed by the research team based on the study aims, the three research questions, relevant policy background, and prior literature on rational intravenous infusion use and qualitative interview methods. Before formal data collection, the guide was pilot-tested with a small number of participants representing key stakeholder groups to assess clarity, relevance, and the flow of questions. During the early stage of data collection, the research team reviewed interview experiences and made minor iterative refinements to the wording and order of prompts to improve clarity and probing. These refinements did not change the core interview topics or the overall focus of the study. To improve reporting completeness and transparency in accordance with the COREQ checklist, additional interview-context information was specified. Interviews were conducted by trained members of the research team with professional backgrounds in hospital pharmacy or clinical practice and experience in hospital quality-improvement work. Before data collection, interviewers received uniform training on the study aims, interview guide, informed consent procedures, confidentiality, and neutral probing techniques. Individual interviewer credentials and gender were not systematically recorded at the interview level and therefore could not be analysed as interview-level characteristics. Interviews were conducted in a quiet area of the participating hospital, and no persons other than the participant and interviewer(s) were present during the interviews. Patient interviews were used primarily to elicit patients’ expectations, perceived value of intravenous infusions, concerns about risks and inconvenience, acceptability of non-intravenous alternatives, and communication needs. Patient data were not treated as direct evidence for clinical prescribing criteria, pathway development, or system-level policy design; rather, they were used to clarify demand-side factors that should be considered when interpreting and operationalising strategies proposed mainly by healthcare professionals. Data analysis Interview data were analysed using a hybrid qualitative approach and managed in NVivo 12 Plus. Inductive thematic analysis provided the overarching analytic framework, whereas open, axial, and selective coding, drawn from grounded theory traditions, were used as procedural tools rather than as components of a full grounded theory study. Accordingly, the purpose of the analysis was to develop empirically grounded themes that addressed the research questions, not to produce a formal or substantive grounded theory. To make the analytic process more transparent, codes were developed into categories and final themes through a three-step, iterative procedure. First, during open coding, two researchers independently read the transcripts line by line and assigned descriptive codes to meaning units that captured participants’ perceptions, experiences, or suggested strategies. Initial codes were then compared, discussed, and refined; overlapping codes were merged, while distinct codes were retained in a working coding list with brief definitions and exemplar quotations. Second, during axial coding, related codes were grouped into higher-order categories by examining conceptual similarities, differences, and relationships within and across stakeholder groups. For example, meaning units such as patients describing infusions as working faster, physicians reporting patient requests for infusion, and physicians expressing concern about patient dissatisfaction were initially coded as perceived rapid effect, patient demand for infusion, and avoidance of patient conflict. These codes were then grouped into the category of patient expectations and perceived treatment value. Third, during selective coding, categories were compared against the research questions and integrated into final themes. In this example, the category concerning patient expectations, together with categories related to insurance audit pressure and clinical condition assessment, contributed to the final theme of clinical decision-making shaped by patient expectations and insurance-related pressures. A similar process was used to develop the remaining themes, with categories checked across physicians, nurses, pharmacists, and patients to ensure that each theme reflected cross-cutting patterns while preserving stakeholder-specific perspectives. Data collection and analysis proceeded iteratively, with saturation assessed at the dataset level when no new themes relevant to the research questions were identified across stakeholder groups or the dataset as a whole. Thus, saturation was assessed across participants rather than within individual interviews, and subgroup sizes reflected the final sample achieved through iterative recruitment and analysis rather than predetermined quotas. Discrepancies were resolved through discussion and consensus. Formal inter-coder agreement statistics were not calculated because the analysis emphasised interpretive consensus. Coding decisions, including reasons for merging, splitting, or renaming codes and categories, were discussed by the research team and documented during analysis. Credibility was enhanced through member checking with selected participants. For strategy-related findings, patient-derived codes were incorporated only when they clarified demand-side barriers and acceptability issues, such as perceived rapid effects of infusion, limited awareness of risks, inconvenience during infusion, willingness to follow medical advice, and the need for clearer explanations. In contrast, clinical and system-level strategies, including prescribing criteria, clinical pathway development, pharmacist involvement, policy alignment, and interdepartmental collaboration, were grounded primarily in the accounts of physicians, nurses, and pharmacists. To avoid treating the Results section as a set of author-generated recommendations, strategy-related findings were coded only when participants explicitly described possible improvement approaches or when they identified implementation needs that were directly linked to such approaches. These codes were compared across stakeholder groups and then grouped into broader data-derived categories, such as education and communication, policy alignment, pharmacist capacity building, pathway development, and interdepartmental collaboration. The Results section reports these categories as participant-informed directions emerging from the data, whereas the Discussion interprets their implications for practice and policy. Reflexivity The research team recognised that recruitment during study team-led educational outreach activities may have influenced participation and increased the risk of socially desirable responses. To reduce this risk, interviewers emphasised voluntary participation, confidentiality, de-identification, and the absence of right or wrong answers. The team also recognised that researchers’ backgrounds in pharmacy, clinical practice, and regional quality-improvement activities could shape data interpretation, particularly by directing attention toward rational drug use, policy implementation, pharmacist roles, and system-level management strategies. During analysis, two researchers independently coded transcripts, compared coding decisions, discussed alternative interpretations, searched for disconfirming accounts across stakeholder groups, and documented reasons for merging, splitting, or renaming codes and categories. Member checking with selected participants was used to assess whether the interpreted themes were consistent with participants’ accounts. These procedures helped keep the analysis grounded in the interview data while acknowledging that qualitative interpretation is shaped by researchers’ positionality. Research quality assurance To ensure research quality, we trained the team to standardise data collection and analysis methods. We recorded and transcribed the interviews for thorough review. During analysis, we maintained the originality and authenticity of the data to avoid personal bias. Ethical considerations The study was approved by the institutional review board of The Affiliated Hospital of Southwest Medical University (protocol number KY2024279). All participants provided written informed consent. To protect participant privacy, study data were coded to remove all identifying information. Results The analysis identified five overarching themes: (1) uneven awareness and interpretation of policy expectations regarding standardised and rational intravenous infusion use, rather than a reduction-specific target; (2) clinical decision-making shaped by patient expectations and insurance-related pressures; (3) pharmacists’ important but constrained role in implementation; (4) diverse patient preferences and limited awareness of risks; and (5) multilevel improvement needs related to education, policy alignment, pathway development, and interdepartmental collaboration. Table 1 presents the key findings by participant group, and Table 2 outlines strategy-related findings that emerged from participant accounts during coding. Each theme is described in detail below with representative quotations. In Table 2, patient perspectives should therefore be interpreted as evidence on demand-side barriers and communication needs, whereas clinical and system-level strategies were derived primarily from healthcare professional accounts. The strategy-related findings summarised in Table 2 were derived from participant accounts during coding and are reported as data-derived directions rather than prescriptive recommendations. Consistent with qualitative reporting conventions, numerical expressions are not used to imply statistical representativeness; subgroup patterns are described narratively to indicate variation in views across participants. Taken together, the themes were not independent findings but reflected a connected implementation process. Policy expectations for rational and standardised infusion use were filtered through local clinical routines, patient expectations, insurance audit concerns, and professional role boundaries. A central tension was that physicians described clinical severity as the formal basis for prescribing decisions, yet also reported pressure from patients who associated infusions with faster or stronger treatment and from insurance audits that could penalise admissions perceived as insufficiently intensive. Nurses and pharmacists recognised potential benefits of reducing unnecessary infusions, but their ability to influence practice depended on communication with physicians, patient trust, staffing capacity, and institutional support. Patient accounts helped explain why education and communication were repeatedly proposed: reducing unnecessary infusion use was not only a technical prescribing issue, but also a process of reshaping expectations about what constitutes appropriate inpatient treatment. Table 1Summary of major findings by participant groupParticipant groupMajor themeMain concerns / suggestedfocusPhysiciansInfusion use was influenced by disease severity, patient expectations, and insurance audit pressureNeed clearer policy guidance and better alignment between clinical judgment, patient expectations, and insurance audit standardsNursesReducing unnecessary infusions may reduce workload and improve care efficiencyNeed more training and stronger patient communication to support acceptance of appropriate non-intravenous alternativesPharmacistsStrong support for rational infusion use and active prescription oversightNeed clearer responsibilities, more staffing support, and greater authority in multidisciplinary implementationPatientsMany perceived infusions as faster and more effective, with limited awareness of risksNeed better education about risks and non-intravenous alternatives; patient expectations acted as a demand-side influence on implementation Table 1Summary of major findings by participant groupParticipant groupMajor themeMain concerns / suggestedfocusPhysiciansInfusion use was influenced by disease severity, patient expectations, and insurance audit pressureNeed clearer policy guidance and better alignment between clinical judgment, patient expectations, and insurance audit standardsNursesReducing unnecessary infusions may reduce workload and improve care efficiencyNeed more training and stronger patient communication to support acceptance of appropriate non-intravenous alternativesPharmacistsStrong support for rational infusion use and active prescription oversightNeed clearer responsibilities, more staffing support, and greater authority in multidisciplinary implementationPatientsMany perceived infusions as faster and more effective, with limited awareness of risksNeed better education about risks and non-intravenous alternatives; patient expectations acted as a demand-side influence on implementation Summary of major findings by participant group Table 2Summary of data-derived strategy-related findings for reducing unnecessary intravenous infusion useIdentified issueData-derived strategy-related findingHow participant perspectives informed the findingUneven awareness of policy expectations for standardised and rational intravenous infusion use among healthcare staffNeed for regular training and improved dissemination of relevant policies, clinical guidance, and implementation requirementsPhysicians, nurses, pharmacistsStrong patient preference for intravenous infusion and limited awareness of risksNeed for patient education and communication regarding risks, benefits, and non-intravenous alternativesPhysicians and nurses suggested patient education and communication; patients clarified demand-side barriers, perceived benefits, risk awareness, acceptability of alternatives, and communication needs.Insurance audit pressure influencing prescribing behaviourNeed for better alignment between insurance audit standards and clinical practice, supported by clearer policy guidancePhysicians, pharmacistsLimited pharmacist capacity and unclear role boundaries in implementationNeed for pharmacist training and expanded pharmacist involvement in prescription review and clinical pathwaysPharmacists, physiciansLack of clear criteria for intravenous infusion use in common conditionsNeed for condition-specific infusion criteria and greater availability of oral or topical alternatives where appropriatePhysicians, pharmacistsNeed for coordinated action across departments and institutionsNeed for multidisciplinary collaboration among clinical staff, pharmacy departments, hospital management, and relevant authoritiesPhysicians, nurses, pharmacists Table 2Summary of data-derived strategy-related findings for reducing unnecessary intravenous infusion useIdentified issueData-derived strategy-related findingHow participant perspectives informed the findingUneven awareness of policy expectations for standardised and rational intravenous infusion use among healthcare staffNeed for regular training and improved dissemination of relevant policies, clinical guidance, and implementation requirementsPhysicians, nurses, pharmacistsStrong patient preference for intravenous infusion and limited awareness of risksNeed for patient education and communication regarding risks, benefits, and non-intravenous alternativesPhysicians and nurses suggested patient education and communication; patients clarified demand-side barriers, perceived benefits, risk awareness, acceptability of alternatives, and communication needs.Insurance audit pressure influencing prescribing behaviourNeed for better alignment between insurance audit standards and clinical practice, supported by clearer policy guidancePhysicians, pharmacistsLimited pharmacist capacity and unclear role boundaries in implementationNeed for pharmacist training and expanded pharmacist involvement in prescription review and clinical pathwaysPharmacists, physiciansLack of clear criteria for intravenous infusion use in common conditionsNeed for condition-specific infusion criteria and greater availability of oral or topical alternatives where appropriatePhysicians, pharmacistsNeed for coordinated action across departments and institutionsNeed for multidisciplinary collaboration among clinical staff, pharmacy departments, hospital management, and relevant authoritiesPhysicians, nurses, pharmacists Summary of data-derived strategy-related findings for reducing unnecessary intravenous infusion use The categories in Table 2 therefore summarise how participants framed possible directions for improvement; they are not presented as evaluated interventions or formal recommendations. We engaged 60 participants who met the inclusion criteria for the interviews. This group included 20 patients (mean age 52 years; 50% female), 10 nurses (mean age 40 years; 90% female), 10 physicians (mean age 45 years; 30% female), and 20 pharmacists (10 clinical pharmacists and 10 heads of pharmacy departments; mean age 44 years; 60% female). All healthcare professionals had more than three years of work experience. For patients, the current length of hospital stay at the time of interview was summarised as median and range to contextualise their ongoing inpatient experience and exposure to current infusion-related decision-making. The demographic, professional, and current hospitalisation characteristics of the participants are presented in Table 3. Table 3Demographic, professional, and current hospitalisation characteristics of participantsRoleNumber of Participants (n)Age, mean (range), yearsFemale (%)Practice years/current hospital stay, summary statistic (range)Professional/patient backgroundPatients2052 (40–65)503 days, median (range: 3–30 days)Currently hospitalised adultsNurses1040 (30–50)908 years, mean (range: 3–15 years)Nursing majorPhysicians1045 (35–55)3010 years, mean (range: 5–20 years)Internal medicine or related fieldsPharmacists2044 (35–55)6011 years, mean (range: 4–25 years)Pharmacy major (10 clinical pharmacists, 10 heads of pharmacy departments)* For healthcare professionals, this variable refers to years of professional practice and is presented as mean (range). For patients, it refers to the current length of hospital stay at the time of interview and is presented as median (range), not years since first hospitalisation or lifetime hospitalisation exposure Table 3Demographic, professional, and current hospitalisation characteristics of participantsRoleNumber of Participants (n)Age, mean (range), yearsFemale (%)Practice years/current hospital stay, summary statistic (range)Professional/patient backgroundPatients2052 (40–65)503 days, median (range: 3–30 days)Currently hospitalised adultsNurses1040 (30–50)908 years, mean (range: 3–15 years)Nursing majorPhysicians1045 (35–55)3010 years, mean (range: 5–20 years)Internal medicine or related fieldsPharmacists2044 (35–55)6011 years, mean (range: 4–25 years)Pharmacy major (10 clinical pharmacists, 10 heads of pharmacy departments)* For healthcare professionals, this variable refers to years of professional practice and is presented as mean (range). For patients, it refers to the current length of hospital stay at the time of interview and is presented as median (range), not years since first hospitalisation or lifetime hospitalisation exposure Demographic, professional, and current hospitalisation characteristics of participants * For healthcare professionals, this variable refers to years of professional practice and is presented as mean (range). For patients, it refers to the current length of hospital stay at the time of interview and is presented as median (range), not years since first hospitalisation or lifetime hospitalisation exposure Perspective of physicians The surveyed physicians had diverse perceptions about reducing unnecessary intravenous infusion use within the broader policy framework of standardised and rational intravenous infusion use. Notably, several physicians did not describe the national medical quality and safety improvement goal in terms of its formal emphasis on standardised and rational use. Instead, during interviews, this broader policy objective was often simplified or interpreted by participants as “reducing infusion use” or reducing unnecessary infusion. This finding should therefore be interpreted cautiously as suggesting gaps in policy awareness, dissemination, or shared terminology, rather than as evidence that clinicians were unaware of a distinct national target focused solely on reduction. One physician remarked, “I am not very clear about the specific requirements and strategies for reducing intravenous infusion use in hospitalised patients.” This statement suggests uncertainty about policy implementation requirements and local communication of relevant guidance, despite the availability of national guidelines. In contrast, those who received relevant training demonstrated a better understanding of strategies to reduce unnecessary intravenous infusion use among hospitalised patients within the broader goal of rational and standardised infusion practice, emphasising the critical role of education. In clinical decision-making, several physicians described the severity of the patient’s condition as the primary basis for choosing intravenous infusions. One physician emphasised, “For critically ill patients, intravenous infusions are a crucial intervention to rapidly address their medical needs.” Furthermore, patients’ reliance on intravenous infusions also affects physicians’ prescribing decisions. Another doctor noted, “Patients expect to receive infusion treatments. Sometimes, even when it is not medically necessary, I comply with their expectations to avoid conflicts with patients.” Medical insurance policies significantly influence the prescribing decisions of physicians. Several physicians expressed concerns, stating, “During medical insurance audits, if patients have not received infusion treatments, they may be judged as having been admitted under low standards. This could result in economic penalties for the hospital.” This situation may compel physicians to opt for intravenous infusions even when such treatments are not clinically justified. Physicians have asserted that enhanced education and clear policy guidance are essential for promoting more rational intravenous infusion use. As one physician remarked, “We need more education and clear policy guidance to help us make better decisions in clinical care.” These findings indicate that prescribing decisions are influenced not only by clinical assessment but also by perceived patient expectations and institutional audit pressures, highlighting policy–practice tensions in promoting standardised and rational intravenous infusion use. This tension linked the physician theme directly with the patient theme: patients’ perceived value of infusion could reinforce prescribing pressure, while audit concerns created an institutional incentive that did not always align smoothly with efforts to reduce unnecessary use. Perspective of nursing staff Some nursing staff also did not clearly distinguish between the formal policy emphasis on standardised and rational intravenous infusion use and local shorthand descriptions of “reducing infusion use.” One nursing staff member remarked, “I knew that intravenous infusions have one of the highest incidences of adverse reactions. However, I was not informed that reducing intravenous infusion use had been designated as a national goal for enhancing medical quality and safety.” This quotation was interpreted as reflecting incomplete policy dissemination and inconsistent understanding of how national rational-use goals were communicated locally, rather than as confirmation of a national reduction-only target. Several nursing staff believed that reducing unnecessary intravenous infusion use among hospitalised patients could reduce their workload. This could ultimately lead to more time and energy for providing meticulous nursing services, thereby enhancing patient satisfaction. One nurse stated, “By reducing unnecessary infusions, we can allocate more time and energy to providing more meticulous nursing services, thereby enhancing patient satisfaction.” Nursing staff play a crucial role in patient education and trust-building. One nurse mentioned, “We need to patiently explain to patients why, in some cases, oral medications might be a better therapeutic choice.” Another nurse emphasised, “Through education, we help patients understand their treatment plans. Establishing trust makes them more willing to accept our recommendations.” This suggests that nursing staff, through education and communication, play an important role in promoting patients’ acceptance of and trust in appropriate treatment plans. In the cross-stakeholder analysis, nurses therefore appeared as an implementation bridge between clinical decisions and patient acceptance: they were not the primary prescribers, but their communication work could mediate the tension between patient expectations and attempts to promote non-intravenous alternatives. Perspective of pharmacists Pharmacists acknowledged the importance of reducing unnecessary intravenous infusion use among hospitalised patients; however, they described numerous challenges in implementing this practice. One pharmacist remarked, “Changing medical habits faces multiple obstacles, such as physicians’ prescribing practices and patient expectations.” In their efforts to assist physicians with rational drug treatment, pharmacists also assume supervisory responsibilities. A clinical pharmacist expressed his dilemma: “We need to ensure the rational use of drugs while also supporting the work of physicians; this dual role sometimes places us in a challenging position.” Despite these challenges, pharmacy departments were actively working to reduce unnecessary intravenous infusion use among hospitalised patients. One pharmacist emphasised, “This effort requires support from hospital management and close coordination among departments.” However, there is a disparity between the ideal and the reality. Another pharmacist noted, “In practice, the pharmacy department often bears primary responsibility for reducing unnecessary intravenous infusion use among hospitalised patients. This significantly increases the challenges we face.” Additionally, a senior pharmacist highlighted a current limitation, stating, “Developing clinical pathways is an effective strategy for reducing unnecessary intravenous infusion use among hospitalised patients. However, limited clinical knowledge among pharmacists means that only a few are involved in creating these pathways.” This observation underscores the necessity for enhanced engagement and targeted educational initiatives aimed at increasing pharmacists’ participation and effectiveness in this critical area. To reduce reliance on intravenous infusions, hospitals have compiled a list of conditions that do not necessitate intravenous medication. The pharmacy department has optimised drug supplies to ensure appropriate routes of administration. A pharmacist stated, “We have increased the supply of oral and topical medications.” Pharmacists described specialised drug management as a potentially useful approach to minimising unnecessary intravenous infusions. This includes targeted management of antimicrobials, proton pump inhibitors (PPIs), traditional Chinese medicine injections, parenteral nutrition, antiemetics, and more. The pharmacy department has established target values for the utilisation of intravenous infusions across clinical departments and linked these targets to departmental performance metrics. One pharmacist noted, “This enhances motivation to reduce unnecessary intravenous infusion use among hospitalised patients, but it also requires ongoing communication with clinical departments to achieve consensus.” Pharmacists regularly analyse data to identify and address issues promptly. Another pharmacist remarked, “Continuous monitoring helps reduce unnecessary intravenous infusion use among hospitalised patients.” Pharmacists played a crucial role in reducing unnecessary intravenous infusion use; however, they face challenges due to insufficient staffing. A pharmacy director noted, “The shortage of clinical pharmacists in secondary medical institutions limits the depth and breadth of our work.” Staff in these institutions often lack confidence in their clinical skills. This may result from fewer opportunities for professional training and continuing education, along with the limited types and complexities of cases encountered in daily practice. This lack of confidence may adversely affect communication with physicians and diminish pharmacists’ initiative and decisiveness in ensuring rational drug use. Pharmacists actively engage in patient education, using materials like popular science videos to enhance understanding of treatment plans. One pharmacist remarked, “This helps patients make wiser treatment choices.” Pharmacists recognise that continuous learning is critical for improving their professional communication skills. One pharmacist observed, “A deep understanding of diseases and pharmacology enables us to communicate more effectively with physicians.” While dedicated, pharmacists also face significant pressures and challenges. One senior pharmacist stated, “We love our work, but we also need to manage stress and prevent professional burnout.” These findings suggested that pharmacists’ work is shaped not only by technical responsibilities in rational drug use management, but also by role ambiguity within hospital hierarchies, as they are expected to supervise prescribing while maintaining collaborative relationships with physicians. This pharmacist-related tension intersected with the physician and policy themes because pharmacists were expected to support rational use through review, monitoring, and pathway development, yet their influence depended on professional authority, clinical confidence, staffing levels, and acceptance by clinical departments. Perspective of patients Patients exhibited a range of attitudes towards intravenous infusions; many described a willingness to follow their doctors’ recommendations. Some patients preferred intravenous infusions because of their perceived rapid therapeutic effects, saying, “If I’m hospitalised, why not have an infusion? I think infusions are more effective.” Patients differed in their awareness of the potential risks and side effects of intravenous infusions. Some patients were unaware of the risks, claiming, “I always thought infusions were a safe and effective treatment method.” Conversely, those who recognised the risks expressed various concerns, saying, “I’ve heard that infusions can cause reactions and other side effects.” Additionally, some patients expressed dissatisfaction with the inconvenience of infusions, noting, “You can’t move freely when you’re on an infusion, which is very inconvenient.” These patient accounts suggest that efforts to reduce intravenous infusion use may be hindered by the perceived symbolic and therapeutic value of infusions, particularly when awareness of risks and alternatives remains limited. Thus, patient interviews informed the education and communication components of the strategy-related findings by identifying expectations, concerns, and acceptability issues that healthcare professionals would need to address; they were not used to generate clinical prescribing criteria or system-level implementation strategies. These accounts also help explain why physicians reported pressure to prescribe infusions and why nurses and pharmacists emphasised education and communication as implementation strategies; patient expectations functioned as a demand-side force that interacted with professional and institutional pressures rather than as an isolated preference. Discussion This study explored stakeholder perceptions of reducing unnecessary intravenous infusion use and the contextual barriers to implementation. Rather than reflecting a single clinical or behavioural issue, the findings show how policy expectations, patient expectations, professional roles, and institutional pressures interact in routine practice. Because this was a qualitative interview study, the implications are presented as participant-reported concerns and data-informed interpretations, not as causal conclusions or tested recommendations. By foregrounding these multilevel relationships, this study adds to evidence on rational infusion use in China beyond prevalence- and outcome-focused research. Cross-theme integration: policy-to-practice dynamics Across themes, national expectations for standardised and rational infusion use were translated through local routines, payment and audit concerns, professional authority, and patient beliefs about treatment value. The findings therefore indicate a policy-to-practice gap rather than a simple knowledge deficit. Physicians described balancing clinical indications with patient requests and audit concerns; nurses translated treatment decisions into patient-facing communication; and pharmacists monitored medication use while depending on staffing, role clarity, and departmental acceptance. These interdependencies suggest that reducing unnecessary infusion use requires coordinated alignment of incentives, professional roles, and patient communication, rather than isolated changes in prescribing behaviour. Implementation science helps explain this pattern and strengthens the interpretative layer of the findings. In CFIR terms, national policies and insurance audit requirements, including audit pressures linked to DRG/DIP payment reform, represent outer-setting influences, whereas hospital indicators, workflow, staffing, and multidisciplinary communication reflect inner-setting conditions [10]. From the COM-B perspective, barriers involved capability (limited policy understanding, variable pharmacist confidence, and limited patient knowledge), opportunity (audit pressure, staffing constraints, and unclear authority), and motivation (patient beliefs in faster effects and clinicians’ desire to avoid conflict) [11]. Education alone is therefore unlikely to be sufficient; sustainable change may require combined strategies that address capability, opportunity, and motivation at multiple levels. Policy awareness and educational needs of healthcare professionals Participants described uneven awareness of policy expectations for standardised and rational intravenous infusion use as a barrier to reducing unnecessary or irrational infusion. This should not be interpreted as evidence that clinicians were unaware of a formal reduction-specific national target. Rather, it suggests inconsistent understanding of current policy language and local implementation requirements, including how reduction of unnecessary infusion fits within a broader rational-use framework. Clearer dissemination of updated guidance and targeted in-service education may therefore help staff apply policy expectations in routine care, although training alone is unlikely to change prescribing behaviour. Insurance audit pressure and clinical decision-making Physicians perceived their prescribing decisions as shaped by clinical judgment, patient expectations, and concerns about medical insurance audits. Some participants worried that audit criteria could be interpreted in ways that conflict with efforts to reduce unnecessary infusion use. These accounts point to a perceived need for better alignment between rational-use policy, insurance audit practice, and frontline clinical decision-making, but they should be read as participant-reported concerns rather than proof of a direct policy effect on prescribing. Pharmacists’ role, capacity, and integration into rational infusion management Participants viewed pharmacists as important contributors through prescription review, medication monitoring, feedback on infusion indicators, optimisation of non-intravenous alternatives, and participation in clinical pathways or medication-specific stewardship. Their contribution, however, depended on their position within hospital hierarchies. Pharmacists were expected to identify inappropriate intravenous administration and support prescribing decisions, but their influence was shaped by clinical confidence, staffing, communication with physicians, and institutional recognition of their authority. Pharmacist-related improvement should therefore be understood as a capacity and integration issue. Participants described staffing shortages, limited training opportunities, and variable confidence, particularly in secondary hospitals [12], as well as the need for clearer responsibilities and multidisciplinary support. Previous studies suggest that clinical pharmacy involvement in pathway management under the DRG/DIP system may support more appropriate medication use [13–15]. In China, these findings also reflect broader challenges related to role ambiguity, limited professional authority, and evolving expectations of pharmacists [16–18]. Although recent policies have promoted pharmaceutical services and professional development [19, 20], the legal and institutional framework remains incomplete, which may constrain further integration into hospital medication management [21]. DRG/DIP reform and clinical pathways For international readers, Diagnosis-Related Group (DRG) payment classifies inpatient cases into groups with similar diagnoses, procedures, and expected resource use, whereas Diagnosis-Intervention Packet (DIP) payment classifies cases using diagnosis- and procedure-based combinations and point values to determine reimbursement. In China, both reforms are intended to move hospital payment away from traditional fee-for-service toward bundled or prospective payment and are accompanied by medical insurance audits that may influence hospital management indicators and clinicians’ perceptions of appropriate resource use. Participants’ accounts suggested possible tensions among payment reform, clinical pathway management, and infusion-related prescribing. DRG/DIP payment reform may encourage rational prescribing, while standardised pathways or explicit medication-use criteria may reduce variation in infusion decisions. In this study, these points are interpreted as policy and implementation implications informed by participant concerns and relevant literature, rather than as direct evidence that DRG/DIP reform or pathways will reduce intravenous infusion use. Medication-specific stewardship of commonly overused intravenous drugs Participants described unnecessary intravenous infusion use as shaped by clinician perceptions, patient expectations, and institutional management processes. Proton pump inhibitors (PPIs) provide one example of how medication-specific stewardship could operationalise rational-use goals through drug categories and prescribing criteria. Inappropriate use of intravenous PPIs has been reported in international clinical settings [22–24]. In China, national guidance has strengthened the rational clinical use of PPIs [25], and hospital-based quality-improvement initiatives have shown potential to improve prescribing practice [26]. This example should be interpreted as an illustration of possible stewardship approaches, not as evidence from the present interviews that targeting PPIs alone would reduce unnecessary infusion use. Patient education and communication Patients’ expectations of intravenous infusions, together with limited awareness of risks and alternatives, may act as demand-side barriers to reducing unnecessary use. This finding underscores the importance of patient education and clinician–patient communication. Previous research has also documented limited public understanding of intravenous infusion; in one survey of 663 hospitalised patients, only 39.73% demonstrated relevant knowledge [27]. Accessible materials, including brochures, videos, and WeChat-based information, may help improve patient understanding and support more rational choices [28]. In this study, patient perspectives helped specify education and communication needs rather than clinical or administrative strategies. Multidisciplinary and interdepartmental collaboration Participants described reducing unnecessary intravenous infusion use as difficult for any single professional group or department to achieve. Coordination among physicians, nurses, pharmacists, patients, hospital managers, and administrative bodies may be needed across prescribing, dispensing and review, administration, communication, and management feedback. Previous research indicates that multidisciplinary interventions may help reduce infusion use among hospitalised patients and improve rational medication management [29]. Participant-informed approaches included joint training, clearer role delineation, communication mechanisms, patient education, continuous quality monitoring, and appropriate incentives. These approaches require future evaluation before they can be treated as evidence-based interventions. Implementation implications and practical relevance Translating participant-informed directions into practice will require coordinated action across policy interpretation, institutional governance, professional practice, and patient communication. Pathways, pharmacist participation, nurse-led communication, and patient education may be more feasible when hospital management aligns resources, audit requirements, and quality-improvement indicators. Future evaluations could examine inpatient infusion rates, use of non-intravenous alternatives, staff training coverage, patient awareness, pathway adherence, and consistency between audit standards and rational-use guidance. These outcomes were not assessed in the present qualitative study. The subsequent local decline in infusion use provides contextual support for continued evaluation but cannot be attributed to this study. Transferability should be interpreted in relation to context. China-specific elements include national policy language on standardised and rational infusion use, DRG/DIP payment reform, hospital performance indicators, insurance audit interpretations, high baseline inpatient infusion use, and the evolving authority of pharmacists in hospitals. More broadly transferable mechanisms include patient expectations of intravenous therapy, clinicians’ balancing of clinical appropriateness with external pressures, variable authority of pharmacists and nurses as implementation agents, and the potential value of multidisciplinary communication, audit-feedback, pathway development, and patient education. International readers should therefore view the findings as adaptable implementation mechanisms that require local contextualisation, rather than as directly transferable intervention prescriptions. Future quality-improvement interventions should define target behaviours at multiple levels, including physicians’ prescribing decisions, pharmacists’ review and feedback, nurses’ communication with patients, managers’ alignment of indicators and resources, and patients’ acceptance of appropriate alternatives. Mapping these behaviours to capability, opportunity, and motivation may help institutions choose coherent strategies, such as training, role clarification, audit-feedback, and risk-benefit communication. Study limitations This study has several limitations. First, nuances may have been lost when interview data were translated from Chinese into English. Second, the geographically concentrated sample may limit transferability. Third, recruitment during study team-supported educational outreach may have introduced selection bias because participants who were more accessible, engaged with quality-improvement activities, or willing to discuss infusion management may have been more likely to participate. The outreach context may also have encouraged socially desirable or policy-concordant responses, despite efforts to emphasise voluntary participation, confidentiality, de-identification, and the absence of right or wrong answers. Fourth, although participants were recruited from multiple hospitals, the study was designed to capture diverse perspectives rather than formal between-hospital comparisons. Fifth, researcher interpretation may have influenced coding and theme development. The team’s pharmacy and quality-improvement backgrounds may have made pharmacist roles, rational drug use, and policy implementation particularly salient. Independent coding, consensus discussion, reflexive examination of alternative interpretations, documentation of coding decisions, and member checking were used to mitigate this risk, but researcher bias cannot be fully excluded. Finally, the study explored stakeholder perceptions, contextual barriers, and participant-informed directions, rather than the effectiveness of specific interventions. Conclusion This study examined stakeholder perceptions, challenges, and participant-informed directions related to reducing unnecessary intravenous infusion use within the broader framework of standardised and rational infusion use in Chinese medical institutions. Patient perspectives primarily clarified demand-side expectations, acceptability, and communication needs. The findings identify interrelated barriers involving policy interpretation, patient expectations, professional roles, clinical practice, and institutional pressures. They also suggest directions for future quality-improvement work, including education, clearer clinical guidance, stronger pharmacist involvement, better clinician-patient communication, and interdepartmental coordination. These data-derived findings may inform efforts to promote more rational infusion use, but they should not be interpreted as evidence that any specific intervention is effective. Future research should evaluate the effectiveness of targeted interventions in practice.