PDF
Citation
Print
I. Introduction
Emergency medical services, as defined by the Emergency Medical Act of 2008, range from the initial recognition of an emergency condition to ensuring that patients receive timely and appropriate care, whether during ambulatory transfer or within hospital settings [1]. A well-structured emergency system integrates community operations, hospital-based care, notification centers, and network management. This integration builds community confidence, ensures timely service, optimizes treatment decisions, minimizes delays, and reduces disability or death rates [2].
In line with Thailand’s health reform policy, integrating health databases is paramount. Such integration improves emergency referrals, treatment, and reimbursement processes while ensuring effective access to medical services. A key element is the implementation of an electronic referral system to streamline emergency care coordination across hospitals. This advancement is particularly crucial in regions like Chiang Mai, northern Thailand’s largest province with over 1.7 million residents [3].
Few studies have evaluated Chiang Mai’s electronic referral system, although initial findings are positive. Suwatmakin et al. [4] reported that the system increased appropriate referrals, reduced investigation waiting times, decreased registration times, and lowered referral costs. Iamthongin [5] found that it enhanced communication, improved data access, reduced workloads, and enabled seamless data capture and transfer within the hospital network. However, these evaluations primarily relied on qualitative methods using interviews and questionnaires with hospital staff. Building on these findings, this study employs quantitative methods to provide empirical evidence on the effectiveness of HIS. SANSAI in Chiang Mai.
HIS.SANSAI was developed to address several healthcare challenges in Chiang Mai, including the rising demand for emergency services, the complexity of inter-hospital patient transfers, and the need for efficient patient information management. This web-based application designed to facilitate the seamless transfer of emergency patients was initially implemented in 2020 in a referral hospital in the Sansai district. It was later integrated into the broader Chiang Mai Health Information System (CMHIS) framework [6].
Unlike the previous referral program, which was restricted to generating referral documents on Windows computers, HIS.SANSAI offers comprehensive patient data accessible via any internet-connected device. This digital system improves emergency referrals and treatment by providing complete patient information to the referral hospital before the patient’s arrival. With HIS.SANSAI, specialists can quickly decide whether to accept referrals or propose alternative treatments, streamlining the approval process [4].
This study aimed to evaluate HIS.SANSAI—a digital solution for transferring emergency patients between referring (sending) and referral (receiving) hospitals in Chiang Mai. Specifically, we assessed its impact on user experiences and health outcomes. The findings offer valuable insights for other regions and countries considering similar systems.
1. HIS.SANSAI Referral Process
Figure 1 compares the non-HIS.SANSAI (former) and HIS. SANSAI (current) referral processes. The non-HIS.SANSAI process relied on the Thai Refer Program combined with the LINE application and/or telephone to transfer patient information and coordinate between hospitals. However, since the Thai Refer Program was limited to Windows computers and primarily used for creating referral documents [4,5,7], it often led to incomplete data transmission. This frequently necessitated additional information requests by referral doctors, resulting in inadequate care preparation and potentially longer service times from referral receipt to emergency room discharge [7].
In contrast, HIS.SANSAI consolidates all details into a single web link accessible via any internet-connected device, ensuring comprehensive patient data transmission. This allows medical staff to prepare more effectively, thereby reducing service time. The process complies with Thailand’s Personal Data Protection Act (PDPA) and international standards, including HL7 FHIR, which facilitates secure, seamless data exchange between diverse healthcare systems [8,9]. Accessing patient data via the referral form requires encoding and identity verification by both the referring and referral (or mid-level referral) hospitals [4,6].
II. Methods
1. Study Design
This study used a multimethod approach, combining a cross-sectional survey to measure user experiences and perceptions of system effectiveness with a quantitative analysis of HIS.SANSAI’s impact on health outcomes.
2. User Experience Survey
A survey instrument was developed to assess the system capabilities and work processes of HIS.SANSAI. The survey evaluated two dimensions using a 10-point scale: system capabilities and work processes. It was developed following a comprehensive literature review and consultations with HIS. SANSAI experts, including the developer team and experienced personnel from both referring and referral hospitals. Two experts familiar with HIS.SANSAI and IT systems reviewed the questionnaire prior to its implementation. Due to the questionnaire’s specificity and the limited number of users, a pilot test was conducted; participants reviewed all questions and provided feedback to ensure relevance and alignment with study objectives.
The survey was administered to a purposive sample, including staff from a mid-level referral hospital and referring hospitals (see Supplement A). Respondents rated their preferences and perceptions on a 10-point scale, with scores ranging from 1 (least preferred) to 10 (most preferred). Descriptive statistics, including arithmetic mean and standard deviation, were used to analyze the survey data. Ethical approval was obtained from the Human Research Ethics Committee of Naresuan University, Thailand (COA No. 410/2021).
3. Data Source
To evaluate HIS.SANSAI’s efficiency and outcomes, we analyzed de-identified data from a mid-level referral hospital for the fiscal years 2019 and 2021 (October–September), during which HIS.SANSAI was used alongside the previous system. Data integration combined electronic health records and paper-based documentation, extracting details such as gender, age, diagnosis, and referral date/time. This information was merged with HIS.SANSAI emergency referral data and linked to patients’ medical histories, diagnostic details, referral status, health parameters, and discharge information. For patients referred via the previous system, data were obtained from paper records. The analysis focused on four common ailments: acute appendicitis, end-stage renal disease (ESRD), moderate head injury (MHI), and ischemic stroke. The sample comprised 723 patients, categorized by emergency department discharge status: 23 with life-threatening conditions requiring resuscitation, 146 with serious illnesses needing emergency assistance, and 554 with urgent conditions.
4. Analytical Approach
Multiple regression analysis was employed to examine whether HIS.SANSAI significantly reduced service duration while controlling for factors affecting service time. Logistic regression was then used to assess the system’s influence on treatment outcomes following emergency department discharge.
The multiple regression model was as follows:
where time represents the service duration (in minute) from referral receipt at the mid-level referral hospital to emergency room discharge, X is a matrix of independent variables, β̂ is the vector of coefficients, and ε is the vector of error terms. Independent variables included:
sex = sex of patients, 1 is male and 0 is otherwise.
resus = status at referral, 1 is resuscitate and 0 is otherwise.
emer = status at referral, 1 is emergency and 0 is otherwise.
age = age of patient (year).
Se_dx = the presence of comorbidities, 1 if there were comorbidities and 0 if there were no comorbidities.
SBP = systolic blood pressure (mmHg).
UD = presence of underlying disease, 1 if an underlying disease was present and 0 if no underlying disease existed.
ach = alcohol consumption, 1 if consumed and 0 if not consumed.
dummy = referral process, 1 if referred via HIS.SANSAN and 0 if not referred via HIS.SANSAI.
To ensure accuracy, we examined 723 observations for outliers. Samples with service times exceeding 400 minutes were removed, as such durations fall outside the normal range for emergency referrals and likely represent exceptional circumstances (see Supplement A). After data cleaning, 594 observations were used for the multiple regression analysis.
Logistic regression was conducted on 630 non-outlier patients: 305 with acute appendicitis, 184 with MHI, 98 with ischemic stroke, and 43 with ESRD. The dependent variable, logit_discharge, was coded as 1 for intensive care patients (“Admit” and “Refer to other hospitals”) and 0 for non-intensive care patients (“Refer back” and “Return home”). Independent variables included sex, age, referral status (resus and emer), comorbidities (Se_dx), SBP, UD, and dummy, as defined above.
The logistic regression model can be written as follows:
where Li is the log of the odds ratio, BXi is a linear function of the independent variables (X) and the coefficient (B), and ui is the error of the unobservable index of sample i.
The estimated coefficients indicate the impact of a one-unit change in an independent variable on the log odds ratio and can be used to estimate the average marginal effect on the likelihood of severe illness upon discharge.
III. Results
1. Users’ Preferences and Perceptions
The survey sample included 53 participants: 20 staff from a mid-level referral hospital (comprising 14 emergency nurses, 1 delivery room nurse, 2 emergency medical staff, 2 emergency medical technicians, and 1 public health officer) and 33 staff from three affiliated referring hospitals (comprising 31 emergency nurses, 1 public health officer, and 1 inpatient affairs officer).
Table 1 summarizes the survey characteristics. Staff from the mid-level referral hospital were predominantly female (95%), mostly aged 21–30 years (60%), with 75% holding a bachelor’s degree, and 70% working as emergency nurses. Staff from referring hospitals were also mostly female (84.80%), had a broader age range (33.33% aged 21–30 years and 30.30% aged 41–50 years), over 96% held higher education degrees, and 93.94% served as emergency nurses. Comprehensive survey results are provided in Supplement A.
Table 2 presents the mean scores and satisfaction levels for HIS.SANSAI. At the referral hospital, respondents reported high overall satisfaction with system capabilities (mean: 7.84) and work processes (mean: 7.72). The highest satisfaction was for the system’s ability to “reduce coordinated time when transferring emergency patients” (mean: 8.00).
In contrast, staff at referring hospitals expressed moderate satisfaction with system capabilities (mean: 5.90) and work processes (mean: 6.22). Although they recognized the system’s ability to “reduce mistakes in patient information transfer” (mean: 7.03), they reported low satisfaction with the ease of recording data (mean: 4.27). These findings suggest that HIS.SANSAI has a more positive impact on user satisfaction at referral hospitals compared to referring hospitals.
At the referral hospital, male respondents reported higher satisfaction than females regarding system capability (mean: 8.6 vs. 7.8) and work processes (mean: 9.00 vs. 7.65). In contrast, at referring hospitals, male and female satisfaction levels were similar for both system capability (5.72 vs. 5.93) and work processes (6.13 vs. 6.23).
Regarding age, older respondents at the referral hospital reported higher satisfaction with both system capabilities and work processes than younger respondents. However, at referring hospitals, respondents aged 31–40 and 41–50 years reported lower satisfaction compared to those aged 21–30 and over 50 years (see Supplement A).
2. Health Outcomes
Table 3 shows changes in referral patterns over three fiscal years, with increasing utilization of the HIS.SANSAI system for various conditions. In fiscal year 2021, referrals via HIS. SANSAI exceeded those through the previous system (non-HIS.SANSAI), indicating greater reliance on the new system.
Multiple regression analysis results (Table 4) revealed that HIS.SANSAI significantly impacted service time—the duration from referral receipt to emergency department discharge. Across all selected conditions, HIS.SANSAI was associated with an 18.21-minute increase in service time after controlling for age, comorbidities, and condition severity. This effect was particularly pronounced for acute appendicitis cases, which experienced a 19.88-minute increase, while referral times for MHI and ischemic stroke showed no significant change.
Table 4 shows no significant overall difference in service time between male and female patients after controlling for other factors. However, further analysis revealed that HIS. SANSAI increased service time for male patients by 22.15 minutes compared to the previous system, with no significant change for females. Moreover, the system significantly extended service time by 18.01 minutes for patients aged 21–59 years and 19.36 minutes for those 60 and older, while no significant change was observed for patients aged 20 or younger.
Logistic regression analysis (Table 5) assessed the impact of HIS.SANSAI on treatment outcomes after emergency department discharge. Across all selected diseases, HIS. SANSAI was associated with a 20.27% lower likelihood of admission or referral for intensive care, after controlling for other factors.
However, disease-specific analysis showed a statistically significant effect only for ischemic stroke cases, where patients referred through HIS.SANSAI had a 13.16% lower likelihood of requiring intensive care compared to those referred via other systems.
IV. Discussion
The HIS.SANSAI emergency referral system, a web-based application, was introduced to improve the transfer and recording of emergency patient data between mid-level referral and referring hospitals in Chiang Mai, Thailand. Previous studies highlight potential benefits of electronic referral systems, such as reduced follow-ups, improved quality, shorter waiting times, better access, enhanced patient confidentiality, and stronger integration between primary and specialized care, thereby improving efficiency [10,11]. However, their impact on health outcomes remains unclear [11].
Our survey found that referral hospital staff were satisfied with HIS.SANSAI’s capabilities, while referring hospital staff expressed lower satisfaction, particularly regarding ease of use. Similarly, referral hospitals rated the work processes more positively than referring hospitals, which rated them only moderately satisfactory. These results are not surprising given that the survey was conducted shortly after HIS.SANSAI’s implementation, when referring hospital staff were using both the old and new systems and had limited experience. The study suggests prioritizing training initiatives, developing a comprehensive user manual, and implementing an awareness campaign to emphasize the program’s benefits.
To assess HIS.SANSAI’s efficiency, we focused on reducing emergency department service time (or length of stay [LOS]) and improving health outcomes post-discharge. LOS is a common indicator of emergency department performance [12] and is closely related to service quality [13]. Research shows an inverse correlation between LOS and emergency room occupancy, linked to improved service quality metrics [13]. Service quality is influenced by patients’ perceptions, reduced delays in care—including consultations and pending investigations—and better survival outcomes [13–15]. HIS.SANSAI provides comprehensive emergency patient data, potentially reducing redundant tests and service times. However, the evidence does not support a reduction in service time; instead, it increased for acute appendicitis cases and remained unchanged for MHI and ischemic stroke. The increase in service time for acute appendicitis after HIS. SANSAI’s 2020 implementation [4] may reflect new practice standards requiring patients to fast for 2–8 hours before an X-ray prior to surgery. In 2021, the Royal College of Anesthesiologists of Thailand published a guideline on preoperative fasting in the Thai Journal of Anesthesiology, aligning with American Society of Anesthesiologists guidelines established in 2000 and updated in 2017 to minimize litigation risk [16–18].
Due to data limitations, this study may have omitted factors influencing emergency department service time. Bickell et al. [19] found that higher staffing levels correlate with shorter diagnosis times and overall service time. Additionally, diagnostic impressions and computed tomography (CT) scan requirements also affect service time [4].
Regarding health outcomes, electronic referral systems are expected to reduce diagnostic errors and improve outcomes [20], by providing comprehensive patient data that enhances clinical decision-making in emergency departments [21] and facilitates better treatments, especially before reaching the hospital [22]. Our study found that, after emergency department discharge, patients with ischemic stroke referred via HIS.SANSAI were 13.16% less likely to be admitted than those referred by the former system, reflecting the positive impact of HIS.SANSAI.
A primary objective of HIS.SANSAI is to enable referral hospital physicians to screen and filter non-urgent cases, allowing mild cases to be treated at referring hospitals. HIS.SANSAI also facilitates consultations, enabling more patients to receive care at referring hospitals and ensuring that referral hospitals handle only urgent cases, potentially increasing admission rates. However, our study found a different outcome: ischemic stroke patients referred via HIS. SANSAI had a 13.16% lower admission rate after emergency care discharge compared to those referred by the previous system. This unexpected result likely reflects HIS.SANSAI’s efficiency in streamlining transfers, which increased referral volume for both urgent and non-urgent cases, thereby lowering admission rates. Data from a mid-level referral hospital indicated a 35% increase in referred patients from 2019 to 2020 and a 15% decrease in patients referred to higher-level hospitals [23,24]. Moreover, comprehensive data from HIS. SANSAI allowed the referral hospital to better prepare for incoming patients by arranging essential procedures such as CT scans, enabling some patients to be referred back to their initial facilities [25].
These findings suggest that HIS.SANSAI improves referral efficiency and optimizes resource utilization at referral hospitals by reducing unnecessary admissions. However, it also challenges referral volume management, underscoring the need to continuously refine HIS.SANSAI protocols to ensure effective allocation of healthcare resources.
A major challenge of our study was its narrow scope due to data constraints. During the 2019–2021 implementation of HIS.SANSAI, incomplete system data required integration with paper records and the hospital’s computer system, limiting the dataset to cases with successful data linkage. Consequently, the sample size was relatively small (305 observations for acute appendicitis, 184 for moderate head injury, 98 for ischemic stroke, and 43 for end-stage renal disease), potentially limiting the generalizability of our findings. Additionally, the study focused solely on select aspects of user preferences, service time, and discharge outcomes compared to the former system. Available data did not include post-discharge mortality or the implicit benefits of safeguarding patient data privacy. A more exhaustive study might reveal additional benefits, such as expedited access to diagnostic tools like CT scans, reduced waiting time costs, financial savings from fewer hospital visits, and minimized coordination time. Lower admission rates or more referrals to lower-level hospitals may further reduce expenses. Future research will explore the rationale and financial impact, and further validate HIS.SANSAI’s operational efficiency and improvements in health outcomes through economic evaluations, including cost-benefit and cost-effectiveness analyses.
XML Download