Design, implementation, and evaluation of neonatal abstinence syndrome care management web application: a multimethod study

Abstract

The Finnegan scoring tool is the clinical standard for assessing neonatal abstinence syndrome (NAS), but manual calculation and documentation pose challenges. Digital decision-support systems may enhance accuracy, streamline workflows, and improve therapeutic interventions. This study aimed to design, implement, and evaluate a neonatal abstinence syndrome decision support web application and record the care management based on the Finnegan scoring system. This study was conducted using a multi-method approach and included design, implementation, and evaluation phases. In the system design phase, the waterfall model was used. The second and third phases of the study, implementation and evaluation, were conducted as a randomized controlled trial. A total of 142 infants diagnosed with neonatal abstinence syndrome were randomly divided into two groups: the intervention group, which used a web-based application that automatically calculates the Finnegan score and provides therapeutic advice, and the control group, which relied on manual calculation and routine care. Data were collected from birth to 5 days post-natal age. Although Finnegan scores decreased significantly over time in both groups (p < 0.001), the between-group comparison showed no statistically significant difference in mean score reduction (p = 0.18). However, clinically meaningful differences were observed: documentation adherence was significantly higher in the intervention group (96.7% vs. 91.4%, p < 0.001), and the frequency of pharmacological interventions was 36% lower in the intervention group compared to controls (p < 0.001). User satisfaction scored ≥ 3.5/5 across all domains. The NAS care management web application improved documentation accuracy, reduced medication interventions, and demonstrated high user satisfaction. It offers a clinically viable tool to standardize and enhance NAS care in neonatal settings.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-54221-9.

Introduction

As a global crisis, the number of addicted pregnant women and its impact on their fetuses is increasing at an alarming rate¹. The World Health Organization reports that the incidence of drug-related complications in pregnant mothers and their infants is between 5 and 10%². This widespread prevalence of poly-substance abuse during pregnancy has created a global health crisis, with prescription drug misuse nearly doubling over the past decade⁴. Over the past decade, the incidence of neonatal abstinence syndrome has increased almost fivefold⁵. Among all types of drug/substance abuse, drug use during pregnancy is associated with numerous adverse outcomes for mothers and infants, such as increased risk of maternal death, placental abruption, preterm delivery, hospitalization for more than seven days, premature birth, low birth weight, congenital anomalies, and increased risk of intrauterine and postnatal growth retardation⁵. In fact, many of these drugs can cross the placenta and affect the developing fetus and the newborn after birth¹. Statistics show that 30 to 80% of newborns exposed to narcotics during pregnancy require intensive care and treatment⁶. Furthermore, the incidence of neonatal abstinence syndrome has increased almost fivefold over the past decade⁷. While this syndrome is primarily associated with maternal substance abuse, certain pharmacological agents and their derivatives may also induce similar withdrawal symptoms in infants⁸. While the exact pathophysiology is not fully understood, the onset and severity of withdrawal in neonates are highly variable and influenced by a complex interplay of maternal, genetic, and environmental factors^8–10. Clinically, affected infants exhibit a range of central nervous system, autonomic, and gastrointestinal symptoms that can be severe and potentially life-threatening, necessitating prompt and accurate clinical assessment^11,12.

In this context, an important and sometimes less well-considered point is that all infants who have been exposed to drugs in utero or who may have been exposed should be assessed with a standard tool such as the Finnegan tool to determine the severity of symptoms and, consequently, appropriate care management¹³. In this regard, there are two major challenges facing nurses and physicians in neonatal intensive care units. The first challenge is accurate and consistent scoring of drug withdrawal symptoms based on the standard Finnegan score, which is difficult and time-consuming to calculate, and the second serious challenge is providing and adjusting care and medications based on the Finnegan score and recording all of these, which is also difficult and time-consuming. If nurses and physicians want to manage these two important challenges effectively, they must either regularly consult protocols and reference books, which is very time-consuming, or rely on their memory, which also increases the likelihood of mistakes and errors. This is occurring in a situation in which we are facing a shortage of professional labor and a high workload for nurses and specialists in the NICUs. Therefore, the use of electronic and intelligent clinical decision support systems has been recommended¹⁴.

Clinical decision support systems, as a type of electronic system used in clinical care, have applications in the care process¹⁵. The need for decision support systems is undeniable, because they are most active in the field of health care, requiring fundamental and important decision-making processes¹⁶. Clinical decision support systems are designed to actively support the care team in the process of reviewing and updating new clinical evidence¹⁷. They also provide dynamic predictions and allow decision-makers to re-evaluate predictions in the light of new clinical information¹⁸. In recent years, clinical decision support systems have become increasingly important for ensuring patient safety and supporting all stages of clinical decision-making¹⁹. Studies on the effectiveness of clinical decision support systems have reported improved process outcomes such as increased acceptance of relevant rules and principles, better documentation of care management, improved teamwork and interpersonal communication, and cost savings²⁰. Also, recent studies have shown that the use of these systems can improve the knowledge of healthcare providers and affect the quality of documentation and patient safety^21,22.

Given the challenges associated with manual Finnegan scoring and the need for protocol-aligned care management, digital clinical decision support systems offer a promising solution to standardize and streamline NAS care. Therefore, this study aimed to design, implement, and evaluate a web application for decision support and care management of infants with neonatal abstinence syndrome based on the Finnegan scoring tool. We hypothesized that, compared to routine manual calculation, the integration of this web application into clinical workflow would result in a reduction in unnecessary pharmacological interventions as the primary clinical benefit. Furthermore, we hypothesized that the application would positively influence secondary outcomes, including increased documentation adherence, a more consistent downward trend in Finnegan scores over the 5-day period, and a high level of user satisfaction among the nursing staff.

Methods

This multi-method study was conducted between January and August 2025, encompassing three phases: system design, implementation, and evaluation. In the system design phase, the waterfall model²³, which is one of the most popular methods for software engineering, was used (Fig. 1).

Fig. 1.

Web application development steps based on the waterfall model.

Phase 1: Web application design

The application design process followed a structured, stepwise approach to ensure clinical accuracy and usability:

• Evidence extraction: To ensure the application was evidence-based and reproducible, the scientific information and decision-support logic were extracted from validated literature. Specifically, the scoring items were derived from the original Finnegan scales²⁴. Furthermore, the pharmacological and non-pharmacological care protocols integrated into the application were derived from standard neonatal clinical guidelines, specifically the Canadian Paediatric Society Practice Guideline for the management of neonatal abstinence syndrome²⁵.

• Wireframing: The conceptual framework of the application was initially drafted in a Word document. This included mapping out the user interface (application pages), defining the required data fields for each page, establishing the navigational flow between screens, and specifying the output formats.

• Programming: The wireframed model was converted into a functional web application in collaboration with a computer programming expert.

• Pilot testing and refinement: The beta version was offered to three volunteer, experienced nurses for a one-week pilot test to identify any bugs or usability issues. Using their feedback, and with the assistance of the programmer, the application was modified, and the final version was prepared¹⁴.

Phase 2: Web application implementation

Study design and setting

Implementation phase in this study was conducted as randomized controlled trial with two groups (control and intervention), in a Neonatal Intensive Care Unit (NICU) at Mahdiyeh hospital in Tehran, Iran. This setting as a national referral center for high-risk pregnancies and neonates, provides comprehensive tertiary care and serves a socioeconomically and culturally diverse population. Although this phase employed a randomized design, the study is officially classified as quasi-experimental by the Iranian Registry of Clinical Trials (IRCT) because the intervention does not involve direct manipulation of the participants’ physiological state.

Participants and eligibility criteria

The participants in this study included neonates and nurses in NICU units at Mahdiyeh Hospital, a referral center in Tehran, Iran. A consecutive sampling method was utilized to select neonates diagnosed with substance/drug withdrawal syndrome at birth. Eligibility was determined by a set of predefined inclusion and exclusion criteria. Inclusion criteria included infants within the first 5 days of life who did not have acute surgical complications or were not in the end stages of life. Participants who met the initial criteria were enrolled. However, during the 5-day study period, if any infant became clinically unstable for reasons unrelated to NAS (e.g., requiring resuscitation or advanced critical care), they were withdrawn from the study to prevent confounding the results. The study process follows the CONSORT flow diagram (Fig. 2). The completed CONSORT checklist is available in the Supplementary File. Additionally, nurses employed in the neonatal wards and NICUs, who had at least one year of experience and were directly responsible for caring for infants with drug withdrawal syndrome, were recruited using a census method. Written informed consent was obtained from all participating nurses, as well as from the parents or legal guardians of the enrolled neonates.

Fig. 2.

CONSORT flow diagram.

Study outcomes

The outcomes were defined and assessed using specific instruments tailored to the study objectives:

Primary Outcome: The primary clinical outcome was the proportion of infants who required pharmacological interventions during the 5-day study period. This binary outcome (medication administered vs. no medication) was objectively documented by the nursing staff. It was hypothesized that this proportion would be significantly lower in the intervention group compared to the control group.

Secondary Outcomes: The secondary outcomes included:

1. Documentation Adherence: Defined as the percentage of scheduled Finnegan assessments actually completed and recorded over the 5 days. We hypothesized that automated logging would result in higher adherence in the intervention group.

2. Longitudinal Finnegan Scores: The trend and changes in the severity of neonatal abstinence syndrome over time. We hypothesized that the application would support a more consistent downward trend in scores.

3. User Satisfaction: The level of satisfaction and acceptance of the web application among the nursing staff. We hypothesized that the application would achieve a high level of acceptance and satisfaction among the nursing staff.

Measurement instruments

To measure these outcomes, the following instruments were used:

1. Finnegan Scoring Tool: A modified version of the Finnegan Tool that includes 21 items. The variables under investigation were categorized into three broad groups: (1) Central nervous system disorders, including 8 items on (crying, sleep, Moro reflex, tremor, muscle tone, myoclonic jerks, and seizures); (2) Metabolic, vasomotor, and respiratory alterations, including 8 items on (sweating, hyperthermia, frequent yawning, skin mottling, nasal congestion, sneezing, flaring of the nostrils, and respiratory rate); and (3) Gastrointestinal dysfunctions, including 5 items on (excessive sucking, poor feeding, milk regurgitation into the mouth, and projectile vomiting)^23,24. Nurses performed Finnegan assessments and scoring three times a day during each shift.

2. Mobile Application Rating Scale (MARS): Used to measure the “User Satisfaction” outcome. This standardized 23-item scale evaluates five domains: Engagement (5 items), Functionality (4 items), Aesthetics (3 items), Information (7 items), and Subjective Quality (4 items). Items are scored on a 5-point Likert scale, with higher scores indicating greater user satisfaction²³. This was completed by the nurses at the end of the 5-day study period.

3. Demographic and Clinical Questionnaire: A researcher-made checklist was used by the attending nurses to extract baseline data from medical records, including maternal characteristics (age, type of substance use) and neonatal characteristics (gestational age, birth weight, gender). To ensure its accuracy, the content validity of the questionnaire was confirmed by a panel of 10 nursing and neonatal experts, yielding a Content Validity Ratio (CVR) of 0.85 and a Content Validity Index (CVI) of 0.90. Its reliability was also established through a pilot study, demonstrating acceptable internal consistency with a Cronbach’s alpha of 0.82.

Sample size calculation

The sample size was calculated by utilizing the proportion of infants requiring pharmacological interventions as the primary clinical outcome. Using G*Power software (version 3.1.9.7) and a comparison of two proportions formula, we assumed a 30% absolute difference in medication intervention rates between the control and intervention groups (based on preliminary clinical estimates of 40% vs. 10%). Considering a significance level (α) of 0.05, a 1:1 allocation ratio, and a statistical power of 0.90, the required sample size was calculated to be 64 infants per group. To account for potential dropouts, this was increased to 72 infants per group (total N = 144). A higher power of 0.90 was deliberately chosen to minimize the risk of Type II error, given the clinical vulnerability of the neonatal population.

Randomization and allocation concealment

A total of 144 eligible participants were randomly assigned in a 1:1 allocation ratio to either the intervention (n = 42) or control (n = 42) group using a computer-generated randomization sequence. Allocation concealment was ensured via sequentially numbered, sealed opaque envelopes prepared by an independent researcher not involved in participant recruitment or assessment. The recruitment and randomization process continued until this number was reached to ensure the target sample size for analysis could be met, as detailed in the CONSORT flow diagram (Fig. 2).

Intervention group

Participants allocated to the intervention group utilized a custom-built, web-based clinical decision support application. The platform automatically calculated the total Finnegan score based on nurse input, generated corresponding pharmacological and non-pharmacological care recommendations, and provided visual trend graphs of the infant’s status over time. Accessible via three user levels (nurse, supervisor, and administrator), the system allowed the care team to deliver protocol-aligned interventions. During the 5-day study period, nurses in this group completed all assessment steps with the same frequency as the control group (three times daily) using the application’s automated scoring and assistant suggestions, with the system automatically recording all process data in real-time for subsequent analysis.

Control group

Participants in the control group underwent standard clinical care according to the institutional protocol. During this phase, the steps of the newborn’s Finnegan scoring process and the care plan from birth to 5 days of age were calculated manually by the nurse three times a day (once per shift) and recorded in the file.

Follow-up period and data collection

The data collection and evaluation period spanned the first five days of life for each enrolled neonate. During this timeframe, clinical data were collected to evaluate both the implementation and the effectiveness of the application. The primary outcome data (proportion of pharmacological interventions) and secondary outcomes (documentation adherence and Finnegan score trends) were extracted and compared between the intervention and control groups. Implementation metrics, including system usage logs and user satisfaction, were also recorded at the end of the 5-day period. The study methodology and participant flow are documented in accordance with the CONSORT 2010 guidelines (Fig. 2), and the protocol received ethical approval from the appropriate institutional review board.

Statistical analysis

All statistical analyses were performed using SPSS software (version 26.0, IBM Corp., Armonk, NY, USA). A two-sided p-value of less than 0.05 was considered statistically significant for all analyses.

Descriptive statistics were used to summarize the demographic and clinical characteristics of the neonates and nurses. The normality of continuous variables was assessed using the Shapiro-Wilk test. Normally distributed data are presented as mean ± standard deviation (SD), while non-normally distributed data are presented as median (interquartile range, IQR). Categorical variables are presented as frequencies and percentages.

To compare the baseline demographic characteristics between the intervention and control groups, independent samples t-tests (for normally distributed continuous data), Mann-Whitney U tests (for non-normally distributed data), and Chi-square tests (for categorical variables) were employed.

For the effectiveness evaluation (Phase 3), the Mann-Whitney U test was used to compare the mean Finnegan scores between the two groups at specific time points (days 3 to 5). To analyze the longitudinal trend of Finnegan scores over the 5-day period, a Generalized Estimating Equation (GEE) model was utilized to account for the repeated measures and within-subject correlations over time. Additionally, an independent samples t-test was used to compare the overall mean decrease in Finnegan scores between the two groups.

To compare the primary clinical outcome, the proportion of pharmacological interventions between the intervention and control groups, a Chi-square test was utilized. Furthermore, to evaluate the implementation outcomes, documentation adherence rates (percentage of completed assessments) were compared between the groups using the Chi-square test. Finally, descriptive statistics (mean and standard deviation) were used to present the user satisfaction scores derived from the Mobile Application Rating Scale (MARS).

Ethical consideration

This study was approved by the Ethics Committee of Tehran University of Medical Sciences (Code: IR.TUMS.FNM.REC.1403.114) and adhered to the Declaration of Helsinki. Written informed consent was obtained from all participating nurses, and parental consent was secured for neonatal data collection. All procedures followed institutional and international ethical guidelines.

Clinical trial number

This study was conducted as a quality improvement initiative and did not require clinical trial registration under institutional guidelines.

Phase 3: web application evaluation

The evaluation of the web application was conducted across two distinct dimensions: (1) comparing the clinical effectiveness of the Finnegan scoring and care management between the intervention and control groups; and (2) measuring the level of nurses’ satisfaction with the application using the Mobile App Rating Scale (MARS).

Results

The results are reported in three sections: web application design and implementation, as well as its evaluation.

Results of phase 1: Web application design

Users log in to the program via the address http://finnegan.ehms.ir by registering, receiving a username and password, and receiving a confirmation email. The program is designed in Persian and English and has three user levels (nurse, supervisor, and admin) (Web Appendix a).

In this application, general information about the Finnegan tool and interventions is entered by the supervisor, and information about the infants is entered by the nurses. Once a neonate’s details have been recorded, it is possible to retrieve or edit the information by first and last name. (Web Appendix b).

By selecting the neonate’s name or entering information, you can enter the page to record information related to the Finnegan score. By selecting the options, the system calculates the total score and displays it at the top of the page (Web Appendix c).

Next, pharmacological and non-pharmacological interventions are displayed, and the user can select one or more items, all of which are automatically recorded in the program. Nurse users can also view a graph of changes in Finnegan scores for each infant over time by selecting the infant’s name and clicking on the graph option. This graph is designed to display the total Finnegan score and each dimension score in different colors over time (Web Appendix d).

Also, the report on Finnegan scores and the interventions for the patient can be exported in the reporting section and can be seen on the same messages page. Finally, the user satisfaction questionnaire can be viewed and completed in the same program, and the output of the results can be reported (Web Appendix e).

Results of Phase 2: Web application implementation

A total of 142 neonates diagnosed with neonatal abstinence syndrome completed the study and formed the final analyzed cohort (N = 70 in the intervention group, N = 72 in the control group). This final number was reached after the initial randomization, with subsequent withdrawals (n = 2 in the intervention group due to clinical deterioration unrelated to NAS) detailed in the CONSORT flow diagram (Fig. 2). Baseline demographic and clinical characteristics did not differ between the groups (Table 1).

Table 1.

Maternal and neonatal demographic and clinical variables.

Variables	Intervention (N = 70) n (%)	Control (N = 72) n (%)	P-value
Gender
Male	34 (48.57)	38 (52.77)	0.723
Female	36 (51.43)	34 (47.23)
Department
Infants	48 (68.57)	58 (80.55)	0.246
NICU	22 (31.43)	14 (19.45)
Delivery type
Cesarean	22 (31.43)	24 (33.34)	0.864
Natural	48 (68.57)	48 (66.66)
Substance use
Narcotics	34 (48.57)	28 (38.89)	0.534
Medication	14 (20.00)	16 (22.22)
Both	22 (31.43)	28 (38.89)
Weight (g)	2934 ± 405	2950 ± 683	0.806
Gestational Age (weeks)	38.83 ± 1.382	39.20 ± 1.232	0.239
Maternal Age (years)	29.79 ± 5.008	29.86 ± 6.043	0.960

Regarding the adoption and adherence to the web application, system logs were analyzed. Data showed that the application was utilized to record the Finnegan score in 96.7% of the expected shifts in the intervention group, compared to 91.4% documentation adherence in the control group over the 5 days. This 5.3% difference was statistically significant (p < 0.001). The remaining 3.3% in the intervention group represents specific shifts where the nursing staff did not use the application (non-adherence).

Results of phase 3: Evaluation of the web application

The Mann-Whitney test showed that the mean Finnegan score on days 3 to 5 was higher in the control group than in the intervention group. However, the independent t-test indicated that the overall mean decrease in Finnegan scores (2.79 in the intervention group vs. 2.46 in the control group) was not statistically significant (p = 0.422).

To further analyze the longitudinal trend of scores over time, a Generalized Estimating Equation (GEE) model was utilized. The GEE model demonstrated that Finnegan scores decreased significantly over time across both groups (p < 0.001), though the difference in the rate of decrease between the two groups did not reach statistical significance (p = 0.183).

Regarding the primary clinical outcome, the chi-square test revealed a significant difference in pharmacological interventions. The proportion of drug interventions used in the control group was 36% higher than in the intervention group (p < 0.001), indicating that the application significantly reduced unnecessary medication use. Furthermore, implementation evaluation metrics showed that documentation adherence was significantly higher in the intervention group (96.7% of shifts) compared to the control group (91.4% of shifts) over the 5 days (p < 0.001).

User satisfaction evaluation

User satisfaction with the web application was evaluated using the Mobile Application Rating Scale (MARS). The total user satisfaction score across all dimensions was above 3.5 out of 5 points, indicating overall positive feedback. The highest-rated domain was “Information” (mean = 4.24 ± 0.35), and the lowest-rated domain was “Subjective Quality” (mean = 3.09 ± 0.55). Detailed satisfaction scores for all domains are presented in Table 2.

Table 2.

MARS questioner scores results.

MARS	Rank	Mean	Standard deviation	Minimum	Maximum
Attractiveness	2	3.9043	0.34060	3.00	4.80
Performance	1	4.1232	0.34199	3.50	4.75
Beauty	1	4.1063	0.41022	3.00	5.00
Information	1	4.2435	0.34788	3.40	5.00
Subjective quality	3	3.0906	0.54734	1.75	4.50
Total Score	–	3.8936	0.18914	3.43	4.25

Discussion

Symptom management and neonatal abstinence syndrome care management have been of interest for many years, and in this regard, policies have been made²⁶, many clinical care guidelines have been developed²⁷, and various assessment tools, symptom management models, and treatment strategies have been proposed^28,29. However, healthcare providers are constantly faced with difficulties due to the difficult and complex assessment process. Therefore, it is necessary to provide a solution to facilitate critical care, such as decision-support applications. Therefore, in this study, a decision-support and care management web application for infants with neonatal abstinence syndrome was designed, implemented, and evaluated based on the Finnegan scoring tool.

The decision-support system for symptom management and care of infants with neonatal abstinence syndrome was designed as a web application with the aim of facilitating usability and widespread access. This application, which can be run on any Internet-connected device, including mobile phones, laptops, and computers, and does not require installation, removes common barriers to use and provides easy access to information and services. This feature can help attract more users. This feature is very valuable, especially in emergency situations and the need for immediate advice. Designing web applications with these features emphasizes the importance of facilitating access to technology in neonatal care. This focus on user needs and the effort to create an efficient tool aligns with the approach used in our previous project¹⁴. In this web application, by entering information about the neonate, accurate assessment and rapid scoring are performed, providing reliable data for treatment decisions. A graph of changes in the Finnegan score for each neonate and the score of each dimension can be viewed in different colors over time and is saved for later comparisons, which helps create a continuous monitoring process.

This visual analysis helps caregivers identify specific patterns in the infant’s condition, enabling them to make timely treatment decisions that ultimately improve the quality of care. However, the success of this method depends on the accuracy of data entries, and any errors in entering information can have a negative impact on the results³⁰. Therefore, it is of particular importance to pay attention to the appropriate training of the care team in the use of this web application and to ensure the quality of data.

In addition, the suggestion of pharmacological and non-pharmacological interventions tailored to the scoring results makes this application a key device in treatment planning. Immediately after scoring and determining the status of the newborn based on the Finnegan scoring tool, the application intelligently offers a set of specific, predefined interventions. These strategies can include pharmacological interventions or non-pharmacological approaches that target the infant’s condition. Tailoring these strategies to the specific condition of each infant provides a new level of individualization of treatment that specifically meets the goal of caring for and promoting the infant’s health status^31,32. This way, care providers can quickly make the right treatment and care decisions using efficient data and accurate analytics. This method not only helps increase transparency in the treatment process but also allows for faster recovery of infants and reduces side effects caused by inappropriate or late treatments or care.

Based on the findings of the present study, the use of this application on the third to fifth day after birth reduced the mean Finnegan score and improved it in infants participating in the intervention group compared to the control group. Also, the rate of change in Finnegan scores decreased significantly over the 5-day period, indicating the effectiveness of the application and the consistency of its positive effects during the acute phase of care. It is important to note that the gradual decrease in Finnegan scores observed in both groups over time is primarily driven by the natural physiological clearance of the substance from the infant’s body following birth²⁵. However, clinical guidelines emphasize that timely and accurate clinical management is still essential during this period to prevent severe complications and support the infant through the withdrawal process²⁶. Therefore, while the natural course dictates the overall downward trend, the application’s role was to ensure that this process was supported by accurate, protocol-aligned decision-making. This finding is probably due to the good experience of nurses and physicians in caring for these infants. Also, given that Finnegan scoring using the designed application does not require manual calculation by the nurse, the results showed that the number of patients assessed was higher in the group that used the application, meaning that better monitoring of the newborns was performed. Although the available studies are limited for better comparison, in line with the results of the present study is a study that designed a decision-support platform for determining morphine dosage in neonates³³. Another study that developed and designed a clinical decision-making application for opioid prescribing can also be mentioned (34). Increased frequency of assessment and more accurate monitoring can lead to early detection of changes in the infant’s condition and faster and more effective treatment and care decisions. In addition, the reduced workload resulting from the elimination of manual calculations allows nurses to devote more time to direct care of infants, which in turn can help improve the overall quality of care. This is also supported by previous studies^14,30,31,33.

The results of our study regarding the adjustment of medication therapy with Finnegan score between the intervention and control groups were significant. In this study, the proportion of medication interventions used in the control group was significantly higher than in the intervention group. A possible explanation for this difference could be that in the intervention group, infants were assessed more frequently and with higher accuracy due to the ease of use of the application. More accurate assessment helped the care provider team to correctly identify the need for medication interventions based on the actual Finnegan score and avoid unnecessary medication administration. It is also possible that manual calculations of Finnegan scores in the control group were less accurate due to fatigue or human error, leading to over-prescription of medication. In any case, given the high vulnerability of infants to adverse medication reactions (35), a cautious approach to prescribing medication only when necessary is of particular importance (36).

Overall, the main users of this app were NICU nurses, and the survey results indicated a high level of satisfaction with using the application. The highest level of satisfaction was related to the “information” dimension, which is to be expected given the focus of the application’s design and implementation on this aspect. This finding emphasizes the importance of easy and fast access to information and shows that this decision-support system, especially in the field of improving users’ knowledge and awareness, can have a positive impact on their performance, a finding that has been emphasized by several studies^31,32.

Also, the lowest level of satisfaction among nurses in the present study was assigned to the “subjective quality” dimension. Since this dimension asked about the willingness to pay for this program and in most cases, users were willing to use the program without cost, this result was not unexpected. This finding is also consistent with a recent review that evaluated mobile applications for hemodialysis using the Mobile Application Rating Scale (MARS)²⁵. Furthermore, in another study that evaluated and analyzed smartphone applications in people who use alcohol and illicit drugs, not only was the average score for this dimension very low, but also a low average was obtained in all dimensions based on MARS scoring, and users were not very satisfied with using the available applications (37). Therefore, it is very important to pay attention to user satisfaction in these types of applications, because satisfaction leads nurses to consider using the application as a useful and efficient tool that not only does not increase their workload but also helps improve the quality of care and save time¹⁴. Ultimately, these findings emphasize that investing in the development and deployment of decision support systems can lead to improved quality of care, increased staff job satisfaction, and improved infant health.

Conclusion

The present study shows that using neonatal abstinence syndrome care management web applications greatly helps nurses. This system helps nurses make more precise, protocol-aligned clinical decisions and provide more effective care by reducing the average Finnegan score, reducing the need for medication interventions, increasing the frequency of close monitoring, and providing comprehensive information. High user satisfaction also shows that a user-friendly design tailored to their needs plays an important role in the acceptance and success of projects implementing new technologies in the health field. Overall, these results confirm that decision support systems can serve as an efficient tool in improving the quality of care and increasing the satisfaction of healthcare professionals, especially in situations where we are faced with resource and time constraints.

Limitations

Despite the app’s promising performance, several limitations should be acknowledged. First, during the infants’ hospitalization, all infants were breastfed according to hospital guidelines. It is possible that variations in the serum level of the drug/substance in the mother’s blood affected its concentration in breast milk, potentially confounding the withdrawal scores to some extent. Second, there is a possibility of human error in documentation, as nurses may have occasionally missed recording some of the administered pharmacological and non-pharmacological interventions. Although exact frequencies were not measured, this limitation was likely more prevalent in the control group’s manual records compared to the automated logging in the intervention group. Third, the small number of nurses participating in the user satisfaction evaluation (n = 14) limits the generalizability of the satisfaction scores. Finally, regarding the study design, a significant limitation is the inability to blind the nursing staff to the group allocation, which is inherent to implementation trials involving digital health tools. The presence of the research team and the heightened focus on Finnegan scoring likely created a Hawthorne effect. It is highly probable that the control group received a higher quality of care, more closely aligned with standard protocols, than they would have under absolute routine conditions. Consequently, this improvement in the control group’s baseline performance may have narrowed the gap between the two groups, leading to an underestimation of the true clinical effectiveness of the web application.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

RB and MV spearheaded the research design and conceptual framework. They developed the study’s multimethod methodology and supervised the technical architecture and design of the web application. Additionally, they managed project coordination, secured funding, and provided overall supervision. MB authored the initial draft of the manuscript and later revised it for technical precision and scientific accuracy. EEG and AG conducted clinical evaluations, user needs assessments and contributed to the design of tailored educational materials for the application’s instructional components. MV performed advanced statistical analyses of evaluation data, ensured methodological rigor, and supported resource curation. All authors critically reviewed the manuscript, provided intellectual input, and approved the final version. MV acted as the corresponding author, overseeing correspondence and revisions.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Data availability

The designed web application can be accessed at the http://finnegan.ehms.ir address in any type of browser. The admin’s web application username and password will be sent to you to view different parts of the application.When you enter the web application by clicking on the flag icon, the language of the web application will be changed to English. You can view all parts of the web application based on what is explained in the article. Of course, because this web application was designed and implemented in Iran, the information about the patients is included in the Persian part of the web application. The statistical data was sent to you as part of the research data.

Declarations

Competing interests

The authors declare no competing interests.