This quantitative study has a non-experimental, or observational, cross-sectional design. This study used secondary data derived from the metadata of the UI Library collection, especially the undergraduate theses, master’s theses, and dissertations of students from the Faculty of Medicine (FK), Faculty of Dentistry (FKG), Faculty of Public Health (FKM), Faculty of Nursing (FIK), and Faculty of Pharmacy (FF) published from 2011–2020 ( Rahmi, 2023). The bibliometric analysis performed in this study adopted the analytical steps of Bradford’s Law regarding the distribution of scientific papers, namely (1) the number of published final assignments, (2) the viewing of sources by faculty, study program, and type of work to identify those that produce the most publications; and (3) the creation of groups based on the topics discussed.

The UI Library Automation and Digital Archive (LONTAR) automation system obtained secondary data in collection metadata. Data were downloaded in March 2022. The data analysis continued until May 2022.

The population used in this research consisted of UI Health Sciences students (Faculty of Medicine, Faculty of Dentistry, Faculty of Public Health, Faculty of Nursing, and Faculty of Pharmacy) who submitted final projects that were published in 2011–2020 and stored in the thesis and dissertation collection of the Universitas Indonesia Library published in 2011–2020 because their final projects are related to SDG 3, Good Health and Well-being. Moreover, the data included in the sample had to meet both inclusion and exclusion criteria. Inclusion criteria consist of the final project for the UI Library collection completed by a UI Health Sciences student (Faculty of Medicine, Faculty of Dentistry, Faculty of Public Health, Faculty of Nursing, and Faculty of Pharmacy); published in the years 2011–2020; and in the form of an undergraduate thesis, master’s thesis, or dissertation. The exclusion criteria are missing data (incomplete); duplication of data; or final work submitted for a specialist or professional program.

In total, there were 13,603 final projects produced by UI Health Sciences students, 10,443 of which were published in the years 2011–2020; 5476 were undergraduate theses, master’s theses, or dissertations, and 2674 of these had missing data (incomplete), duplicated data, or were in the form of specialist and professional program final works.

Secondary data were collected by downloading the metadata from the UI Library’s final project collection of undergraduate theses, master’s theses, and dissertations published in 2011–2020 by UI Health Sciences students through the UI library’s LONTAR automation system. The data included the authors’ names, counsellors’ names, titles (Indonesian and English), faculty, year of publication, type of work, abstract, study program, subject topic, and keywords.

Researchers collected data by searching the back-office automation system LONTAR UI Library. The search strategy was: 1. A search in the LONTAR back office was carried out in the “Processing > Add Data” menu; 2. A keyword search was done in “264b Publisher Name”; 3. The keywords used were the names of the faculties with collections of final assignments related to SDG 3, including “Faculty of Medicine”, “Faculty of Dentistry”, “Faculty of Public Health”, “Faculty of Nursing”, and “Faculty of Pharmacy”; 4. The type of collection searched was the final project collection, including “UI—Undergraduate Thesis (Open)”, “UI—Undergraduate Thesis (Membership)”, “UI—Master Thesis (Open)”, “UI—Master Thesis (Membership)”, “UI—Dissertation (Open)”, “UI—Dissertation (Membership)”, and “Download”; and 5. After the search results appeared, the data were downloaded in Microsoft Excel (.xlsx) file using Excel version 365.

Data processing consisted of data cleaning and data coding. Data cleaning was done by removing missing data (for example, work lacking a title, topic subject), eliminating duplicated data, and correcting writing errors (for example, the year “2104” was updated to “2014”). After the data cleaning process, 3059 final assignments remained, which were then put into categories (coded) according to the 13 targets of SDG 3. The categories used to code the data were: 1. Maternal Health; 2. Baby and Toddler Health; 3. Infectious Diseases; 4. Non-Communicable Diseases and Mental Health; 5. Drug and Alcohol Abuse; 6. Traffic Accidents; 7. Access to Reproductive Health Services; 8. Access to Universal Healthcare; 9. Hazardous Chemicals and Pollution; 10. Tobacco Control (Cigarettes); 11. Drug and Vaccine Development; 12. Health Personnel (HR); and 13. Health Risk Management.

Two “raters” performed the data coding process to improve its reliability: an undergraduate Library Science student and an undergraduate Public Health student. The level of agreement between them serves as an indicator of the quality of the categorisation. Cohen’s Kappa was used to assess the level of agreement between the two raters. The following is an interpretation of the Kappa values ( Warrens, 2015).

0.00–0.20 = Low agreement

0.21–0.40 = Sufficient agreement

0.41–0.60 = Moderate agreement

0.61–0.80 = Substantive agreement

0.81–1.00 = Almost perfect agreement

Cohen’s Kappa was measured using GNU PSPP version 1.5.3 software. The Kappa value deter-mined in the test was 0.52, so the raters showed a moderate level of agreement.

Data analysis was performed through descriptive analysis, science mapping, and multivariate analysis.

The descriptive analysis was conducted to visualise the distribution of the variables studied, including the faculty, publication year, type of work, subject topic, and SDG 3 target. The descriptive analysis was performed using IBM SPSS Statistics Standard 28.0 software and Tableau Public version 2022.1.2.

Science mapping was performed to visualise the co-occurrence network of the relationships between topic subjects in the selected work ( Ding et al., 2014). VOSviewer software version 1.6.18 was used.

Multivariate analysis was performed using a multinomial logistic regression with predictive models. Multinomial logistic regression was used because the dependent variable in this study (SDG 3) was categorical (13 categories). This type of analysis allows a model consisting of several independent variables that are considered the best to predict the occurrence of the dependent variable to be created ( Field, 2018). The multivariate analysis was performed using IBM SPSS Statistics Standard 28.0 software.

The Technical Implementation Unit (UPT) of the UI Library was established on 5th March, 1983, according to the Decree of the Minister of Education and Culture No. 0130/O/1983 regarding the Organization and Work Procedures of Universitas Indonesia Article 4 and Article 131-136, which explains the functions, duties, and organisation of the UI Library UPT. In 1987, the UI Library, formerly the UPT Central Library, moved to a new building on the UI Depok Campus. The Central Library then served to coordinate 12 faculty libraries ( Perpustakaan UI, 2018).

In 2011, the UI Library moved into a new building dubbed “The Crystal of Knowledge”. This building was constructed according to one of the points of the UI’s strategic plan at the time: integration into the field of facilities. The process of integrating collections into the UI Library UPT started in March 2012. The Central Library, Faculty of Humanities Library, Faculty of Engineering Library, Faculty of Mathematics and Natural Sciences Library, Faculty of Nursing Library, and Faculty of Law Library followed. Several other faculties still retain separate libraries but have moved some of their collections to new buildings, namely, the Faculty of Computer Science Library, Faculty of Psychology Library, Faculty of Social and Political Sciences Library, Faculty of Public Health Library, and Faculty of Economic Library ( Perpustakaan UI, 2018).

The UI library aims to collect, select, organise, and provide access to various information and knowledge resources for UI Citizens. Therefore, the UI Library collects and preserves library materials that are intellectual works. In addition, the UI Library follows developments in and adopts information and communication technology to improve its services. This is reflected in the vision of the UI Library for 2015–2019. Specifically, “In 2019, Universitas Indonesia Library will become a reference for national and regional university libraries sourced from the intellectual works of UI Citizens, and quality e-resources owned and supported by modern facilities owned by Universitas Indonesia.” The UI Library’s mission is ( Perpustakaan UI, 2015): (1) to provide quality access for UI residents and the public to information and knowledge resources, with excellent service based on information and communication technology; (2) to support the research conducted by UI Citizens by providing information and knowledge resources based on information and communication technology; (3) to provide community services for the utilisation of information and knowledge resources mainly via UIANA collections based on information and communication technology at the national and regional scales; and (4) to build entrepreneurship by offering empowering information resources and knowledge based on information technology.

To help realise its vision, the UI Library has support services, collections, and facilities. UI Library Services are grouped into circulation services, reference services, and information technology services. Circulation services include membership activation, borrowing and returning printed books, and a free library certificate (SKBP). Reference services consist of information literacy training, literature searching, guidance on the use of reference sources, e-resource delivery services (EDS), document similarity checking, library utilisation studies, manuscript services, broadcast information, special collection services and journals, and research tools services. Finally, information technology services consist of website access and management services, access to e-resources, and LONTAR development ( Perpustakaan UI, 2022).

In addition to these services, the UI Library has various collections to meet users’ information needs. At the end of 2021, the UI Library collections included 3,626,797 titles (3,761,309 total copies). The collections include e-resource collections (88.9%), UIANA collections (5.1%), textbooks (4.4%), printed magazines and journals (1.1%), reference books (0 4%), multimedia collections (0.1%), and ancient manuscripts and classic books (0.1%) ( Perpustakaan UI, 2022).

As the data above show, after e-resources (e-books, proceedings, e-journals, online videos, and other electronic resources owned and subscribed to by UI), one of the largest collections in the UI Library is the UIANA collection. The UIANA collection is grey literature consisting of undergraduate theses, master’s theses, dissertations, non-seminar papers, research reports, proceedings, and other scientific works from the UI Academic Civitas ( Tyasmara & Susetyo-Salim, 2018). This collection can be accessed in both printed and digital forms.

Distribution of final project publications by year of publication

The analysis results presented in Table 1 show the project distribution by publication year. Most of the UI students’ final project publications that were related to SDG 3, Good Health and Well-being, were published in 2019 (23.1%), 2020 (14.7%), and 2018 (13.1%). The number of publications from year to year tended to increase, as shown in Table 2. The average growth rate of the number of publications per year, from 2011–2020, was 29%.

Distribution of final project publications by faculty

As the data analysis presented in Table 2 shows, most of the UI students’ final assignments related to SDG 3 came from the Faculty of Public Health (35.9%) and the Faculty of Medicine (26.8%). Final project publications from other faculties included the Faculty of Nursing (19.2%), the Faculty of Pharmacy (13.9%), and the Faculty of Dentistry (4.2%). In terms of the growth rate in the number of publications, the Faculty of Dentistry has had the highest average annual growth rate (84%), followed by the Faculty of Medicine (57%), the Faculty of Public Health (50%), the Faculty of Nursing (48%), and the Faculty of Pharmacy (43%).

Distribution of final project publications by study program

Based on the results of the data analysis presented in Table 2 most of the publications of the UI students’ final assignments related to SDG 3 came from students completing the Doctoral Education Bachelor Program (14.17%), Master of Public Health Sciences (11.18%), and Bachelor of Pharmacy (10, 7%). As Table 2 shows, the epidemiology master’s degree program has the highest average publication growth rate per year (168%), followed by the medical science doctoral degree (141%) and the medical education master’s degree (119%) programs.

Distribution of final project publications by type of work

The analysis results presented in Table 3 show the distribution by type of work. Most of UI students’ published final assignments related to SDG 3 were undergraduate theses (54%). Master’s theses accounted for 42% of the total, while dissertations accounted for as much as 4%. The number of publications almost always increased yearly for each type of work. Nevertheless, the highest average growth rate in the number of publications per year was in the dissertation collection—as much as 207%.

The topics for UI student thesis publications were chosen by the librarian based on the Library of Congress subject heading. The results of the analysis presented in Figure 1 show that the most common subjects are industrial safety (appearing in 49 publications), breastfeeding (38 publications), industrial hygiene (27 publications), nursing care (25 publications), and tuberculosis (25 publications). In terms of development, the number of subjects studied most frequently changed from year to year. This is demonstrated in Figure 2. In 2011, the most commonly studied subjects were Bacillus thuringiensis, diabetes mellitus, and Dengue haemorrhagic fever, while in 2020, the most frequently researched subjects were nursing care, health education, and reproductive health.

SDG 3, Good Health and Well-being, has 13 targets, which are shown in Table 4. Of these 13 targets, most of the UI students’ final project publications had topics related to Target 4, Non-Communicable Diseases and Mental Health (18%), followed by Target 11, Drug and Vaccine Development (17.8%), and Target 13, Health Risk Management (13.2%). Meanwhile, topics related to Target 6 (Traffic Accidents), Target 5 (Abuse of Narcotics and Alcohol), Target 7 (Access to Reproductive Health Services), and Target 10 (Tobacco/Cigarette Control) appear less frequently in UI students’ final assignments. In addition, 14.2% of publications in the “Other” category are not directly related to SDG 3 targets, including topics related to forensics, diagnostics, nutrition, health promotion, and others.

In terms of the growth rate of publications, as Table 5 shows, Target 1 (Maternal Health) has undergone the greatest average annual growth (152%), followed by Target 2 (Infant and Toddler Health) and Target 8 (Universal Healthcare Access). The analysis results presented in Table 6 show which topics have been most frequently studied in each faculty and program. Target 4 (Non-Communicable Diseases and Mental Health) and Target 11 (Development of Drugs and Vaccines) are the most frequently studied topics in all faculties. Table 7 shows the distribution of final project publications related to SDG 3 based on the type of work.

Figure 3 shows the results of the word co-occurrence analysis of the topic subject (index keywords) and keywords (author keywords) of UI students’ final project publications related to SDG 3 that were published in 2011–2020. The analysis results led to the following 17 clusters: Cluster 1 – Health Services and Performance of Health Workers; Cluster 2 – Diabetes Prevention and Management; Cluster 3 – HIV/AIDS; Cluster 4 – Non-Communicable Diseases and Mental Health; Cluster 5 – Occupational Health and Safety; Cluster 6 – Dental Health and Mental Health; Cluster 7 – Maternal and Child Health; Cluster 8 – Youth Health; Cluster 9 – Child Development; Cluster 10 – Respiratory Tract Infections; Cluster 11 – Adolescent Reproductive Health; Cluster 12 – Infectious Diseases; Cluster 13 – Maternal Health; Cluster 14 – Health Promotion; Cluster 15 – Drug Abuse; Food Hygiene and Sanitation; Cluster 16 – Cancer; and Cluster 17 – Risk Management.

Each cluster associated with SDG 3 targets has links to the following targets:

Cluster 1: Target 8, Target 12

Cluster 2: Target 4, Target 11

Cluster 3: Target 3, Target 7

Cluster 4: Target 4, Target 11

Cluster 5: Target 13

Cluster 6: Target 4, Target 11, Target 14

Cluster 7: Target 1, Target 2

Cluster 8: Target 7, Target 11, Target 14

Cluster 9: Target 2, Target 14

Cluster 10: Target 3, Target 9

Cluster 11: Target 7

Cluster 12: Target 3, Target 11

Cluster 13: Target 1, Target 11

Cluster 14: Target 10, Target 14

Cluster 15: Target 5; Target 14

Cluster 16: Target 4, Target 11

Cluster 17: Target 13

The results of the statistical test that are presented in Table 8 show p-values below the significance level (α) of 0.05, which means that there are significant relationships between the faculty, publication year, and type of work and the topic of UI students’ final project publications in the field of good health and well-being. The coefficient of determination (R2) value of 0.562 indicates that the above factors can explain 56.2% of the variation in the dependent variable, the topic of UI students’ final project publications related to good health and well-being. Other factors can explain the rest of the variation.

A multinomial logistic regression analysis was used to find the variables with the most influence on UI students’ final project publications related to SDG 3. Table 9 presents the regression analysis results, where “Other Topics” was used to compare the relationships between the 13 topics related to SDG 3 and the year of publication, faculty, and work variables. Meanwhile, the Faculty of Pharmacy (FF) and dissertation were used as the comparison groups for the faculty and type of work variables.

Table 9 presents the regression coefficient (B) and odds ratio (OR) values. These analysis results reveal which topics are more common in a particular faculty compared to other faculties: The greater the B and OR values in a group, the greater the frequency of topic discussion compared to other groups. For example, the group with the largest B and OR scores for maternal health is the Faculty of Public Health (FKM), suggesting that maternal health is most frequently studied by FKM students. The same was found for infant and toddler health, infectious diseases, traffic accidents, access to reproductive health services, hazardous chemicals and pollution, health personnel (HR), and health risk management.

Meanwhile, the Faculty of Pharmacy (FF) excels in the topic of drug and vaccine development, compared with other faculties. Faculty of Nursing (FIK) students have mostly researched non-communicable diseases and mental health, narcotics and alcohol abuse, and tobacco control (cigarettes) compared with other faculties. The complete results are presented in Table 9.

In terms of publication year, the analysis results presented in Table 9 show whether a topic tends to be studied more or less frequently throughout the year. A positive B value for publication year indicates a directly proportional relationship between the frequency of discussion and the year.

Conversely, a negative B value indicates that the frequency of discussion is inversely proportional to the year. This information affirms that non-communicable diseases and mental health, narcotics and alcohol abuse, access to universal health services, hazardous chemicals and pollution, tobacco control (cigarettes), and drug and vaccine development tend to be researched more frequently as the year progresses, while the remaining topics (maternal health, infant and toddler health, infectious diseases, traffic accidents, access to reproductive health services, health personnel, and health risk management) tend to be less and less researched.

In terms of the type of work, several topics are studied more by doctoral students (S-3) than undergraduate (S-1) or master’s (S-2) students, including narcotics and alcohol abuse, access to health services, reproduction, and the development of drugs and vaccines, maternal health, infant and toddler health, non-communicable diseases and mental health, access to universal health services, tobacco control (cigarettes), and health personnel (HR) were most commonly studied by master’s students. For undergraduate students, the most frequently studied topics included infectious diseases, traffic accidents, hazardous chemicals and pollution, and health risk management.

The prediction model for UI students’ final project publications related to good health and well-being involved the analysis of the relationship between the subject of publication and the SDG 3 target discussed in the publication. This analysis was conducted using a simple linear regression method to determine the relationships between two or more variables and was intended to predict the value of the dependent variable (SDG 3 target) via the independent variable (publication subject). The dependent variable in a linear regression must be numerical, so the SDG 3 targets were assigned numbers 1–14.

As the statistical test results presented in Table 10 show, the p-value was smaller than the significance level (α) of 0.05, which indicates a significant relationship between the publication subject and UI students’ final project publication topics related to good health and well-being. Furthermore, the correlation value (R) of 0.977 indicates a strong, positive relationship, and the coefficient of determination (R ²) of 0.954 indicates that the subject can explain 95.4% of the variation in the dependent variable, namely the Topic of Publication of UI Students’ Final Projects Related to Good Health and Well-being.

The regression equation model cannot be displayed in Table 10 because 2261 topic subjects were used as independent variables. The model was applied as a topic predictor, as shown in Figure 2. These predictors allow librarians to enter the subject of publication to find the SDG 3 target that is associated with the topic. For example, Figure 4 shows that a final project with the subject “abortion” was coded “1” for the SDG 3 target “maternal health”.

This study used secondary data from the UI Library collection, acquired through the LONTAR system. During data processing, information bias can occur, especially during the categorisation process based on the 13 SDG 3 targets. Meanwhile, of the 5476 documents that met the inclusion criteria, 2802 (51%) had missing or duplicated data or were in the form of final works of specialist and professional programs. The number of missing data points may have caused selection bias in this study.

In total, the final project publications (undergraduate theses, master’s theses, and dissertations) of UI students from 2011–2020 and related to good health and well-being accounted for 2674 titles or 1.8% of all undergraduate theses, master theses, and dissertations in the UI Library. Of the five faculties that belong to the health sciences cluster, the Faculty of Public Health (FKM), the Faculty of Medicine (FK), and the Faculty of Nursing (FIK) have published the most. This may be because the number of students in these three faculties (9% of the total UI students) exceeds the number of students in the Faculty of Dentistry (FKG) and the Faculty of Pharmacy (FF) (only 3% of the total UI students). In addition, the FK, FKM, and FIK have more study programs than the FKG and FF.

The publication year data shows that the number of publications significantly increased in 2019, to almost double that of the previous year, with the average growth rate per year reaching 29% (see Figure 5). This shows that research interest in good health and well-being as an SDG did not gain much traction until 2019. A similar, significant upward trend was reported in the Punnakitikashem and Hallinger (2019) study on sustainable health management, which showed that the number of publications almost doubled in 2013 compared to the previous year. These trends suggest that research on good health and well-being could be researched more and more frequently in the years to come.

However, the number of final project publications produced in 2020 was less than that in 2019. This is due to librarians’ decreased productivity in processing the final project collection during the implementation of the work-from-home (WFH) system during the COVID-19 pandemic. One of the reasons for this decline in productivity was inadequate equipment and internet network access. This increased the number of final assignments by UI students published in 2020 that have not been processed. To overcome this and given the control of COVID-19 cases, the UI Library implemented a full work-from-office (WFO) system in 2022. In addition, the UI Library plans to recruit student staff (Wiradha) to assist with the processing of final project publications.

Most of the publications were theses. Meanwhile, dissertations only accounted for 3.5% (94 titles) of all publications. Dissertations were only included from about 52.6% of the estimated number of doctoral program graduates from the health sciences cluster in 2011–2020. Therefore, the UI Library needs to increase the acceptance of scientific works from doctoral programs. The UI library could work with the UI Directorate of Education (Dirpen) to form regulations that require students of all levels to upload their final assignments to the UI Library and disseminate this policy to all faculties at UI.

The analysis results show that the 30 topics studied were dominated by those related to access to universal health services (nursing care, public health, national health insurance, hospital administration, medical care, and health insurance), non-communicable diseases (diabetes mellitus, obesity, blood pressure, and fatigue), infectious diseases (tuberculosis, malaria, and Escherichia coli), maternal and child health (breastfeeding, child development, anaemia, and nutrition), and health risk management (industrial safety, industrial hygiene, and occupational health and safety).

Access to universal health care was in the top five most-discussed topics every year from 2017 to 2020. This aligns with the findings that Ghanbari et al. (2021) reported that researchers’ interest in studying universal health service access has increased continuously from 2009 to 2019. In addition, Sweileh (2020) shows that universal health service access was the most frequently studied SDG 3 target from 2015–2019. Therefore, continuing to increase interest in researching this topic, especially in developing countries, will be useful to understand the challenges and obstacles inhibiting the achievement of universal health service access and the improvement of global health overall ( Ghanbari et al., 2021).

Regarding SDG 3 targets, UI students most often studied Target 4, Non-Communicable Diseases and Mental Health. This aligns with the results of basic health research showing that the prevalence of non-communicable diseases, such as diabetes mellitus (1.5%) and heart disease (1.5%), in Indonesia is higher than the incidence of infectious diseases, such as pulmonary tuberculosis (0.42%) and hepatitis (0.39%). As far as mental health issue was concerned, the prevalence of severe mental health disorder was cited at 1.7% of Indonesian community ( Hartini et al., 2018).

Meanwhile, topics that have not been widely researched include Target 6 (Traffic Accidents), Target 5 (Abuse of Narcotics and Alcohol), and Target 10 (Tobacco Control). This also conforms to Sweileh’s results (2020) showing that these topics, especially drug and alcohol abuse, were not studied much compared with other topics according to the listings in the Scopus database from 2015–2019. The stigma against narcotics users can make researching the topic challenging. Narcotics abuse can increase the risk of other health problems, such as HIV/AIDS and hepatitis ( Sweileh, 2020). Librarians must be competent in educating users about alternative research and sampling methods for hard-to-reach communities. In addition, librarians need insight into appropriate secondary data sources for research related to SDG 3.

The word co-occurrence analysis using the VosViewer software identified the five words with the most links to other topics: “knowledge”, “teenager”, “hypertension”, “industrial safety”, and “attitude” (see Figure 6).

The words “knowledge” and “attitude” indicate that publications related to SDG 3 mostly focus on promotive and preventative efforts. It has been proved that promotive efforts can support the enabling factor to increase self-determination and control among people regarding their health while prevention efforts focus on avoiding the increase of risk factors towards disease, therefore for early detection can be done in reducing the incident and damage related to people’s health ( Helfer et al., 2020). Furthermore, the words associated with “knowledge” and “attitude” indicate that pregnant and lactating women, children, and adolescents are the most frequently studied populations. The health problems studied in the adolescent population have been quite diverse but are dominated by non-communicable diseases and mental health. The number of studies on non-communicable diseases is also indicated by the term “hypertension”, which is one of the words that is most frequently related to other topics. Hypertension is a risk factor for various non-communicable diseases, including diabetes mellitus. In addition to these topics, industrial safety is also widely studied. This topic is also related to health risk management and occupational safety and health. Its presence in this analysis indicates that, in addition to the populations of pregnant women, breastfeeding women, children, and adolescents, UI students study industry and the workplace broadly.

Publication year

The results of the multivariate analysis show a significant relationship between the year of publication and the topic chosen. Non-communicable diseases and mental health, drug and alcohol abuse, access to universal health services, hazardous chemicals and pollution, tobacco control (cigarettes), and drug and vaccine development have tended to be researched more frequently over the years, while the remaining topics (maternal health, infant and toddler health, infectious diseases, traffic accidents, access to reproductive health services, health personnel, and health risk management) have tended to be researched progressively less. Librarians can consider this information when selecting collection procurement activities. Collections related to topics that tend to be researched more frequently each year can be prioritised, adopting one of the benefits of bibliometrics: assistance with decision-making in developing collections ( Grace et al., 2019).

Faculty

The results of the multivariate analysis showed a significant relationship between faculty and publication topic. Among topics related to SDG 3, FKM students generated more publications on related topics than other faculties, except for drug and vaccine development, which was mainly covered by FF students, and non-communicable diseases and mental health, abuse of narcotics and alcohol, and tobacco control (cigarettes), which were mainly covered by FIK students. These topics were also researched by FK and FKG students. However, students in those faculties also studied topics that were not directly related to SDG 3, such as anatomy, physiology, forensics, and dental and bone health.

Type of work

The results of the multivariate analysis showed a significant relationship between the type of work and the topic chosen. Several topics were identified as being studied more frequently by postgraduate students (S-2 and S-3) than undergraduate students (S-1), including narcotics and alcohol abuse, access to reproductive health services, the development of drugs and vaccines, maternal health, infant and toddler health, non-communicable diseases and mental health, access to universal health services, tobacco control (cigarettes), and health personnel (HR). Other topics (infectious diseases, traffic accidents, hazardous chemicals and pollution, and health risk management) have been mainly studied by undergraduate students. One of the UI Library’s services to which the results of this analysis can be applied is the Lecturer and Postgraduate Special Reading Room Service. In this service, one librarian can consult with a user on their literature search. In addition, librarians can use the results of this analysis to provide information to meet users’ needs, especially those in postgraduate programs.

Subject

The results of the multivariate analysis show a significant relationship between the subject and the topic chosen. In addition, using the linear regression analysis that examined the relationship between the subject and the topic of publication, a linear regression model could be made to predict the topic of publication based on the subject. This offers another of the benefits of bibliometrics, namely that, according to Zipf’s Law, the analysis of the occurrence of words in documents can contribute to identifying index makers ( Latief, 2014). The US National Library of Medicine (NLM) has implemented word occurrence analysis to aid in indexing. The NLM uses an algorithm with a text ranking approach to implement automatic medical subject heading (MeSH) indexing.