Along with the worldwide demographic change in the increased number of population aged over 65, Indonesia was no exception. Consequences of the prolonged lifeexpectancy and changes in disease pattern, the care recipients suffering from the chronic diseases have increased. However, the sufficient health care system in the long-term care settings, the home-health care in particular, was not yet established in Indonesia.   The first objective of this research was to observe the current home-based care services in Indonesia, and then determine quality of services, care recipients satisfaction, care recipients’ preference. The main objective was to develop a home-health care model with the concept of Coordination Unit.   The study design was the qualitative research with the mix-method approach consisting of interviews. More specifically, the mixture of the exploratory research approach and applied policy research approach was used with the exploratory approach were applied. The purposive sampling technic was applied, and data are collected through interviews, and analyzed by the framework and thematic analysis.   The result of the research was that no sufficient home-health care services were observed and some of care services were not accordance with the laws and regulations concerning “health care” that might lead to some in miss-treatment. Furthermore, no integrative network was found within the current home-health care. The most importantly, the care recipients’ satisfaction levels were not considered as high, due to the quality of care and care providers. The result of care recipients’ experiences, preferences and satisfaction levels as well as expatriates indicated the necessity of establishment of the care coordination units that possibly increase efficiency, sufficiency in the home-health care, and also improve the care recipients’ satisfaction level as the coordination units undergo to function as a bridge among care recipients, healthcare sectors and non-healthcare sectors. Furthermore, the coordination units establish cooperation and partnerships with appropriate health professionals and healthcare facilities to get the community integrated home-health care, and functions as the center of healthcare information and data of care recipients. Despite of some difficulties such as the health human resources and the competency level of care providers, the community health centers (Puskesmas) have great potentials to play roles as the coordination unit with a new position of “Care Manager.”  This study concluded that the home-health care model developed in this study that was the community-integrated care system with the coordination unit was applicable with high potential. Especially, the home-health care model working together with the Indonesian Health asset such as “Puskesmas,” and or “programs of public health efforts in non-communicable diseases” seems the most practical for Indonesia. Key Words: Care-Recipients-Satisfaction, Home-Health-Care, Care-Coordination-Unit

 

 

 

 

Industri petrokimia merupakan sektor berisiko tinggi yang membutuhkan penerapan sistem manajemen keselamatan dan budaya keselamatan yang matang. Penelitian ini bertujuan mengembangkan model maturitas budaya keselamatan yang sesuai untuk industri petrokimia di Indonesia. Model ini mengadaptasi teori Hudson, Fleming, Parker et al., dan Filho, serta menggambarkan lima tingkat kematangan budaya keselamatan, dari tingkat “Dasar” hingga “Berkelanjutan.”. Melalui pendekatan multidimensi, dikembangkan kerangka dan instrumen penilaian yang valid dan reliabel dengan lima dimensi utama: Komitmen, Komunikasi, Informasi, Keikutsertaan Karyawan, dan Pembelajaran Organisasi. Penelitian ini menemukan bahwa semua perusahaan dalam sampel telah mencapai tingkat “Berkelanjutan,” khususnya pada aspek Komitmen dan Pembelajaran Organisasi. Namun, Keikutsertaan Karyawan masih menjadi aspek yang perlu ditingkatkan. Hasil juga menunjukkan bahwa perusahaan multinasional dan penanggung jawab keselamatan menunjukkan pemahaman budaya keselamatan yang lebih baik. Model yang dikembangkan dapat digunakan sebagai alat praktis untuk menilai dan meningkatkan strategi keselamatan berkelanjutan di sektor petrokimia, mendorong keterlibatan aktif pekerja, serta memperkuat efektivitas sistem manajemen keselamatan proses.

The petrochemical industry is a high-risk sector requiring a mature implementation of safety management and safety culture. This study aims to develop a safety culture maturity model tailored to the Indonesian petrochemical industry. The model adapts the theoretical frameworks of Hudson, Fleming, Parker et al., and Filho, and describes five levels of safety culture maturity, from "Basic" to "Sustainable." A multidimensional approach was used to develop a valid and reliable assessment framework and instrument comprising five key dimensions: Commitment, Communication, Information, Employee Participation, and Organizational Learning. Findings show that all sampled companies have reached the “Sustainable” level, particularly in Commitment and Organizational Learning. However, Employee Participation remains an area needing improvement. The study also reveals that multinational companies and safety officers demonstrate a stronger understanding of safety culture. The developed model serves as a practical tool for evaluating and improving sustainable safety strategies in the petrochemical sector, enhancing employee engagement, and strengthening the effectiveness of process safety management systems.

Beban Diabetes Melitus Tipe 2 (DMT2) semakin tinggi di Indonesia dengan prevalensi 11,7% menurut Survei Kesehatan Indonesia 2023. Lebih dari 70% penyandang DMT2 belum terdiagnosis, yang menyebabkan keterlambatan pengobatan dan peningkatan risiko komplikasi. Teknologi pembelajaran mesin (machine learning) berpotensi digunakan untuk deteksi dini risiko DMT2 secara efisien dan berbasis data. Penelitian ini bertujuan mengembangkan, memvalidasi, dan menguji kelayakan model skrining risiko DMT2 berbasis pembelajaran mesin untuk populasi Indonesia. Studi ini menggunakan desain observasional dan dilaksanakan dari Oktober 2024 hingga Januari 2025. Penelitian terdiri dari tiga fase. Fase pertama mencakup pengembangan model prediksi menggunakan algoritma pembelajaran mesin dengan data dari dua sumber Kementerian Kesehatan, yakni kohor Penyakit Tidak Menular (PTM) (n=5.010) dan Aplikasi Sehat Indonesiaku (ASIK) (n=8.147.932). Fase kedua adalah validasi eksternal model menggunakan data ASIK tahun 2023 (n=24.735.050). Fase ketiga merupakan uji coba lapangan di dua Puskesmas di DKI Jakarta yang mencakup uji diagnostik (n=100) serta penilaian kelayakan sistem tenaga kesehatan (n=30) menggunakan instrumen System Usability Scale (SUS). Fase 1 menunjukkan bahwa model LightGBM pada data ASIK memiliki performa terbaik (AUC 0,90; sensitivitas 0,70; spesifisitas 0,88), sementara model CatBoost pada data kohor PTM menunjukkan AUC 0,76. Seleksi fitur mengidentifikasi lima variabel utama (usia, tekanan darah sistolik, aktivitas fisik, riwayat keluarga DMT2, konsumsi sayur/buah) sebagai penentu utama risiko. Fase 2 mencakup validasi eksternal dengan data ASIK 2023 menghasilkan AUC 0,90 (LightGBM) dan AUC 0,73 (CatBoost). Fase 3 uji coba lapangan menunjukkan hasil diagnostik yang sejalan dengan skor risiko model, dengan sensitivitas 0,80 dan spesifisitas 0,39 pada cut-off HbA1c 6,5% dan sensitivitas 0,79 dan spesifisitas 0,54 pada cutoff HbA1c 5,7%. Evaluasi oleh tenaga kesehatan menghasilkan skor SUS sebesar 72,5. Sebagai kesimpulan, Model skrining DMT2 berbasis pembelajaran mesin terbukti akurat dan layak diimplementasikan.

Type 2 Diabetes Mellitus (T2DM) burden is increasing in Indonesia, with a national prevalence of 11.7% according to the 2023 Indonesia Health Survey. More than 70% of individuals with T2DM remain undiagnosed, leading to delayed treatment and increased complications. Machine learning has the potential to enhance early detection of T2DM through an efficient, data-driven risk screening process. This study aims to develop, validate, and assess the feasibility of a machine learning-based T2DM risk screening model tailored for the Indonesian population. This study employed an observational design and was conducted from October 2024 to January 2025. The study consisted of three phases. The first phase involved the development of predictive models using machine learning algorithms with two datasets from the Ministry of Health: the PTM Cohort (n=5,010) and the ASIK registry (n=8,147,932). The second phase was an external validation using the ASIK 2023 dataset (n=24,735,050). The third phase was a field trial at two primary health centers (Puskesmas) in Jakarta, which included diagnostic testing of patients (n=100) and usability assessment by healthcare workers (n=30) using the System Usability Scale (SUS). In Phase 1, the LightGBM model trained on ASIK data achieved the best performance (AUC 0.90; sensitivity 0.70; specificity 0.88), while the CatBoost model trained on PTM Cohort data yielded an AUC of 0.76. Feature selection identified five main predictors (age, systolic blood pressure, physical activity, family history of T2DM, and fruit/vegetable consumption). In Phase 2, external validation with ASIK 2023 data confirmed high model performance with AUC 0.90 (LightGBM) and AUC 0.73 (CatBoost). In Phase 3, field testing showed diagnostic performance consistent with model-based risk scores, with sensitivity 0.80 and 0.39 at the HbA1c cut-off of 6.5% and sensitivity 0.79 and specificity 0.54 at the HbA1c cut-off of 5.7%. Usability testing by healthcare workers resulted in an average SUS score of 72.5. In conclusion, the machine-learning-based T2DM screening model has demonstrated high accuracy and is feasible for implementation.