Study design and study setting
This research is a cross-sectional study conducted among slum-dwellers with type 2 diabetes in Tabriz city in 2022. Inclusion criteria for the study were being over 18 years old, not disabled physically or mentally, and living in slums for at least 5 years. Other types of diabetes were excluded from the study. Tabriz is a metropolis located in northwest Iran with almost 2 million inhabitants. In recent years, Tabriz has witnessed an increase in the population of slums, according to municipal officials. In Tabriz city, 382124 people lived in slums at the time of the study, according to the official statistics. In recent years, health complexes have been launched in Tabriz city slums. Launching the health complexes in slums has made diabetic patients identified increase from 6 cases to 621 cases . These results show that slum-dwellers with this silence disease have been neglected even in terms of disease diagnosis, until a few years ago. The Tabriz city health system has consisted of some health complexes. Each health complex controls some health service centers. Four health complexes are located in the slums of Tabriz (Chamran, Azadi, Imam Khomeini, Hakim). Each complex controls 3-4 health service centers. In general, there are 15 health service centers in slums.
According to the Vice-Chancellor for Health of Tabriz University of Medical Sciences, there are 13155 patients with type 2 diabetes living in Tabriz's slums. To calculate the appropriate sample size, we used Morgan's table . Nearly 375 patients were required based on this table. To increase the power of the study, we selected 400 patients with type 2 diabetes. In Tabriz, slums are located in geographically different areas. To reduce selection bias, health service centers located in each area should be selected. Therefore, we conducted four simple random samplings to select the health service centers under each complex's supervision. Based on the diabetic population in any health complex, some health care centers were selected. Two, three, one, and one center were selected from Chamran, Azadi, Imam Khomeini, and Hakim, respectively. Non-compliance of some interviewees was predicted so more patients in each stratify were considered. Fig 1 shows the process of sampling.
Iran's integrated health system, known as SIB, records, maintains, and updates information on health status. Patients' primary data, such as addresses and phone numbers, were gathered from the SIB system in health service centers. Patients with type 2 diabetes were categorized according to their gender and ten-year-old age. Then, stratified systematic random sampling was used to select patients from these groups.
Data collection tools, data analysis, and ethical considerations
To collect information about diabetic patients' health care needs, a literature review was conducted. However, the study information on healthcare utilization and costs of diabetes in Brazil was the most useful to create a questionnaire about diabetes health care . To specify the essential health care services for diabetic patients, we used the IraPEN (Iran’s package of essential NCDs) instructions. IraPEN package has characterized essential health services and time intervals for them. So, we compared the rate of utilization with the recommended advice by IraPEN for patients with type 2 diabetes. The obtained information was discussed with medical specialists. According to experts, the questionnaire covers the study objectives. Additionally, to demographic information, the questionnaire asked about outpatient services utilization and frequency in the last four weeks, inpatient care utilization and frequency in the past year, and health care utilization and receiving consultation in the past three months. The utilization questions were answered with yes and no, or frequency of use. The complete questionnaire is accessible [see S-1]. Data were collected by two trained individuals. The data collectors were responsible for asking the questions, and the questionnaire was completed concurrently. This was due to the likelihood of illiteracy in the slum dwellers’ population. We trained data collectors not to ask directional questions. As a result, we were relatively confident that the study was not biased. Because of the specific process of data collection and the low probability of data loss, our approach for missing data is to exclude cases pairwise. After collecting the data, SPSS version 22 was used to perform descriptive and analytical statistics on the data. We used a binary logistic regression model to examine the effect of sociodemographic variables on utilization. In the following, we used the generalized model to predict the number of uses of outpatient services and inpatient care. Participants' informed consent was obtained, and the respondents were informed that participation was entirely voluntary. We informed the participants that the study's findings would be published in a scientific journal.