Study area and period
The study was conducted in Bahir Dar city; located 565Km far from Addis Ababa, the capital city of Ethiopia, at Amhara national regional state, North West Ethiopia. In Bahir Dar city there are two public referral hospitals, one primary hospitals, ten health center and four private hospitals. And this study was conducted in the two public referral hospitals, namely: Felege Hiwot comprehensive specialized referral hospital (FHCSH) and Tibebe Ghion specialized teaching hospital (TGSTH). Each of this hospital can be expected to serve for more than 10 million populations coming from Bahir Dar city, west Gojjam zone, east Gojam zone, awi zone, north and south wollo zones, south& north Gondar zones, partial part of Benshangul Gumuz and Oromia region. FHCSH has currently a total of 1431 man power in each discipline with 500 formal beds, 11 wards, 39 clinical and non-clinical departments /service unit / providing Diagnostic, curative, Rehabilitation and preventive service at outpatient &inpatient based. Similarly TGSTH is a teaching hospital under Bahir Dar University College of medicine and health sciences that has 459 bed capacity and with around 14 outpatient departments.
Apart from other services both referral hospitals provide diabetic treatment services by nurse practitioners, pediatrics residents and pediatricians.
The study period address from1stJanuary, 2016 to February 30 /2021.
Study design
An institution based retrospective follow up study was employed.
Source population
The source population were all type 1 diabetes mellitus children<15 years old who had follow up at diabetes clinic of the two referral hospitals.
Study population
The study population were all type 1 diabetes mellitus children <15 years old who were on follow up during the study period.
Study unit
All type one diabetic children’s chart that were selected randomly for investigation.
Inclusion criteria
Children age less than 15 years old and diagnosed with T1DM with regular follow up and had at least one HbA1c and/or a three month consecutive measurements of fasting blood sugar (FBS) with clear date of diagnosis between January 1/2016 to February 30/2021 were included.
Exclusion criteria
Children’s medical record/chart with incomplete information (such as HbA1c/average FBG and other relevant predictors like age with date of diagnosis, sex, treatment modality, frequency of follow up visit and last visit health condition of the children), those having less than 3 month follow up during the study period and those cases transferred in with unclear date of diagnosis from other institution were excluded from the study.
Sample size determination
Sample size was determined by double proportion formula after taking of predictors associated to optimal glycemic control from previous study conducted by retrospective cohort design (50)with the help of epi info version 7 by considering the following statistical assumptions: 95% Confidence Interval (CI), power 80%,percent of outcome in unexposed group 8.93%,risk ratio 0.253, marginal error 5%(50) .The calculated total sample size is 378, then by adding 10% for data incompleteness from the client chart, the final sample size became 416.
Sampling technique and procedure
The study participants were selected from the registration book. The medical records of children who were on follow up with type one diabetes mellitus from January 2016 to February 2021 were selected. A total of 721 children were recorded from the registration book of the two referral hospitals (sampling frame). Of which 416 cards were sampled using a simple random sampling technique by a computer generating method. Finally, cards that fulfilled the criteria were reviewed.
Dependent variables
Time to first optimal glycemic control
Independent variables
Socio demographic (age, gender, Residence); Institutional related variable (frequency of clinic visit); Diabetic related variables (duration of diabetes, diabetes related complication.); Comorbidities (preceding infections and other pathology) and treatment related variables (insulin therapy and adherence, noncompliance and other self-monitoring practice)
Age of the participants, frequency of glycemic control, body mass index and duration of diabetes were categorized in to groups in order to alien with the other literatures(36,40,50)
Operational definitions
Optimal glycemic control: Optimal glycemic control is defined as the three consecutive month HbA1c <7.5% and/or average FBG of 80–150 mg/dl with more or less stringent glycemic goals for individual clients based on age/life expectancy, comorbid condition, advanced complication, hypoglycemia unawareness and individual patient considerations (6- 8,80).
Event: Achieving first optimal glycemic control during the study period
Survival time: The time starting from date of diagnosis to first optimal glycemic control was determined for each participant
Censoring: Patients died, lost to follow up, transferee out, and complete the follow up period without achieving optimal glycemic control
Time to event: Time between diagnosis up to achieving first optimal glycemic control or censoring with measure of interest in month
Carbohydrate counting: Practicing healthy diet at home by non-refined carbohydrate utilization and eating consistent amount of food regularly with application of food pyramid as a meal planning tool to optimize blood sugar level (35).
Data collection procedure
The data were collected from patients chart that visit Felege Hiwot comprehensive specialized referral hospital and Tibebe Ghion specialized teaching hospital. Data that were relevant to measure the association between times to first optimal glycemic control among diabetic children were collected by two BSc nurses supervised by one senior nurse having second degree in public health.
Patient records were retrieved using their medical registration number identified in the total DM case load in the logbook of registration follow up form. Then medical registration number (MRN) of all diabetic pediatric patient were sorted. After that, the sample selection mechanism was simple random sampling technique, in which each of the patients had equal chance of being selected to be part of study.
A structured data extraction tool adapted by considering study variables such as socio demographic, personal and clinical predictors from patients’ charts.
Data quality assurance
Training was given for data collectors and supervisors about the objective and process of data collection by the principal investigator. Pretest was done on 5 % of sample size. Then pretested data abstraction tool/check list that comprises of questions to measure the relevant variables were used to collect the necessary data from the patient medical chart by those trained data collectors. Data quality was also assured by designing proper data abstraction tool and through continuous supervision. All collected data were checked for completeness and clarity.
Data processing and statistical analysis
The collected data was coded, enter, cleaned and stored into Epi-data version 3.1 and exported into STATA 14.0 statistical software for analysis. Descriptive statistics were presented with frequency tables, Kaplan Meier (KM) plots and median survival times. Months are used as a time scale to calculate time to first optimal glycemic control. The outcome of each participant was dichotomized in to censured or event (first optimal glycemic control)
Kaplan-Meier technique was used to measure survival experience of diverse groups of patients by using survival curves. Log-rank test was used to assess significant difference among survival distributions of groups for equality. After performing the Cox-proportional hazard regression, model goodness-of-fit was checked by Cox Snell residuals & assumptions was checked by using Shenfield residual test and graphically by using log minus log function survival curves.
Bivariable analysis was performed to calculate crud hazard ratio (CHR) and to screen out potentially significant independent variables at p value < 0.25 level of significance.
Association between the significant independent variables and the time to first optimal glycemic control was assessed using multivariable Cox Proportional Hazard (PH) model.
Adjusted hazard ratio (AHR) and 95% CI for HR were used to test significance and interpretation of results.
Variables with p-value < 0.05 were considered as statistically associated with the time to first optimal glycemic control in months.
Ethical considerations
Ethical clearance was obtained from the institutional review board (IRB) of Bahir Dar University (IRB number 01-008).Written supportive letter was taken from pediatrics department of the hospitals on behalf of the patients. This study had no any danger or negative consequences for the study participants. Medical record numbers were used for the data collection and personal identifiers of the client were not used in this research report. Access to collected information was limited to the principal investigator and confidentiality had preserved throughout the time.