Design and Study site
A cross-sectional analytic study was done at the Aga Khan Hospital Dar es Salaam and Muhimbili National Hospital, Dar es Salaam in Tanzania. The Aga Khan Hospital is a not for profit, tertiary care facility and the biggest private hospital in Tanzania. Its antenatal care clinic attends to averagely 300 women a week. Muhimbili hospital is the national referral hospital providing tertiary care services to referral cases and the population surrounding it. Antenatal care attendance in this hospital averages 1000 women per week.
Participants
Total sample size was 358. Sample size was determined using a standard formula on the basis of 22% magnitude of screening for GDM from a study in a similar setting (11), 20% precision, 80% power, 5% non-response rate and type I error of 5%. Each of the 2 study sites provided half of the total sample. Convenient, consecutive sampling was used to achieve the required sample.
The study population comprised of pregnant women above 18 years of age, who had already received some antenatal care and who were above the gestational age at which screening for GDM is routinely performed. Included were pregnant women attending the antenatal care specialist clinics, aged 18 years and above, possession of antenatal card, gestational age of 30 or more weeks calculated using menstrual dates or earliest ultra sound scan, and who attended a 1st or 2nd trimester antenatal care visit at the study site. Women were excluded if they had pre-existing self-reported diagnosis of diabetes mellitus, severe illness, mental illness, participated in another study or were taking corticosteroids, antidepressants, diuretics or beta blockers.
Data collection
During the routine health talks all women were educated about GDM and given details about the study. Informed, written and signed consent was obtained from willing participants. A mini check list confirmed eligibility of participants.
Data collection tools were an interviewer-administered questionnaire and data extraction sheet. Participants were interviewed using a pre-tested, structured questionnaire in the language of choice between Kiswahili and English. The questionnaire identified patient demographic characteristics, risk factors for GDM, socio-economic factors associated with screening for GDM, past obstetric and medical history and tests for glycemic status and related information regarding screening for GDM. Additional information including weight, height, body mass index, blood pressure was obtained and verified from the antenatal cards, patient files and laboratory reports. This information was recorded on the data collection sheet. Where information from the interview and patient records differed, the latter were used.
Women were considered screened for GDM if WHO 2013 guidelines (3), Tanzania Diabetic association guidelines (10) or Tanzania ministry of health guidelines 2017 (9) had been followed. All participants found not to have been screened for GDM were offered the 75g OGTT. The glucometer SD GlucoNavii® GDH blood glucose monitoring system and SD BIOSENSOR INC lancets were used. This apparatus had a coefficient of variation 5% above 5.55mmol/l and within standard deviation 4mg/dl for readings below 5.55mmol/l. For those with medical insurance, the cost of the test was covered by the insurer. Cash paying clients were offered a discounted price at the Aga Khan hospital and it was offered free of charge for those who could not afford. Interpretation of results was according to WHO 2013 guidelines (3). Women diagnosed with GDM or pre-diabetic states were informed and referred for appropriate care.
Data were collected on possible determinants of screening for GDM which were either socio-economic factors or risk factors of GDM. The socio-economic factors included: education status/ literacy, marital status, employment status, women’s autonomy, private health insurance status, distance to hospital, journey time to health facility, mode of transport, ability to pay for the service/ financial status, husband support, family support, obstetric history, parity, awareness of GDM or its risks, health care providers informing the women about the test and help with household work.
The risk factors for GDM included: glycosuria, Body Mass Index (BMI) >25 at time of interview, hypertension in current pregnancy, history of chronic hypertension, history of pregnancy induced hypertension (PIH), history of GDM, history of pre-diabetic states, family history of hypertension, family history of diabetes mellitus, history of multiple pregnancy, history of delivering a macrosomic (≥4kg of weight) baby, clinical or ultrasound diagnosis of large for gestational age in the ongoing pregnancy, history of excessive weight gain (≥5kg) since 18 years of age, race, history of pregnancy loss and grand-multiparity (≥ 5 pregnancies).
Data Analysis
Data was entered into Microsoft excel 2007 and cleaned. The data was then transferred to SPSS 20 for analysis. We defined “magnitude” of screening for GDM as the proportion of participants attending ANC who had received any test aimed at screening for GDM. Various tests done during screening were categorized and frequencies and percentages for each calculated and recorded. The percentage of participants who had not initially screened but were diagnosed with GDM by OGTT in the study was the magnitude of undiagnosed GDM. Socio-economic determinants of screening for GDM and risk factors for GDM were evaluated for association with screening for GDM using the chi-square test and associated p-values. Factors found to be statistically significant by chi square test were further analyzed using logistic regression for possible association with screening for GDM.