Gestational Diabetes Mellitus: A Study of Women Who Fail to Attend Appointments

A retrospective review was conducted on all pregnant women who were diagnosed with GDM and who had received diabetes education at the Diabetes Centre at Liverpool Hospital between 1 January and 31 December 2011. Demographic data, including age, parity, BMI, ethnicity, family history of diabetes, history of previous GDM, glucose tolerance test results, A1C, and need for insulin therapy, were documented. The women's scores on the Edinburgh Depression Scale (EDS) at the first antenatal visit were also noted. Pregnancy outcomes including prematurity (< 37 weeks), birth weight, birth weight percentile, Apgar scores, fetal death, and admission to the neonatal intensive care unit (NICU) were recorded. Macrosomia was defined by a birth weight of > 90th percentile.

Attendance at diabetes-related services during pregnancy was recorded in the hospital appointment system. Women who failed to attend appointments with two clinicians scheduled on the same day were counted as having missed one appointment. Women who failed to attend two or more diabetes-related clinic appointments were identified as “multi-FTA.” The “well-attended” group was composed of women who attended all scheduled appointments or missed no more than one appointment throughout their pregnancy.

At Liverpool Hospital, a tertiary referral center located in the south-western part of Sydney, Australia, women diagnosed with GDM were booked into a diabetes group seminar conducted by a diabetes nurse educator (DNE) and dietitian at the Diabetes Centre. During the 2-hour session, these women were educated about blood glucose monitoring and given advice on medical nutritional therapy. The DNE also discussed the implications of GDM on the developing fetus and informed the women about their future risk of developing type 2 diabetes.

These women were followed up at the high-risk antenatal clinic, a multidisciplinary clinic staffed by midwives, obstetricians, renal physicians, endocrinologists, DNEs, and a dietitian. Most of the subsequent appointments were coordinated such that the women would see the obstetrics and diabetes teams at the same clinic on the same day. The frequency of diabetes-related appointments was dependent on the women's glycemic profile and dietary-related issues. Women who required insulin therapy would attend an extra education session at the Diabetes Centre to learn how to self-inject. Interpreters were readily available when required.

Continuous variables were expressed as median and interquartile range, whereas categorical variables were expressed as percentages. A non-parametric method was used to compare all continuous variables between binary outcomes. χ² Testing was used to compare categorical variables between binary outcomes. Differences were considered significant at the level of P < 0.05.

All variables were first tested in a univariate model. Only variables with a P < 0.25 were included in the multivariate model. To assess the associations between binary outcome variable and potential clinical factors, a logistic regression model was used. For predicting multi-FTA, these variables were ethnicity, parity, previous GDM, requirement for insulin therapy, BMI, and A1C at diagnosis. For predicting macrosomia, these variables were ethnicity, parity, need for insulin therapy, BMI, and multi-FTA.

The study was approved by the Sydney South West Area Health Service Human Research Ethics Committee.

Between January and December 2011, 366 women with GDM received diabetes education at the Liverpool Hospital Diabetes Service. The women had a median of eight diabetes-related clinic appointments during pregnancy. Among these women, 37.2% (136 women) missed at least one appointment during their pregnancy, whereas 21.9% (80 women) missed appointments at least twice (multi-FTA group). Twenty-two women missed at least four of their diabetes-related appointments.

Women in the multi-FTA group had greater BMIs, were less likely to be nulliparous, had a significantly higher rate of previous GDM, and were likely to be from a non-European background (Table 1). There were no differences between the multi-FTA and well-attended groups in the need for interpreters, scores on the EDS, or total number of diabetes-related appointments during pregnancy. In terms of their metabolic profiles, women in the multi-FTA group had higher fasting glucose levels on their glucose tolerance test at the time of GDM diagnosis. On multivariate logistic regression analysis, previous GDM and certain non-Anglo-European ethnicities (South Asians, Middle Easterners, and others) were independent predictors for multi-FTA (Table 2).

At 36 weeks' gestation, 272 women had a repeat A1C measurement. Those in the multi-FTA group had poorer glycemic control, as indicated by a greater proportion of women in this group having an A1C > 5.5% (Table 3). More of them required insulin therapy. In terms of pregnancy outcomes, women in the multi-FTA group gave birth to neonates with significantly larger birth weight and higher birth weight percentile, and a greater proportion were macrosomic (Table 3). There was a trend toward higher rates of neonate admission into the NICU in the multi-FTA group, but mode of delivery, proportion of premature births, neonates with low Apgar scores, and peri-natal deaths were not significantly different compared to those in the compliant group. Using multivariate logistic regression analysis taking into account factors such as maternal BMI, parity, ethnicity, and insulin use, women in the multi-FTA group had an increased risk of macrosomia (odds ratio [OR] 1.98, 95% CI 0.93–4.18, P = 0.076).

The reasons behind nonattendance of antenatal services were complex and involved factors relating to women's perceived value of the services, clinic set-up, and women's level of understanding of their condition and how it affects fetal well-being.^2,3  Campbell et al.⁴  reported a no-show rate of 28% at their high-risk obstetric clinic. A number of studies have identified several sociodemographic factors related to late initiation or inadequate adherence to antenatal care. These have included high parity, low income, and membership in a minority ethnic group.^2,3,5–7  In the literature, there has not been any study specifically assessing the clinic attendance of women with GDM.

Because unforeseen circumstances could preclude women from attending one of the many appointments scheduled during pregnancy, we chose women who missed at least two appointments as our reference FTA group (multi-FTA). In our cohort, a substantial number of women with GDM (more than one-fifth) missed their diabetes-related appointments two or more times during their pregnancy.

We demonstrated that women with GDM who frequently missed their appointments were more likely to have previous GDM and were often from non-European ethnic groups. Women with children (parity > 0) were more likely to miss appointments, but this association was no longer significant in multivariate analysis. We also showed that depression scores on EDS did not differ between those who attended the program fully and those in the multi-FTA group.

We observed that women with previous GDM had a greater risk of multi-FTA. These women could be more confident with GDM management from their previous experience, but it was not clear if that led to a lower threshold for missing appointments. For women with previous GDM and stable diabetes control, one might argue that these women could be reviewed at the clinic less frequently, and phone consultations could be employed more readily in these women.

Although women from a non-Anglo/European background had a higher multi-FTA rate, this was not related to the need for interpreters. Apart from language barriers, cultural differences and level of health literacy in these women could possibly affect the likelihood for nonattendance to appointments. These women would need more time and greater resources to help them understand their condition better and adhere to recommendations. The “others” ethnic group consisted of women from the Pacific Islands, Africa, and South America, and they had a significantly high nonattendance rate. However, because of the small number of women in each of these groups, separate analysis for each group was not valid.

The important question was whether attendance at diabetes-related appointments mattered in terms of pregnancy outcomes. In our study, those in the multi-FTA group had poorer glycemic control by 36 weeks, and their risk of macrosomia was higher. There was no difference in other pregnancy outcomes, although there was a trend toward increased risk of admission to the NICU. One other study,⁶  which was not limited to women with GDM, demonstrated that poor attendance at antenatal clinic appointments was associated with lower 5-minute Apgar scores of neonates, higher risk for fetal and neonatal death, and higher risk of placental abruption and intrauterine infections.

At our institution, women with GDM attended a mean of eight diabetes-related appointments throughout their pregnancy. Although most of the diabetes-related appointments were made in conjunction with obstetrics visits, attending all of the appointments could still be challenging, especially for women who worked or had young children at home. Our analysis showed no association between the number of diabetes-related appointments and the FTA rate. Nevertheless, it would be difficult to assess the impact to a woman's GDM management if she missed two out of a total of four appointments compared to missing two out of 10 appointments.

The most common reasons women gave for nonattendance included lack of transportation, unsuitable appointment times, need to care for a sick child or relative, and lack of available child care facilities.^2,4  Apart from reinforcing the importance of good diabetes control on pregnancy outcomes, improving the clinic set-up for these women could be vital. Strategies such as offering more flexible appointment times, maintaining shorter waiting times at the clinic, and providing a more child-friendly environment at the clinic could improve clinic attendance, but these strategies would demand greater health resources. To reduce the total number of clinic appointments, phone consultations could be employed more frequently in women with stable diabetes control and those with previous GDM. However, face-to-face reviews were still invaluable in that they provided more comprehensive assessment of women's diabetes control and nutritional status. A study using a telemedicine system⁸  was highly effective in enhancing communication between women with GDM and clinicians, but pregnancy outcomes were not different.

There were a few limitations in our study. Information such as the education level and socioeconomic status of the women, which could affect the FTA rate, was not available. The findings of this study also could not be extrapolated to other models of care (e.g., clinics in the private sector) where the service set-up, such as the scheduling of appointments and waiting times, is different. The socioeconomic background, level of health literacy, and perception of health care could also differ between women who choose private versus public antenatal services. Finally, other facets of compliance to GDM management, including adherence to self-monitoring of blood glucose, medical nutrition therapy, and insulin therapy, were not assessed in this study.

In summary, this was the first study assessing factors associated with multi-FTA of GDM appointments at a public hospital. At our institution, we found that women with previous GDM and those belonging to certain ethnic groups had higher rates of multi-FTA. Because effective diabetes management affects both maternal and fetal well-being, each institution should review its FTA rates and formulate strategies to improve attendance of these women.