Women's Health Issues 24-4 (2014) e373–e379
www.whijournal.com
Original article
Understanding Why Some Women with a History of Gestational Diabetes Do Not Get Tested for Diabetes Kathryn A. Paez, RN, PhD a,*, Emma M. Eggleston, MD, MPH b, Susan J. Griffey, DrPH, BSN c, Brandy Farrar, PhD a, Jacquelyn Smith c, Jennifer Thompson, MPP b, Matthew W. Gillman, MD, SM b a
Health Policy and Research, Health and Social Development, American Institutes for Research, Silver Spring, Maryland Obesity Prevention Program, Department of Population Medicine Harvard Medical School and Harvard Pilgrim Health Care Institute, Boston, Massachusetts c Social & Scientific Systems, Inc., Silver Spring, Maryland b
Article history: Received 12 October 2013; Received in revised form 11 April 2014; Accepted 21 April 2014
a b s t r a c t Background: The proportion of women with previous gestational diabetes mellitus (GDM) receiving postpartum diabetes testing is far less than desired. Even in health care systems with high testing rates, some women remain untested. We explored what helps and what hinders women to obtain recommended testing. Methods: In this mixed methods study, we recruited 139 patients with a history of GDM in their most recent pregnancy (6 months to 4.5 years before study enrollment) from a delivery system that had instituted a quality improvement program to increase postpartum diabetes testing rates. We determined whether they had received a postpartum diabetes test according to American Diabetes Association guidelines. Using survey data, we ran logistic regression models to assess correlates of testing status, and we conducted in-depth interviews with 22 women to provide greater context to their survey responses. Results: Of the 139 women, 21 women (15%) did not complete recommended diabetes testing. From the survey data, women who visited a primary care provider had 72% (95% CI, 0.09–0.83) lesser odds of not having been tested. From the qualitative interviews, difficulty fitting testing around work and caregiver demands were the most common reasons for not testing. Untested women interpreted providers’ reassurances that diabetes would resolve after delivery and lack of reminders to reschedule missed appointments and to complete diabetes testing as indicators that their physicians were not concerned about their diabetes risk. Conclusions: Among hard-to-reach women, multiple demands on their time were common explanations for not receiving a postpartum diabetes test. Consistent messages regarding long-term diabetes risk during pregnancy, access to postpartum primary care and convenient lab appointments, and systematic reminders to providers and patients are approaches that, in combination, may influence more resistant women to test. Copyright Ó 2014 by the Jacobs Institute of Women’s Health. Published by Elsevier Inc.
A diagnosis of gestational diabetes mellitus (GDM)d carbohydrate intolerance identified during pregnancydis a major health event in a woman’s life. During pregnancy, GDM is
associated with hypertension, preeclampsia, and caesarean delivery. Her child is at risk for macrosomia, premature delivery, birth trauma, and, in the long run, obesity (Casey, Lucas, Mcintire,
The authors are grateful for the financial support from the National Institute of Diabetes and Digestive and Kidney Diseases under contract number GS10F0381, purchase number HHSN276201000389U. We would like to acknowledge Sheryl Rifas-Shiman, MPH (Senior Research Analyst, Obesity Prevention Program, Department of Population Medicine, Harvard Medical School and Harvard Pilgrim Health Care Institute), for analyzing the diabetes testing and survey data, and Karen McCowen, MD (Assistant Professor, Endocrinologist, Harvard Vanguard Medical Associates and Harvard Medical School during the study), for serving as the HVMA liaison to the study and assisting with
verification of the diabetes testing data. The authors declare that they have no competing interests. Kathryn Paez, the corresponding author of this manuscript, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. * Correspondence to: Kathryn A. Paez, RN, PhD, Health Policy and Research American Institutes for Research, 10720 Columbia Pike, Suite 500 Silver Spring, MD 20901-4400. Phone: 301 592 2229; fax: 301 5939433. E-mail address: [email protected] (K.A. Paez).
1049-3867/$ - see front matter Copyright Ó 2014 by the Jacobs Institute of Women’s Health. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.whi.2014.04.008
e374
K.A. Paez et al. / Women's Health Issues 24-4 (2014) e373–e379
& Leveno, 1997; Johns, Olynik, Mase, Kreisman, & Tildesley, 2006; Regnault Gillman, Rifas-Shiman, Eggleston & Owen, 2013). In addition to posing immediate risks to the pregnancy, GDM also signals that a woman may develop diabetes in the future. Women with a history of GDM have seven times greater risk for future development of type 2 diabetes than a woman without GDM (Bellamy, Casas, Hingorani, & Williams, 2009). In a recent study, 51% of 174 women followed over the 5-year postpartum period experienced some form of abnormal glucose tolerance; 30% developed diabetes (Ekelund, Shaat, Almgren, Groop, & Berntorp, 2010). Physicians, nurses, and other clinical staff who support women during and after a GDM pregnancy are in a position to make a difference in these women’s lives by advising them in the steps they can take to delay or prevent diabetes. In particular, ongoing testing is important to detecting prediabetes and diabetes so that treatment to reduce the risk of diabetes or diabetesrelated complications can be promptly initiated (Ratner et al., 2008). The American Diabetes Association (ADA) recommends that all women with GDM receive a 75-g oral glucose tolerance test (OGTT) or fasting plasma glucose at 6 to 12 weeks postpartum to identify ongoing abnormal glycemia (ADA, 2010). The ADA also recommends that an OGTT, hemoglobin A1C, or fasting blood sugar (FBS) be performed at a minimum of every 3 years. Despite the benefits of testing and dissemination of the ADA guidelines, the proportion of women with previous GDM receiving diabetes testing in the 6-month postpartum period is far less than desired. Researchers studying postpartum diabetes testing rates in the United States have reported widely varying rates from 3% in a cross-sectional study of a Medicaid-eligible population (Hale et al., 2012) to as many as 72% in a managed health care organization where researchers sent patients and physicians reminders and standardized order sets were established (Clark, Graham, Karovitch, & Keely, 2009; Ferrara, Peng & Kim, 2009; Shea et al. 2011; Vesco et al., 2012). But even with reminders and automated ordering, the rate of diabetes testing remains less than ideal and some women go untested. A variety of factors contribute to women not receiving a diabetes test (Henderson et al., 2012; Keely, Clark, Karovitch, & Graham, 2010). Sometimes physicians fail to order a diabetes test or disagree with recommendations for testing (BentleyLewis, Levkoff, Stuebe, & Seely, 2008). Inconsistencies in recommendations for postpartum diabetes testing by the ADA, American College of Obstetrics and Gynecology, and the World Health Organization may contribute to confusion, causing some women not to be tested at all (Hunsberger, Donatella, Lindsay, & Rosenberg, 2012). Lack of coordination between obstetricians and primary care internists or family physicians can create uncertainty about who is responsible for monitoring diabetes risk (Bentley-Lewis et al., 2008). Women may not return for a postpartum visit or follow through even if the test is ordered and when follow-up reminders are sent. In an era where some providers have made significant strides in increasing postpartum diabetes testing rates, we were unable to identify any studies that provided insight into why the most resistant women do not return for diabetes testing. The purpose of this study was to explore what helps and what hinders women with a previous history of GDM to obtain the recommended diabetes testing. Our particular focus was on women, who despite the best efforts by providers, still do not complete testing. Using survey and in-depth interview data gathered from women with previous GDM who were participating in an intervention study, we identified individual and
health system factors associated with the receipt of postpartum diabetes testing. Methods This mixed methods study was an adjunct study to the Avoiding Diabetes After Pregnancy Trial (ADAPT), which evaluated the effectiveness of an intervention designed to increase compliance with ADA recommended testing and reduce diabetes risk through weight loss. Researchers recruited patients with a history of GDM in their most recent completed pregnancy, 6 months to 4.5 years before study enrollment, from Harvard Vanguard Medical Associates (HVMA), a multispecialty group medical practice located in the greater Boston area. The study was approved by the Harvard Pilgrim Health Care Human Studies Committee. The initial goal of ADAPT was to conduct two randomized, controlled trials, one to increase testing rates among those not already tested and the other to effect weight loss among those who were overweight. Because HVMA successfully introduced an initiative to increase testing rates concomitant with recruitment, the sample size was too small to conduct the testing trial. Thus, in our survey we compared women who were not tested (i.e., eligible for the testing trial) with women recruited to the weight loss trial who had already been tested. Recruitment took place between December 2011 and May 2012. Of the 395 women who were eligible, 139 agreed to participate in ADAPT. Of the 139 women, 118 had been tested for diabetes according to ADA guidelines and 21 had not been tested. For the qualitative part of the study, we divided women into four groups by testing and intervention status and then randomly selected 26 ADAPT participants. Selected participants were invited to participate by email followed by a phone call. Three women refused to participate and one woman failed to respond, leaving a sample of 22 for the interviews. To identify eligible patients, a member of the research team searched HVMA’s electronic medical records (EMR) for women with a history of GDM. GDM was identified in the EMR by an outpatient ICD9 code 648.8X with either a prescription for test strips or two or more abnormal values on an antepartum 100-g OGTT. In addition, the health record database was used to determine the testing status and body mass index of patients with a history of GDM. In cases with OGTT values but no diagnosis codes or diagnosis codes but no 100-g OGTT, chart abstraction was performed by an endocrinologist member of the study team (K.M.) to confirm the women met criteria for GDM. Women were excluded from the study if they were younger than 18, were no longer a patient of HVMA, had a history of type 1 or type 2 diabetes, lacked internet or telephone access, or had significant mental health disorders that could interfere with informed consent. Primary care physician approval was needed to participate in the study. Postpartum diabetes testing was defined as women who received an OGTT or a FBS at 6 to 12 weeks postpartum or an OGTT, FBS, or hemoglobin A1C after 12 weeks postpartum. To assess testing status, OGTT, FBS, and hemoglobin A1C tests after the index pregnancy were identified in the HVMA EMR by identification of the following CPPT codes: 80048, 80053, 82947, 82948, 82950, 82951, and 83036. A manual search was done to validate the EMR data when the coding was unclear. We drew from three sources of data for this study: 1) The EMR, 2) a survey administered to participants at the beginning of the study (pre-intervention), and 3) semistructured interviews conducted after participants completed the intervention. The
K.A. Paez et al. / Women's Health Issues 24-4 (2014) e373–e379
e375
baseline survey was distributed by email and participants received a $25 gift card for completing it. All interviews were conducted by phone. Participants received an additional $25 gift card for participating in the interview. The study was approved by all institutional review boards of the research team. Women signed informed consents both for participation in ADAPT as a whole and separately for the interviews, as well as an authorization form compliant with the Health Insurance Portability and Accountability Act of 1996 to approve access to the EMR.
initial conceptual model, and discussed their impressions. One researcher coded the data with the second researcher reviewing the assigned codes and discussing any differences in interpretation of the data until both researchers were in agreement. The researchers then worked together in an iterative process to develop memos describing themes and subthemes and electing exemplar quotes that best represented the data.
Measures
Descriptive Statistics
The survey included questions about the women’s sociodemographic status, health and lifestyle, management of diabetes during pregnancy, pregnancy history, family history, and health-seeking behaviors. The survey included the Edinburgh Postpartum Depression Scale, a validated 10-item questionnaire that measures depressive symptoms in women after giving birth (Cox, Holder, & Sagovsky, 1987), and the personal control scale, a subset of the Michigan Diabetes Research and Training Center’s Risk Perception Survey for Developing Diabetes, adapted for women with histories of GDM, and a single item assessing perceived 10-year risk of diabetes (Michigan Diabetes Research and Training Center, 2014). To provide greater depth and context to responses that were collected in the baseline survey, a semistructured interview guide was developed. The interview guide included open-ended questions with prompts about women’s experiences with diabetes during the qualifying pregnancy and with diabetes testing and weight management postpartum, family members’ history of and experiences with diabetes, and the women’s personal concerns about receiving a diagnosis of diabetes in the future. The first two interviews were attended by all three researchers. A subsequent debriefing was held to discuss any potential useful adjustments to the interview guide. All researchers agreed that the questions on the guide solicited relevant and detailed responses. Thus, no modifications were made to the guide.
Of the 139 women who enrolled in the study, 21 (15%) did not complete diabetes testing in the immediate postpartum period or thereafter; 118 (85%) women were tested (Table 1). On average, 2.7 years had passed since participants’ GDM pregnancy. Mean age was 38 (SD, 4.2) years, and mean (SD) number of children under 12 years old in each woman’s home was 1.8 (0.07). Most (78%) of the women in the study were married; the sample was educated and diverse. Almost 70% of the women (compared with 46.3% of the Massachusetts population overall, according to the 2007–2011 American Community Survey) had a college degree and 46% (compared with 19.6% of the Massachusetts population overall, according to the 2010 U.S. Census [2012]) were non-White.
Data Analysis Respondent’s testing status (taken from the EMR) was merged with the survey data and analyzed using SAS 9.3 (SAS Institute, Inc., Cary, NC). First, we assessed bivariate relationships between testing status and sociodemographic characteristics, barriers to getting tested, risk for developing type 2 diabetes, medical care utilization, and lifestyle behaviors. Next, we conducted logistic regression models to assess the influence of each of several variables, chosen a priori or on the basis of differences in the bivariate analysis, on the likelihood of receiving testing according to ADA guidelines. We assessed the odds of not being testing for diabetes to approximate relative risk of failing to test because this condition was rarer among women than being tested for diabetes. Because of the small sample size, our ability to run multivariable regression was limited and a parsimonious model was needed. Therefore, we adjusted covariates with propensity score to correct for biases in the sample. Next, we ran a logistic regression model with the exposure of interest and the propensity score as independent variables to estimate the risk of having been tested. For the qualitative portion of the study, we audiotaped the interviews and transcribed them verbatim and then imported the data into NVivo for analysis. The codebook was developed in an iterative fashion with multiple rounds of revision after two members of the team reviewed each transcript, reviewed the
Results
Factors Associated with Postpartum Diabetes Testing Status Survey findings A number of factors were correlated with testing status (p < .10). Women who had received the recommended diabetes testing were more likely to be married (81%) than the nontested group (62%). The tested group also had a greater proportion of women who were college graduates (72% vs. 52%). Regarding the items assessing barriers to getting tested, women who had been tested breastfed 2 months longer on average than women who had not been tested. Tested women were also less likely find it difficult to get physical activity because of lack of childcare (45% vs. 76%). More women in the tested group (79%) reported that they believed they had a moderate to high chance of getting diabetes over the next 10 years than did women in the nontested group (52%). Last, more women in the tested group (81%) reported that they saw an internal medicine or family medicine provider for routine preventive care than did women in the nontested group (62%). In Table 2, we present results from unadjusted and adjusted logistic regression models, which identify correlates of not receiving a diabetes tested. In the adjusted analysis using a propensity score, we found that women who had difficulty getting childcare for physical activity were 3.6 times greater odds (95% CI, 1.2–10.8) of not testing than those women who did have access to childcare for exercise. Women who identified as black had 3.5 times greater odds (95% CI, 1.0–12.4) of not being tested than women identifying as White. Seeing an internal medicine or family medicine provider for preventive care lowered the odds (odds ratio, 0.28; 95% CI, 0.09–0.83) of not testing compared with those who did not see a primary care provider. Insulin use also decreased the odds of not being tested but the confidence interval included 1.0 (odds ratio, 0.62; 95% CI, 0.20–1.91). College education was not associated with not receiving a diabetes test in the unadjusted and adjusted regression models. Qualitative findings Barriers to testing. Five of the 22 interviewees did not receive recommended diabetes testing and 8 women discussed barriers
e376
K.A. Paez et al. / Women's Health Issues 24-4 (2014) e373–e379
Table 1 Comparison of Women with a History of Gestational Diabetes Mellitusy Who Did Not Receive a Postpartum Diabetes Test* with Those Who Were Tested Domain and Characteristic Sociodemographics Age (y), mean (SD) College graduate, n (%) Yes No Married, n (%) Yes No Race/ethnicity, n (%) White Asian Black Other Barriers to diabetes testing Reasons it is hard to get physical activity, n (%) Lack childcare Too much effort Work 20 hours per week No. of children 1 time per month, n (%) Time since delivery, years, mean (SD)
Total (n ¼ 139)
Not Tested (n ¼ 21)
Tested (n ¼ 118)
p-Value
38.2 (4.2)
38.4 (5.0)
38.2 (4.1)
.87
96 (69) 43 (31)
11 (52) 10 (48)
85 (72) 33 (22)
.07
108 (78) 31 (22)
13 (62) 8 (38)
95 (81) 23 (19)
.06
74 29 20 14
(54) (21) (15) (10)
11 2 6 2
(52) (10) (29) (10)
63 27 14 12
(55) (23) (12) (10)
.18
69 43 95 1.8 6.2 6.8
(50) (31) (68) (0.7) (4.7) (5.5)
16 7 17 1.8 4.5 5.4
(76) (33) (81) (0.7) (4.4) (4.3)
53 36 78 1.8 6.5 7.0
(45) (31) (66) (0.7) (4.7) (5.6)
.01 .83 .18 .73 .07 .21
51 (37) 104 (75)
5 (24) 11 (52)
46 (39) 93 (79)
.18 .01
5.7 (0.9) 96 (73)
5.9 (0.8) 14 (74)
5.7 (0.9) 82 (73)
.45 .92
13 14 4 2.4
95 76 11 2.8
.06 .84 .19 .20
108 90 15 2.7
(78) (65) (11) (1.3)
(62) (67) (19) (1.4)
(81) (64) (9) (1.3)
Abbreviations: OB/GYN, obstetrics/gynecology; SD, standard deviation. * Postpartum diabetes testing: Oral glucose tolerance test (OGTT) or a fasting blood sugar (FBS) at 6–12 weeks postpartum or an OGTT, FBS, or hemoglobin A1C after 12 weeks postpartum. y Baseline survey data from 139 women participating in the Avoiding Diabetes After Pregnancy Trial (ADAPT). z Edinburgh Depression Scale is on 0–30 scale: 13 indicates possible depression. x Personal Control Over Diabetes Prevention scale consists of four questions with responses on a 7-point Likert scale; range is 4–28. A higher score indicates a perception of greater personal control over preventing diabetes.
to testing (Table 3). The explanations that these women gave for not getting tested paralleled and extended the findings from the survey and regression models. Four out of the five nontested interviewees said that they found it challenging to fit in the testing around their work schedules or caregiving demands. Another commonly reported challenge from nontested interviewees was lack of follow-up from providers. Although most respondents took responsibility for not keeping appointments for follow-up testing, several pointed out that their doctors did not “press them” about it or follow up when appointments were missed. In some instances, interviewees reported that they had expected to be contacted by their providers after missing appointments, but were not. In other instances, interviewees reported that their providers made it seem as if their diabetes would resolve after delivery and that further testing or behavioral and lifestyle management was not necessary. In addition to the challenges that nontested interviewees reported as keeping them from getting tested, the data also suggest that some women who self-monitor their blood glucose levels are less likely to get tested for diabetes. Several nontested interviewees seemed to be using self-testing as a replacement for formal testing. When this home testing resulted in normal levels, interviewees believed that there was no need for formal glucose testing. One nontested interviewee respondent jokingly clarified that she was not “playing doctor” by self-testing her blood glucose levels.
Facilitators to testing. Seventeen of the 22 interviewees had been tested for diabetes and 11 women discussed facilitators to receiving a postpartum diabetes test (Table 4). In the interview discussions, several of these women also noted that they were receiving ongoing follow-up testing after the initial postpartum test as well (Table 4). The most common impetus to testing was related to health system factors. Specifically, women talked about their testing as being a part of the routine preventive primary care that they received annually or biannually. Several women noted that their providers ordered diabetes tests and discussed the results as they related to their ongoing risk of getting diabetes in the future during these checkups. Thus, routine preventive care, especially in the primary care setting, seems to stimulate appropriate diabetes testing postpartum. Another theme that emerged from women who had been tested was that their doctors expressed that it was especially important that they engage in routine, ongoing testing because of their health history. For example, two interviewees noted that their doctors stressed the importance of their continued testing because they had a family history of diabetes. Another interviewee shared that her doctor ordered ongoing tests for her because she tested as having prediabetes even before her experience with GDM. It seems that when a woman has an underlying increased risk, consistent recommendation from her doctor for follow-up testing is influential.
K.A. Paez et al. / Women's Health Issues 24-4 (2014) e373–e379 Table 2 Characteristics Predicting the Odds of Not Having Received a Postpartum Diabetes Test* Characteristic
Sociodemographics
College graduate No 1.0 (ref) 1.0 (ref) Yes 0.4 (0.2, 1.1) 0.6 (0.2, 1.6) Race/ethnicity White, reference 1.0 (ref) 1.0 (ref) Asian 0.4 (0.1, 2.0) 0.5 (0.1, 2.7) Black 2.5 (0.8, 7.8) 3.5 (1.0, 12.4) Other 1.0 (0.2, 4.9) 1.3 (0.2, 7.3) Hard to get physical activity because do not have childcare No, reference 1.0 (ref) 1.0 (ref) Yes 3.9 (1.3, 11.4) 3.6 (1.2, 10.8) Insulin during pregnancy No 1.0 (ref) 1.0 (ref) Yes 0.5 (0.2, 1.4) 0.6 (0.2, 1.9) Saw internal medicine or family medicine provider No 1.0 (ref) 1.0 (ref) Yes 0.4 (0.1, 1.1) 0.3 (0.1, 0.8)
Barriers
Risk
Utilization
Unadjusted Odds Ratio (95% CI)
Adjustedy Odds Ratio (95% CI)
Domain
* Postpartum diabetes testing: Oral glucose tolerance test (OGTT) or a fasting blood sugar (FBS) at 6–12 weeks postpartum or an OGTT, FBS, or hemoglobin A1C after 12 weeks postpartum. Baseline survey data from 139 women participating in the Avoiding Diabetes After Pregnancy Trial. y Independent variables in the logistic regression model include the characteristic of interest and a propensity score. The propensity score, developed from all other characteristics listed in the table, adjusts for bias in the sample.
Another common motivation that interviewees reported for getting tested was the desire to “know their status.” In these discussions, interviewees expressed that it was better to know whether their blood sugar level was elevated rather than not know. These women expressed discomfort with not knowing whether a potential risk was real or not. In addition to these common themes, there were three other factors that were associated with getting tested that only one person each discussed. One woman described getting tested as a way of proactively managing her health so that she did not become ill and unable to take care of her children. She worried that, if she got diabetes, she would feel so physically sick that she would be unable parent successfully. Similarly, one woman who intended to have more children in the future believed that it was important to “keep an eye” on her blood sugar to avoid any complications that might jeopardize her ability to do this successfully. Not only was she engaging in routine testing, this woman also intended to get tested earlier than recommended once she became pregnant again so that she could stay informed of her health status. Last, one woman reported that self-testing motivated her to get tested. She was monitoring her blood sugar at home and, after seeing several results that indicated elevated blood glucose levels, decided to get tested formally. Because in other women normal results from self-testing signaled that there was no need for formal testing, self-testing can be both a facilitator and a barrier to testing, depending on its results. Discussion In this mixed method study, we examined a variety of sociodemographic, perceptual, and health care-related factors that could influence whether women with a previous history of GDM received a postpartum diabetes test according to ADA guidelines.
e377
Our study provides insights into why women who are most resistant to testing fail to receive a diabetes test after a pregnancy with GDM. Among these hard-to-reach women, our survey revealed that having follow-up primary care was associated with increased testing rates. Women identifying as black were much less likely to be tested than white women. Difficulty finding childcare to exercise, a proxy for support systems that enable or inhibit women to seek care, was associated with receipt of testing. The in-depth interviews reinforced survey findings by providing evidence that a lack of childcare, childcare demands, and an inability to find time owing to work were common explanations for not receiving a postpartum diabetes test. Women whose GDM was more severe either because they perceived it to be so or took insulin during pregnancy were more apt to be tested. Providers’ reassurance that diabetes would resolve after delivery sent a message that postpartum testing was unnecessary; women whose primary care providers discussed diabetes risk believed testing was important. Some women continued to self-monitor their blood glucose immediately after delivery and believed normal readings indicated a laboratory test was not needed. The high testing rate among study participants may be partially attributed to a quality improvement program that HVMA began in late 2010, just before we began ADAPT recruitment. HVMA set a goal that all women diagnosed with GDM would complete an OGTT within 8 weeks postpartum. HVMA obstetric practices implemented a standardized order set and nurses followed up monthly by phone and mail with women who had not been tested. Studies of similar programs have found that patient and provider reminders are an effective approach to increasing postpartum diabetes testing rates (Carson, Frank, & Keely, 2013). Nevertheless, the testing rate was not 100%, and our study provides information on modifiable factors that could decrease the remaining gap. Physicians’ perception of a woman’s severity of diabetes risk may be an important factor in determining how they interact with women who have had GDM. Our findings suggest that women who are more difficult to reach require education from providers about their postpartum diabetes risk and an explanation that self-monitoring is not a diagnostic tool. It is possible that when women had multiple risk factors for diabetes such as a family history, prediabetes before pregnancy, and insulin use during pregnancy, primary care physicians were more likely to emphasize diabetes risk reduction and the need for routine postpartum diabetes testing. As shown in our qualitative data, patient–physician interactions clearly influenced women’s attitudes about their diabetes risk and the need for testing. Our study confirms what has been reported in the literature about barriers to diabetes testing after a pregnancy with GDM and adds to the literature by providing insights into how providers can influence women to test or not. In a survey of Canadian women who had a history of GDM, time pressures were the most frequent reason for failing to be tested (Keely et al., 2010). In a university hospital setting, interviews with women around the time of the first postpartum visit found that feelings of emotional stress owing to adjusting to a new baby and the fear of receiving a diabetes diagnosis were key barriers while access to child care and desire for a checkup were among the key facilitators to seek out care (Bennett et al., 2011). Both of these studies reported on cases where women failed to test because they substituted self-monitoring of glucose for laboratory testing. Limitations to this study should be considered in interpreting the findings. The small sample size may have precluded identifying some associations of interest. Generalizability to all practices or all women with GDM may be limited by the relatively high
e378
K.A. Paez et al. / Women's Health Issues 24-4 (2014) e373–e379
Table 3 Findings From In-Depth Interviews Among 22 Participants in the Avoiding Diabetes after Pregnancy Trial: Barriers to Testing (n ¼ 8*) Barrier
No. of Women Reporting*
Sociodemographic-related barriers Difficulty coordinating childcare logistics to attend an appointment
7 6
Lack of insurance
1
Inflexible work schedules incompatible with lab hours
1
Fear of being diagnosed
1
Health system–related barriers Lack of follow-up from physicians
5 5
Self-assessments of health status
3
Time for OGTT test too long.
2
Availability of convenient lab hours
1
*
Exemplar Quote
“I don’t think there was anything that made me hesitate other than, you know, life with a newborn and two other children. And, again, not looking to make excuses, but my husband works – you know, usually at least I’d say 80 hours a week. And it’s sometimes a little difficult to try to schedule something not knowing if I’m going to have childcare. And I hate canceling things at the last moment because he’s not available to be home. So I think it was just one of those things I intended to do. And there wasn’t anything I intentionally put off for fear of hearing that I had diabetes. It just truly was a case of expecting at some point to have the time to do it in.” “This year I haven’t been tested yet. We have some issues with insurance right now, so I’m going to wait until January to do all my testing.” “The only thing that I was hesitant about is just the time of going in and trying to work with my schedule. That was the only thing. Because I was working. And I couldn’t really go to the doctor as much because of my schedule. That was the only thing that really stopped me from being active about getting it checked out.” “I know I should do it. But it’s almost like – this is going to sound awful – but you want honest answers. Like, no news is good news. And my family is like, oh my gosh, if I had that, what would I do? It’s awful.It’s, like, oh my gosh, I don’t want to have it. And so, I guess, in my mind, it’s been, if I don’t get checked, maybe I won’t develop it.” “No. I mean, that wasn’t really followed up on. I can’t recall it being, like, a subject matter. Basically, it was on my plate to follow through. But it wasn’t as though I received a call or a request from my primary [care doctor].” Q. “And how long did you check your blood glucose? I mean, did you do it until you lost weight, or...?” A. “No, I would say the first 6 weeks, I think, I was checking it. That’s more my reason why I didn’t go for my 6-week testing, because I already knew it was gone.” “I mean, they wanted me to do, you know, the 6 weeks. But after that, where I told them, you know, I didn’t have the time and stuff. Because you have to drink the drink, wait a while, you know? And then take the blood work. ” “The appointments that were available to me, even when I was pregnant, were late at night, like 6:30, 7:00 at night, and that really didn’t work for my schedule.”
Barriers were mentioned in 20 instances by 8 women.
education level of the participants as well as their obtaining care in a group medical practice with a well-developed EMR and reminder systems. Strengths included the racial/ethnic diversity of the population and the mixed methods approach in which the qualitative and quantitative results complemented each other.
Implications for Policy and/or Practice The findings of this study may inform the development of effective strategies to increase testing, particularly among women who are most difficult to reach, and will help providers
Table 4 Facilitators to Testing (N ¼ 16*) Facilitators
No. of Interviewees Reporting*
Health system–related facilitators Routine preventive care
11 7
Increased provider vigilance owing to family history
2
Increased provider vigilance owing to pre-diabetes risk
1
Desire to “know status”/ uncomfortable with not knowing
3
Subsequent pregnancy
1
Desire to be healthy for kids
1
Self-assessment of health status
1
Abbreviation: PCP, primary care physician. * Facilitators were mentioned in 16 instances by 11 women.
Exemplar Quote
“So going forward after I had my gestational diabetes, after my pregnancy, every year they, my doctors required me to get tested for gestational diabetes.” Q: “So do you think [your provider] felt it was a pretty routine test? A: “Yeah, ‘routine’ as far as he knew of my history, and my mom has diabetes, so that was also another concerning part with the test.” “I was already being tested every time I went to the doctor’s, because I was already overweight pre-pregnancy. So, it was something that my PCP was doing every time I came in for an annual physical anyway.” “I know that there is risk associated with people who had gestational diabetes that, within a certain time period, it’s likely that the person will develop diabetes. And so, I just want to know, as years pass, how I’m doing.” “But in the back of my mind, I’m kind of thinking, you know, eventually I’ll probably have another child. So I’ll definitely be cognizant of it. And we’ll probably get tested prior to the six-month mark of the pregnancy, you know, to keep it in check, to see if it recurs, kind of thing.” Q: “What do you think is your primary motivation for [getting tested]? A: “I just want to be healthy for my kids.” “And, I think earlier this summer, when my weight was probably at the highest it’s ever been, I did get myself tested. Because of that, but I was like, hmm, I should get myself tested.”
K.A. Paez et al. / Women's Health Issues 24-4 (2014) e373–e379
to recognize the characteristics of women who are least likely to get tested and thus who may require special attention and targeted outreach. Our study demonstrates the importance of health system factors and the provider–patient relationship in influencing women to be tested for type 2 diabetes after a pregnancy complicated by GDM. There are several implications of our findings. The delivery system should make efforts to effect a hand-off from obstetric to primary care after delivery. Systematic reminders to women who do not receive timely diabetes testing signal to women that testing is important. From the first evidence of GDM to after delivery, both obstetric and primary care providers should reinforce to women with GDM that their elevated risk for developing type 2 diabetes requires proactive follow-up care. Woman should be encouraged to make an appointment after delivery with a primary care provider so that a routine schedule for diabetes testing in a laboratory setting (rather than selfmonitoring) can be initiated. In addition, health care providers, employers, and families need to consider the multiple demands on women’s time and make it feasible for them to obtain the recommended testing. References American Diabetes Association (ADA). (2010). Standards of medical care in diabetes-2012. Diabetes Care, 35, s11–s63. Bellamy, L., Casas, J. P., Hingorani, A. D., & Williams, D. (2009). Type 2 diabetes mellitus after gestational diabetes: A systematic review and meta-analysis. Lancet, 373, 1773–1779. Bennett, W. L., Ennen, C. S., Carrese, J. A., Hill-Briggs, F., Levine, D., Nicholson, W. K., et al. (2011). Barriers to and facilitators of postpartum follow-up care in women with recent gestational diabetes mellitus: A qualitative study. Journal of Women’s Health, 20, 239–245. Bentley-Lewis, R., Levkoff, S., Stuebe, A., & Seely, E. W. (2008). Gestational diabetes mellitus: Postpartum opportunities for the of type 2 diabetes mellitus. Endocrinology & Metabolism, 4, 552–558. Carson, M. P., Frank, M. I., & Keely, E. (2013). Postpartum testing rates among women with a history of gestational diabetes: Systematic review. Primary Care Diabetes, 7, 177–186. Casey, B. M., Lucas, M. J., Mcintire, D. D., & Leveno, K. J. (1997). Pregnancy outcomes in women with gestational diabetes compared with the general obstetric population. Obstetrics & Gynecology, 90, 869–873. Clark, H. D., Graham, I. D., Karovitch, A., & Keely, E. (2009). Do postal reminders increase postpartum screening of diabetes mellitus in women with gestational diabetes mellitus? A randomized clinical trial. American Journal of Obstetrics and Gynecology, 200, e631–e637, 634. Cox, J. L., Holden, J. M., & Sagovsky, R. (1987). Detection of postnatal depression. Development of the 10-item Edinburgh Postnatal Depression Scale. British Journal of Psychiatry, 150, 782–786. Ekelund, M., Shaat, N., Almgren, P., Groop, L., & Berntorp, K. (2010). Prediction of postpartum diabetes in women with gestational diabetes mellitus. Diabetologia, 53, 452–457. Ferrara, A., Peng, T., & Kim, C. (2009). Trend in postpartum diabetes screening and subsequent diabetes and impaired fasting glucose among women with histories of gestational diabetes mellitus. Diabetes Care, 32, 2269–2274. Hale, N. L., Probst, J. C., Liu, J., Martin, A. B., Bennett, K. J., & Glover, S. (2012). Postpartum screening for diabetes among Medicaid-eligible South Carolina women with gestational diabetes. Women’s Health Issues, 22, e163–e169. Henderson, C. E., Kavookjian, J., Leitstein, H., McKoy, J. M., Murage, W. J., & Lipman, R. D. (2012). Window of opportunity: Postpartum screening of women with gestational diabetes for early detection of prediabetes and type 2 diabetes. Open Diabetes Journal, 5, 25–28.
e379
Hunsberger, M. L., Donatella, R. J., Lindsay, K., & Rosenberg, K. D. (2012). Physician care patterns and adherence to postpartum glucose testing after gestational diabetes mellitus in Oregon. PLOS One, 7, e47052. Johns, K., Olynik, C., Mase, R., Kreisman, S., & Tildesley, H. (2006). Gestational diabetes mellitus outcome in 394 patients. Journal of Obstetrics and Gynaecology Canada, 2, 122–127. Keely, E., Clark, H., Karovitch, A., & Graham, I. (2010). Screening for type 2 diabetes after gestational diabetes. Canadian Family Physician, 56, 558–563. Michigan Diabetes Research and Training Center. (2014). Risk Perception Survey for Developing Diabetes (RPS-DD), adapted for women with histories of gestational diabetes. Available at: http://www.med.umich.edu/mdrtc/profs/ survey.html. Ratner, R. E., Christophi, C. A., Metzger, B. E., Dabelea, D., Bennett, P. H., PiSunyer, X., et al. (2008). Diabetes Prevention Program Research Group. Prevention of diabetes in women with a history of gestational diabetes: effects of metformin and lifestyle interventions. Journal of Clinical Endocrinology Metabolism, 93, 4774–4779. Regnault, N., Gillman, M. W., Rifas-Shiman, S. L., Eggleston, E., & Oken, E. (2013). Sex-specific associations of gestational glucose tolerance with childhood body composition. Diabetes Care, 36, 3045–3053. Shea, A. K., Shah, B. R., Clark, H. D., Malcolm, J., Walker, M., Karovitch, A., et al. (2011). The effectiveness of implementing a reminder system into routine clinical practice: Does it increase postpartum screening in women with gestational diabetes? Chronic Disease in Canada, 331, 58–62. U.S. Census. (2012). State and county quick facts. Available at: http://quickfacts. census.gov/qfd/states/00000.html. Vesco, K. K., Dietz, P. M., Bulkey, J., Bruce, F. C., Callaghan, W. M., England, L., et al. (2012). A system-based intervention to improve postpartum diabetes screening among women with gestational diabetes. American Journal of Obstetric Gynecology, 207, 283.e1–283.e6.
Author Descriptions Kathryn A. Paez, RN, PhD, is a principal researcher in Health Policy and Research at the American Institutes for Research. Her research interests concern improving quality of care and access to care to improve population health.
Emma M. Eggleston, MD, MPH, is an instructor in the Department of Population Medicine at Harvard Medical School and the Harvard Pilgrim Health Care Institute. Her clinical and research interests are in diabetes, pregnancy, and vascular disease.
Susan J. Griffey, DrPH, BSN, directs the Evaluation Center at Social & Scientific Systems. She researches public health interventions and programs in the US and in global development settings in primary and reproductive health care, HIV/AIDS, and chronic diseases.
Brandy Farrar, PhD, is a researcher in Health Policy and Research at the American Institutes for Research. The majority of Ms. Farrar’s research involves evaluating the effectiveness, viability, and impact of innovative programs designed to improve quality, access, and capacity of health care services.
Jacquelyn Smith, MA, is a senior survey manager at Social & Scientific Systems. She applies her 20 years of experience in survey methodology and survey operational management to numerous public health studies in both acute and chronic disease topics.
Jennifer Thompson, MPP, is a project manager in the Obesity Prevention Program in the Department of Population Medicine at the Harvard Pilgrim Health Care institute and Harvard Medical School. Her projects focus on diabetes, early childhood nutrition, and obesity.
Matthew Gillman, MD, SM, is a Professor and Director of the Obesity Prevention Program, which seeks to lessen the burden of obesity and its consequences, in the Department of Population Medicine at the Harvard Pilgrim Health Care Institute and Harvard Medical School.
www.whijournal.com
Original article
Understanding Why Some Women with a History of Gestational Diabetes Do Not Get Tested for Diabetes Kathryn A. Paez, RN, PhD a,*, Emma M. Eggleston, MD, MPH b, Susan J. Griffey, DrPH, BSN c, Brandy Farrar, PhD a, Jacquelyn Smith c, Jennifer Thompson, MPP b, Matthew W. Gillman, MD, SM b a
Health Policy and Research, Health and Social Development, American Institutes for Research, Silver Spring, Maryland Obesity Prevention Program, Department of Population Medicine Harvard Medical School and Harvard Pilgrim Health Care Institute, Boston, Massachusetts c Social & Scientific Systems, Inc., Silver Spring, Maryland b
Article history: Received 12 October 2013; Received in revised form 11 April 2014; Accepted 21 April 2014
a b s t r a c t Background: The proportion of women with previous gestational diabetes mellitus (GDM) receiving postpartum diabetes testing is far less than desired. Even in health care systems with high testing rates, some women remain untested. We explored what helps and what hinders women to obtain recommended testing. Methods: In this mixed methods study, we recruited 139 patients with a history of GDM in their most recent pregnancy (6 months to 4.5 years before study enrollment) from a delivery system that had instituted a quality improvement program to increase postpartum diabetes testing rates. We determined whether they had received a postpartum diabetes test according to American Diabetes Association guidelines. Using survey data, we ran logistic regression models to assess correlates of testing status, and we conducted in-depth interviews with 22 women to provide greater context to their survey responses. Results: Of the 139 women, 21 women (15%) did not complete recommended diabetes testing. From the survey data, women who visited a primary care provider had 72% (95% CI, 0.09–0.83) lesser odds of not having been tested. From the qualitative interviews, difficulty fitting testing around work and caregiver demands were the most common reasons for not testing. Untested women interpreted providers’ reassurances that diabetes would resolve after delivery and lack of reminders to reschedule missed appointments and to complete diabetes testing as indicators that their physicians were not concerned about their diabetes risk. Conclusions: Among hard-to-reach women, multiple demands on their time were common explanations for not receiving a postpartum diabetes test. Consistent messages regarding long-term diabetes risk during pregnancy, access to postpartum primary care and convenient lab appointments, and systematic reminders to providers and patients are approaches that, in combination, may influence more resistant women to test. Copyright Ó 2014 by the Jacobs Institute of Women’s Health. Published by Elsevier Inc.
A diagnosis of gestational diabetes mellitus (GDM)d carbohydrate intolerance identified during pregnancydis a major health event in a woman’s life. During pregnancy, GDM is
associated with hypertension, preeclampsia, and caesarean delivery. Her child is at risk for macrosomia, premature delivery, birth trauma, and, in the long run, obesity (Casey, Lucas, Mcintire,
The authors are grateful for the financial support from the National Institute of Diabetes and Digestive and Kidney Diseases under contract number GS10F0381, purchase number HHSN276201000389U. We would like to acknowledge Sheryl Rifas-Shiman, MPH (Senior Research Analyst, Obesity Prevention Program, Department of Population Medicine, Harvard Medical School and Harvard Pilgrim Health Care Institute), for analyzing the diabetes testing and survey data, and Karen McCowen, MD (Assistant Professor, Endocrinologist, Harvard Vanguard Medical Associates and Harvard Medical School during the study), for serving as the HVMA liaison to the study and assisting with
verification of the diabetes testing data. The authors declare that they have no competing interests. Kathryn Paez, the corresponding author of this manuscript, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. * Correspondence to: Kathryn A. Paez, RN, PhD, Health Policy and Research American Institutes for Research, 10720 Columbia Pike, Suite 500 Silver Spring, MD 20901-4400. Phone: 301 592 2229; fax: 301 5939433. E-mail address: [email protected] (K.A. Paez).
1049-3867/$ - see front matter Copyright Ó 2014 by the Jacobs Institute of Women’s Health. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.whi.2014.04.008
e374
K.A. Paez et al. / Women's Health Issues 24-4 (2014) e373–e379
& Leveno, 1997; Johns, Olynik, Mase, Kreisman, & Tildesley, 2006; Regnault Gillman, Rifas-Shiman, Eggleston & Owen, 2013). In addition to posing immediate risks to the pregnancy, GDM also signals that a woman may develop diabetes in the future. Women with a history of GDM have seven times greater risk for future development of type 2 diabetes than a woman without GDM (Bellamy, Casas, Hingorani, & Williams, 2009). In a recent study, 51% of 174 women followed over the 5-year postpartum period experienced some form of abnormal glucose tolerance; 30% developed diabetes (Ekelund, Shaat, Almgren, Groop, & Berntorp, 2010). Physicians, nurses, and other clinical staff who support women during and after a GDM pregnancy are in a position to make a difference in these women’s lives by advising them in the steps they can take to delay or prevent diabetes. In particular, ongoing testing is important to detecting prediabetes and diabetes so that treatment to reduce the risk of diabetes or diabetesrelated complications can be promptly initiated (Ratner et al., 2008). The American Diabetes Association (ADA) recommends that all women with GDM receive a 75-g oral glucose tolerance test (OGTT) or fasting plasma glucose at 6 to 12 weeks postpartum to identify ongoing abnormal glycemia (ADA, 2010). The ADA also recommends that an OGTT, hemoglobin A1C, or fasting blood sugar (FBS) be performed at a minimum of every 3 years. Despite the benefits of testing and dissemination of the ADA guidelines, the proportion of women with previous GDM receiving diabetes testing in the 6-month postpartum period is far less than desired. Researchers studying postpartum diabetes testing rates in the United States have reported widely varying rates from 3% in a cross-sectional study of a Medicaid-eligible population (Hale et al., 2012) to as many as 72% in a managed health care organization where researchers sent patients and physicians reminders and standardized order sets were established (Clark, Graham, Karovitch, & Keely, 2009; Ferrara, Peng & Kim, 2009; Shea et al. 2011; Vesco et al., 2012). But even with reminders and automated ordering, the rate of diabetes testing remains less than ideal and some women go untested. A variety of factors contribute to women not receiving a diabetes test (Henderson et al., 2012; Keely, Clark, Karovitch, & Graham, 2010). Sometimes physicians fail to order a diabetes test or disagree with recommendations for testing (BentleyLewis, Levkoff, Stuebe, & Seely, 2008). Inconsistencies in recommendations for postpartum diabetes testing by the ADA, American College of Obstetrics and Gynecology, and the World Health Organization may contribute to confusion, causing some women not to be tested at all (Hunsberger, Donatella, Lindsay, & Rosenberg, 2012). Lack of coordination between obstetricians and primary care internists or family physicians can create uncertainty about who is responsible for monitoring diabetes risk (Bentley-Lewis et al., 2008). Women may not return for a postpartum visit or follow through even if the test is ordered and when follow-up reminders are sent. In an era where some providers have made significant strides in increasing postpartum diabetes testing rates, we were unable to identify any studies that provided insight into why the most resistant women do not return for diabetes testing. The purpose of this study was to explore what helps and what hinders women with a previous history of GDM to obtain the recommended diabetes testing. Our particular focus was on women, who despite the best efforts by providers, still do not complete testing. Using survey and in-depth interview data gathered from women with previous GDM who were participating in an intervention study, we identified individual and
health system factors associated with the receipt of postpartum diabetes testing. Methods This mixed methods study was an adjunct study to the Avoiding Diabetes After Pregnancy Trial (ADAPT), which evaluated the effectiveness of an intervention designed to increase compliance with ADA recommended testing and reduce diabetes risk through weight loss. Researchers recruited patients with a history of GDM in their most recent completed pregnancy, 6 months to 4.5 years before study enrollment, from Harvard Vanguard Medical Associates (HVMA), a multispecialty group medical practice located in the greater Boston area. The study was approved by the Harvard Pilgrim Health Care Human Studies Committee. The initial goal of ADAPT was to conduct two randomized, controlled trials, one to increase testing rates among those not already tested and the other to effect weight loss among those who were overweight. Because HVMA successfully introduced an initiative to increase testing rates concomitant with recruitment, the sample size was too small to conduct the testing trial. Thus, in our survey we compared women who were not tested (i.e., eligible for the testing trial) with women recruited to the weight loss trial who had already been tested. Recruitment took place between December 2011 and May 2012. Of the 395 women who were eligible, 139 agreed to participate in ADAPT. Of the 139 women, 118 had been tested for diabetes according to ADA guidelines and 21 had not been tested. For the qualitative part of the study, we divided women into four groups by testing and intervention status and then randomly selected 26 ADAPT participants. Selected participants were invited to participate by email followed by a phone call. Three women refused to participate and one woman failed to respond, leaving a sample of 22 for the interviews. To identify eligible patients, a member of the research team searched HVMA’s electronic medical records (EMR) for women with a history of GDM. GDM was identified in the EMR by an outpatient ICD9 code 648.8X with either a prescription for test strips or two or more abnormal values on an antepartum 100-g OGTT. In addition, the health record database was used to determine the testing status and body mass index of patients with a history of GDM. In cases with OGTT values but no diagnosis codes or diagnosis codes but no 100-g OGTT, chart abstraction was performed by an endocrinologist member of the study team (K.M.) to confirm the women met criteria for GDM. Women were excluded from the study if they were younger than 18, were no longer a patient of HVMA, had a history of type 1 or type 2 diabetes, lacked internet or telephone access, or had significant mental health disorders that could interfere with informed consent. Primary care physician approval was needed to participate in the study. Postpartum diabetes testing was defined as women who received an OGTT or a FBS at 6 to 12 weeks postpartum or an OGTT, FBS, or hemoglobin A1C after 12 weeks postpartum. To assess testing status, OGTT, FBS, and hemoglobin A1C tests after the index pregnancy were identified in the HVMA EMR by identification of the following CPPT codes: 80048, 80053, 82947, 82948, 82950, 82951, and 83036. A manual search was done to validate the EMR data when the coding was unclear. We drew from three sources of data for this study: 1) The EMR, 2) a survey administered to participants at the beginning of the study (pre-intervention), and 3) semistructured interviews conducted after participants completed the intervention. The
K.A. Paez et al. / Women's Health Issues 24-4 (2014) e373–e379
e375
baseline survey was distributed by email and participants received a $25 gift card for completing it. All interviews were conducted by phone. Participants received an additional $25 gift card for participating in the interview. The study was approved by all institutional review boards of the research team. Women signed informed consents both for participation in ADAPT as a whole and separately for the interviews, as well as an authorization form compliant with the Health Insurance Portability and Accountability Act of 1996 to approve access to the EMR.
initial conceptual model, and discussed their impressions. One researcher coded the data with the second researcher reviewing the assigned codes and discussing any differences in interpretation of the data until both researchers were in agreement. The researchers then worked together in an iterative process to develop memos describing themes and subthemes and electing exemplar quotes that best represented the data.
Measures
Descriptive Statistics
The survey included questions about the women’s sociodemographic status, health and lifestyle, management of diabetes during pregnancy, pregnancy history, family history, and health-seeking behaviors. The survey included the Edinburgh Postpartum Depression Scale, a validated 10-item questionnaire that measures depressive symptoms in women after giving birth (Cox, Holder, & Sagovsky, 1987), and the personal control scale, a subset of the Michigan Diabetes Research and Training Center’s Risk Perception Survey for Developing Diabetes, adapted for women with histories of GDM, and a single item assessing perceived 10-year risk of diabetes (Michigan Diabetes Research and Training Center, 2014). To provide greater depth and context to responses that were collected in the baseline survey, a semistructured interview guide was developed. The interview guide included open-ended questions with prompts about women’s experiences with diabetes during the qualifying pregnancy and with diabetes testing and weight management postpartum, family members’ history of and experiences with diabetes, and the women’s personal concerns about receiving a diagnosis of diabetes in the future. The first two interviews were attended by all three researchers. A subsequent debriefing was held to discuss any potential useful adjustments to the interview guide. All researchers agreed that the questions on the guide solicited relevant and detailed responses. Thus, no modifications were made to the guide.
Of the 139 women who enrolled in the study, 21 (15%) did not complete diabetes testing in the immediate postpartum period or thereafter; 118 (85%) women were tested (Table 1). On average, 2.7 years had passed since participants’ GDM pregnancy. Mean age was 38 (SD, 4.2) years, and mean (SD) number of children under 12 years old in each woman’s home was 1.8 (0.07). Most (78%) of the women in the study were married; the sample was educated and diverse. Almost 70% of the women (compared with 46.3% of the Massachusetts population overall, according to the 2007–2011 American Community Survey) had a college degree and 46% (compared with 19.6% of the Massachusetts population overall, according to the 2010 U.S. Census [2012]) were non-White.
Data Analysis Respondent’s testing status (taken from the EMR) was merged with the survey data and analyzed using SAS 9.3 (SAS Institute, Inc., Cary, NC). First, we assessed bivariate relationships between testing status and sociodemographic characteristics, barriers to getting tested, risk for developing type 2 diabetes, medical care utilization, and lifestyle behaviors. Next, we conducted logistic regression models to assess the influence of each of several variables, chosen a priori or on the basis of differences in the bivariate analysis, on the likelihood of receiving testing according to ADA guidelines. We assessed the odds of not being testing for diabetes to approximate relative risk of failing to test because this condition was rarer among women than being tested for diabetes. Because of the small sample size, our ability to run multivariable regression was limited and a parsimonious model was needed. Therefore, we adjusted covariates with propensity score to correct for biases in the sample. Next, we ran a logistic regression model with the exposure of interest and the propensity score as independent variables to estimate the risk of having been tested. For the qualitative portion of the study, we audiotaped the interviews and transcribed them verbatim and then imported the data into NVivo for analysis. The codebook was developed in an iterative fashion with multiple rounds of revision after two members of the team reviewed each transcript, reviewed the
Results
Factors Associated with Postpartum Diabetes Testing Status Survey findings A number of factors were correlated with testing status (p < .10). Women who had received the recommended diabetes testing were more likely to be married (81%) than the nontested group (62%). The tested group also had a greater proportion of women who were college graduates (72% vs. 52%). Regarding the items assessing barriers to getting tested, women who had been tested breastfed 2 months longer on average than women who had not been tested. Tested women were also less likely find it difficult to get physical activity because of lack of childcare (45% vs. 76%). More women in the tested group (79%) reported that they believed they had a moderate to high chance of getting diabetes over the next 10 years than did women in the nontested group (52%). Last, more women in the tested group (81%) reported that they saw an internal medicine or family medicine provider for routine preventive care than did women in the nontested group (62%). In Table 2, we present results from unadjusted and adjusted logistic regression models, which identify correlates of not receiving a diabetes tested. In the adjusted analysis using a propensity score, we found that women who had difficulty getting childcare for physical activity were 3.6 times greater odds (95% CI, 1.2–10.8) of not testing than those women who did have access to childcare for exercise. Women who identified as black had 3.5 times greater odds (95% CI, 1.0–12.4) of not being tested than women identifying as White. Seeing an internal medicine or family medicine provider for preventive care lowered the odds (odds ratio, 0.28; 95% CI, 0.09–0.83) of not testing compared with those who did not see a primary care provider. Insulin use also decreased the odds of not being tested but the confidence interval included 1.0 (odds ratio, 0.62; 95% CI, 0.20–1.91). College education was not associated with not receiving a diabetes test in the unadjusted and adjusted regression models. Qualitative findings Barriers to testing. Five of the 22 interviewees did not receive recommended diabetes testing and 8 women discussed barriers
e376
K.A. Paez et al. / Women's Health Issues 24-4 (2014) e373–e379
Table 1 Comparison of Women with a History of Gestational Diabetes Mellitusy Who Did Not Receive a Postpartum Diabetes Test* with Those Who Were Tested Domain and Characteristic Sociodemographics Age (y), mean (SD) College graduate, n (%) Yes No Married, n (%) Yes No Race/ethnicity, n (%) White Asian Black Other Barriers to diabetes testing Reasons it is hard to get physical activity, n (%) Lack childcare Too much effort Work 20 hours per week No. of children 1 time per month, n (%) Time since delivery, years, mean (SD)
Total (n ¼ 139)
Not Tested (n ¼ 21)
Tested (n ¼ 118)
p-Value
38.2 (4.2)
38.4 (5.0)
38.2 (4.1)
.87
96 (69) 43 (31)
11 (52) 10 (48)
85 (72) 33 (22)
.07
108 (78) 31 (22)
13 (62) 8 (38)
95 (81) 23 (19)
.06
74 29 20 14
(54) (21) (15) (10)
11 2 6 2
(52) (10) (29) (10)
63 27 14 12
(55) (23) (12) (10)
.18
69 43 95 1.8 6.2 6.8
(50) (31) (68) (0.7) (4.7) (5.5)
16 7 17 1.8 4.5 5.4
(76) (33) (81) (0.7) (4.4) (4.3)
53 36 78 1.8 6.5 7.0
(45) (31) (66) (0.7) (4.7) (5.6)
.01 .83 .18 .73 .07 .21
51 (37) 104 (75)
5 (24) 11 (52)
46 (39) 93 (79)
.18 .01
5.7 (0.9) 96 (73)
5.9 (0.8) 14 (74)
5.7 (0.9) 82 (73)
.45 .92
13 14 4 2.4
95 76 11 2.8
.06 .84 .19 .20
108 90 15 2.7
(78) (65) (11) (1.3)
(62) (67) (19) (1.4)
(81) (64) (9) (1.3)
Abbreviations: OB/GYN, obstetrics/gynecology; SD, standard deviation. * Postpartum diabetes testing: Oral glucose tolerance test (OGTT) or a fasting blood sugar (FBS) at 6–12 weeks postpartum or an OGTT, FBS, or hemoglobin A1C after 12 weeks postpartum. y Baseline survey data from 139 women participating in the Avoiding Diabetes After Pregnancy Trial (ADAPT). z Edinburgh Depression Scale is on 0–30 scale: 13 indicates possible depression. x Personal Control Over Diabetes Prevention scale consists of four questions with responses on a 7-point Likert scale; range is 4–28. A higher score indicates a perception of greater personal control over preventing diabetes.
to testing (Table 3). The explanations that these women gave for not getting tested paralleled and extended the findings from the survey and regression models. Four out of the five nontested interviewees said that they found it challenging to fit in the testing around their work schedules or caregiving demands. Another commonly reported challenge from nontested interviewees was lack of follow-up from providers. Although most respondents took responsibility for not keeping appointments for follow-up testing, several pointed out that their doctors did not “press them” about it or follow up when appointments were missed. In some instances, interviewees reported that they had expected to be contacted by their providers after missing appointments, but were not. In other instances, interviewees reported that their providers made it seem as if their diabetes would resolve after delivery and that further testing or behavioral and lifestyle management was not necessary. In addition to the challenges that nontested interviewees reported as keeping them from getting tested, the data also suggest that some women who self-monitor their blood glucose levels are less likely to get tested for diabetes. Several nontested interviewees seemed to be using self-testing as a replacement for formal testing. When this home testing resulted in normal levels, interviewees believed that there was no need for formal glucose testing. One nontested interviewee respondent jokingly clarified that she was not “playing doctor” by self-testing her blood glucose levels.
Facilitators to testing. Seventeen of the 22 interviewees had been tested for diabetes and 11 women discussed facilitators to receiving a postpartum diabetes test (Table 4). In the interview discussions, several of these women also noted that they were receiving ongoing follow-up testing after the initial postpartum test as well (Table 4). The most common impetus to testing was related to health system factors. Specifically, women talked about their testing as being a part of the routine preventive primary care that they received annually or biannually. Several women noted that their providers ordered diabetes tests and discussed the results as they related to their ongoing risk of getting diabetes in the future during these checkups. Thus, routine preventive care, especially in the primary care setting, seems to stimulate appropriate diabetes testing postpartum. Another theme that emerged from women who had been tested was that their doctors expressed that it was especially important that they engage in routine, ongoing testing because of their health history. For example, two interviewees noted that their doctors stressed the importance of their continued testing because they had a family history of diabetes. Another interviewee shared that her doctor ordered ongoing tests for her because she tested as having prediabetes even before her experience with GDM. It seems that when a woman has an underlying increased risk, consistent recommendation from her doctor for follow-up testing is influential.
K.A. Paez et al. / Women's Health Issues 24-4 (2014) e373–e379 Table 2 Characteristics Predicting the Odds of Not Having Received a Postpartum Diabetes Test* Characteristic
Sociodemographics
College graduate No 1.0 (ref) 1.0 (ref) Yes 0.4 (0.2, 1.1) 0.6 (0.2, 1.6) Race/ethnicity White, reference 1.0 (ref) 1.0 (ref) Asian 0.4 (0.1, 2.0) 0.5 (0.1, 2.7) Black 2.5 (0.8, 7.8) 3.5 (1.0, 12.4) Other 1.0 (0.2, 4.9) 1.3 (0.2, 7.3) Hard to get physical activity because do not have childcare No, reference 1.0 (ref) 1.0 (ref) Yes 3.9 (1.3, 11.4) 3.6 (1.2, 10.8) Insulin during pregnancy No 1.0 (ref) 1.0 (ref) Yes 0.5 (0.2, 1.4) 0.6 (0.2, 1.9) Saw internal medicine or family medicine provider No 1.0 (ref) 1.0 (ref) Yes 0.4 (0.1, 1.1) 0.3 (0.1, 0.8)
Barriers
Risk
Utilization
Unadjusted Odds Ratio (95% CI)
Adjustedy Odds Ratio (95% CI)
Domain
* Postpartum diabetes testing: Oral glucose tolerance test (OGTT) or a fasting blood sugar (FBS) at 6–12 weeks postpartum or an OGTT, FBS, or hemoglobin A1C after 12 weeks postpartum. Baseline survey data from 139 women participating in the Avoiding Diabetes After Pregnancy Trial. y Independent variables in the logistic regression model include the characteristic of interest and a propensity score. The propensity score, developed from all other characteristics listed in the table, adjusts for bias in the sample.
Another common motivation that interviewees reported for getting tested was the desire to “know their status.” In these discussions, interviewees expressed that it was better to know whether their blood sugar level was elevated rather than not know. These women expressed discomfort with not knowing whether a potential risk was real or not. In addition to these common themes, there were three other factors that were associated with getting tested that only one person each discussed. One woman described getting tested as a way of proactively managing her health so that she did not become ill and unable to take care of her children. She worried that, if she got diabetes, she would feel so physically sick that she would be unable parent successfully. Similarly, one woman who intended to have more children in the future believed that it was important to “keep an eye” on her blood sugar to avoid any complications that might jeopardize her ability to do this successfully. Not only was she engaging in routine testing, this woman also intended to get tested earlier than recommended once she became pregnant again so that she could stay informed of her health status. Last, one woman reported that self-testing motivated her to get tested. She was monitoring her blood sugar at home and, after seeing several results that indicated elevated blood glucose levels, decided to get tested formally. Because in other women normal results from self-testing signaled that there was no need for formal testing, self-testing can be both a facilitator and a barrier to testing, depending on its results. Discussion In this mixed method study, we examined a variety of sociodemographic, perceptual, and health care-related factors that could influence whether women with a previous history of GDM received a postpartum diabetes test according to ADA guidelines.
e377
Our study provides insights into why women who are most resistant to testing fail to receive a diabetes test after a pregnancy with GDM. Among these hard-to-reach women, our survey revealed that having follow-up primary care was associated with increased testing rates. Women identifying as black were much less likely to be tested than white women. Difficulty finding childcare to exercise, a proxy for support systems that enable or inhibit women to seek care, was associated with receipt of testing. The in-depth interviews reinforced survey findings by providing evidence that a lack of childcare, childcare demands, and an inability to find time owing to work were common explanations for not receiving a postpartum diabetes test. Women whose GDM was more severe either because they perceived it to be so or took insulin during pregnancy were more apt to be tested. Providers’ reassurance that diabetes would resolve after delivery sent a message that postpartum testing was unnecessary; women whose primary care providers discussed diabetes risk believed testing was important. Some women continued to self-monitor their blood glucose immediately after delivery and believed normal readings indicated a laboratory test was not needed. The high testing rate among study participants may be partially attributed to a quality improvement program that HVMA began in late 2010, just before we began ADAPT recruitment. HVMA set a goal that all women diagnosed with GDM would complete an OGTT within 8 weeks postpartum. HVMA obstetric practices implemented a standardized order set and nurses followed up monthly by phone and mail with women who had not been tested. Studies of similar programs have found that patient and provider reminders are an effective approach to increasing postpartum diabetes testing rates (Carson, Frank, & Keely, 2013). Nevertheless, the testing rate was not 100%, and our study provides information on modifiable factors that could decrease the remaining gap. Physicians’ perception of a woman’s severity of diabetes risk may be an important factor in determining how they interact with women who have had GDM. Our findings suggest that women who are more difficult to reach require education from providers about their postpartum diabetes risk and an explanation that self-monitoring is not a diagnostic tool. It is possible that when women had multiple risk factors for diabetes such as a family history, prediabetes before pregnancy, and insulin use during pregnancy, primary care physicians were more likely to emphasize diabetes risk reduction and the need for routine postpartum diabetes testing. As shown in our qualitative data, patient–physician interactions clearly influenced women’s attitudes about their diabetes risk and the need for testing. Our study confirms what has been reported in the literature about barriers to diabetes testing after a pregnancy with GDM and adds to the literature by providing insights into how providers can influence women to test or not. In a survey of Canadian women who had a history of GDM, time pressures were the most frequent reason for failing to be tested (Keely et al., 2010). In a university hospital setting, interviews with women around the time of the first postpartum visit found that feelings of emotional stress owing to adjusting to a new baby and the fear of receiving a diabetes diagnosis were key barriers while access to child care and desire for a checkup were among the key facilitators to seek out care (Bennett et al., 2011). Both of these studies reported on cases where women failed to test because they substituted self-monitoring of glucose for laboratory testing. Limitations to this study should be considered in interpreting the findings. The small sample size may have precluded identifying some associations of interest. Generalizability to all practices or all women with GDM may be limited by the relatively high
e378
K.A. Paez et al. / Women's Health Issues 24-4 (2014) e373–e379
Table 3 Findings From In-Depth Interviews Among 22 Participants in the Avoiding Diabetes after Pregnancy Trial: Barriers to Testing (n ¼ 8*) Barrier
No. of Women Reporting*
Sociodemographic-related barriers Difficulty coordinating childcare logistics to attend an appointment
7 6
Lack of insurance
1
Inflexible work schedules incompatible with lab hours
1
Fear of being diagnosed
1
Health system–related barriers Lack of follow-up from physicians
5 5
Self-assessments of health status
3
Time for OGTT test too long.
2
Availability of convenient lab hours
1
*
Exemplar Quote
“I don’t think there was anything that made me hesitate other than, you know, life with a newborn and two other children. And, again, not looking to make excuses, but my husband works – you know, usually at least I’d say 80 hours a week. And it’s sometimes a little difficult to try to schedule something not knowing if I’m going to have childcare. And I hate canceling things at the last moment because he’s not available to be home. So I think it was just one of those things I intended to do. And there wasn’t anything I intentionally put off for fear of hearing that I had diabetes. It just truly was a case of expecting at some point to have the time to do it in.” “This year I haven’t been tested yet. We have some issues with insurance right now, so I’m going to wait until January to do all my testing.” “The only thing that I was hesitant about is just the time of going in and trying to work with my schedule. That was the only thing. Because I was working. And I couldn’t really go to the doctor as much because of my schedule. That was the only thing that really stopped me from being active about getting it checked out.” “I know I should do it. But it’s almost like – this is going to sound awful – but you want honest answers. Like, no news is good news. And my family is like, oh my gosh, if I had that, what would I do? It’s awful.It’s, like, oh my gosh, I don’t want to have it. And so, I guess, in my mind, it’s been, if I don’t get checked, maybe I won’t develop it.” “No. I mean, that wasn’t really followed up on. I can’t recall it being, like, a subject matter. Basically, it was on my plate to follow through. But it wasn’t as though I received a call or a request from my primary [care doctor].” Q. “And how long did you check your blood glucose? I mean, did you do it until you lost weight, or...?” A. “No, I would say the first 6 weeks, I think, I was checking it. That’s more my reason why I didn’t go for my 6-week testing, because I already knew it was gone.” “I mean, they wanted me to do, you know, the 6 weeks. But after that, where I told them, you know, I didn’t have the time and stuff. Because you have to drink the drink, wait a while, you know? And then take the blood work. ” “The appointments that were available to me, even when I was pregnant, were late at night, like 6:30, 7:00 at night, and that really didn’t work for my schedule.”
Barriers were mentioned in 20 instances by 8 women.
education level of the participants as well as their obtaining care in a group medical practice with a well-developed EMR and reminder systems. Strengths included the racial/ethnic diversity of the population and the mixed methods approach in which the qualitative and quantitative results complemented each other.
Implications for Policy and/or Practice The findings of this study may inform the development of effective strategies to increase testing, particularly among women who are most difficult to reach, and will help providers
Table 4 Facilitators to Testing (N ¼ 16*) Facilitators
No. of Interviewees Reporting*
Health system–related facilitators Routine preventive care
11 7
Increased provider vigilance owing to family history
2
Increased provider vigilance owing to pre-diabetes risk
1
Desire to “know status”/ uncomfortable with not knowing
3
Subsequent pregnancy
1
Desire to be healthy for kids
1
Self-assessment of health status
1
Abbreviation: PCP, primary care physician. * Facilitators were mentioned in 16 instances by 11 women.
Exemplar Quote
“So going forward after I had my gestational diabetes, after my pregnancy, every year they, my doctors required me to get tested for gestational diabetes.” Q: “So do you think [your provider] felt it was a pretty routine test? A: “Yeah, ‘routine’ as far as he knew of my history, and my mom has diabetes, so that was also another concerning part with the test.” “I was already being tested every time I went to the doctor’s, because I was already overweight pre-pregnancy. So, it was something that my PCP was doing every time I came in for an annual physical anyway.” “I know that there is risk associated with people who had gestational diabetes that, within a certain time period, it’s likely that the person will develop diabetes. And so, I just want to know, as years pass, how I’m doing.” “But in the back of my mind, I’m kind of thinking, you know, eventually I’ll probably have another child. So I’ll definitely be cognizant of it. And we’ll probably get tested prior to the six-month mark of the pregnancy, you know, to keep it in check, to see if it recurs, kind of thing.” Q: “What do you think is your primary motivation for [getting tested]? A: “I just want to be healthy for my kids.” “And, I think earlier this summer, when my weight was probably at the highest it’s ever been, I did get myself tested. Because of that, but I was like, hmm, I should get myself tested.”
K.A. Paez et al. / Women's Health Issues 24-4 (2014) e373–e379
to recognize the characteristics of women who are least likely to get tested and thus who may require special attention and targeted outreach. Our study demonstrates the importance of health system factors and the provider–patient relationship in influencing women to be tested for type 2 diabetes after a pregnancy complicated by GDM. There are several implications of our findings. The delivery system should make efforts to effect a hand-off from obstetric to primary care after delivery. Systematic reminders to women who do not receive timely diabetes testing signal to women that testing is important. From the first evidence of GDM to after delivery, both obstetric and primary care providers should reinforce to women with GDM that their elevated risk for developing type 2 diabetes requires proactive follow-up care. Woman should be encouraged to make an appointment after delivery with a primary care provider so that a routine schedule for diabetes testing in a laboratory setting (rather than selfmonitoring) can be initiated. In addition, health care providers, employers, and families need to consider the multiple demands on women’s time and make it feasible for them to obtain the recommended testing. References American Diabetes Association (ADA). (2010). Standards of medical care in diabetes-2012. Diabetes Care, 35, s11–s63. Bellamy, L., Casas, J. P., Hingorani, A. D., & Williams, D. (2009). Type 2 diabetes mellitus after gestational diabetes: A systematic review and meta-analysis. Lancet, 373, 1773–1779. Bennett, W. L., Ennen, C. S., Carrese, J. A., Hill-Briggs, F., Levine, D., Nicholson, W. K., et al. (2011). Barriers to and facilitators of postpartum follow-up care in women with recent gestational diabetes mellitus: A qualitative study. Journal of Women’s Health, 20, 239–245. Bentley-Lewis, R., Levkoff, S., Stuebe, A., & Seely, E. W. (2008). Gestational diabetes mellitus: Postpartum opportunities for the of type 2 diabetes mellitus. Endocrinology & Metabolism, 4, 552–558. Carson, M. P., Frank, M. I., & Keely, E. (2013). Postpartum testing rates among women with a history of gestational diabetes: Systematic review. Primary Care Diabetes, 7, 177–186. Casey, B. M., Lucas, M. J., Mcintire, D. D., & Leveno, K. J. (1997). Pregnancy outcomes in women with gestational diabetes compared with the general obstetric population. Obstetrics & Gynecology, 90, 869–873. Clark, H. D., Graham, I. D., Karovitch, A., & Keely, E. (2009). Do postal reminders increase postpartum screening of diabetes mellitus in women with gestational diabetes mellitus? A randomized clinical trial. American Journal of Obstetrics and Gynecology, 200, e631–e637, 634. Cox, J. L., Holden, J. M., & Sagovsky, R. (1987). Detection of postnatal depression. Development of the 10-item Edinburgh Postnatal Depression Scale. British Journal of Psychiatry, 150, 782–786. Ekelund, M., Shaat, N., Almgren, P., Groop, L., & Berntorp, K. (2010). Prediction of postpartum diabetes in women with gestational diabetes mellitus. Diabetologia, 53, 452–457. Ferrara, A., Peng, T., & Kim, C. (2009). Trend in postpartum diabetes screening and subsequent diabetes and impaired fasting glucose among women with histories of gestational diabetes mellitus. Diabetes Care, 32, 2269–2274. Hale, N. L., Probst, J. C., Liu, J., Martin, A. B., Bennett, K. J., & Glover, S. (2012). Postpartum screening for diabetes among Medicaid-eligible South Carolina women with gestational diabetes. Women’s Health Issues, 22, e163–e169. Henderson, C. E., Kavookjian, J., Leitstein, H., McKoy, J. M., Murage, W. J., & Lipman, R. D. (2012). Window of opportunity: Postpartum screening of women with gestational diabetes for early detection of prediabetes and type 2 diabetes. Open Diabetes Journal, 5, 25–28.
e379
Hunsberger, M. L., Donatella, R. J., Lindsay, K., & Rosenberg, K. D. (2012). Physician care patterns and adherence to postpartum glucose testing after gestational diabetes mellitus in Oregon. PLOS One, 7, e47052. Johns, K., Olynik, C., Mase, R., Kreisman, S., & Tildesley, H. (2006). Gestational diabetes mellitus outcome in 394 patients. Journal of Obstetrics and Gynaecology Canada, 2, 122–127. Keely, E., Clark, H., Karovitch, A., & Graham, I. (2010). Screening for type 2 diabetes after gestational diabetes. Canadian Family Physician, 56, 558–563. Michigan Diabetes Research and Training Center. (2014). Risk Perception Survey for Developing Diabetes (RPS-DD), adapted for women with histories of gestational diabetes. Available at: http://www.med.umich.edu/mdrtc/profs/ survey.html. Ratner, R. E., Christophi, C. A., Metzger, B. E., Dabelea, D., Bennett, P. H., PiSunyer, X., et al. (2008). Diabetes Prevention Program Research Group. Prevention of diabetes in women with a history of gestational diabetes: effects of metformin and lifestyle interventions. Journal of Clinical Endocrinology Metabolism, 93, 4774–4779. Regnault, N., Gillman, M. W., Rifas-Shiman, S. L., Eggleston, E., & Oken, E. (2013). Sex-specific associations of gestational glucose tolerance with childhood body composition. Diabetes Care, 36, 3045–3053. Shea, A. K., Shah, B. R., Clark, H. D., Malcolm, J., Walker, M., Karovitch, A., et al. (2011). The effectiveness of implementing a reminder system into routine clinical practice: Does it increase postpartum screening in women with gestational diabetes? Chronic Disease in Canada, 331, 58–62. U.S. Census. (2012). State and county quick facts. Available at: http://quickfacts. census.gov/qfd/states/00000.html. Vesco, K. K., Dietz, P. M., Bulkey, J., Bruce, F. C., Callaghan, W. M., England, L., et al. (2012). A system-based intervention to improve postpartum diabetes screening among women with gestational diabetes. American Journal of Obstetric Gynecology, 207, 283.e1–283.e6.
Author Descriptions Kathryn A. Paez, RN, PhD, is a principal researcher in Health Policy and Research at the American Institutes for Research. Her research interests concern improving quality of care and access to care to improve population health.
Emma M. Eggleston, MD, MPH, is an instructor in the Department of Population Medicine at Harvard Medical School and the Harvard Pilgrim Health Care Institute. Her clinical and research interests are in diabetes, pregnancy, and vascular disease.
Susan J. Griffey, DrPH, BSN, directs the Evaluation Center at Social & Scientific Systems. She researches public health interventions and programs in the US and in global development settings in primary and reproductive health care, HIV/AIDS, and chronic diseases.
Brandy Farrar, PhD, is a researcher in Health Policy and Research at the American Institutes for Research. The majority of Ms. Farrar’s research involves evaluating the effectiveness, viability, and impact of innovative programs designed to improve quality, access, and capacity of health care services.
Jacquelyn Smith, MA, is a senior survey manager at Social & Scientific Systems. She applies her 20 years of experience in survey methodology and survey operational management to numerous public health studies in both acute and chronic disease topics.
Jennifer Thompson, MPP, is a project manager in the Obesity Prevention Program in the Department of Population Medicine at the Harvard Pilgrim Health Care institute and Harvard Medical School. Her projects focus on diabetes, early childhood nutrition, and obesity.
Matthew Gillman, MD, SM, is a Professor and Director of the Obesity Prevention Program, which seeks to lessen the burden of obesity and its consequences, in the Department of Population Medicine at the Harvard Pilgrim Health Care Institute and Harvard Medical School.
