Objective
We sought to determine whether our process improvement program led to increased postpartum diabetes screening rates among women with gestational diabetes mellitus (GDM).
Study Design
In early 2009, we conducted obstetrics department staff education sessions, revised GDM patient care protocols, and developed an electronic system to trigger reminder calls to patients who had not completed diabetes mellitus screening by 3 months postpartum. We then evaluated the rates of postpartum glucose test order entry and completion for women with GDM delivering from July 2009 through June 2010 (n = 179) and July 2007 through June 2008 (n = 200).
Results
After the program’s implementation, the proportion of women receiving an order for a postpartum glucose test within 3 months of delivery increased from 77.5-88.8% ( P = .004), and test completion increased from 59.5-71.5% (hazard ratio, 1.37; 95% confidence interval, 1.07–1.75).
Conclusion
Rates of postpartum diabetes testing can be improved with system changes and reminders.
Women who have had gestational diabetes mellitus (GDM) have a 35-60% risk of developing diabetes mellitus in the next 10-20 years. In the immediate postpartum period, up to 10% of women who had GDM will be diagnosed with diabetes and an additional 12-36% will have impaired fasting glucose or impaired glucose tolerance. Postpartum screening of women with pregnancies complicated by GDM is recommended by both the American Congress of Obstetricians and Gynecologists and the American Diabetes Association to identify women with glucose abnormalities and provide appropriate counseling and referrals. Previous studies of postpartum diabetes screening among women with GDM-affected pregnancies have noted test completion rates ranging from 14-61%. In a previous study at Kaiser Permanente Northwest (KPNW), we used electronic medical record data to evaluate the proportion of women with a GDM-affected pregnancy who received an order for a postpartum fasting glucose or oral glucose tolerance test (OGTT) and the proportion of women with GDM who completed a postpartum glucose test. We found that 79% of affected women received an order for a postpartum glucose test, but only 58% of the women completed screening. Order receipt and test completion rates varied substantially by clinical practice site within KPNW; order receipt rates ranged from 29–94% and test completion rates among women who received an order ranged from 42–85%, suggesting some clinics could improve their postpartum order entry and test completion rates with additional training and improvements in clinical systems.
In an effort to improve postpartum screening rates within the KPNW health system, we implemented a 2-tiered process improvement program. The first step targeted obstetrics department staff (clinicians, nurses, and medical assistants) with the goal of increasing both order entry and postpartum test completion rates within the first 90 days of delivery. The second step directly targeted women with GDM-affected pregnancies who had not completed testing within the first 90 days to further improve the rate of postpartum testing beyond the immediate postpartum period. This manuscript describes our process improvement project and results of postpartum diabetes testing before and after implementation. Our objective was to determine whether our process improvement program increased the number of women who received orders for testing and the number who completed testing.
Materials and Methods
KPNW GDM care
KPNW is a large, nonprofit, prepaid, federally certified, Joint Commission–accredited, group practice health maintenance organization with 470,000 members in western Oregon and Washington state. Most members receive prenatal care at 1 of 8 practice sites; 1 site serves as a maternal-fetal medicine referral practice. Screening for GDM is a component of usual care for pregnant women at KPNW. During the study period (June 2007 through December 2010), the standard of care was to screen all pregnant women between 24-28 weeks’ gestation with a 50-g, 1-hour glucose challenge test (GCT). Screening is also performed in early pregnancy for women with risk factors such as previous history of GDM. Those with 1-hour values ≥130 mg/dL are offered a 100-g, 3-hour OGTT. In the preintervention period (deliveries occurring July 2007 through June 2008), the National Diabetes Data Group (NDDG) criteria (fasting glucose >105 mg/dL, 1-hour >190 mg/dL, 2-hour >165 mg/dL, 3-hour >145 mg/dL) were used to establish a laboratory diagnosis of GDM. Standard clinical practice at that time was to diagnose a woman as having GDM if she had a fasting glucose >105 mg/dL alone, any 2 abnormal values on the 3-hour OGTT, or a 1-hour GCT ≥200 mg/dL. During the postintervention period (deliveries occurring July 2009 through June 2010), KPNW switched to the Carpenter and Coustan cut-points (fasting glucose >95 mg/dL, 1-hour >180 mg/dL, 2-hour >155 mg/dL, 3-hour >140 mg/dL) for the 3-hour OGTT. Because the switch occurred midway during our postimplementation period, we considered only NDDG cut-points and the results of the 1-hour GCT for establishing the diagnosis of GDM for this analysis.
Within KPNW, a fasting plasma glucose (FPG) test is the usual method of postpartum screening for diabetes mellitus; however, clinicians may order a 2-hour OGTT if they prefer.
Design
We used a pre-/postintervention design to assess the efficacy of our process improvement project. The study included 3 main periods from June 1, 2007, through Dec. 31, 2010: preimplementation (June 1, 2007, through Dec. 31, 2008), implementation (Jan. 1, 2009, through May 31, 2009), and postimplementation (June 1, 2009, through Dec. 31, 2010). Figure 1 outlines the study design. The preimplementation cohort included all women with GDM-affected pregnancies who delivered from July 1, 2007, through June 30, 2008. The postimplementation cohort included all women with GDM-affected pregnancies who delivered from July 1, 2009, through June 30, 2010. The implementation phase was from Jan. 1, 2009, through May 31, 2009, during which time provider education (Jan. 1, 2009, through April 1, 2009) and system changes took place.
Process improvement intervention
The goal of the project was to increase the proportion of women for whom a postpartum glucose test was ordered (provider behavior) and to increase the proportion of women who completed the screening test (patient behavior). We employed multiple strategies including revising the nursing protocol for care of pregnant women with GDM, enhancing the electronic medical record system, educating clinical staff, and providing additional reminders to women who did not complete the test within 3 months of delivery.
The GDM nurse protocol was revised as part of our project to allow the GDM nurse care managers to place the order for the postpartum FPG, increasing the likelihood for order placement. As an enhancement to the electronic medical record system, we revised electronic order entry sets (a set of commonly placed orders that have been grouped together for quick identification and order entry) for women with diabetes (used primarily by nurses and clinicians) and for women attending postpartum visits (used primarily by clinicians and medical assistants) to include a check box for the postpartum fasting glucose testing as a reminder to order the FPG for women with GDM. In addition, the postpartum order entry set was revised to trigger a second reminder to clinicians to order a test or remind women to complete the test. During the intervention period, all GDM nurse care managers were encouraged to make at least 1 postpartum reminder call.
Staff education
After revision of the protocol and electronic order entry sets, the first author (K.K.V.) gave a series of presentations on postpartum screening for women with recent GDM at regional department and local clinic meetings. In addition, a half-day diabetes training program was provided to the obstetrics department nursing staff. We also provided the clinic staff a handout to give to women with GDM that describes the importance of follow-up screening and the long-term risks of developing diabetes mellitus.
Additional screening reminders
The intervention included development of an electronic algorithm that automatically searched the electronic data systems and created a list identifying women who had not completed their glucose test by 3 months postpartum. We set an a priori goal for our GDM nurse care manager (K.J.H.) to call each woman who appeared on the tracking list up to 3 times to remind her to go to the laboratory to complete her screening. If after the third attempt the patient could not be reached, our nurse would send her an e-mail if she had an active e-mail account, otherwise, she would send a letter by US mail. All attempts to contact participants ended 1 month prior (Nov. 30, 2010) to the end of the follow-up period (Dec. 31, 2010).
Process and outcome evaluation
We evaluated the intervention process by assessing the proportion of the obstetrics department staff (medical assistants, nurses, and clinicians) that attended at least 1 educational meeting and we collected and summarized satisfaction ratings of the attendees regarding the presentations. The date and type (telephone contact, letters, e-mails, and verbal in-person reminders) of all glucose test screening reminders were abstracted from medical records. Reminders were categorized as occurring within the first 3 months after delivery or >3 months’ postpartum.
Two main outcomes were evaluated: (1) clinician orders for either a fasting blood glucose test or 2-hour 75-g OGTT placed from 1 month before through 3 months after delivery; and (2) completed fasting glucose or 2-hour OGTT tests performed between 14 days after delivery and Dec. 31, 2008, for the preimplementation period and Dec. 31, 2010, for the postimplementation period. Test results were defined as normal (fasting <100 mg/dL or 2 hours after a 75-g glucose load <140 mg/dL), impaired (fasting 100-125 mg/dL or 2-hour 140-200 mg/dL), or diabetes (fasting >125 mg/dL or 2-hour >200 mg/dL). For women who had both a fasting and a 2-hour glucose test, the most abnormal result of the 2 is reported.
We compared women in both study periods by demographic and clinical characteristics. Prepregnancy body mass index, maternal age, parity, race/ethnicity, the woman’s primary language (English, yes/no), having a clinical visit with a nutritionist during pregnancy, use of insulin or glyburide during pregnancy, trimester of entry into prenatal care (initiated prenatal care in first trimester, yes/no), having a postpartum visit within 3 months of delivery, practice site where the woman received her postpartum visit, and infant birthweight (>4000 g, yes/no) were obtained from KPNW electronic medical records.
Differences in pre- and postimplementation cohort characteristics were assessed using t tests and χ 2 tests for independence. The primary independent variable was the period (preimplementation vs postimplementation). Cox proportional hazards models were used to evaluate test completion. None of the tested covariates was significantly related to period. Therefore, testing of the intervention effect on orders placed and tests completed was done without controlling for covariates. For Cox models, the time variable was calculated as time from the date of delivery to the date of test completion. Women who did not complete glucose testing were censored at the end of their respective study periods (Dec. 31, 2008, for preimplementation and Dec. 31, 2010, for postimplementation), or at the date a new pregnancy was identified, or at time of discontinuation of insurance coverage through KPNW if these occurred first.
We used software (SAS, version 9.2; SAS Institute Inc, Cary, NC) for all statistical programming. This study was approved by the institutional review boards of the Centers for Disease Control and Prevention and KPNW.
Results
Staff education
Approximately 72% (64/89) of clinicians, 90% (45/50) of registered nurses, and 66% (35/53) of medical assistants within the obstetrics department attended at least 1 education meeting. Mean satisfaction with the educational meetings was 4.82 (SD = 0.39) on a scale of 1 (poor) to 5 (outstanding).
Primary outcome data
From the 5250 live and stillbirths occurring within the preimplementation period and the 4765 in the postimplementation period, we identified 379 women with GDM (n = 200 pre, n = 179 post), yielding a GDM prevalence of 3.8% in both periods. There were no significant differences in patient characteristics between the 2 time periods ( Table 1 ).
| Chracteristic | Preimplementation period, n = 200, % | Postimplementation period, n = 179, % | P value a |
|---|---|---|---|
| Method of diagnosis | |||
| 1-h OGCT >200 mg/dL | 14.0 | 12.3 | .89 |
| 2 abnormal values on 3-h OGTT | 63.5 | 64.8 | |
| Fasting glucose >105 mg/dL | 22.5 | 22.9 | |
| Timing of test during pregnancy | .76 | ||
| <20 wk’ gestation | 14.5 | 13.4 | |
| ≥20 wk’ gestation | 85.5 | 86.6 | |
| Prepregnancy BMI | |||
| <25.0 | 25.0 | 24.0 | .83 |
| 25.0-29.9 | 21.5 | 24.0 | |
| ≥30.0 | 41.5 | 38.0 | |
| Missing | 12.0 | 14.0 | |
| Age, y | .19 | ||
| <35 | 70.0 | 63.7 | |
| ≥35 | 30.0 | 36.3 | |
| Parity | .63 | ||
| 0 | 41.5 | 38.0 | |
| 1 | 28.0 | 31.8 | |
| ≥2 | 29.5 | 27.9 | |
| Unknown | 1.0 | 2.2 | |
| Race/ethnicity | .27 | ||
| Non-Hispanic white | 51.5 | 46.9 | |
| Black/other | 7.5 | 3.9 | |
| Hispanic | 9.5 | 12.9 | |
| Asian | 17.0 | 16.2 | |
| Missing | 14.5 | 20.1 | |
| English not primary language | .66 | ||
| Yes | 13.5 | 15.1 | |
| No | 86.5 | 84.9 | |
| Nutrition visit during pregnancy | .09 | ||
| Yes | 75.5 | 82.7 | |
| No | 24.5 | 17.3 | |
| Insulin or glyburide during pregnancy | .07 | ||
| Yes | 26.0 | 34.6 | |
| No | 74.0 | 65.4 | |
| Initiated prenatal care in first trimester | .69 | ||
| Yes | 84.0 | 85.5 | |
| No | 16.0 | 14.5 | |
| Attended postpartum clinic visit | .93 | ||
| Yes | 92.5 | 92.7 | |
| No | 7.5 | 7.3 | |
| Site of postpartum clinic visit | .25 | ||
| Clinic 1 | 22.5 | 26.3 | |
| Clinic 2 | 7.0 | 8.4 | |
| Clinic 3 | 9.0 | 9.5 | |
| Clinic 4 | 18.5 | 15.1 | |
| Clinic 5 | 11.0 | 4.5 | |
| Clinic 6 | 8.5 | 9.5 | |
| Clinic 7 | 13.0 | 14.5 | |
| Other clinic or non-KP provider | 3.0 | 5.0 | |
| No visit/no clinic | 7.5 | 7.3 | |
| Infant birthweight >4000 g | .67 | ||
| Yes | 10.5 | 13.4 | |
| No | 85.0 | 82.7 | |
| Missing | 4.5 | 3.9 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
