A system-based intervention to improve postpartum diabetes screening among women with gestational diabetes.

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.

Extent of maternal morbidity in a managed care population in georgia.

BACKGROUND: Although maternal deaths are among the most tragic events related to pregnancy, they are uncommon in the US and, therefore, inadequate indicators of a woman's pregnancy-related health. Maternal morbidity has become a more useful measure for surveillance and research. Traditional attempts to monitor maternal morbidity have used hospital discharge data, which include data only on complications that resulted in hospitalisation, underestimating the frequency and scope of complications. METHODS: To obtain a more accurate assessment of morbidity, we applied a validated computerised algorithm to identify pregnancies and pregnancy-related complications in a defined population enrolled in a health maintenance organisation in the south-eastern US. We examined the most common morbidities by pregnancy outcome and maternal characteristics. RESULTS: We identified 37 741 pregnancies; in half (50.7%), at least one complication occurred. The five most common were urinary tract infections, anaemia, mental health conditions, pelvic and perineal complications, and obstetrical infections. Complications were more likely in women with low socio-economic status (SES), and among non-Hispanic Black women compared with non-Hispanic White women. Multivariable models stratified by race/ethnicity indicated that in pregnancies among non-Hispanic White women, low SES had a modest effect on the odds of having preexisting medical conditions [adjusted odd ratio (AOR) 1.3 [95% confidence interval (CI) 1.2, 1.5]] or having any morbidity (AOR 1.3 [95% CI 1.2, 1.4]). Low SES had little effect on complications among non-Hispanic Black women. CONCLUSION: Our findings suggest that comprehensive health insurance coverage may lessen the unfavourable impact of socio-economic disadvantage on the risk of maternal morbidity.

Association between obesity during pregnancy and increased use of health care.

BACKGROUND: In the United States, obesity during pregnancy is common and increases obstetrical risks. An estimate of the increase in use of health care services associated with obesity during pregnancy is needed. METHODS: We used electronic data systems of a large U.S. group-practice health maintenance organization to identify 13,442 pregnancies among women 18 years of age or older at the time of conception that resulted in live births or stillbirths. The study period was between January 1, 2000, and December 31, 2004. We assessed associations between measures of use of health care services and body-mass index (BMI, defined as the weight in kilograms divided by the square of the height in meters) before pregnancy or in early pregnancy. The women were categorized as underweight (BMI <18.5), normal (BMI 18.5 to 24.9), overweight (BMI 25.0 to 29.9), obese (BMI 30.0 to 34.9), very obese (BMI 35.0 to 39.9), or extremely obese (BMI > or =40.0). The primary outcome was the mean length of hospital stay for delivery. RESULTS: After adjustment for age, race or ethnic group, level of education, and parity, the mean (+/-SE) length of hospital stay for delivery was significantly (P<0.05) greater among women who were overweight (3.7+/-0.1 days), obese (4.0+/-0.1 days), very obese (4.1+/-0.1 days), and extremely obese (4.4+/-0.1 days) than among women with normal BMI (3.6+/-0.1 days). A higher-than-normal BMI was associated with significantly more prenatal fetal tests, obstetrical ultrasonographic examinations, medications dispensed from the outpatient pharmacy, telephone calls to the department of obstetrics and gynecology, and prenatal visits with physicians. A higher-than-normal BMI was also associated with significantly fewer prenatal visits with nurse practitioners and physician assistants. Most of the increase in length of stay associated with higher BMI was related to increased rates of cesarean delivery and obesity-related high-risk conditions. CONCLUSIONS: Obesity during pregnancy is associated with increased use of health care services.

Maternal morbidity rates in a managed care population.

OBJECTIVE: To identify and estimate prevalence rates of maternal morbidities by pregnancy outcome and selected covariates during the antepartum, intrapartum, and postpartum periods in a defined population of pregnant women. METHODS: We used electronic data systems of a large, vertically integrated, group-model health maintenance organization (HMO) to develop an algorithm that searched International Classification of Diseases, 9th Revision, Clinical Modification, codes for 38 predetermined groups of pregnancy-related complications among women enrollees of this HMO between January 1, 1998, and December 31, 2001. RESULTS: We identified 24,481 pregnancies among 21,011 women. Although prevalence and type of morbidity varied by pregnancy outcome, overall, 50% of women had at least one complication. The most common complications were anemia (9.3%), urinary tract infections (9.0%), mental health conditions (9.0%), hypertensive disorders (8.5%), and pelvic and perineal trauma (7.0%). CONCLUSION: A range of mild-to-severe pregnancy complications were identified using linked inpatient and outpatient databases. The most common complications we found usually do not require hospitalization so would be missed in studies that use only hospitalization data. Our data allowed examination of a broad scope of conditions and severity. These findings increase our understanding of the extent of maternal morbidity. LEVEL OF EVIDENCE: II.

Postpartum screening for diabetes after a gestational diabetes mellitus-affected pregnancy.

OBJECTIVE: To estimate trends in postpartum glucose testing in a cohort of women with gestational diabetes mellitus (GDM). METHODS: A validated computerized algorithm using Kaiser Permanente Northwest automated data systems identified 36,251 live births or stillbirths from 1999 through 2006. The annual percentage of pregnancies complicated by gestational diabetes with clinician orders for and completion of a fasting plasma glucose (FPG) test within 3 months of delivery was calculated. Logistic regression with generalized estimating equations was used to test for statistically significant trends. RESULTS: The percentages of pregnancies affected by GDM increased from 2.9% in 1999 to 3.6% in 2006 (P<.01). Clinician orders for postpartum tests increased from 15.9% in 1999 to 79.3% in 2004 (P<.01), and then remained stable through 2006. Completed FPG tests increased from 9.0% in 1999 to 57.8% in 2004 (P<.01), and then remained stable through 2006. No oral glucose tolerance tests were ordered. From 2004 to 2006, the practice site where women received care was the factor most strongly associated with the clinician order, but it was not predictive of test completion. Among women with clinician orders, those who were Asian or Hispanic or who attended the 6-week postpartum examination were more likely to complete the test than their counterparts. CONCLUSION: Postpartum glucose testing in women with GDM-affected pregnancies increased over time. However, even in recent years, 42% of women with GDM-affected pregnancies failed to have a postpartum FPG test, and no test was ordered for 21% of GDM-affected pregnancies.

Clinically identified maternal depression before, during, and after pregnancies ending in live births.

OBJECTIVE: This study estimated the prevalence of diagnosed depression and treatment among women before, during, and after pregnancies ending in live births. METHOD: A previously validated algorithm identified health plan members with at least one pregnancy between Jan. 1, 1998, and Dec. 31, 2001. Women with a pregnancy ending in one or more live births and continuously enrolled from 39 weeks before pregnancy through 39 weeks after pregnancy were eligible. Maternal depression was identified from the medical records. Depression treatment included antidepressant medication and/or mental health visits. The authors examined the prevalence of depression and treatments received. RESULTS: Among 4,398 continuously enrolled women with eligible pregnancies ending in live births, 678 (15.4%) had depression identified during at least one pregnancy phase; 8.7%, 6.9%, and 10.4% had depression identified before, during, and/or after pregnancy, respectively. Among women with identified depression during the 39 weeks before pregnancy, 56.4% also had a depression diagnosis during pregnancy. Of women identified with depression during the 39 weeks following pregnancy, 54.2% had depression diagnoses either during or preceding pregnancy. Most women diagnosed with depression received antidepressant medications and/or had at least one mental health visit. Having at least one mental health visit did not vary before, during, or after pregnancy; however, antidepressant use was lower during pregnancy than before or after pregnancy. CONCLUSIONS: Approximately one in seven women was identified with and treated for depression during 39 weeks before through 39 weeks after pregnancy, and more than half of these women had recurring indicators for depression.

Complications at the end of life in ovarian cancer.

Women dying of ovarian cancer vary considerably in their complications and in the types of health care they receive. The objective of this study was to describe the complications of ovarian cancer, other than pain, and their treatment at the end of life. This study used a cohort of 421 enrollees in three nonprofit managed-care organizations who died with ovarian cancer during 1995-2000. Data were collected from abstraction of paper and electronic medical records. Proportions of women experiencing complications and undergoing treatments were calculated. Logistic regression was used to evaluate the association of patient characteristics with the probability of receiving an intervention for complications. The most common complications recorded in the medical record were fatigue or weakness (75%), nausea or vomiting (71%), constipation (49%), edema of the extremities (44%), and anemia (34%). The prevalence of major complications was as follows: ascites, 28%; bowel obstruction, 12%; pleural effusion, 10%; bladder obstruction, 3%; and disordered nutrition that required support with parenteral nutrition, 9%. Patients may not always have received interventions for major complications; for example, pleural effusion apparently was left untreated in almost half of the women with this problem. After adjustment, women who died at younger ages were more likely to receive an intervention, compared to older women (odds ratio for each decade of age, 0.71, 95% confidence interval=0.53, 0.94, P for trend=0.02). The study, which preceded the establishment of palliative care programs, suggests that care given to ovarian cancer patients at the end of life may be inadequate.

Development of an algorithm to identify pregnancy episodes in an integrated health care delivery system.

OBJECTIVE: To develop and validate a software algorithm to detect pregnancy episodes and maternal morbidities using automated data. DATA SOURCES/STUDY SETTING: Automated records from a large integrated health care delivery system (IHDS), 1998-2001. STUDY DESIGN: Through complex linkages of multiple automated information sources, the algorithm estimated pregnancy histories. We evaluated the algorithm's accuracy by comparing selected elements of the pregnancy history obtained by the algorithm with the same elements manually abstracted from medical records by trained research staff. DATA COLLECTION/EXTRACTION METHODS: The algorithm searched for potential pregnancy indicators within diagnosis and procedure codes, as well as laboratory tests, pharmacy dispensings, and imaging procedures associated with pregnancy. PRINCIPAL FINDINGS: Among 32,847 women with potential pregnancy indicators, we identified 24,680 pregnancies occuring to 21,001 women. Percent agreement between the algorithm and medical records review on pregnancy outcome, gestational age, and pregnancy outcome date ranged from 91 percent to 98 percent. The validation results were used to refine the algorithm. CONCLUSIONS: This pregnancy episode grouper algorithm takes advantage of databases readily available in IHDS, and has important applications for health system management and clinical care. It can be used in other settings for ongoing surveillance and research on pregnancy outcomes, pregnancy-related morbidities, costs, and care patterns.

Building a virtual cancer research organization.

BACKGROUND: The Cancer Research Network (CRN) comprises the National Cancer Institute and 11 nonprofit research centers affiliated with integrated health care delivery systems. The CRN, a public/private partnership, fosters multisite collaborative research on cancer prevention, screening, treatment, survival, and palliation in diverse populations. METHODS: The CRN's success hinges on producing innovative cancer research that likely would not have been developed by scientists working individually, and then translating those findings into clinical practice within multiple population laboratories. The CRN is a collaborative virtual research organization characterized by user-defined sharing among scientists and health care providers of data files as well as direct access to researchers, computers, software, data, research participants, and other resources. The CRN's research management Web site fosters a high-functioning virtual scientific community by publishing standardized data definitions, file specifications, and computer programs to support merging and analyzing data from multiple health care systems. RESULTS: Seven major types of standardized data files developed to date include demographics, health plan eligibility, tumor registry, inpatient and ambulatory utilization, medication dispensing, laboratory tests, and imaging procedures; more will follow. Data standardization avoids rework, increases multisite data integrity, increases data security, generates shorter times from initial proposal concept to submission, and stimulates more frequent collaborations among scientists across multiple institutions. CONCLUSIONS: The CRN research management Web site and associated standardized data files and procedures represent a quasi-public resource, and the CRN stands ready to collaborate with researchers from outside institutions in developing and conducting innovative public domain research.

Disparities and survival among breast cancer patients.

BACKGROUND: Although rates of survival for women with breast cancer have improved, the survival disparity between African American and white women in the United States has increased. PURPOSE: To determine whether this survival disparity persists in an insured population with access to medical care. METHODS: In this retrospective cohort study, we extracted data from the tumor registries of six nonprofit, integrated health care delivery systems affiliated with the Cancer Research Network and assessed the survival of African American (n = 2276) and white (n = 18 879) female enrollees who were diagnosed with invasive breast cancer from January 1, 1993, through December 31, 1998. Cox proportional hazards regression was used to estimate the death rate among African American women relative to that of white women after adjustment for potential explanatory factors including stage at diagnosis, tumor characteristics, and treatment. RESULTS: Five-year survival was lower for African American women (73.8%) than for white women (81.6%). African American women were less likely to have tumor characteristics with good prognosis. Controlling for age at diagnosis, stage, grade, tumor size, and estrogen and progesterone receptor status, the adjusted hazard rate ratio for African American women was 1.34 (95% confidence interval = 1.22 to 1.46). Similar risks were found among women ages 20-49 and 50 and older. Controlling for treatment slightly lowered the hazard rate ratio to 1.31 (95% confidence interval = 1.20 to 1.43). CONCLUSIONS: Among women with invasive breast cancer, being insured and having access to medical care does not eliminate the survival disparity for African American women.

Issues in pooling administrative data for economic evaluation.

Managed care, in particular the health maintenance organization (HMO), now dominates US healthcare delivery, and economic evaluation is receiving increasing attention as a management tool that can be tailored to its perceived business needs. This encourages use of HMO administrative data as an efficient source of resource utilization and cost measures. Use of administrative data coincides with growing research interest in multisite analyses that increase external validity. The best alternative to a nationally representative data set is to pool administrative data from multiple sites within one database. However, pooling administrative data is problematic because HMO data sources reflect differences in systems of care, costing, and coding. This paper describes issues inherent in the pooling of HMO administrative cost data for use in multisite economic evaluations. We describe the attributes of administrative data that are relevant to costing and discuss their implications for multisite economic evaluations. We then briefly describe our experience with pooling multisite cost data, discuss lessons learned, and offer suggestions for researchers working with such data, followed by concluding comments. Multisite administrative data provide unique opportunities to conduct population-based clinical and economic research.

Newborn size among obese women with weight gain outside the 2009 Institute of Medicine recommendation.

OBJECTIVE: To estimate risk of delivering macrosomic, large-for-gestational-age and small-for-gestational-age neonates in obese women with gestational weight gain outside the 2009 Institute of Medicine recommendation (11-20 pounds). METHODS: In a retrospective cohort study, we evaluated 2,080 obese women (body mass index 30 or higher) with singleton pregnancies that resulted in term live births within one health maintenance organization between 2000 and 2005; women with diabetes or hypertensive disorders were excluded. Gestational weight gain was categorized as less than 0, 0 to less than 11, 11-20 (referent), greater than 20-30, greater than 30-40, and greater than 40 pounds and as above, below, or within Institute of Medicine recommendations. We conducted multivariable logistic regression to estimate the odds of large for gestational age and small for gestational age (birth weights greater than the 90th percentile and less than the 10th percentile for gestational age, respectively) and macrosomia (greater than 4,500 g) adjusting for potential confounders. RESULTS: Eighteen percent gained below, 25% within, and 57% above Institute of Medicine recommendations. Prevalence of macrosomia, large for gestational age, and small for gestational age were 4.3%, 19.8%, and 4.3%, respectively. Compared with weight gain of 11-20 pounds, weight gain above recommendations did not significantly decrease small-for-gestational-age risk but was associated with increased odds of macrosomia (adjusted odds ratio [OR] 3.36; 95% confidence interval [CI] 1.74-6.51; 6.0% compared with 2.1%) and large for gestational age (adjusted OR 1.80; 95% CI 1.36-2.38; 23.8% compared with 16.6%). Weight gain below recommendations was associated with increased odds of small for gestational age (adjusted OR 3.94; 95% CI 2.04-7.61; 8.8% compared with 2.7%) and decreased odds of large for gestational age (adjusted OR 0.56; 95% CI 0.37-0.84; 11.2% compared with 16.6%). CONCLUSION: Regarding small for gestational age and large for gestational age, there is no benefit of weight gain above Institute of Medicine recommendations. Weight gain below recommendations decreases large for gestational age but increases small-for-gestational-age risk. LEVEL OF EVIDENCE: II.

Excessive gestational weight gain and postpartum weight retention among obese women.

OBJECTIVE: To evaluate the incremental effect of weight gain above that recommended for term pregnancy (15 pounds) on postpartum weight retention at 1 year among obese women. METHODS: In a retrospective cohort study, we identified 1,656 singleton gestations resulting in live births among obese women (body mass index at or above 30 kg/m) between January 2000 and December 2005 in Kaiser Permanente Northwest. Pregnancy weight change (last available predelivery weight minus weight at pregnancy onset) was categorized as less than 0, 0-15, greater than 15 to 25, greater than 25 to 35, and greater than 35 pounds. Postpartum weight change (weight at 1 year postpartum minus weight at pregnancy onset) was defined as less than 0, 0-10, and greater than 10 pounds. RESULTS: Total gestational weight gain was -33.2 (weight loss) to +98.0 pounds (weight gain). Nearly three fourths gained greater than 15 pounds, and they were younger and weighed less at baseline than women who gained 15 pounds or less. Pregnancy-related weight change showed a significant relationship with postpartum weight change. For each pound gained during pregnancy, there was a 0.4-pound increase above baseline weight at 1 year postpartum. In adjusted logistic regression models, the risk of a postpartum weight greater than 10 pounds over baseline was twofold higher for women gaining greater than 15 to 25 pounds compared with women gaining 0-15 pounds (odds ratio [OR] 2.18, 95% confidence interval [CI] 1.54-3.10), fourfold higher for women gaining greater than 25 to 35 pounds (OR 3.91, 95% CI 2.75-5.56), and almost eightfold higher for women gaining greater than 35 pounds (OR 7.66, 95% CI 5.36-10.97). CONCLUSION: Incremental increases in gestational weight gain beyond the current recommendation for obese women substantially increase the risk of weight retention at 1 year postpartum. LEVEL OF EVIDENCE: II.

Social support among women who died of ovarian cancer.

GOALS OF WORK: We investigated the effects of social support in the last 6 months of life for women who died of ovarian cancer. MATERIALS AND METHODS: The study population included women enrolled in one of three Managed Care Organizations who died of ovarian cancer (1995-2000). Information was collected on demographics, living environment, presence of escorts to oncology encounters, comorbidities, medications, outpatient and inpatient encounters, and referrals to home health and hospice. Two characteristics of social support were examined: living with others and being escorted to one or more oncology visits. RESULTS: Of 421 subjects, both aspects of social support were known for 345 (82%). Of these, 227 (66%) lived with others and were escorted, 33 (10%) lived with others but were never escorted, 59 (17%) lived alone but were escorted, and 26 (8%) lived alone and never were accompanied. Women living alone were less likely to be taking a psychotropic medication (57% vs 70%, p = 0.021) and were somewhat less likely to receive hospice referral (42% vs 53%, p = 0.054). Women who were never escorted had fewer outpatient encounters (12.60 vs 15.77, p = 0.033) and were less likely to be referred to home health (18% vs 30%, p = 0.046). CONCLUSIONS: This study indicates that social support has some beneficial effects on receipt of personal health services. Friends and family may act as proponents for the patient in obtaining services. Health care professionals should be encouraged to assess the cancer patient's social situation and identify areas where help may be needed.

Pooling multisite administrative data for economic analysis.

Economic evaluations, such as cost-offset analyses, are receiving increased attention by US health insurers and payers. Administrative data collected by health maintenance organizations (HMOs) are considered an efficient source of utilization and cost measures for multisite economic analyses that increase external validity. However, pooling administrative data are problematic because HMO data sources reflect differences in systems of care, costing and coding. This paper describes issues arising from the pooling of HMO administrative cost data for use in multisite economic evaluations. We describe the attributes of administrative data that are relevant to costing and discuss their implications. We then briefly describe our experience with pooling these data, discuss lessons learned and offer suggestions for researchers working with such data.

Risk adjustment using automated ambulatory pharmacy data: the RxRisk model.

OBJECTIVES: Develop and estimate the RxRisk model, a risk assessment instrument that uses automated ambulatory pharmacy data to identify chronic conditions and predict future health care cost. The RxRisk model's performance in predicting cost is compared with a demographic-only model, the Ambulatory Clinical Groups (ACG), and Hierarchical Coexisting Conditions (HCC) ICD-9-CM diagnosis-based risk assessment instruments. Each model's power to forecast health care resource use is assessed. DATA SOURCES: Health services utilization and cost data for approximately 1.5 million individuals enrolled in five mixed-model Health Maintenance Organizations (HMOs) from different regions in the United States. STUDY DESIGN: Retrospective cohort study using automated managed care data. SUBJECTS All persons enrolled during 1995 and 1996 in Group Health Cooperative of Puget Sound, HealthPartners of Minnesota and the Colorado, Ohio and Northeast Regions of Kaiser-Permanente. MEASURES RxRisk, an algorithm that classifies prescription drug fills into chronic disease classes for adults and children. RESULTS: HCCs produce the most accurate forecasts of total costs than either RxRisk or ACGs but RxRisk performs similarly to ACGs. Using the R(2) criteria HCCs explain 15.4% of the prospective variance in cost, whereas RxRisk explains 8.7% and ACGs explain 10.2%. However, for key segments of the cost distribution the differences in forecasting power among HCCs, RxRisk, and ACGs are less obvious, with all three models generating similar predictions for the middle 60% of the cost distribution. CONCLUSIONS: HCCs produce more accurate forecasts of total cost, but the pharmacy-based RxRisk is an alternative risk assessment instrument to several diagnostic based models and depending on the nature of the application may be a more appropriate option for medical risk analysis.

Using risk-adjustment models to identify high-cost risks.

BACKGROUND: We examine the ability of various publicly available risk models to identify high-cost individuals and enrollee groups using multi-HMO administrative data. METHODS: Five risk-adjustment models (the Global Risk-Adjustment Model [GRAM], Diagnostic Cost Groups [DCGs], Adjusted Clinical Groups [ACGs], RxRisk, and Prior-expense) were estimated on a multi-HMO administrative data set of 1.5 million individual-level observations for 1995-1996. Models produced distributions of individual-level annual expense forecasts for comparison to actual values. Prespecified "high-cost" thresholds were set within each distribution. The area under the receiver operating characteristic curve (AUC) for "high-cost" prevalences of 1% and 0.5% was calculated, as was the proportion of "high-cost" dollars correctly identified. Results are based on a separate 106,000-observation validation dataset. MAIN RESULTS: For "high-cost" prevalence targets of 1% and 0.5%, ACGs, DCGs, GRAM, and Prior-expense are very comparable in overall discrimination (AUCs, 0.83-0.86). Given a 0.5% prevalence target and a 0.5% prediction threshold, DCGs, GRAM, and Prior-expense captured $963,000 (approximately 3%) more "high-cost" sample dollars than other models. DCGs captured the most "high-cost" dollars among enrollees with asthma, diabetes, and depression; predictive performance among demographic groups (Medicaid members, members over 64, and children under 13) varied across models. CONCLUSIONS: Risk models can efficiently identify enrollees who are likely to generate future high costs and who could benefit from case management. The dollar value of improved prediction performance of the most accurate risk models should be meaningful to decision-makers and encourage their broader use for identifying high costs.