PS2-28: The Association of Body Composition Indices and Hormonal Contraceptive Use in Adolescent and Young Adult Women

Background/Aims Hormonal contraception is a popular contraceptive choice among sexually-active women. Yet, its association with body composition in younger women is not well understood. We compared body composition measures by duration of Depo-Provera (DMPA) use and dose and duration of oral contraceptive (OC) use in adolescent and young adult women. Methods Study participants were Group Health Cooperative members. DMPA use (new, prevalent, or none) was collected in 170 adolescents aged 14-18 years, and in 440 women aged 18-39 years. OC use was gathered in 301 adolescents aged 14-18 years, and 305 women aged 19-30 years. Among OC users, ethinyl estradiol (EE) dose (=30mcg vs. <30mcg) and months of use were also collected. For all participants, weight, BMI, and truncal and total fat and lean mass were estimated using DEXA. Mean differences in baseline body composition by dose and duration of hormone contraception use were compared cross-sectionally using ANOVA. Results Adolescents who were prevalent DMPA users had higher baseline truncal and total fat mass compared to adolescent non-DMPA users. For example, baseline mean (SE) total fat mass was 23.3 (1.2) kilograms for prevalent DMPA users, but was 19.9 (0.9) kilograms for non-DMPA users (p<0.05). No mean differences in the body composition measures were observed between new DMPA vs. non- DMPA adolescent users or among young women, regardless of DMPA use. With OC use, baseline BMI and truncal and total fat mass differed by EE dose in young adult women. Baseline mean (SE) total fat mass, in kilograms, for women using =30mcg EE doses, <30mcg EE doses, and for non-OC users was 20.8 (0.9), 19.0 (0.9), and 22.2 (0.9), respectively (p=0.02). No differences were noted between EE dose and body composition indices in adolescents. For both adolescents and young women, baseline body composition was not associated with duration of OC use. Conclusions Our results suggest that DMPA use is associated with fat mass in adolescents and that EE dose in OCs is associated with fat mass in young adult women. Analysis of follow-up data to determine if these relationships continue to be seen longitudinally is needed.     

PS2-61: Establishment of a Cohort of Women to Study the Effect of Cervical Procedures on Reproductive Health Outcomes

Background/Aims Persistent oncogenic human papillomavirus (HPV) infection may lead to precancerous lesions, which health providers may remove using surgical procedures before the lesions progress to cervical cancer. This paper describes the methods used to establish a retrospective cohort of Kaiser Permanente Northwest (KPNW) members to test hypotheses that surgical procedures involving the cervix affect subsequent reproductive health and pregnancy outcomes. Methods The HMORN Virtual Data Warehouse (VDW) was used to obtain health plan enrollment, diagnosis, procedure, pharmacy, and demographic data for 14- to 53-year-old female KPNW members during the study years, 1998-2009. A list of CPT and ICD-9 codes was used to identify women who were exposed to destructive or excisional procedures of interest. A group of unexposed women was selected by frequency-matching on age. Contraceptive data for all women in the cohort was pulled to account for time on and off various contraceptive methods, including: oral contraceptives, intrauterine devices (IUDs), implantable and injectable hormonal methods. Results An age-matched cohort of 86,898 women was retrospectively established to compare the reproductive health outcomes of 4,138 women who were exposed to surgical cervical procedures to a 1:20 age-matched group of unexposed women. The demographic and contraceptive method use characteristics were similar for the exposed and unexposed members of the cohort. However, a greater proportion of exposed women than unexposed women used contraceptives at some point during the study period. Conclusions We used the VDW tables to establish a retrospective cohort of women in which the effect of cervical procedures related to precancerous cervical lesions on reproductive health outcomes may be studied. An analytic dataset has been set up to perform Cox proportional hazards regression analyses adjusting for contraceptive use.     

PS2-24: 60 Seconds or Less: Factors Influencing the Likelihood of a Mental Health Discussion During Periodic Health Exams

Background/Aims The majority of patients with mental health concerns turn to their primary care physician (PCP). We investigate whether periodic health exams (PHEs) may be a good opportunity for these patients to receive mental health services. We examine the impact of patients' need for mental health care, competing demands from biomedical issues, and the availability of behavioral health providers on the probability of discussing mental health. Methods The study uses audio recordings of 308 unique patients' PHEs with 59 PCPs. Visits took place in 22 clinics of an integrated delivery system in metropolitan Detroit between 2007-2009. Administrative data and electronic medical records spanning 12-months before the PHE are also used. Patients are aged 50 to 80; all were deemed to be potentially in need of mental health services because of their Personal Health Questionnaire (PHQ2) score, prior diagnosis of mental illnesses, use of behavioral health services, or psychotropic medication use. We coded the audio to capture visit contents and time spent on each "topic" (defined as an issue that had at least two complete exchanges between patient and physician). We examined the probability of a mental health discussion using a logit generalized estimating equation (GEE) regression. Results The median length of patient-physician interaction was 26 minutes. The median number of "topics" was 19. Biomedical discussions were present in all visits; 38% (118 of 308) of the visits included a mental health discussion. The median length of a mental health discussion was 47 seconds. Logit results suggest the likelihood of mental health discussion increased if the patient had a PHQ2 =3 (odds ratio [OR]=5.30, p<0.01), was on psychotropic medication (OR=2.47, p<0.05), had a mental health diagnosis in prior year (OR=2.52, p<0.05), and was female (OR=1.88, p<0.05). The likelihood decreased if the clinic has co-located behavioral health (OR=0.47, p<0.05). Percent time spent on biomedical issues did not significantly impact the probability of a mental health discussion. Conclusions Many discussions occurred during PHEs with patients potentially needing mental health care. Only 38% (118 of 308) of all visits contained mental health discussions, and in those that did more than half lasted less than a minute.     

Intima-Media Thickness Measurements in the Fetus and Mother During Pregnancy: A Feasibility Study

Fetal intima-media thickness (IMT) has been suggested as a marker of pre-clinical atherosclerosis, and maternal IMT could be altered through dynamic circumstances related to pregnancy. We investigated the feasibility of measurement of IMT at four pre-defined fetal and four pre-defined maternal arterial locations to determine vascular changes that could be associated with impaired vascular function. IMT was measured from the first to third trimester (12–34 wk), in 38 low-risk pregnancies. We imaged a 10-mm region of interest using a Mindray (Shenzhen, China) high-resolution ultrasound machine with automated IMT measurement software. Fetal abdominal aorta IMT was measurable during the second trimester in 71% and during the third trimester in 100% of the case, and umbilical artery IMT was measurable in 50% and 82% of cases during the second and third trimesters, respectively. Fetal IMT measurements were not possible during the first trimester. It was not often feasible to measure the IMT of the fetal common carotid artery, fetal renal artery and maternal iliac artery (maximal 20% of cases). Maternal common carotid artery, abdominal aorta and uterine artery IMTs were measurable throughout pregnancy. There was a significant relation between gestational age and IMT in the umbilical artery (p = 0.03) and a significant relation between body mass index and IMT in the maternal common carotid artery (p = 0.01). IMT measurements are feasible in some maternal and fetal vessels of interest. Further studies are underway to obtain more insight into vascular development during normal and pathologic pregnancies.     

Spouses' Experiences During Pregnancy and the Postpartum: A Program of Research and Theory Development

A program of nursing research was established to test a theory proposing that wives and husbands have similar pregnancy-related experiences. The research was guided by a conceptual framework of the family as a living open system. Findings were conflicting from three studies that investigated the relationship between spouses' strength of identification and similarities in changes in various body image components during and after pregnancy; taken together the findings suggested that spouses do not have similar patterns of change in their body images during pregnancy and the postpartum. Two other studies investigated the relationship between spouses' strength of identification and similarities in their reports of physical and psychological symptoms during pregnancy and the postpartum. In these studies the spouses reported similar physical and psychological symptoms during pregnancy and the postpartum. There was no evidence, however, in any of the studies of a relationship between spouses' strength of identification and similarities in their pregnancy-related experiences. The validity of the theory of similar pregnancy-related experiences and the credibility of the conceptual framework of the family as an open system are questioned.     

PS2-39: Key Challenges and Decisions in the Development of HMO Death Data

Background The mortality status of an HMO's members, whether a person is alive or dead, is critically important to health care research. The uses of mortality status include determination of causes of death, death rates, and for selection of study cohorts. Unlike other clinical activity and health statuses, mortality status cannot be determined solely with a health plan's internal data, even for well-defined populations, because health plan members often die outside of the plan's care delivery and claims payment systems. For such members, determination of mortality status and date of death requires linkage of health plan data to external sources of mortality data, such as state death certificate or federal Social Security data. This person-level statistical linkage is a complex, multi-step process that involves many decisions, assumptions, and choosing of priorities. Even when death data from internal systems is available, internal data sources can disagree, likewise requiring complex decisions to determine a member's "true" mortality status. Methods This presentation highlights key challenges and decisions points in the development of death data logic at Kaiser Permanente Northern California (KPNC). Results Key decisions in KPNC's death data logic include: handling internal membership data with conflicting personal identifiers, e.g., multiple Social Security Numbers associated with a single Medical Record Number; ranking, grading, and selecting from multiple matches between internal and external data; finding members with multiple MRNs; using utilization contact dates and membership activity dates to evaluate death records, judging reliability of internal health plan data, setting thresholds and criteria for confidence scoring, selecting between multiple possible dates of death, chart validation of final death data. Conclusions The development of death data logic at KPNC is a complex process that requires intricate decisions, careful analysis, and a higher degree of discretionary judgment than is typical of building other data areas. The quality of final results depends upon validation and iterative improvements to the process.     

PS2-48: De-duping the VDW: A Protocol for a Identifying Common Enrollees Across Sites Without Sending PHI

Background/Aims Early on in VDW history, the sheer geographical distribution of HMORN member sites made it exceedingly unlikely that the same person would be found in more than one site's data. Thus our multi-site research has proceeded without much fear of double-counting anyone. But as the HMORN expands its membership to new organizations, this assumption becomes harder to justify. We now have 3 sites operating in Minnesota, for instance, and 2 each in Massachusetts, Northern California, Idaho, and Washington. Methods This presentation will propose a protocol for creating a dataset at HMORN sites listing the local MRNs of people who also appear in another site's data. Most notably-the protocol does not call for transmitting any unencrypted PHI. The core of the method involves the repeated use of a commutative cipher to encrypt identifiers, each site using its own key, and then comparing the results for overlaps. The end result of the protocol is that each site will know which of its MRNs represent people who also appear in other sites' data, and which site(s) those are. The presentation will describe the data flows, algorithms and processes necessary to create this data. Results The protocol described can be run once to assess the actual overlap, and-if it is significant enough-periodically to create this data at the sites. Once created, the data could be made part of the VDW, allowing HMORN projects to easily assure that each person in each distributed cohort appeared in only one site's file. Conclusions Being able to say that we have actually investigated population overlap and either found it in fact to be negligible, or that we have a method for de-duplicating people is far preferable to hand-waving and bare assertions that any overlap must be insignificant. As the HMORN grows and VDW collaborators become more numerous, we should expect this question to take on more and more significance in the eyes of both our funders and the consumers of our research.     

PS2-04: Knowledge of Heart Attack and Stroke Symptomology: A Cross-sectional Comparison of Rural and Non-rural US Adults

Background/Aims In the US heart disease is the number one cause of death and stroke is the leading cause of adult disability. Early treatment, received within three-hours after a heart attack or a stroke, is critical to lowering the risk of disability and/or death. Understanding the symptoms and signs of heart attacks and strokes are important not only in saving lives, but also in preserving quality of life. Findings from recent research have yielded that the prevalence of cardiovascular disease risk factors are higher in rural populations, suggesting that adults living in rural locales may be at higher risk for heart attack and/or stroke. Knowledge of heart attack and stroke symptomology as well as calling 911 as the appropriate first response to suspected heart attack or stroke are essential first steps in seeking care. This study sought to examine the knowledge of heart attack and stroke symptoms among rural adults in comparison to non-rural adults living in the US. Methods Using multivariate techniques, a cross-sectional analysis of an amalgamated multi-year Behavioral Risk Factor Surveillance Survey (BRFSS) database was performed. The dependent variable for this analysis was low heart attack and stroke knowledge score. The covariates for the analysis were: age, sex, race/ethnicity, annual household income, attained education, health insurance status, health care provider (HCP), timing of last routine medical check-up, medical care deferment, self-defined health status and geographic locale. Results Logistic regression analysis revealed that those US adults who had low composite heart attack and stroke knowledge scores were more likely to be rural (OR=1.218 95%CI 1.216-1.219) rather than non-rural residents. Furthermore, those with low scores were more likely to be: male (OR=1.353 95%CI 1.352-1.354), >65 years of age (OR=1.369 95%CI 1.368-1.371), African American (OR=1.892 95%CI 1.889-1.894), not educated beyond high school (OR=1.400 955CI 1.399-1.402), uninsured (OR=1.308 95%CI 1.3-6-1.310), without a HCP (OR=1.216 95%CI 1.215-1.218), and living in a household with <$50,000 annual income(OR= 1.429 95%CI 1.428-1.431). Discussion Analysis identified clear disparities between the knowledge levels US adults have regarding heart attack and stroke symptoms. These disparities should inform education endeavors focused on improving knowledge of MI and stroke symptoms.     

PS2-49: The VDW Census File: Strengths, Issues, and Recommendations for the Future

Background/Aims The VDW Census file contains information about our members such as income and education. The information is estimated based on data collected by the decennial census for the geographic areas around our members' addresses (identified by geocodes).The VDW Census file was initially developed based on 2000 census data. Data from the 2010 census is now being released, and in many cases, the 2010 census data structure looks different than the 2000 census. For example, income and education were no longer collected in the 2010 census, but have been moved to a supplementary survey (American Community Survey, or ACS) collected over a different time period. Methods The VDW Census workgroup reviewed the changes in the 2010 census data and compared the available information to the data in the existing VDW Census file. Many of the fields from the 'long-form' in the 2000 census were moved to the ACS. Where possible, the workgroup mapped the existing VDW Census fields to fields with similar, if not equivalent, data in the new census data structure. The workgroup also considered whether new data available from the ACS might be of interest to HMORN investigators. Where major changes in information availability were anticipated, the workgroup used online surveys to solicit feedback from a larger group of HMORN analysts and investigators. Results Our primary goal was to identify changes to the specification that would allow updating the Census file to 2010 data. Our secondary goal was to identify that would allow sites to maintain multiple sets of census data to accommodate projects with different time periods. To accommodate multiple years of census data and reduce file sizes, we recommend splitting the VDW Census file into two files: one tracking members' geocoded addresses, and another containing the census data for all available geocodes. Conclusions We found significant changes in the 2010 census data that require considerable changes to the VDW Census file. Once agreement has been reached on the new Census specification, a subcommittee will develop code to build a new VDW Census data file and recommendations on building a file containing members' geocodes over time.     

PS2-54: The Flow of Tumor Data: A Technical Review of Tumor Data Production, Submission and Use

Background The Virtual Data Warehouse tumor registry enables important multi-site research in cancer prevention, treatment and outcomes, as well as health communications and quality. Ideally, data are collected from hospitals where the tumors are biopsied, staged, and reported to the state. These data can be difficult to obtain for HMOs such as Kaiser Permanente in the Mid-Atlantic States (KPMAS) that do not own their own hospitals. In such cases, the HMO must request tumor data from the state. This process is particularly complex at KPMAS, which represents three independent jurisdictions (Maryland, Virginia and the District of Columbia). Our end goal is to develop a KPMAS tumor registry that integrates data from electronic health record (EHR) systems and state tumor registries. A key intermediate objective was to develop a comprehensive understanding of tumor registry development at more advanced HMORN sites in order to efficiently build the KPMAS tumor registry. Methods We used a multi-pronged approach to develop the technical structure for the KPMAS tumor registry. First, we surveyed other HMORN sites that have previously successfully developed tumor registries. Second, we reached out to existing tumor registry efforts within KPMAS to reduce duplication, capitalize on existing work in this area, and document the workflow: from data capture through state reporting and integration into our electronic health record. Third, we developed relationships with key tumor registry managers in all three KPMAS jurisdictions to identify the processes required for obtaining tumor data corresponding to our members. During this process, we requested both a data dictionary and a set of dummy data from each of the external jurisdictions. Results Using the data dictionaries and dummy data provided by the external tumor registries, we developed a technical solution for housing and managing tumor related information from multiple sources. Key considerations involved integrating data from external sources with data from the EHR and removing duplicate data that results from integrating data from three external agencies that cover a highly mobile geographic area. Conclusions Integrating tumor data from multiple sources involves both technical challenges and requires that a complete understanding of the data meaning.     

Smoking behavior and related cancers: the role of CYP2A6 polymorphisms

Smoking exerts complex central and peripheral nervous system, behavioral, cardiovascular, and endocrine effects in humans and is a primary risk factor for various cancers. Nicotine, a major constituent of tobacco, is the compound that is responsible for the development and maintenance of tobacco dependence. The absorbed nicotine is rapidly and extensively metabolized to inactive cotinine by CYP2A6 in human livers, which has a major impact on nicotine clearance. Progress has been made in understanding the relationship between the inter-individual variability in nicotine metabolism and genetic polymorphisms of CYP2A6. Recent findings have increased our knowledge concerning ethnic differences in the allele frequencies of the CYP2A6 variants, nicotine metabolism, and cancer risk. In this review, the potential associations between the CYP2A6 polymorphisms and smoking behavior or the risk of cancer are also discussed.     

PS2-21: Validity and Reliability of the Family Economic Impact Interview: A New Instrument to Measure Costs Related to Child Mental Health Conditions

Background/Aims Care of children with mental health and developmental conditions (MHC) is often demanding and expensive, imposing both financial and time costs on families. Because costs to families are relatively difficult to quantify, most studies have focused on studying family costs using health system data which are more readily available. Few studies have collected cost data directly from families. Methods This study presents results of psychometric analyses of a new instrument that directly measures family costs, the Family Economic Impact Interview (FEII). Reliability and validity of the FEII was evaluated as part of a developmental study which recruited a random sample of 70 parents of children with MHC between the ages of 3 and 12 years old from a large integrated health system. Parents were invited to participate in two interviews within a two week time period. Kappa statistics were calculated to evaluate the agreement between categorical variables. Results We used standard cut off values (values over .75 - excellent agreement, values between .40 and .75 - good agreement, and values less than .40 - poor agreement). Reliability of parents' reports of time spent in activities related to MHC were also good. In 20 of 24 time use categories, kappas ranged from .60 to 1.00. Validity of the FEII was assessed by examining the relationship between the degree of emotional burden a parent perceived from MHC and the amount of family costs measured by the FEII. We used regression analyses to examine whether there was a significant relationship between the main FEII outcomes and family emotional burden. Discussion As hypothesized, we found significant positive association between emotional burden of care and family costs (p<.001). Preliminary analyses of the FEII indicate that a wide range of parents' can complete the FEII, that parents' can reliably report a variety of family costs including out-of-pocket financial costs and time costs, and that many families with children with MHC experience a wide range of costs - many of which have not been measured in previous studies.