Gender & Health Call for Papers Assessing & Improving Quality of Care by Gender Chair: Amal Khoury, University of Florida Tuesday, June 28 • 8:30 am – 10:00 am ●Are Older Women More Likely to Get Inappropriate Drugs? Arlene Bierman, M.D., MS, Mary Jo V. Pugh, Ph.D. , RN, Irfan Dhalla, M.D., Megan Amuan, MPH, B. Graeme Fincke M.D., Amy Rosen, Ph.D., Bei- Hung Chang Sc.D., Dan R. Berlowitz, M.D., MPH Presented By: Arlene Bierman, M.D., MS, OWHC Chair in Women's Health, Inner City Health Research Unit, St. Michael's Hospital, 30 Bond Street( 70 Richmond Street East, 4th floor), Toronto, M5B 1W8; Tel: (416) 864-6060 X2894; Fax: (416) 864-5485; Email: arlene.bierman@utoronto.ca Research Objective: Prior studies have suggested that older women are more likely to receive inappropriate prescriptions than older men. A better understanding of gender differences in inappropriate prescribing may lead to the development of effective improvement interventions. Therefore, in this study, we assessed gender differences in rates of inappropriate prescribing before and after adjusting for potentially appropriate indications and examined gender differences in correlates of inappropriate drug use. Study Design: Retrospective national Veterans Health Administration (VA) cohort study using linked pharmacy, inpatient, and outpatient data. We determined the unadjusted and diagnosis-adjusted prevalence of 33 potentially inappropriate medications among men and women using the Beers criteria: overall, by individual drug, and in three categories grouped by potential indication: always avoid, rarely appropriate, and some indications according to the Zahn classification. Using logistic regression we determined the odds ratios of women receiving inappropriate drugs compared to men for individual agents and by drug category. Race/ethnicity, psychiatric comorbidity, physical comorbidity, number of medications, and care characteristics (long-term care, medical-surgical and psychiatric admissions, receipt of geriatric care, and proportion of visits in primary care) were included as covariates. Logistic regression was used to identify patient and care characteristics associated with inappropriate prescribing for men and women. Population Studied: Veterans 65 years or older, who had having at least one VA patient health care encounter in both fiscal year (FY)99 and FY00 (N=965,756; 946,641 men; 19,115 women. Principal Findings: Women were more likely than men to receive inappropriate medications overall and in all three categories, even after adjusting for diagnoses that may have justified the prescription. The majority of use remained inappropriate after adjusting for potential indications. For example, in the “some indications” category, 22.6% of women and 18.1% of men received a potentially inappropriate medication. After adjusting for diagnoses that may have justified the prescription, 18.2% of women and 15.6% of men received a potentially inappropriate medication. Women were more likely to receive 16 of 33 medications (psychotropic drugs, analgesics, anticholinergic agents) and men more likely to receive 3 of the 33. In adjusted models, women were more likely to receive any inappropriate drug (OR 1.23 95%CI 1.18, 1.27). The OR for always avoid drugs was 1.73 95%CI 1.55, 1.94. Psychiatric comorbidity was associated with inappropriate prescribing for men but not for women. Conclusions: Older women veterans were more likely to receive inappropriate medications than older men in the VA, even after accounting for a liberal set of indications, and even though some drugs such as indomethacin and dypiridamole are more likely to be prescribed to men. Analgesic, psychotropic, and anticholinergic medications contribute to the higher rates of inappropriate prescribing among older women. Despite the higher proportion of men in the VA, our findings suggest that there are gender differences in correlates of inappropriate prescribing. Implications for Policy, Delivery, or Practice: Efforts to improve the quality of medication management in the elderly should address gender differences in prescribing patterns. Studies of inappropriate prescribing should examine gender differences whenever possible. Primary Funding Source: VA ●Cervical Cancer Screening Practices in the United States Since the Release of National Guidelines on Genital Human Papillomavirus (HPV) Testing: Results from a Recent National Clinician Survey Kathleen Irwin, M.D., MPH Presented By: Kathleen Irwin, M.D., MPH, Chief, Health Services Research, Division of STD Prevention, Centers for Disease Control and Prevention, 1600 Clifton Road, Mailstop E-80, Atlanta, GA 30333; Tel: (404) 639-8979; Fax: (404) 6398607; Email: kirwin@cdc.gov Research Objective: Persistent infection with oncogenic types of genital human papillomavirus (HPV) causes cervical intraepithelial neoplasia (CIN) and cancer which result in 4000 US deaths and >$1 billion in cancer screening, diagnosis, and treatment costs annually. Of the ~3 million American women diagnosed with borderline Pap test results (ASC-US) each year, HPV-infected women are at highest risk of developing CIN and cancer. In 2001, CDC and several national clinical organizations recommended using a recently licensed HPV test to guide triage of HPV-infected women with ASC-US Pap results to colposcopy, an expensive, invasive procedure not readily available to all women screened for cervical cancer. HPV testing is not recommended to triage patients with higher-grade Pap abnormalities because colposcopy is warranted regardless of HPV test results. By 2003, two national organizations had endorsed HPV testing as an optional adjunct to Pap testing for women 30+ years to distinguish HPV-infected women who warrant frequent followup Pap tests from non-HPV-infected women at low risk of cervical cancer whose Pap testing intervals could be lengthened. We conducted a national clinician survey to evaluate the influence of new guidelines on screening practices. Study Design: In 2004, the Centers for Disease Control and Prevention surveyed nationally representative samples of primary care clinicians identified through national registries. The mail survey addressed Pap and HPV testing practices. We weighted categorical analyses for differences in sampling fraction and non-response by clinician specialty. Population Studied: Primary care physicians practicing internal, adolescent, or family/general medicine or obstetrics/gynecology; nurse-midwives; nurse-practitioners; and physician assistants. Principal Findings: Of the 3365 primary care clinicians who responded (adjusted response rate=81%), 87% reported providing Pap tests. Of these, 59% reported using HPV tests for patients with borderline or abnormal Pap results. Of clinicians reporting HPV testing of such patients, 84% reported usually/always ordering HPV tests for patients with borderline (ASC-US) Pap results, as endorsed by national guidelines. However, 57% to 78% of these clinicians reported usually/always ordering HPV tests for patients with highergrade Pap abnormalities, a practice not endorsed by guidelines. Only 21% of Pap test providers reported ordering HPV tests as an adjunct to Pap tests, of whom few reported usually/always testing women aged 30+ (12%) or <30 (15%). Obstetrician/gynecologists and nurse-midwives and clinicians with on-site colposcopy were more likely than other clinicians to report using HPV tests for cancer screening indications that were and were not recommended by guidelines. Conclusions: Most Pap test providers report ordering HPV tests to guide colposcopy triage of women with borderline Pap results, but many also report providing HPV tests to patients with higher-grade abnormalties who merit colposcopy regardless of HPV test results. Few Pap test providers report using HPV tests as an adjunct to Pap tests, suggesting that HPV testing has not resulted in lengthening Pap test intervals among women at low cancer risk. Implications for Policy, Delivery, or Practice: Survey findings should be used to update clinical training materials, decision support tools, and patient education materials to promote consistency with new guidelines. Primary Funding Source: CDC poverty level) below which a woman was eligible for Medicaid. This policy parameter varies greatly across states and over time. In addition, we control for the annual state-level welfare caseloads and unemployment rates as well as for the various individual characteristics and state and time fixed effects that might have had an impact on maternal morbidity. Population Studied: All of our analyses are performed separately for whites and blacks. Also, in an attempt to verify the causality of the relationship between Medicaid eligibility and health, we adopt a difference-in-difference type of approach and estimate our models separately for two treatment groups (low-income single and married mothers) and a control group (high-income married pregnant women). Principal Findings: We find that the eligibility changes lead to a higher utilization of prenatal care, less anemia, and possibly less hypertension among low-income women. In 1997, the expenditures for hospitalizations due to anemia and pregnancy-related hypertension were over $962 million and $1,237 million, respectively. Our results therefore suggest that the Medicaid expansions may indeed have lead to significant health improvements and cost savings for one of its key target groups - the pregnant women themselves. Conclusions: If maternal health is considered an important outcome, our results suggest that the conclusions regarding the effectiveness of the Medicaid eligibility expansions may need to be reevaluated. Implications for Policy, Delivery, or Practice: When designing programs to improve birth outcomes, policymakers need to be aware of the effects of the implemented policies on both the infants and their mothers. Viewed from this broader perspective, interventions targeted at low-income pregnant women – such as the Medicaid eligibility expansions of the 1990’s - may be more effective than initially thought. Primary Funding Source: No Funding ●The Forgotten Beneficiary of the Medicaid Expansions Andrea Kutinova, MA (Ph.D. student), Karen Smith Conway, Ph.D. Presented By: Patick Roohan, MS, Director, New York State Department of Health, Bureau of Quality Management and Outcomes Research, OMC, 1938 Corning Tower, Albany, NY 12237; Tel: (518) 486-9012; Fax: (518) 486-6098; Email: pjr02@health.state.ny.us Research Objective: The objective of this study was to analyze the relationship of surgical and hospital volume on three-year and five-year survival for women with breast cancer. Study Design: Three datasets were used for this study, the New York State Cancer Registry, the Statewide Planning and Research Cooperative System (SPARCS) inpatient database and the SPARCS outpatient surgery database. Records of women diagnosed with breast cancer from 1995 through 1998 were identified on the Cancer Registry and matched separately to the SPARCS inpatient and outpatient datasets. Matching the Cancer Registry to the SPARCS inpatient files yielded a match rate of 87% for the four year period. Due to limited matching data elements on the outpatient surgery file in 1995 and 1996, the Cancer Registry and SPARCS outpatient match was conducted for the years 1997 and 1998, and also had a match rate of 87%. Two analyses emerged: a five-year survival study of inpatient surgeries for the years 1995-1998, and a three-year survival study of inpatient and outpatient Presented By: Andrea Kutinova, MA (Ph.D. student), University of New Hampshire, 4 Library Way, Durham, NH 03824; Tel: (603)862-7071; Email: kutinova@cisunix.unh.edu Research Objective: Empirical evidence on the effectiveness of the Medicaid eligibility expansions of the 1990’s is mixed. While it is generally believed that the expansions succeeded in promoting prenatal care utilization, the effects on pregnancy outcomes seem to be, at best, modest. So far, however, a key beneficiary – the mother – has been left completely out of the analysis. Without considering the effects of the Medicaid expansions on maternal health, any discussion of the policy effectiveness is incomplete. Therefore, in this paper, we contribute to the ongoing debate by focusing on the effects the eligibility changes had on pregnant women themselves. Study Design: We use the Natality Detail Files from 1990 to 1996 to study the improvements in maternal morbidity during pregnancy and delivery attributable to the Medicaid eligibility increases of the 1990’s. Our key explanatory variable of interest is the income cutoff (as a percent of the federal ●The Relationship of Surgeon and Hospital Volume with Long-Term Survival for Women with Breast Cancer Patrick Roohan, MS surgeries for the years 1997-1998. The Cancer Registry included the following information used in modeling: date of death (through 2001), age, race/ethnicity, socioeconomic status and stage of cancer. Type of insurance, type of surgery (breast conserving surgery, mastectomy), excision of lymph nodes (yes, no), comorbidity, surgeon volume and hospital volume were derived from the SPARCS inpatient and outpatient data. The Cox proportional hazard model was used to evaluate the contributions of surgeon and hospital volume, while adjusting for risk factors outlined above. Population Studied: Women in New York State diagnosed with breast cancer in 1995 through 1998, who had breast surgery. Principal Findings: In the five-year survival study after adjusting for socio-demographic and clinical factors, women who went to a surgeon with an annual breast cancer surgical volume of 1-5 were 1.36 times more likely to have a poorer outcome in five years compared to women who went to a surgeon that performed over 46 surgeries per year. Similar increased risk of poorer survival exists for surgeons that perform 6-10 (Risk ratio 1.34), 11-20 (RR=1.27), 21-45 surgeries per year (RR=1.21) compared to surgeons that perform 46+ surgeries annually. Hospital volume, after adjusting for surgeon volume and other factors was not significant. In the three-year study, women who had surgery performed by a surgeon with annual volumes of 1-6 (RR=1.39) and 7-12 (RR=1.20) had poorer outcomes compared to women who went to a surgeon with an annual volume of over 25 surgeries per year. Similar to the five-year study, hospital volume after adjusting for other factors, was not significant. Conclusions: In both studies presented, long-term survival from breast cancer increases with increased surgeon volume. This may be due to “practice makes perfect”, the association of high-volume providers and comprehensive cancer centers, and/or increased surgical skill for some high volume surgeons. Implications for Policy, Delivery, or Practice: Women with breast cancer should consider the expertise of the surgeon in breast surgery in her decision-making process. Primary Funding Source: New York Community Trust ●Gender Differences in Prescribing Drugs Potentially Harmful to Elderly Managed Care Enrollees Lok Wong, MHS, Russell Mardon, Ph.D., Arlene Bierman, M.D. Presented By: Lok Wong, MHS, Senior Health Care Analyst, Quality Measurement, National Committee for Quality Assurance, 2000 L Street NW, Suite 500, Washington, DC 20036; Tel: (202)955-1784; Fax: (202)955-3599; Email: wong@ncqa.org Research Objective: Previous population-based studies have found older women to be more likely to receive potentially inappropriate drugs than older men. However, it is not known whether these gender differences in prescribing patterns exist within Medicare managed care plans. Therefore we evaluated whether elderly women enrolled in managed care are more likely than men to receive drugs potentially harmful to the elderly. Study Design: Retrospective pharmacy claims data analysis in 9 health plans in the U.S. Calculations included percentages of elderly enrolled throughout the year who received at least one drug, or at least two different drugs potentially harmful in the elderly. Rates were reported by plan, age, gender, across the study population and various high-risk drug categories. Potentially harmful drugs are classified as never appropriate, rarely appropriate or sometimes indicated by Zahn (2001). Population Studied: Over 433,000 Medicare enrollees ages 65 and older in 9 health plans in different regions across the United States in 2002 and 2003. Principal Findings: Nearly half a million elderly enrollees received more than 3 million prescriptions of drugs potentially harmful in the elderly, of which more than 500,000 are never or rarely appropriate in the elderly. Most Medicare enrollees included in the study were female (63%). Overall, women were more likely than men to receive drugs potentially harmful in the elderly, and across a variety of high-risk drug classifications. Women were more likely than men to receive at least one never appropriate drug (8.3% vs. 6%) or rarely appropriate (drugs 28.6% vs. 25.7%) as defined by Zahn. Women were also more likely to receive at least two different drugs considered never appropriate or rarely appropriate (6.9% vs. 4.2%). Prescribing rates were slightly higher in women for sometimes indicated drugs (30.4% vs. 29.3%). Differences are statistically significant. Similar results and patterns were found in 2003 data. Conclusions: Older women enrolled in Medicare managed care plans appear to be more likely to receive drugs potentially harmful in the elderly. High rates of harmful prescribing are of concern given the majority of Medicare enrollees are women. Further research is needed to develop drug-risk classification systems that can determine if there are gender differences in exposure to patient harms, and if differences in prescribing rates are affected by differential disease burden and drug usage. Implications for Policy, Delivery, or Practice: This study highlights the need to be aware of potential gender differences in potential drug harms exposed to managed care enrollees. A better understanding of the factors associated with these gender differences is needed. Interventions aimed at improving the quality of medication management for Medicare managed care enrollees will need address gender differences in prescribing patterns. Primary Funding Source: CMS Related Posters Poster Session B Monday, June 27 • 6:15 pm – 7:30 pm ●Searching for Women’s Health Centers: Yield of Internet vs. Traditional Telephone Contacts for Identifying Research Sites Nichole Bayliss, BA, Bevanne Bean-Mayberry, M.D., MHS, Elizabeth M. Yano, Ph.D., Judith Navratil, BA, Carol S. Weisman, Ph.D., Sarah Hudson Scholle, DrPH Presented By: Nichole Bayliss, BA, Research Assistant, VA Pittsburgh Center for Health Equity Research and Promotion, University Drive (151-C), Pittsburgh, PA 15240; Tel: (412)6886000 x815113; Fax: (412)688-6527; Email: nichole.bayliss@med.va.gov Research Objective: In anticipation of identifying and recruiting sites for participation in health services research studies, we examined and compared Internet vs. traditional telephone contact strategies to determine their efficiency in identifying non-federal, hospital-based Women’s Health Centers or Programs (WHC/P). Study Design: Using the American Hospital Association’s (AHA’s) Annual Survey of Hospitals for 2001, we identified the national sample of hospitals reporting the presence of a WHC/P, defined as “an area set aside for coordinated education and treatment services specifically for and promoted by women that may or may not include obstetrics but include a range of services other than OB,” (n=2,214). We excluded both federal comprehensive women’s health centers and centers of excellence resulting in 2,113 hospitals. We selected a random sample of 500 hospitals proportionately stratified by control/ownership (private=375; public=125). By convenience, sites were then contacted using one of four methods: (1) web search only (n=53), (2) single call (n=150), (3) web search with maximum of three calls (n=249), or (4) unlimited calls and web search (n=48). Population Studied: Hospital based women's health centers. Principal Findings: Over half (264/500, 53%) of the hospitals that self-reported having a WHC/P through the AHA survey could not be verified through these screening strategies. Among hospitals with completed screenings (236/500, 47%), only 173 hospitals confirmed that they had a WHC/P as defined by the AHA (173/236, 73%). Of the 173 hospitals reporting a WHC, only 24 (24/173, 14%) reported delivery of primary care services. Single calls led to the most indeterminate findings (n=117/150, 78%), and unlimited calls augmented by a web search led to the least indeterminate findings (n=11/48, 23%). Unlimited calls with a web search also led to the highest proportion of sites reporting no WHC/P (n=21/48, 44%); however, this method also led to the highest proportion of sites reporting a WHC with comprehensive primary care services (n=7, 15%). Web searches identified the most WHC/P (n=31/53, 58%), but could only find evidence of primary care services for one site. Conclusions: Multiple telephone calls augmented by a web search yielded the most complete information and identified the most comprehensive centers, but this process was timeconsuming, requiring at least two attempts per site and one- quarter requiring at least 7 contacts. It also led to the unexpected result that many hospitals denied having a WHC/P in contradiction to the AHA survey data. Web searches often confirmed a women’s health center but rarely found evidence that the center provided primary care services. Implications for Policy, Delivery, or Practice: Improvements in search and collection methods include refining the AHA question on women’s health centers with hospital identification of specific services provided for women. An organizational definition of comprehensive women’s health care or primary care distinction would diminish the use of “women’s health center” interchangeably when referring to reproductive health, maternal/neonatal health or primary care services. Additionally, revision of hospital based websites with better search options and contact information is key for public access. Easy access to these clinical services via the Internet may prove to be an asset for marketing to the local health consumer, who is often female. Primary Funding Source: Department of Health and Human Services (DHHS) ●Assuring High Quality Primary Care for Women Veterans: Predictors of Success Bevanne Bean-Mayberry, M.D., MHS, Chung-Chou Chang, Ph.D., Melissa McNeil, MD. MPH, Sarah Scholle, DrPH Presented By: Bevanne Bean-Mayberry, M.D., MHS, Physician, Division of General Medicine, VA Pittsburgh/ University of Pittsburgh, University Drive C (151-C), Pittsburgh, PA 15240; Tel: (412)688-6477; Fax: (412)688-6527; Email: Bevanne.Bean-Mayberry@med.va.gov Research Objective: Provider gender, provider specialty, and clinical setting affect quality of primary care delivery for women, but previous research has not looked at these factors in combination. The purpose of this study is to determine whether the separate or combined effects of provider gender, availability of gynecologic services from the provider, and women’s clinic setting improve patient ratings of primary care quality. Study Design: Women veterans receiving care in women’s clinics or traditional primary care at 10 VA medical centers completed a mailed questionnaire (N=1321, 61%) rating four validated domains of primary care (i.e., preference for provider, communication, coordination, and accumulated knowledge). For each domain, summary scores were calculated and dichotomized into perfect score (i.e. maximum score) versus other. Multiple logistic regressions were used to estimate the probability of a perfect score in each domain while controlling for patient characteristics and site. Population Studied: Women Veterans Principal Findings: Female provider was significantly associated with perfect ratings for communication (OR 2.9, 95% CI 1.4, 5.8) and coordination (OR 3.7, 95% CI 1.5, 9.0). Providing gynecological care among female providers was significantly associated with an increased number of domains with perfect ratings: preference for provider (OR 4.0, 95% CI 1.8, 8.7); communication (OR 2.7, 95% CI 1.3, 5.5); and coordination (OR 2.7, 95% CI 1.1, 7.1). Patients who used a women’s clinic and had a female provider who gave routine gynecological care had perfect or nearly perfect ratings for preference for provider (OR 4.7, 95% CI 2.3, 9.7), communication (OR 2.7, 95% CI 1.7, 5.3), and accumulated knowledge (OR 6.1, 95% CI 1.3, 28.5). Conclusions: Provision of gynecologic services at the provider or practice level is associated with improved patient ratings of primary care separate from and in synergy with the effect of female provider. Male and female providers should consider offering routine gynecologic services or working in coordination with a setting that provides gynecological services. Health care evaluations should assess scope of services for provider and practice. Implications for Policy, Delivery, or Practice: Comprehensive outpatient services are associated with higher ratings of quality in primary care. Primary Funding Source: VA ●Contrasting the Program Characteristics at Federally Supported Comprehensive Women’s Health Centers Bevanne Bean-Mayberry, M.D., MHS, Elizabeth Yano, Ph.D., Judith Navratil, BA, Nichole Bayliss, BA, Carol Weisman, Ph.D., Sarah Scholle, DrPH Presented By: Bevanne Bean-Mayberry, M.D., MHS, Physician, Division of General Medicine, VA Pittsburgh/ University of Pittsburgh, University Drive C (151-C), Pittsburgh, PA 15240; Tel: (412)688-6477; Fax: (412)688-6527; Email: Bevanne.Bean-Mayberry@med.va.gov Research Objective: Fragmentation of women’s health care has stemmed from the separation of reproductive care needs from general medical care. To promote the delivery of comprehensive, integrated clinical care for women, the Department of Health and Human Services (DHHS) Office of Women’s Health launched the national Centers of Excellence (CoE), while the VA initiated the Comprehensive Women’s Health Centers (CWHC). The purpose of this study was to compare the organization, staffing, practice setting, and service availability at federally supported specialized women’s health centers. Study Design: 1) Data on the VA women’s health programs were obtained from the VHA Survey of Women Veterans Health Programs and Practices: Senior Clinician Questionnaire. VA hospitals selected had either a CWHC (n=8) or a VA –designated Clinical Center of Excellence in Women’s Health (CCoE, n=4). 2) This VA tool was adapted for use with clinical directors at DHHS CoEs in operation in June 2003 (n=13). 3) Geographic data and academic affiliation were abstracted from the 2001 American Hospital Association hospital survey. Data were reviewed and entered into a database to make descriptive comparisons. Population Studied: Comprehensive outpatient women's centers Principal Findings: All VA and DHHS women’s health programs were located in urban areas and nearly all had academic partnerships, but DHHS sites had mean caseloads that were triple VA caseloads. Women’s health fellowships were common, and all offered educational training to providers in women’s health topics. Preventive screening and general reproductive services (e.g., STD treatment, menopause management) were on site at nearly all facilities; however, the DHHS sites offered more extensive reproductive services (e.g., on site obstetrical and infertility services). The VA sites were unique in mental health care with all offering sexual trauma screening, trauma and rape crisis counseling, and most with on-site mental health providers. Staff variations followed these service differences with DHHS sites having more gynecologists while VA’s had more psychiatrists. Conclusions: Despite the separate history and development, the VA and DHHS specialized women’s health programs share similar organization, education, and clinical services. The wide range of clinical services available demonstrates the commitment to multidisciplinary care. Implications for Policy, Delivery, or Practice: The presence of these federally recognized centers with common structural components presents an opportunity to create VA and non-VA partnerships to evaluate quality of care for women nationally and benchmark the findings in VA and community standards. Primary Funding Source: Department of Health and Human Services ●What Organizational Factors Support the Development of Women's Health Clinics for Primary Care Bevanne Bean-Mayberry, M.D., MHS, Cynthia D. Caffrey, MD, Donna L. Washington, M.D., MPH, Lisa Altman, MD, Andrew B. Lanto, MPH, Elizabeth M. Yano, PhD, MSPH Presented By: Bevanne Bean-Mayberry, M.D., MHS, Physician, Division of General Medicine, VA Pittsburgh/ University of Pittsburgh, University Drive C, Pittsburgh, PA 15240; Tel: (412)688-6477; Fax: (412)688-6527; Email: Bevanne.Bean-Mayberry@med.va.gov Research Objective: Within the Department of Veterans Affairs (VA), there is ongoing debate about the best way to organize care for women veterans. In response to a rapid increase in the number of women veterans, many VA facilities have developed women´s health clinics (WHC) that provide primary care services. Other facilites deliver primary care to male and female veterans in the same clinics. Our objective was to examine health system organizational factors associated with the development of separate VA Women´s Health Clinics for primary care. Such information enables us to understand which environments are conducive to the development of women´s health clinic settings, which may be more appropriate for both primary and gender-specific care in the VA Healthcare System. Study Design: A national, cross-sectional, survey was mailed to primary care (PC) directors at each eligible VA facility. Outpatient health care facilities were compared by administrative patient data and by VA organizational characteristics focused on leadership, staffing, training programs, and affiliations. Each PC director was asked whether their geographically distinct site of care had a women´s health clinic (WHC) that delivered primary care. Multiple logistic regressions were performed to determine the independent factors associated with the development of a separate VA WHC for primary care. Principal Findings: VA facilities offering a separate WHC for primary care were more likely to have (1) separate primary care leadership (OR 4.0, 95% CI 1.6, 9.7) and (2) separate authority to establish administrative PC procedures (OR 2.7, 95% CI 1.4, 5.1). Presence of other allied health staff within the primary care program (e.g., dietitian and pharmacist) and the presence of established primary group teams (organized provider groups for outpatient clinic coverage) showed similar trends for the development of WHC delivering primary care in bivariate analyses but were not statistically significant in multiple logistic regression. Conclusions: Women´s health clinics that deliver primary care to women veteran users of VA health care services have a stronger likelihood of thriving in settings where the primary care programs are independent and have well-organized leadership with practice autonomy. Such findings may be more influential than general female patient volume or workload. Implications for Policy, Delivery, or Practice: More research and practice evaluations are needed to determine the quality of care afforded to women who choose to attend these clinics in contrast to general primary care clinics. Such information will be crucial to designing effective primary care programs for the rapidly growing female veteran population. Primary Funding Source: VA ●Differential Effects of QI on Men’s and Women’s Treatment for Depression Chloe Bird, Ph.D., Cathy Sherbourne, Ph.D., Rob Weiss, Ph.D., Naihua Duan, Ph.D. Presented By: Chloe Bird, Ph.D., Sociologist, Health Unit, RAND, 1776 Main Street, PO Box 2138, Santa Monica, CA 90407-2138; Tel: (310)393 0411 x6260; Fax: (310) 260-8159; Email: chloe@rand.org Research Objective: To examine whether a quality improvement (QI) program for depression care effectively encourages treatment for both men and women, and whether such interventions lead men and women into different types of care. Study Design: A group-level, randomized, controlled trial in 46 primary care practices within six managed care organizations. Clinics were randomized to usual care or one of two gender-neutral QI programs that supported QI teams and provided resources to support medication management (QIMeds) or psychotherapy (QI-Therapy). We conducted intentto-treat analyses; patients were included in the analysis according to their original assignment, irrespective of whether or not they remained in the practice, and whether or not they used intervention resources. Results are presented as odds ratios comparing the interventions to usual care. We compared changes in care for those in each treatment arm to those receiving usual care. Population Studied: 1,187 primary care patients who screened positive for depression and completed at least one patient assessment questionnaire administered at baseline or 6 months later. Principal Findings: Change in receipt of appropriate care in the QI-med intervention versus usual care was significant for both men and women (OR= 2.9, CI 1.1-5.8 for men vs. OR= 2.6, CI 1.7-4.8 for women) and did not differ by gender. However, QI-therapy improved utilization of appropriate care for men (OR= 5.3, CI 2.2-12.7), but not for women (OR= 1.4, CI 0.9-2.3). Men benefited more than women (F/M OR= 0.3; CI 0.1-0.7), because for men QI-therapy was associated with increased odds of utilizing 4 or more specialty mental health visits (OR= 9.0, CI 3.7-21.7) and of receiving any appropriate medication for at least two months (OR= 5.6, CI 2.2-14.1). For women, QI-therapy was associated with increased odds of having 4 or more specialty mental health visits (OR= 3.3, CI 2.0, 5.5), but not of receiving of appropriate medications (OR= .9; CI 0.6-1.6). Women benefited more from QI-meds than QI-therapy (OR= 2.1, CI 1.3-3.3) in terms of their odds of receiving appropriate care. The interventions had similar effects on men’s utilization of any antidepressant or specialty mental health care (QI-meds vs. QI-therapy OR= 0.7, CI 0.31.6). For women, QI-Meds had a larger effect on receiving any antidepressant or specialty mental health care (OR= 2.8, CI 1.8-4.4). Conclusions: QI worked for both men and women. However, men and women responded somewhat differently to the QI interventions. Men in QI-therapy experienced increased odds of receiving both types of care. Females in QI-therapy were already receiving higher levels of medication than men at baseline; QI-therapy brought men up to equivalent levels. Implications for Policy, Delivery, or Practice: Although the interventions studied here were not designed to serve men and women differently, they shed light on an unexpected opportunity to improve men’s quality of care for depression by facilitating their access to mental health specialty care. Strengthening QI interventions for both men and women is a priority for future studies; our findings may afford some clues as to what features may be useful in achieving these aims. Primary Funding Source: NIMH ●What Women Want: Understanding Obesity and Preferences for Primary Care Weight Reduction Interventions among African American and Caucasian Women Carol Blixen, Ph.D., Anisha Singh, M.D., Holly Thacker, M.D., Meng Xu, MS, Edward Mascha, Ph.D. Presented By: Carol Blixen, Ph.D., Associate Staff, Biostatistics and Epidemiology-Wb-4, The Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195; Tel: (216) 444-6254; Fax: (216) 445-7659; Email: blixenc@ccf.org Research Objective: Although African American (AA) women have been shown to have similar weight loss practices as Caucasian (C) women, C women have been more successful than AA women in controlling their weight. However, no studies have been conducted that compare preferences for weight management interventions between AA and C women. The objectives of this study were to explore attitudes and perceptions of obesity, and identify the actual preferences for weight management interventions by AA and C women in the primary care setting. Study Design: From the content areas identified from separate focus groups of AA and C women with a BMI>30, we developed and pre-tested an 18-item questionnaire. We then used the multiple mailing method to send the questionnaires to the study population. Descriptive statistics, including means and standard deviations, medians, quartiles, percents, and relevant confidence intervals, were used to describe the variables. T-tests, or Wilcoxin rank sum tests, or Chi square tests, were used to compare groups on outcomes, using a significance level of 0.05. Population Studied: 464 female patients 18 years and older, with a diagnosis of obesity (ICD–9 code-278.00/278.01) who were being followed in the General Internal Medicine outpatient clinics of the Cleveland Clinic Foundation. Principal Findings: The majority of the women 256 (55%) who returned their questionnaires were primarily (138) AA and C (91). AA women were more likely to be unmarried (65%.7% vs 46.7%, P=0.003) than C women, and have a lower socioeconomic status (P<0.001). Although AA women weighed significantly more than C women (M=245.4; SD=58.6 vs 225.3; SD=51.8, P=0.008), C women expressed having a lower self-esteem because of their weight than did AA women ((80.9% vs 65.7%, P=0.013). C women felt that past weight loss efforts were helped by weight loss programs significantly more than AA women (P< .001), but AA women were more likely to feel that their cultural background contributed to their weight gain than did C women (P=0.001 ). AA women expressed a higher need for one-on-one counseling with their physician (<0.001) as well as team meetings with the dietician, physician, and other women (0.004) than did C women. AA women also felt it was more important for weight loss programs to have information on food common to their culture than did C women (P<0.001). Conclusions: Weight loss programs may be geared towards the needs and preferences of the dominant society and, therefore, may be less attractive or successful with those who are not members of the majority Implications for Policy, Delivery, or Practice: Differences in cultural background and preferences about weight loss interventions have important policy implications for how the U.S. health system provides care to an ever increasing multicultural population with a national epidemic such as obesity. Primary care physicians should recognize that people use the categories and rules of their specific cultures and ethnic groups to frame what they consider to be acceptable and preferable foods and use this information to develop culturally appropriate weight loss menus. In addition, AA women may benefit from shared medical visits with other AA women, that focus on weight loss interventions in a group setting. Primary Funding Source: Cleveland Clinic Foundation ●Women Veterans’ Satisfaction with VA Hospitalization Carron Cherrie, Ph.D., Skai Schwartz, Ph.D., Leigh Mathias, MPH, Margaret Mikelonis, MS, Toni Lawrie, MS Presented By: Carron Cherrie, Ph.D., Project Manager, Nursing Research, VHA, 11605 North Nebraska Avenue, Tampa, FL 33612; Tel: (813) 903-2416; Fax: (813) 558-3994; Email: Carron.cherrie@med.va.gov Research Objective: Background: As women make up just 6.5% of the VA population, few studies have examined their satisfaction with hospitalization. This study examines correlates of women’s satisfaction with VA hospitalization. Study Design: Design: A cross sectional mailed survey. Participants: All women hospitalized during a four-month period in 2003 and 2004 in six VA medical centers were included. Measures: Overall satisfaction was comprised of three questions from the Women Veterans Inpatient Satisfaction Survey pertaining to overall rating of quality of care, complaints and their choice to return for care. Methods: ANOVA models were used to determine the association of each of 42 potential reasons for dissatisfaction with the overall satisfaction score. A step-wise selection procedure was used after adjustment for age group, education, self-rated health and hospital unit. Population Studied: All women discharged from inpatient units of six facilities between February 1, 2003 and August 31, 2003 [N=634] or between February 2, 2004 and June 2, 2004 [N=566] were included. Principal Findings: Results: Forty-nine percent of women had the highest possible score on overall satisfaction. Analysis of baseline and hospital characteristics indicated that younger women, women with graduate degrees, women in poorer health and women in psychiatric units tended to be less satisfied with hospital care. After adjusting for these characteristics 39 of the 41 questions related to specific reasons for dissatisfaction were strongly associated (p<0.001 for each) with the overall satisfaction score. Questions remaining highly significant after stepwise selection included not having one doctor, providers contradicting each other, feeling like a second class citizen and poor courtesy of nurses. Conclusions: Most women are satisfied with hospital care; possible areas for improvement may include identification of one doctor in charge, better provider communication and more respect for the patients. Implications for Policy, Delivery, or Practice: The VISN 8 women veterans’ workgroup submits recommendations for changes in policy to the Executive Leadership Board Health Systems Committee. This streamlined communication process has facilitated approval of changes needed in women’s health care. The annual satisfaction survey has provided a quality assurance tool that evaluates both strengths and weaknesses in service provision, provides time trending and promotes customer participation in shaping service delivery. Nursing care on inpatient units has improved as a result. One example of positive change resulting from implementation of survey recommendations is sensitivity training for nursing staff to improve courtesy and respect for patients. Subsequent to the nurse sensitivity training patient satisfaction ratings for these indicators improved. Primary Funding Source: VISN 8 ●Rurality, Gender and Disparities in Mental Health Treatment Emily Hauenstein, Ph.D., LCP, APRN, BC, Stephen Petterson, Ph.D., Douglas Wagner, Ph.D., Virginia Rovnyak, Ph.D., Elizabeth Merwin, Ph.D., MSN, Barbara Heise, MSN, NP Presented By: Emily Hauenstein, Ph.D., LCP, APRN, BC, Associate Professor, School of Nursing, University of Virginia, Box 800782, Charlottesville, VA 22908; Tel: (434)924-0093; Fax: 434-982-1809; Email: ejh7m@virginia.edu Research Objective: To assess urban-rural and gender differences in mental health treatment rates and pathways to care among non-elderly adults using a refined measure of rurality. Study Design: We use data from the Medical Expenditure Panel Survey (MEPS) to examine the relationship between rurality, gender, and mental health treatment. Three calendar year measures of mental health treatment are examined: any medical visit related to a mental health condition, the number of those visits and any specialized mental health care visit. We use the Urban-Rural Continuum as an ordinal measure of rurality, which classifies metropolitan counties on the basis of the size of the metropolitan area and nonmetropolitan counties on the basis of their urbanization and their adjacency to metropolitan areas. We also assess the sensitivity of core findings to confounding by a number of explanatory factors using logistic and linear regression. The explanatory factors include self-perceived mental health status and physical health status, income-to-needs, insurance status, region, age, race, ethnicity and marital status. Population Studied: The sample consists of adults aged 18 to 64 in the first four panels MEPS, covering a period from 1996 to 2000. The MEPS is a large nationally representative panel survey of households designed to provide estimates of the use of health services, medical expenditures and sources of payment. The sample size is 33,090. Of these respondents, approximately 3% reside in the two most rural types of counties on the urban-rural continuum. Principal Findings: Rural men and women report poorer mental health than do urban dwellers. Despite this, they receive less treatment than do urban men and women. Only 6.2% of rural women received mental health visits while 5.4% of rural men received mental health visits. Urban women are more likely to receive a mental health visit (10.3%) than urban men (5.9%). Marital disruption is a particular salient determinant of mental health treatment for rural women. Among rural men, higher income is strongly associated with mental health treatment. Conclusions: Rural men and women receive substantially less mental health care than urban women, despite greater needs. Implications for Policy, Delivery, or Practice: Both men and women living in rural counties may have unique problems that put them at greater risk for mental health problems. Inadequate services and mental health delivery mechanisms may heighten that risk by reducing overall access to treatment. Primary Funding Source: National Institute of Mental Health ●Are Women's Pregnancy Intentions Associated With Health Care Access? Marianne Hillemeier, Ph.D., MPH, Carol S. Weisman, Ph.D. Presented By: Marianne Hillemeier, Ph.D., MPH, Assistant Professor, Health Policy and Administration, The Pennsylvania State University, 116 Henderson, University Park, PA 16802; Tel: (814)863-0873; Fax: (814)863-2905; Email: mmh18@psu.edu Research Objective: Women’s health status prior to pregnancy is likely to influence their subsequent risk of experiencing pregnancy complications and adverse birth outcomes including preterm birth and low birthweight. Because high quality health care can play a key role in promoting optimal health and well-being, the degree to which preconceptional women have access to such services is of considerable interest. The objective of this study is to determine whether women who intend to become pregnant differ from those who do not in their access to health care. Study Design: A telephone survey using random-digit dialing was administered to a representative sample of 2,000 reproductive aged women ages 18-45. The survey asked about the following reproduction-related and health care access information: 1) pregnancy history; 2) contraceptive use; 3) intent to become pregnant within the next year, at some other time in the future, or not at all; 4) whether respondent had a usual source of care and type of provider; 5) health insurance status; 6) visits to provider and health services received in previous 12 months; and 7) perceived obstacles to obtaining health care. Information was also collected related to socioeconomic status and county of residence, which allowed linkage to county-level information about the availability of health care providers and facilities. Population Studied: The study population includes women living in a 28 county region of central Pennsylvania. This region includes a large and predominantly white rural population, and includes 15 counties that are primarily or entirely rural. There are also several metropolitan areas in the region, in which sizable African American and Latina populations reside. The study region is relatively disadvantaged in terms of health care resources, with only 8 counties having hospitals that offer both obstetric and neonatal intensive care services, and 6 rural counties in the region lacking hospitals with either type of service. Principal Findings: About one in five reproductive-aged women in our sample reported being uninsured at some time during the previous year, and 6% lacked a usual source of health care. Access to care measures varied among groups of women with different pregnancy intention. The most favorable access measures were seen among women not planning to become pregnant who were, for example, least likely to have been uninsured and more likely to have a usual source of care. In contrast, women who were planning to become pregnant either in the next year or at some other time in the future were at higher risk of being uninsured and lacking a usual source of care. Conclusions: Women who plan to become pregnant have disproportionately poorer access to health care than other reproductive-aged women. Implications for Policy, Delivery, or Practice: Because poor preconceptional health status may precipitate adverse pregnancy and birth outcomes, increasing access to health care for reproductive-aged women at risk for pregnancy should be a priority. Primary Funding Source: Pennsylvania Deptartment of Health ●Maternal Morbidity Rates in a Managed Care Population Mark C. Hornbrook, Ph.D., Cynthia Berg, M.D., MPH, Evelyn Whitlock, M.D., MPH, William Callaghan, M.D., MPH, Carol Bruce, MPH, Patricia Dietz, DrPH, MPH, Rachel Gold, Ph.D., MPH Presented By: Mark C. Hornbrook, Ph.D., Chief Scientist, Center for Health Research, Kaiser Permanente Northwest, 3800 North Interstate Avenue, Portland, OR 97227-1110; Tel: (503) 335-6746; Fax: (503) 335-2228; Email: mark.c.hornbrook@kpchr.org Research Objective: To assess the prevalence of pregnancyrelated morbidity in a large community-based population of pregnant women enrolled in an integrated health care delivery system. Most research on maternal and child health has focused on the health of fetuses and newborns, while few researchers have studied the health of pregnant women. Study Design: To ascertain pregnancies in automated encounter and clinical data systems, we developed a computerized episode algorithm to identify pregnancy begin and end dates, pregnancy outcome, duration of gestation, and number of fetuses. Compared to medical records, the algorithm identifies 96 percent of pregnancy episodes. We used ICD-9-CM codes to develop a list of 46 major disease classes that encompass complications directly and indirectly related to pregnancy. The denominator for pregnancy-related morbidity prevalence rates is the number of pregnancies among women of childbearing age in our study population. Population Studied: Among members of Kaiser Permanente Northwest, a prepaid group practice HMO with about 455,000 members, the algorithm searched for evidence of pregnancy episodes for all women of childbearing age, 12 to 55 years old, who were KPNW members for at least 42 continuous days at any time during the study period, January 1, 1998, through December 31, 2001. We included only those episodes fully contained within the study period and the mother’s health plan eligibility period. Principal Findings: Prevalence rates of maternal morbidity ranged from 9 percent for anemia and urinary tract infection to less than 1 percent for tuberculosis. The six most common maternal health problems are anemia, 9.3 percent; urinary tract infections, 9.0 percent; mental health conditions, 9.0 percent; pelvic and/or perineal complications, 7.0 percent; postpartum hemorrhage, 3.8 percent; and asthma, 3.8 percent. Other health problems with less clear relationships to pregnancy included upper respiratory infections, 19.0 percent, and back disorders, 9.8 percent. Out of 24,680 pregnancies, 17 percent ended in therapeutic abortions and 12.5 percent in spontaneous abortions. Over 67.0 percent of pregnancies ended in live births. Ectopic pregnancies accounted for 1.3 percent of episodes, stillbirths accounted for 0.4 percent, and trophoblastic disease accounted for 0.1 percent. Conclusions: Pregnant women experience a variety of mild to severe health problems. Assuring early diagnosis and high quality intervention for pregnancy-related health problems should lead to better outcomes for both mothers and babies. Implications for Policy, Delivery, or Practice: Using computerized administrative and clinical data to retrospectively identify the beginning and end of pregnancy regardless of pregnancy outcome allows for defining a window to explore maternal pregnancy complications. Understanding these complications may facilitate design and improvement of obstetrical care in health care systems. While more than six million women in the US become pregnant each year, efforts to assess the extent and nature of morbidities experienced during pregnancy have been limited to analyzing hospitalization data. These shortcomings are salient because recent changes in medical practice have resulted in increased treatment and better outcomes for maternal complications in outpatient settings. A pregnancy episode framework is needed to measure morbidity profiles in the prepartum, intrapartum, and postpartum phases. Primary Funding Source: CDC ●Predictors of Breast Cancer Screening in African American and White Women Amal Khoury, Ph.D., MPH, Nedra Lisovicz, Ph.D., MPH, Amanda Avis, MPH Presented By: Amal Khoury, Ph.D., MPH, Associate Professor, Health Services Research, Management & Policy, University of Florida, PO Box 100195, Gainesville, FL 32610; Tel: (352)273-6079; Email: akhoury@phhp.ufl.edu Research Objective: Breast cancer is a leading cause of death for American women. Breast cancer mortality rates have decreased in recent years for the population overall. However, disparities exist among subgroups of women. Specifically, African American women are more likely to die from breast cancer than White women, in part because they have lower screening rates and therefore later stage at diagnosis. The objective of this study is to examine factors associated with mammography screening among African American and White, non-Hispanic women in the south. Study Design: We conducted a cross-sectional telephone survey of a representative sample of households in the state of Mississippi with at least one woman 40 years of age and older. Computer-assisted telephone interviewing (CATI) was used. The survey was developed and then pilot tested with a sample of 30 women. An extensive literature review established face validity of the questionnaire. Content validity was established by breast cancer experts. Social cognitive theory was used to guide the analysis. The outcome variable was the proportion of women who reported getting a screening mammogram within one year preceding the survey (for women 50+) and withing the past two years (for women 40 to 49 years of age). Explanatory variables included personal factors (knowledge, attitudes, self-efficacy, perceptions of screening and treatment) and environmental factors (having a usual source of care, perceptions of the healthcare system, social norm). Sociodemographic characteristics were controlled for. Two regression models, one for African American and one for White women, were developed. Population Studied: African American and White, nonHispanic women 40 years of age and older in the state of Mississippi. The final sample included 299 African American and 728 White women for a total sample of $1,027 women. The response rate exceeded 60 percent. Principal Findings: Among women 50+, 68 percent reported getting a screening mammogram in the year preceding the survey. White women appeared more knowledgeable about risk factors and signs and symptoms of the disease. Overall, no differences between the two groups emerged in their awareness of the benefits of screening. Many respondents had negative perceptions of the healthcare system: 62% believed that rich people received better medical care than poor people, and 77% believed that health insurance affected the kind of care that a person received. Also, 44% agreed that hospitals sometimes do not tell patients the truth. Perceptions of racism in healthcare delviery were also assessed and varied by group. The regression analyses are underway to determine to factors associated with screening behavior at the multivariate level. Conclusions: A significant proportion of Mississippian women are not receving screening mammograms according to the national guidelines. Both personal and environmental factors contribute to this problem. Implications for Policy, Delivery, or Practice: Although breast cancer screening rates have increased in recent years and the disparity in screening rates between subgroups of women has narrowed, there is room for further increasing the rates overall and closing the gap between African American and White women. This study results will provide useful information for developing behavioral and healthcare system interventions to achieve this goal. Primary Funding Source: Susan G Komen Breast Cancer Foundation ●Gender Differences in Healthcare Utilization and Expenditures Associated With Chronic Conditions Kristen Kjerulff, Ph.D., Carol S. Weisman, Ph.D., Kevin D. Frick, Ph.D., Jeffrey A. Rhoades, Ph.D., Anne-Marie Dyer, MS Presented By: Kristen Kjerulff, Ph.D., Associate Professor, Health Evaluation Sciences, Penn State College of Medicine, 600 Centerview Drive, Hershey, PA 17033-0855; Tel: (717)5311258; Fax: (717)531-0839; Email: kkjerulf@hes.hmc.psu.edu Research Objective: To investigate gender differences in healthcare utilization and expenditures associated with the presence of one or more chronic conditions among American men and women and to assess the impact of health insurance, income, age, race and health status on these factors. Study Design: Estimates were derived from the 2001 Medical Expenditure Panel Survey (MEPS), a representative national survey of 33,556 non-institutionalized individuals. Weights were used to estimate the number of men and women aged 20 and older who had 1 or more chronic conditions, the number of ambulatory care visits, prescription medications filled, emergency room visits and hospital stays associated with these disorders; the associated expenditures for each type of utilization and overall. We used multiple and logistic regression equations to estimate the effects of health insurance, income, race and health status on these factors, controlling for age and sex-specific conditions (such as prostate cancer or breast cancer). Population Studied: This study included all males (n = 10,573) and females (n = 12,105) aged 20 and older who participated in the 2001 MEPS. Principal Findings: 60.7% of females and 48.1% of males aged 20 and older had one or more chronic conditions in 2001, and 25.11% of females and 4.24% of males reported having a sex-specific condition. The most common chronic conditions for females were pulmonary diseases (25.5%), hypertension (20.0%)and mental disorders (18.6%), while for males the most common chronic conditions were pulmonary diseases (16.8%), hypertension (16.4%) and joint disorders (10.6%). Women had an average of 1.35 chronic conditions, while men had an average of 0.94 chronic conditions. The average overall medical expenditures among women with 1 or more chronic conditions was $4,694.45, the largest component of which was ambulatory care (30.4%). For men with 1 or more chronic conditions, the average overall expenditure was $4,707.52, the largest component of which was in-patient care (35.19%). Multiple regression indicated that even after controlling for age, race, number of chronic conditions, sex-specific conditions, health insurance coverage and health status, total medical expenditures were higher for American women than men. Conclusions: American women are more likely than American men to report having both chronic and sex-specific conditions and have higher rates of healthcare utilization and expenditures associated with these conditions. However, women with chronic conditions are more likely to use ambulatory care and less likely to be hospitalized than men with chronic conditions. Implications for Policy, Delivery, or Practice: It's important that projections of future patterns of healthcare utilization take into account the higher rates of chronic as well as sex-specific conditions among American women in comparison to men, and higher rates of heathcare utilization, particularly in terms of ambulatory care. Primary Funding Source: AHRQ ●Depression & Diabetes: Sex Differences Cheryl Laskowski, DNS, APRN-BC, Benjamin Littenberg, M.D. Presented By: Cheryl Laskowski, DNS, APRN-BC, Assistant Professor, Department of Nursing, University of Vermont, Rowell Room 202, 106 Carrigan Drive, Burlington, VT 05401; Tel: (802)656-0229; Fax: (802)656-8306; Email: cheryl.laskowski@uvm.edu Research Objective: To explore the association between depression and glycemic control in patients with diabetes. Study Design: Cross-sectional survey. Depression was measured by the Patient Health Questionnaire-9 (PHQ), a validated measure of the presence and severity of depression in primary care. Diabetic control was measured by serum A1C. We used linear regression to control for age, education, marital status, duration of diabetes, and antidepressant use. Population Studied: 333 adults (182 women and 151 men) enrolled in the Vermont Diabetes Information System. All subjects were receiving outpatient services in Primary Care settings in Vermont or adjoining states and had been diagnosed with diabetes by their Primary Care Provider. Principal Findings: The mean A1C was 7.23% (median 7; IQR 6.3, 7.9). The mean PHQ score was 3.8 (median 2; InterQuartile Range 0, 5). 68% of subjects had no or minimal findings of depression (PHQ<5; mean A1C 7.15%) 22% were mildly depressed (PHQ 5-9; A1C 7.29%). 11% were moderately or severely depressed (PHQ>9; A1C 7.60%). Neither PHQ nor A1C varied significantly by sex. PHQ was directly associated with A1C by simple linear regression without adjustment (coefficient = +0.036; 95% CI +0.006, +0.065; P=0.015). Exploratory analyses revealed that sex was an important modifier of this relationship. In linear regressions controlling for age, education, marital status, duration of diabetes, and antidepressant use, PHQ was independently associated with A1C in men (0.059; CI +0.008, +0.109; P=0.023) but not women (-0.010; CI -0.051, +0.031; P=0.62). The results were similar using other markers of depression such as self-report. Conclusions: In our study, a 5 point difference in PHQ scores (about one severity category) was associated with a clinically important 0.29% difference in A1C in men. No such relationship was observed in women. Possible explanations for these findings include differential measurement of depression in men and women (instrument bias), unmeasured confounders, or intrinsic psychological or biologic sex differences. The causal direction of the relationship between depression and diabetes, if any, is unclear. Implications for Policy, Delivery, or Practice: Depression is common in people with diabetes. The degree of depression appears to be related to glycemic control in men. Further research is required to examine whether treatment of depression alters glycemic control (or vice versa) in men or women. Primary Funding Source: NIDDK ●National Patterns of Medication Use during Pregnancy Euni Lee, Pharm.D., Ph.D., Sheila R. Weiss, Ph.D., Leah Smith, Mary Maneno, Anthony K. Wutoh, Ph.D., Ilene H. Zuckerman, Pharm.D. Presented By: Euni Lee, Pharm.D., Ph.D., Assistant Professor, Clinical and Administrative Pharmacy Science, Howard University, 2300 4th Street NW, Washington, DC 20059; Tel: (202)806-4919; Fax: (202)806-4478; Email: eunlee@howard.edu Research Objective: To estimate and compare the national prevalence of medication use during pregnancy in two ambulatory care settings (i.e., office-based vs. hospital outpatient department) and to describe the use by therapeutic drug class, and the Food and Drug Administration (FDA) pregnancy risk category. Study Design: Cross-sectional analysis of the 1999 and 2000 National Ambulatory Medical Care Survey (NAMCS) and National Hospital Ambulatory Medical Care Survey (NHAMCS) Population Studied: Visits to office-based physicians and hospital ambulatory care clinics by pregnant women Principal Findings: In 1999 and 2000, pregnant women made a total of 54.5 million (95% CI 49.3-58.9) and 8.6 million (95% CI 7.9-9.3 million) visits to private- and hospital-based ambulatory clinics, respectively. The mean age was 27.7 (SD 10.4) years and 25.4 (SD 11.4) years and most visits were made by women of white race (84% in private vs. 67% in hospital setting). Paytypes varied by setting with public insurance more prevalent in the hospital setting. In addition, visits that were associated with complications of pregnancy were more prevalent in the hospital setting. One half of all pregnant visits were recorded with one or more medications. The prevalence was higher in the hospital care setting (44% vs. 54%). Other than vitamins or mineral supplements, top drug classes were medications for anemia and blood glucose regulators. However, antiasthmatics and diagnostic agents were more prevalent in the hospital care setting and antimicrobial agents were more prevalent in the private care setting. When prescribed medications were classified by the FDA pregnancy risk category according to drug information references, approximately one half of all prescriptions were from FDA class A (47.5% in private vs 54.2% in hospital setting) followed by class C (22.2% in private vs 28.1% in hospital setting). Of all the medications that were prescribed to pregnant women, drugs classified as Class D/X accounted for 5.0% of prescriptions in the private and 2.1% in the hospital setting.(p<0.001) A list of the drugs from Class X include estrogens, medroxyprogesterone, warfarin, methotrexate, triazolam, and simvastatin. Logistic regression analysis adjusted by other visit characteristics showed no difference in patterns of prescribing drugs with pregnancy risk of Class D/ X by the care setting. Conclusions: Many pregnant women require drug therapy because of pregnancy-induced, chronic, or acute conditions. Our findings show that in one half of all visits made by pregnant women one or more medications were prescribed. Although visit characteristics and patterns of prevalent drug forms were different between private- and hospital-based ambulatory care settings, quality of care measured by prescriptions with the FDA high risk drugs was not different between these settings. Implications for Policy, Delivery, or Practice: Drug use in pregnancy has been an importance issue among health care providers. Few data have been available to provide insight of prevalence of medication use among pregnant women in ambulatory care setting. Clear understanding regarding the effects of those medications on fetal development is required and appropriate team efforts are needed to choose medications with less risk during pregnancy. Primary Funding Source: Funds for Academic Excellence, Howard University ●The Postpartum Health of Employed Mothers Patricia McGovern, Ph.D., MPH, RN, Bryan Dowd, Ph.D., Dwenda Gjerdingen, M.D., MS, Cindy Gross, Ph.D., Todd Rockwood, Ph.D., Sally Kenney, Ph.D., Laurie Ukestad, MS Presented By: Patricia McGovern, Ph.D., MPH, RN, Associate Professor, School of Public Health, University of MN, Mayo Building, MMC 807, 420 Delaware Street SE, Minneapolis, MN 55455; Tel: (612)625-7429; Fax: (612)626-4837; Email: pmcg@umn.edu Research Objective: To evaluate the effects of personal, family and employment factors on health at six weeks postpartum. Study Design: This study employs a hybrid model of health and workforce participation and a prospective cohort design to estimate a production function for maternal health. The woman’s health is measured at six weeks after childbirth as a function of personal, family and employment characteristics and choice variables. Women were recruited into the study while hospitalized for childbirth. Data were abstracted from hospital records; women were interviewed in-person at enrollment and by-telephone at six weeks postpartum. TwoStage Least Squares (2SLS) was used to estimate the effects of explanatory variables on maternal postpartum health. Women’s health outcomes were measured with the SF-12 to assess general mental and physical health and with a symptom list developed by the investigators. Population Studied: The population included all women 18 years of age and older, residing in the Twin Cities, Minnesota admitted to the hospital and giving birth in 2001. Women were enrolled from three hospitals where the demographics of the birth mothers were consistent with the population. Sample selection criteria included: having had a live birth of a singleton infant with no serious complications, being 18 years or older, and employed before childbirth. Principal Findings: At 6 weeks postpartum, 716 women (88% of enrollees) were employed and completed an interview. On average, women were 30 years old; 86% were Caucasian, 73% were married, 47% were first time mothers, 7% had returned to work, and 67% were breastfeeding. On the SF-12, postpartum women scored 51.4 (SD:7.2)on the Physical Component Summary (PCS) score (national norm:52.7;SD:9.1), and 49.4 (SD:7.6)on the Mental Component Summary (MCS) score (national norm: 47.2;SD:12.1). Postpartum women scored slightly lower on physical health and slightly higher on mental health than national (non-postpartum) norms with significant differences on z-scores. On average, women reported 6 postpartum symptoms (SD:3.5). The most frequent symptoms reported were breast symptoms(69%), fatigue (64%), decreased desire for sex (52%), headaches(50%)and back or neck pain (45%). Outcomes measures of maternal health included general physical health (PCS), general mental health (MCS) and the number of postpartum symptoms. Multivariate analyses using 2SLS revealed factors associated with better postpartum health on one or more of outcome measures including better pre-pregnancy health, better prenatal moods, greater perceived control, a blue collar/service occupation (vs. professional), more prenatal workplace support, vaginal (vs. C-section) delivery, increased social support, having a girl baby, and not breastfeeding. Conclusions: These mothers continue to experience several childbirth-related symptoms at six weeks after delivery indicating a need for ongoing rest and recovery. Issues for health care providers to consider include counseling women on expected symptoms, functional limitations and appropriate length of Family and Medical Leave in association with delivery type and breastfeeding status and teaching women to ask for support. Implications for Policy, Delivery, or Practice: Women experiencing a greater number of symptoms or poorer postpartum health may need physician authorization for use of intermittent rather than straight-time leave from work under the federal Family and Medical Leave Act. Primary Funding Source: National Institute for Occupational Health and Safety ●Women in the Veterans Health Administration: Medical Conditions, Utilization and Costs of Care Ciaran Phibbs, Ph.D., Susan Frayne, M.D., MPH, Wei Yu, Ph.D., Elizabeth Yano, PhD, Lakshmi Ananth, MS, Samina Iqbal, M.D., Ann Thrailkill, RPN, MSN, CNS Presented By: Ciaran Phibbs, Ph.D., Health Economist, Health Economics Resource Center (152), VA Palo Alto Health Care System, 795 Willow Road, Menlo Park, CA 94025; Tel: (650) 493-5000 x22813; Fax: (650) 617-2639; Email: cphibbs@stanford.edu Research Objective: Historically, men have constituted the vast majority of Veterans Health Administration (VA) patients. With growing numbers of women veterans receiving care in VA, a systematic assessment of their medical profile was needed to guide innovations in VA’s approach to women’s health care. This study examined disease burden, utilization and costs of VA care differ in women veterans, compared to male veteran VA patients. Study Design: Cross-sectional study using fiscal year 2002 centralized VA administrative and clinical files. Patients with missing gender, age, veterans status, or costs of care were excluded. A modified version of AHRQ’s Clinical Classifications Software was used to describe the clinical conditions of patients. Descriptive statistics were used to compare patients by gender and veterans status. Logistic regressions were used to estimate the age-adjusted odds ratios (OR) of the clinical conditions in women veterans versus male veterans. Mean health care utilization and costs were compared by gender and the presence and absence of medical and mental health conditions. Linear regression was used to compare differences in utilization and costs by gender, controlling for age and clinical conditions. Population Studied: All (N=4,444,577) users of VA care in FY 2002. The main analyses focused on the 178,463 women veterans and 3,940,148 male veterans who used VA care. We also characterized the non-veteran women (N=185,040). Principal Findings: Half of the women who received VA care were not veterans, compared to only 3% of males. Among veterans, women were younger, less likely to be married, and more likely to have a service-connected disability. Chronic conditions were common in both men and women veterans with some differences in prevalence. Several conditions were more common in women than in men, most notably osteoporosis (age-adjusted OR 13.6). Mean total annual outpatient encounters per patient was 16.9 for veteran women and 15.1 for veteran men. Mean total annual inpatient days per patient was 1.9 days for women veterans and 2.5 days for male veterans. Mean total annual cost per patient was $3,941 for veteran women and $4,446 for veteran men. Adjusting for age and medical conditions, women veterans had 1% more outpatient encounters, 11% fewer inpatient days, and 3% less total cost. Among veterans with both medical and mental health conditions, women used outpatient services more heavily. Conclusions: There are substantial gender differences in the patients receiving VA care. But, after accounting for differences in age and clinical conditions, the VA costs and utilization are similar across genders. Researchers using VA data need to be careful about non-veterans in the data, especially among women. Implications for Policy, Delivery, or Practice: VA’s efforts to build capacity for the growing women veteran population must account for the spectrum of medical conditions seen in this special population, which cross the life span and include both gender-specific and gender-neutral diseases. To meet the needs of the subset of women veterans with comorbid medical and mental health conditions, who represent high intensity users of outpatient services, delivery systems integrating ambulatory medical and mental health care deserve attention. Primary Funding Source: VA ●Veteran and Non-Veteran Health Care System Use and Diabetes Care Among Women Veterans Usha Sambamoorthi, Ph.D., Patricia Findley, DrPH, MSW, Wenhui Wei, Ph.D. Presented By: Usha Sambamoorthi, Ph.D., Health Scientist, Health Care Knowledge and Management, Veteran Affairs NJ Health Care System, 385 Tremont Avenue, East Orange, NJ 07018; Tel: (973)676-1000 x1512; Fax: (973) 395-7111; Email: ushasambamoorthi@gmail.com Research Objective: This paper describes women who use Veteran Health Administration (VHA) services and compares them to those who use other health systems, examining the impact of VHA system use on utilization of services for diabetes care among veteran women with diabetes. Study Design: Cross-sectional data from the 2003 Behavioral Risk Factor Surveillance Survey were used. Women veterans were identified through a self-reported question on veteran status. We measured quality of care by the total number of completed diabetes process of care measures. Chi-square techniques and mutivariate regressions were used. Population Studied: Veteran women aged 21 and over and not in active military duty (n = 2,494) were included. Principal Findings: Overall, 17% veteran women received either all or some of their health care from VHA facilities. Women who were full VHA users were less likely to be college educated, insured, or healthy. Nearly 7% (n = 194) of veteran women reported physician-diagnosed diabetes. Despite the fact that disadvantaged women were more likely to use VHA services, of those with diabetes, we found those who received all of their health care from the VHA had similar quality of diabetes care and women who sometimes use the VHA were had higher quality of diabetes care, even after adjusting for relevant covariates. Conclusions: Our study highlights underutilization of the VHA by women veterans. Despite the fact that the more vulnerable subgroup of veteran women must rely on the VHA, women veterans get a similar or higher quality of care than the veteran women who do not use VHA care. Implications for Policy, Delivery, or Practice: Known for its overlapping medical, educational, and social missions, the VA is able to provide comprehensive and integrated services. Women who had poor access defined by having a cost barrier to care need special attention to improve their completed diabetes care measures. Primary Funding Source: No Funding Source ●Gender Discrepancies in Level of Disease Progression for Diabetic Patients Jay Shen, Ph.D, Elmer L. Washington, M.D., MPH Presented By: Jay Shen, Ph.D., University Professor, Health Administration, Governors State University, One University Parkway, University Park, IL 60423; Tel: (708)235-2131; Fax: (708) 534-8041; Email: j-shen@govst.edu Research Objective: To assess gender differences in level of progression of diabetes at the point of hospitalization. Study Design: We categorized causes of hospitalization into three tiers. Tier 1 represented those causes being unrelated to diabetes and would be expected to be present in equivalent frequencies among diabetics and non-diabetics. Tier 2 represented those causes (e.g., heart disease, ischemic stroke, pneumonia) of hospitalization for which diabetes is a significant risk factor, and if controlled, would be expected to lower the risk of hospitalization associated with the condition. Tier 3 represented those causes (e.g., diabetic ketoacidosis, diabetic retinopathy, gangrene due to peripheral vascular disease, nonketotic coma, and renal failure caused by Kimmelstein-Wilson related glomerulosclerosis) of hospitalization that never occur in the absence of diabetes. Covariates such as age, race, insurance status, income level, region of residence, and hospital characteristics were controlled in multivariate analyses. Population Studied: 650,735 adults (54% females and 46% males) with type 2 diabetes as principle or secondary diagnosis identified from the 2001 National Inpatient Sample. Principal Findings: 58.3% of females and 52.1% of males belonged to Tier 1, 26.3% of females and 31.7% of males belonged to Tier 2, and 4.2% of females and 5.0% of males belonged to Tier 3. All differences were statistically significant. As compared to females, males were less likely to be admitted as Tier 1 patients (odds ratio (OR), 0.77, [95% confidence interval (CI)], [0.76, 0.78]), but more likely to be admitted as either Tier 2 patients (OR [CI], 1.27 [1.26, 1.29]) or Tier 3 patients (OR [CI], 1.24 [1.21, 1.28]). Conclusions: Male patients with diabetes were more likely to be admitted with clinical conditions consistent with a long term history of poor control. This is consistent with well established differences in care seeking behavior by gender. Implications for Policy, Delivery, or Practice: Historical, social, and behavioral factors contributing to relatively lower levels of care seeking behavior among males need to be addressed in order to improve diabetes control and reduce risk of related diseases and complications among the male diabetic population. Primary Funding Source: GSU Fund ●Social Support as an Explanation for Differential Effects of a Quality Improvement Program for Men and Women Cathy Sherbourne, Ph.D., Chloe Bird, Ph.D., Robert Weiss, Ph.D., Naihua Duan, Ph.D. Presented By: Cathy Sherbourne, Ph.D., Senior Health Policy Analyst, Health, RAND, 1776 Main St. PO Box 2138, Santa Monica, CA 90407-2138; Tel: (310) 393-0411 ext. 7216; Fax: (310) 260-8150; Email: Cathy_Sherbourne@rand.org Research Objective: Gender differences in the impact of primary care quality improvement (QI) programs have been infrequently studied. We found that two gender-neutral QI programs, one that facilitated access to medication management (QI-Meds) and the other that facilitated access to effective psychotherapy for depression (QI-Therapy), help narrow the gender gap that leaves a greater proportion of depressed men untreated compared with women. While QI was effective for both genders, men showed more improvement than women on some outcomes under QITherapy. Here, we explore one potential reason for the beneficial effect of a psychotherapy-oriented intervention for men. Because inadequate social support plays an important role in poor mental health and support is often lower among men, we hypothesized that men with low social support might benefit particularly from an intervention that involved a counseling component. Study Design: Partners in Care (PIC) assigned 46 primary care practices across different sites in the US to either their usual care (UC) for depression or to programs that provided education for providers and patients about depression treatments. PIC provided resources to make it easier to get either medication or cognitive-behavioral psychotherapy if necessary. Analyses were stratified by gender. We compared the trajectory of mental health-related quality-of-life, as measured by the MCS12, for patients with initially low versus high levels of social support within each intervention arm. Interactions between intervention, time and social support were tested. Population Studied: This analysis focused on 1299 primary care patients who screened positive for depression and completed at least one questionnaire over 24 months of follow-up Principal Findings: At baseline men and women had comparable levels of social support; those with higher levels of social support had better mcs12 in each intervention arm than those with low levels of support. There was a significant main effect for social support (t=4.61, p<.0001 for men and t=7.26, p<.0001) but not for interactions between social support and either intervention or intervention*time. For women, the baseline pattern of lower MCS12 for patients with lower social support remained constant over time in each intervention arm. For men with low social support, the QI-Therapy intervention brought mental health scores up to those of men with high levels of social support in the QI-Meds and UC arms. This pattern was maintained over 12-18 months. Conclusions: QI interventions are effective for men and women with low and high levels of social support, but the mental health of women with low support remains lower that those with high support. QI-therapy appeared particularly beneficial to men with low initial social support. Moreover, the benefits relative to usual care or QI-Meds persisted over time. Implications for Policy, Delivery, or Practice: Data on illness trajectories provide information to inform clinical care. For depressed men with low levels of social support a QI program that facilitates access to effective psychotherapy is beneficial. However, QI programs for depressed women with low levels of social support may need to be enhanced further to bring their levels of mental health up to those with high levels of social support. Primary Funding Source: NIMH ●Gender and Diabetes Care Chin-Lin Tseng, DrPH, Susan Frayne, M.D., MPH, Usha Sambamoorthi, Ph.D., Anjali Tiwari, M.D., MS, Mangala Rajan, MBA, Leonard Pogach, M.D., MBA Presented By: Chin-Lin Tseng, DrPH, Health Research Scientist; Assistant Professor, Ceter of Healthcare Lnowledge Management;, New Jersey Health Care System-East Orange VA Medical Center; University of Medicine and Dentistry, 385 Tremont Avenue; #129, East Orange, NJ 07078; Tel: (973)6761000x2028; Fax: (973)395-7114; Email: tseng@njneuromed.org Research Objective: Gender disparities in care have been well documented for a number of conditions. However, gender disparities in diabetes care have not been examined extensively. The study was to assess whether women and men veterans clinical users had similar diabetes care. Study Design: This was a retrospective longitudinal cohort study. The outcome diabetes care variables (fiscal year (FY) 2000) included lab tests (A1c, low-density lipoproteins(LDL), blood pressure(BP), and creatinine), whether these lab values exceeding certain thresholds, as well as indication of eye care visits, foot care visits, and visits to an endocrinologist. Independent variables (FY 1999) were age, race, marital status, comorbid illness burden (physical and mental), number of face to face visits, Veteran priority status, and Medicare enrollment reasons. Multiple logistic regression models were used to study the gender effect on diabetes care measures while controlling for other covariates. The results reported below were all significant at Type I error rate 0.01. Population Studied: We used the Diabetes Epidemiology Cohort dataset from our prior work to identify 235,361 veteran clinical users with diabetes who were aged 65 years or older and were dually enrolled in Medicare fee-for service in FYs 1999 and 2000 alive at the end of FY2000. 3227 (1.4%) were women; 232,134 (98.6%) were men. Principal Findings: Compared to men with diabetes, women with diabetes were older (74.5+-5.2 vs. 72.7+-5.1), more likely to be white (89.5% vs. 82.9%), less likely to be married (27.2% vs. 69.9%). They had fewer physical health conditions (4.8+2.8 vs. 4.9+-2.8), but more likely to have any of the six more prevalent mental health conditions included in this study. Women also had more frequent face-to-face visits then their men counterparts (12.47+-8.6 vs. 11.5+-7.8). No gender differences existed on any of the A1c and creatinine measures. However, women were less likely to have LDL tested (66.4% vs. 68.95), blood pressure taken (68.0% vs. 74%), and more likely to have LDL exceeding 130 (10.7% vs. 8.9%), and blood pressure exceeding thresholds of 130 (34.6% vs. 31.4%) and of 140 (26.2% vs. 22.75). We continued to observe women doing poorly than men on variables that combined absence of tests with bad control of these process measures for LDL and BP. Women had more frequent visits to specialists such as podiatrists (43.3% vs. 36.8%) and endocrinologists (16.2% vs. 13.9%), although there were no gender differences on eye care. These gender differences remained after adjusting for other covariates except for presence of LDL tests, combined measures of absence of tests and bad control of LDL (both 100 and 130 thresholds) and BP measures (the 130 threshold). Conclusions: There were no gender differences on measures of A1c, creatinine, and eye care. Gender differences existed on LDL, BP, and visits to specialists. Implications for Policy, Delivery, or Practice: Our data suggest gender differences only existed on some aspects of diabetes care. It may be due to differential emphasis on various aspects of diabetes care at VHA and requires further investigation. Similar studies may be extended to the civilian population. Primary Funding Source: VA ●Causes and Effects of a Changed Gender Ratio in the Health Professions Leo van der Reis, M.D. Presented By: Leo van der Reis, M.D., Adjunct Professor in Health Care Management; Clinical Professor of Community & Rural Medicine, University of Alabama-Quincy Foundation, 70 San Pablo Avenue, San Francisco, CA 94127; Tel: (415)661 8865; Fax: (415)566 0110; Email: lvdr@comcast.net Research Objective: To determine the short and long term effects of the changes that have taken place in the gender ratio in the health professions since World War II. To describe remedies to offset undesirable aspects of the change in gender ratio. Study Design: Continued monitoring of the scientific and lay press and interviews with students and graduates from professional schools in medicine, dentistry, pharmacy and veterinary medicine. A form has been developed which assesses and tabulates a broad spectrum of opinions and attitudes. Population Studied: Students, matriculants, faculty and administrators of health care professional schools primarily in the United States, but also including Switzerland, the United Kingdom, Australia, New Zealand, Israel and The Netherlands. The foreign data provide valuable comparative information. Principal Findings: A) Persistent " traditional" discrimination against women; B) Little recourse on the part of complainants; C) Persistent skewing of female faculty compensation; D)A lower percentage of opportunities for professional women when viewed against the number of women in the health care professions. Conclusions: In spite of legislative actions, traditional bias against women in the health professions persists. Measures must be taken to remedy the unfortunate and undesirable effects in order to prevent deterioration in the health care services in the United States and world wide. Implications for Policy, Delivery, or Practice: In spite of the increased number of women, the overall number of health care professionals is declining. This phenomenon is related to the impact of the changed gender ratio. Policy makers must consider the related issues in order to prevent deterioration of health care services. Primary Funding Source: Other Foundation ●Enrollment in Medicare HMO by Elderly Women Veterans Iris Wei, DrPH,Dolly John, Jessica Davila, Margaret Byrne, Laura Petersen, Nora Osemene, Raji Sundaravaradan, Robert Morgan Presented By: Iris Wei, DrPH, Michael E. DeBakey VA Medical Center (152), 2002 Holcombe Boulevard, Houston, TX 77030; Tel: (713) 794-8648; Email: wei.iris@med.va.gov Research Objective: Although the Veterans’ Health Administration, VHA, expanded women-specific services in 1992, studies investigating women veterans’ healthcare use are limited. Researches on women veterans’ enrollment in Medicare Health Maintenance Organizations, HMOs, are especially scarce. The present study examines how the availability of Medicare HMO health plans affects women veterans’ decisions to enroll in Medicare HMO, and how this effect varies by health plan or personal characteristics of VHAusing women veterans. Study Design: We pooled VHA user datasets including fiscal years 1996-2001 to create a VHA-user cohort for calendar year, CY2000. We merged the cohort with: CY2000 Medicare denominator data which provided age, sex, race, and state buy-in status; Group Health Plan data which provided enrollees’ Medicare contracts; and Benefits Information Form data, BIF, for contract plan benefits and plan service areas. We identified all BIF contracts, N=261, and the counties where they were active in CY2000, N=817. Enrollment status was defined by having a BIF contract at any time during CY2000. We used logistic regressions to examine the likelihood of enrolling in Medicare HMO by VHA-using women veterans aged >=65, N=24,354, adjusting for age, race, income, VHA priority code, distance to VHA healthcare facilities, metro/nonmetro status of residence, and separately, three county-level plan benefit indicators. The benefit indicators signified whether each county offered a basic plan with $0 monthly premium, prescription drug coverage, or generous benefit package. Population Studied: Medicare-enrolled women veterans aged >=65 who used VHA healthcare between 10/1/1996 and 12/31/2000. Principal Findings: Our data showed 24% Medicare HMO enrollment. The availability of plans with generous benefits, prescription drug coverage, or a $0 monthly premium was positively related to higher HMO enrollment, ORs ranged 1.44 to 1.80 with 95%CI ranged 1.32 to 2.13. Additionally, the availability of >1 plan, farther distance from residence to the VHA inpatient facility, vs. <15 miles, lower service priority, 5 or 7 vs. 1-4 and 6, and Hispanics were associated with higher HMO enrollment, ORs ranged from 1.16 to 1.97 with 95%CI ranged 1.04 to 2.22. In contrast, age >75 years, lower income, <=20K, use of VHA in CY2000, state buy-in, non-metropolitan residence, or farther distance to a VHA outpatient facility were associated with lower HMO enrollment, ORs ranged 0.70 to 0.90 with 95%CI ranged 0.59 to 0.99. Conclusions: Medicare HMO plans with more generous benefits, prescription coverage, or a $0 monthly premium attracted higher enrollment. Conversely, older age, lower income, living farther away from VHA health facilities, and living in non-metropolitan counties were associated with a lower likelihood of HMO enrollment among Medicareenrolled, VHA-using women veterans. Implications for Policy, Delivery, or Practice: Elderly women veterans’ enrollment in Medicare HMO plans can be constrained by person level characteristics such as age, income, where they live, as well as by system level characteristics like number of active HMO plans in county of residence, and the scope of plan benefits. Future studies need to delineate whether service availability or quality of care at the VHA drive women veterans to seek non-VHA care. Such information can better guide VHA's committed efforts to deliver comprehensive healthcare to women veterans. Primary Funding Source: VA ●The Utility of Intimate Partner Violence: Impact on Health Related Quality of Life Eve Wittenberg, Ph.D., MPP, Erika L. Lichter, Sc.D., Michael L. Ganz, Ph.D., Laura A. McCloskey, Ph.D. Presented By: Eve Wittenberg, Ph.D., MPP, Senior Scientist, Institute for Technology Assessment, Massachusetts General Hospital, 101 Merrimac Street 10th floor, Boston, MA 02114; Tel: (617)724-4481; Email: eve@mgh-ita.org Research Objective: To estimate the utility of health states associated with intimate partner violence (IPV) and the independent effect of IPV on health related quality of life. Study Design: In-person interviews were conducted with a convenience sample of 231 abused and non-abused women defined by scores on the Conflict Tactics Scale. Data were collected on health status using the SF-12, which were transformed to utilities using a published algorithm. Linear regression modeling was used to estimate the association between IPV and utility while controlling for other health and demographic factors. Population Studied: A convenience sample of women visiting outpatient hospital departments (obstetrics/gynecology, emergency, primary care, pediatrics and addiction recovery) and freestanding clinics (gynecology) for random reasons in greater Boston. Principal Findings: Utilities estimated from 93 women experiencing varying degrees of IPV showed a median value of 0.58 on a scale of 0 (dead) to 1.0 (perfect health), ranging from 0.64 for less-severe IPV (lower scores on the conflict tactics scale) to 0.53 for more-severe IPV (higher scores). Current IPV was independently associated with lower health related quality of life (HRQOL; as measured by utility), while controlling for depression, comorbid conditions, past IPV, level of education and employment status (n=217; parameter estimate for IPV=-0.14, p=0.003; model adjusted rsquare=0.26). Of these other factors, depression was the largest contributor to lower HRQOL (parameter estimate = - 0.12, p=0.001). The impact of IPV on HRQOL may be most significant at moderate levels of abuse, and less so at lower levels while tapering off at higher ones. Conclusions: Women experiencing IPV have significantly diminished health related quality of life compared with unaffected women and other health conditions. While women who experience IPV have co-existing characteristics and health conditions that diminish health related quality of life, IPV alone has a separate negative association with quality of life that remains after controlling for these other effects. Implications for Policy, Delivery, or Practice: These results provide a previously unidentified comparison between partner violence and other diseases and conditions that affect the health related quality of life of women. The estimated utility values for IPV-related health states can be used in economic analyses of IPV prevention and intervention efforts to determine their cost effectiveness relative to other public health and medical programs. These results inform priorities and resource allocation decisions for efforts to improve health related quality of life. Primary Funding Source: AHRQ