Concordance between medical records and personal interview in ascertaining breast cancer and diabetes status: Analysis of the south Texas women's health project

Accurate ascertainment of risk factors and disease status is vital in public health research for proper classification of research subjects. The two most common ways of obtaining this data is by self-report and review of medical records (MRs). South Texas Women’s Health Project was a case-control study looking at interrelationships between hormones, diet, and body size and breast cancer among Hispanic women 30-79 years of age. History of breast cancer, diabetes mellitus (DM) and use of DM medications was ascertained from a personal interview. At the time of interview, the subject identified her major health care providers and signed the medical records release form, which was sent to the designated providers. The MRs were reviewed to confirm information obtained from the interview.^ Aim of this study was to determine the sensitivity and specificity between MRs and personal interview in diagnosis of breast cancer, DM and DM treatment. We also wanted to assess how successful our low-cost approach was in obtaining pertinent MRs and what factors influenced the quality of MR or interview data. Study sample was 721 women with both self-report and MR data available by June 2007. Overall response rate for MR requests was 74.5%. MRs were 80.9% sensitive and 100% specific in confirming breast cancer status. Prevalence of DM was 22.7% from the interviews and 16% from MRs. MRs did not provide definite information about DM status of 53.6% subjects. Sensitivity and specificity of MRs for DM status was 88.9% and 90.4% respectively. Disagreement on DM status from the two sources was seen in 15.9% subjects. This discordance was more common among older subjects, those who were married and were predominantly Spanish speaking. Income and level of education did not have a statistically significantly association with this disagreement.^ Both self-report and MRs underestimate the prevalence of DM. Relying solely on MRs leads to greater misclassification than relying on self-report data. MRs have good to excellent specificity and thus serve as a good tool to confirm information obtained from self-report. Self-report and MRs should be used in a complementary manner for accurate assessment of DM and breast cancer status.^

Maternal mortality in Texas 2001--2006

Background. Maternal mortality is often used as a measure of health and well being of women across the globe. Improved surveillance efforts at the state level can improve maternal mortality estimates and develop strategies to address the needs of maternal and child health populations. The aims of this study are (1) To provide better estimates of maternal mortality in Texas; (2) To better understand the origin, governance, function, sustainability and impact on policy and practice of maternal mortality review committees at the state level; and (3) To create a comprehensive implementation model for a statewide maternal mortality review committee.^ Methods. AIM I: Analyzed the enhanced surveillance of fetal death and live birth records linked to pregnancy-related and women of childbearing age (15-44 years) deaths records in Texas from 2001-2006. AIM II: Conduct semi-structured telephone interviews of key informants from states with active maternal mortality review committees. AIM III: Develop a comprehensive maternal mortality review committee implementation model for Texas from the results of AIMS I and II. ^ Results. AIM I: Enhanced surveillance methods identified almost 3.5 times more deaths that may be associated with pregnancy than standard methods. The leading cause of pregnancy-associated death from 2001-2006 among all causes, was accidents. The estimated pregnancy-associated mortality ratio for 2001-2006 was 31 maternal deaths per 100,000 live births. Enhanced surveillance confirmed a persistent race/ethnicity trend in maternal mortality. AIM II: Key informant interviews confirmed existing literature on maternal mortality review committees. Sustainability was maintained not only by the funding; but also by the dedication of committee members to conducting reviews and disseminating recommendations to improving quality of care and systems. All statewide committees examined preventability of deaths and provided recommendations to policymakers and stakeholders. Statewide committees also took the initiative to develop and implement programs to align healthcare systems and improve quality of care.^ Conclusion. The comprehensive implementation model for a statewide maternal mortality review committee has the potential to transform the knowledge learned from enhanced surveillance into a systematic effort to evaluate the circumstances surrounding a pregnancy-associated death; influencing policy and practice decisions addressing maternal mortality, women’s health and maternal and child health in Texas.^