Leslie Spangler

Validation of electronic medical record-based phenotyping algorithms: results and lessons learned from the eMERGE network

BACKGROUND Genetic studies require precise phenotype definitions, but electronic medical record (EMR) phenotype data are recorded inconsistently and in a variety of formats. OBJECTIVE To present lessons learned about validation of EMR-based phenotypes from the Electronic Medical Records and Genomics (eMERGE) studies. MATERIALS AND METHODS The eMERGE network created and validated 13 EMR-derived phenotype algorithms. Network sites are Group Health, Marshfield Clinic, Mayo Clinic, Northwestern University, and Vanderbilt University. RESULTS By validating EMR-derived phenotypes we learned that: (1) multisite validation improves phenotype algorithm accuracy; (2) targets for validation should be carefully considered and defined; (3) specifying time frames for review of variables eases validation time and improves accuracy; (4) using repeated measures requires defining the relevant time period and specifying the most meaningful value to be studied; (5) patient movement in and out of the health plan (transience) can result in incomplete or fragmented data; (6) the review scope should be defined carefully; (7) particular care is required in combining EMR and research data; (8) medication data can be assessed using claims, medications dispensed, or medications prescribed; (9) algorithm development and validation work best as an iterative process; and (10) validation by content experts or structured chart review can provide accurate results. CONCLUSIONS Despite the diverse structure of the five EMRs of the eMERGE sites, we developed, validated, and successfully deployed 13 electronic phenotype algorithms. Validation is a worthwhile process that not only measures phenotype performance but also strengthens phenotype algorithm definitions and enhances their inter-institutional sharing.

Glad you asked: participants' opinions of re-consent for dbGap data submission.

No consensus exists about when researchers need additional participant consent (re-consent) to submit existing data to the federal database of Genotypes and Phenotypes (dbGaP). Re-consent for submission of their data to dbGaP was sought from 1,340 study participants, 1,159 (86%) of whom agreed. We invited the first 400 of those who agreed to complete a telephone survey about their reasoning for their consent decision and their satisfaction with the re-consent process; 365 participants completed the survey. Respondents reported that it was very (69%) or somewhat (21%) important that they were asked for their permission. Many respondents considered alternatives to consent, such as notification-only or opt-out, to be unacceptable (67% and 40%, respectively). These results suggest that re-consent for dbGaP deposition may be advisable in certain cases to anticipate and honor participant preferences.

Burden of Illness for Osteoporotic Fractures Compared With Other Serious Diseases Among Postmenopausal Women in the United States

OBJECTIVES To provide a national estimate of the incidence of hospitalizations due to osteoporotic fractures (OFs) in women; compare this with the incidence of myocardial infarction (MI), stroke, and breast cancer; and assess temporal trends in the incidence and length of hospitalizations. PATIENTS AND METHODS The study included all women 55 years and older at the time of admission, admitted to a hospital participating in the US Nationwide Inpatient Sample for an outcome of interest. We performed a retrospective analysis of hospitalizations for OFs (hip, forearm, spine, pelvis, distal femur, wrist, and humerus), MI, stroke, or breast cancer, using the US Nationwide Inpatient Sample, 2000-2011. RESULTS From 2000 to 2011, there were 4.9 million hospitalizations for OF, 2.9 million for MI, 3.0 million for stroke, and 0.7 million for breast cancer. Osteoporotic fractures accounted for more than 40% of the hospitalizations in these 4 outcomes, with an age-adjusted rate of 1124 admissions per 100,000 person-years. In comparison, MI, stroke, and breast cancer had age-adjusted incidence rates of 668, 687, and 151 admissions per 100,000 person-years, respectively. The annual total population facility-related hospital cost was highest for hospitalizations due to OFs (

Vitamin D intake from foods and supplements and depressive symptoms in a diverse Population of Older Women.

5.1 billion), followed by MI (

Vitamin D Supplementation and Depression in the Women's Health Initiative Calcium and Vitamin D Trial

4.3 billion), stroke (

Oral Contraceptive Use and Bone Density Change in Adolescent and Young Adult Women: A Prospective Study of Age, Hormone Dose, and Discontinuation

3.0 billion), and breast cancer (

Evaluating components of dental care utilization among adults with diabetes and matched controls via hurdle models

0.5 billion). CONCLUSION These data provide evidence that in US women 55 years and older, the hospitalization burden of OFs and population facility-related hospital cost is greater than that of MI, stroke, or breast cancer. Prioritization of bone health and supporting programs such as fracture liaison services is needed to reduce this substantial burden.

A Population-Based Study of Periodontal Care Among Those With and Without Diabetes

BACKGROUND Vitamin D may plausibly reduce the occurrence of depression in postmenopausal women; however, epidemiologic evidence is limited, and few prospective studies have been conducted. OBJECTIVE We conducted a cross-sectional and prospective analysis of vitamin D intake from foods and supplements and risk of depressive symptoms. DESIGN Study participants were 81,189 members of the Womens Health Initiative (WHI) Observational Study who were aged 50-79 y at baseline. Vitamin D intake at baseline was measured by food-frequency and supplement-use questionnaires. Depressive symptoms at baseline and after 3 y were assessed by using the Burnam scale and current antidepressant medication use. RESULTS After age, physical activity, and other factors were controlled for, women who reported a total intake of ≥800 IU vitamin D/d had a prevalence OR for depressive symptoms of 0.79 (95% CI: 0.71, 0.89; P-trend < 0.001) compared with women who reported a total intake of <100 IU vitamin D/d. In analyses limited to women without evidence of depression at baseline, an intake of ≥400 compared with <100 IU vitamin D/d from food sources was associated with 20% lower risk of depressive symptoms at year 3 (OR: 0.80; 95% CI: 0.67, 0.95; P-trend = 0.001). The results for supplemental vitamin D were less consistent, as were the results from secondary analyses that included as cases women who were currently using antidepressant medications. CONCLUSIONS Overall, our findings support a potential inverse association of vitamin D, primarily from food sources, and depressive symptoms in postmenopausal women. Additional prospective studies and randomized trials are essential in establishing whether the improvement of vitamin D status holds promise for the prevention of depression, the treatment of depression, or both.

Cross-Sectional Study of Periodontal Care and Glycosylated Hemoglobin in an Insured Population

While observational studies have suggested that vitamin D deficiency increases risk of depression, few clinical trials have tested whether vitamin D supplementation affects the occurrence of depression symptoms. The authors evaluated the impact of daily supplementation with 400 IU of vitamin D(3) combined with 1,000 mg of elemental calcium on measures of depression in a randomized, double-blinded US trial comprising 36,282 postmenopausal women. The Burnam scale and current use of antidepressant medication were used to assess depressive symptoms at randomization (1995-2000). Two years later, women again reported on their antidepressant use, and 2,263 completed a second Burnam scale. After 2 years, women randomized to receive vitamin D and calcium had an odds ratio for experiencing depressive symptoms (Burnam score ≥0.06) of 1.16 (95% confidence interval: 0.86, 1.56) compared with women in the placebo group. Supplementation was not associated with antidepressant use (odds ratio = 1.01, 95% confidence interval: 0.92, 1.12) or continuous depressive symptom score. Results stratified by baseline vitamin D and calcium intake, solar irradiance, and other factors were similar. The findings do not support a relation between supplementation with 400 IU/day of vitamin D(3) along with calcium and depression in older women. Additional trials testing higher doses of vitamin D are needed to determine whether this nutrient may help prevent or treat depression.

Bone Mineral Density Screening Among Women with a History of Breast Cancer Treated with Aromatase Inhibitors

CONTEXT Oral contraceptive (OC) use is common, but bone changes associated with use of contemporary OC remain unclear. OBJECTIVE The objective of the study was to compare bone mineral density (BMD) change in adolescent and young adult OC users and discontinuers of two estrogen doses, relative to nonusers. DESIGN AND SETTING This was a prospective cohort study, Group Health Cooperative. PARTICIPANTS Participants included 606 women aged 14-30 yr (50% adolescents aged 14-18 yr): 389 OC users [62% 30-35 μg ethinyl estradiol (EE)] and 217 age-similar nonusers; there were 172 OC discontinuers. The 24-month retention was 78%. MAIN OUTCOME MEASURE The main outcome measure was BMD measured at 6-month intervals for 24-36 months. RESULTS After 24 months, adolescents using 30-35 μg EE OCs, but not those using lower-dose OCs, had significantly smaller adjusted mean percentage BMD gains than nonusers at the spine [group means (95% confidence interval for between group differences) 1.32 vs. 2.26% (-1.89, -0.13%)] and whole body [1.45 vs. 2.03% (-1.29%, -0.13%)]. Adolescents who discontinued 30-35 μg EE OC showed significantly smaller gains than nonusers at the spine after 12 months [0.51 vs. 1.72% (-2.38%, -0.30%)]. Young adult OC users did not differ from nonusers. However, OC discontinuers of both doses differed significantly from nonusers at the spine 12 months after discontinuation [-1.32% < 30 μg EE, -0.92% 30-35 μg EE vs. +0.27% nonusers (-2.48, -0.54, and -1.94%, -0.55%, respectively)]. Results were similar for mean absolute BMD change (grams per square centimeter). CONCLUSIONS Both OC use and discontinuation were associated with BMD losses/smaller gains relative to nonusers (differences < 2% after 12-24 months for all skeletal sites). The clinical significance of these results regarding future fracture risk is unknown. Study of longer-term trends after discontinuation is needed.