Samantha R. Seals

Survival analysis and regression models

SummaryTime-to-event outcomes are common in medical research as they offer more information than simply whether or not an event occurred. To handle these outcomes, as well as censored observations where the event was not observed during follow-up, survival analysis methods should be used. Kaplan-Meier estimation can be used to create graphs of the observed survival curves, while the log-rank test can be used to compare curves from different groups. If it is desired to test continuous predictors or to test multiple covariates at once, survival regression models such as the Cox model or the accelerated failure time model (AFT) should be used. The choice of model should depend on whether or not the assumption of the model (proportional hazards for the Cox model, a parametric distribution of the event times for the AFT model) is met. The goal of this paper is to review basic concepts of survival analysis. Discussions relating the Cox model and the AFT model will be provided. The use and interpretation of the survival methods model are illustrated using an artificially simulated dataset.

Masked Hypertension and Incident Clinic Hypertension Among Blacks in the Jackson Heart Study

Masked hypertension, defined as nonelevated clinic blood pressure (BP) and elevated out-of-clinic BP may be an intermediary stage in the progression from normotension to hypertension. We examined the associations of out-of-clinic BP and masked hypertension using ambulatory BP monitoring with incident clinic hypertension in the Jackson Heart Study, a prospective cohort of blacks. Analyses included 317 participants with clinic BP <140/90 mm Hg, complete ambulatory BP monitoring, who were not taking antihypertensive medication at baseline in 2000 to 2004. Masked daytime hypertension was defined as mean daytime blood pressure ≥135/85 mm Hg, masked night-time hypertension as mean night-time BP ≥120/70 mm Hg, and masked 24-hour hypertension as mean 24-hour BP ≥130/80 mm Hg. Incident clinic hypertension, assessed at study visits in 2005 to 2008 and 2009 to 2012, was defined as the first visit with clinic systolic/diastolic BP ≥140/90 mm Hg or antihypertensive medication use. During a median follow-up of 8.1 years, there were 187 (59.0%) incident cases of clinic hypertension. Clinic hypertension developed in 79.2% and 42.2% of participants with and without any masked hypertension, 85.7% and 50.4% with and without masked daytime hypertension, 79.9% and 43.7% with and without masked night-time hypertension, and 85.7% and 48.2% with and without masked 24-hour hypertension, respectively. Multivariable-adjusted hazard ratios (95% confidence interval) of incident clinic hypertension for any masked hypertension and masked daytime, night-time, and 24-hour hypertension were 2.13 (1.51–3.02), 1.79 (1.24–2.60), 2.22 (1.58–3.12), and 1.91 (1.32–2.75), respectively. These findings suggest that ambulatory BP monitoring can identify blacks at increased risk for developing clinic hypertension.

Masked Hypertension and Cardiovascular Disease Events in a Prospective Cohort of Blacks: The Jackson Heart Study

Masked hypertension, defined as nonelevated clinic blood pressure (BP) with elevated out-of-clinic BP, has been associated with increased cardiovascular disease (CVD) risk in Europeans and Asians. Few data are available on masked hypertension and CVD and mortality risk among blacks. We analyzed data from the Jackson Heart Study, a prospective cohort study of blacks. Analyses included participants with clinic-measured systolic/diastolic BP <140/90 mm Hg who completed ambulatory BP monitoring after the baseline examination in 2000 to 2004 (n=738). Masked daytime (10:00 AM–8:00 PM) hypertension was defined as mean ambulatory systolic/diastolic BP ≥135/85 mm Hg. Masked nighttime (midnight to 6:00 AM) hypertension was defined as mean ambulatory systolic/diastolic BP ≥120/70 mm Hg. Masked 24-hour hypertension was defined as mean systolic/diastolic BP ≥130/80 mm Hg. CVD events (nonfatal/fatal stroke, nonfatal myocardial infarction, or fatal coronary heart disease) and deaths identified through December 2010 were adjudicated. Any masked hypertension (masked daytime, nighttime, or 24-hour hypertension) was present in 52.2% of participants; 28.2%, 48.2% and 31.7% had masked daytime, nighttime, and 24-hour hypertension, respectively. There were 51 CVD events and 44 deaths during a median follow-up of 8.2 and 8.5 years, respectively. CVD rates per 1000 person-years (95% confidence interval) in participants with and without any masked hypertension were 13.5 (9.9–18.4) and 3.9 (2.2–7.1), respectively. The multivariable adjusted hazard ratio (95% confidence interval) for CVD was 2.49 (1.26–4.93) for any masked hypertension and 2.86 (1.59–5.13), 2.35 (1.23–4.50), and 2.52 (1.39–4.58) for masked daytime, nighttime, and 24-hour hypertension, respectively. Masked hypertension was not associated with all-cause mortality. Masked hypertension is common and associated with increased risk for CVD events in blacks.

Incident Cardiovascular Disease Among Adults With Blood Pressure <140/90 mm Hg

Background: Data from before the 2000s indicate that the majority of incident cardiovascular disease (CVD) events occur among US adults with systolic and diastolic blood pressure (SBP/DBP) ≥140/90 mm Hg. Over the past several decades, BP has declined and hypertension control has improved. Methods: We estimated the percentage of incident CVD events that occur at SBP/DBP <140/90 mm Hg in a pooled analysis of 3 contemporary US cohorts: the REGARDS study (Reasons for Geographic and Racial Differences in Stroke), the MESA (Multi-Ethnic Study of Atherosclerosis), and the JHS (Jackson Heart Study) (n=31 856; REGARDS=21 208; MESA=6779; JHS=3869). Baseline study visits were conducted in 2003 to 2007 for REGARDS, 2000 to 2002 for MESA, and 2000 to 2004 for JHS. BP was measured by trained staff using standardized methods. Antihypertensive medication use was self-reported. The primary outcome was incident CVD, defined by the first occurrence of fatal or nonfatal stroke, nonfatal myocardial infarction, fatal coronary heart disease, or heart failure. Events were adjudicated in each study. Results: Over a mean follow-up of 7.7 years, 2584 participants had incident CVD events. Overall, 63.0% (95% confidence interval [CI], 54.9−71.1) of events occurred in participants with SBP/DBP <140/90 mm Hg; 58.4% (95% CI, 47.7−69.2) and 68.1% (95% CI, 60.1−76.0) among those taking and not taking antihypertensive medication, respectively. The majority of events occurred in participants with SBP/DBP <140/90 mm Hg among those <65 years of age (66.7%; 95% CI, 60.5−73.0) and ≥65 years of age (60.3%; 95% CI, 51.0−69.5), women (61.4%; 95% CI, 49.9−72.9) and men (63.8%; 95% CI, 58.4−69.1), and for whites (68.7%; 95% CI, 66.1−71.3), blacks (59.0%; 95% CI, 49.5−68.6), Hispanics (52.7%; 95% CI, 45.1−60.4), and Chinese-Americans (58.5%; 95% CI, 45.2−71.8). Among participants taking antihypertensive medication with SBP/DBP <140/90 mm Hg, 76.6% (95% CI, 75.8−77.5) were eligible for statin treatment, but only 33.2% (95% CI, 32.1−34.3) were taking one, and 19.5% (95% CI, 18.5−20.5) met the SPRINT (Systolic Blood Pressure Intervention Trial) eligibility criteria and may benefit from a SBP target goal of 120 mm Hg. Conclusions: Although higher BP levels are associated with increased CVD risk, in the modern era, the majority of incident CVD events occur in US adults with SBP/DBP <140/90 mm Hg. While absolute risk and cost-effectiveness should be considered, additional CVD risk-reduction measures for adults with SBP/DBP <140/90 mm Hg at high risk for CVD may be warranted.

White-Coat Effect Among Older Adults: Data From the Jackson Heart Study.

Many adults with elevated clinic blood pressure (BP) have lower BP when measured outside the clinic. This phenomenon, the “white‐coat effect,” may be larger among older adults, a population more susceptible to the adverse effects of low BP. The authors analyzed data from 257 participants in the Jackson Heart Study with elevated clinic BP (systolic/diastolic BP [SBP/DBP] ≥140/90 mm Hg) who underwent ambulatory BP monitoring (ABPM). The white‐coat effect for SBP was larger for participants 60 years and older vs those younger than 60 years in the overall population (12.2 mm Hg, 95% confidence interval [CI], 9.2–15.1 mm Hg and 8.4 mm Hg, 95% CI, 5.7–11.1, respectively; P=.06) and among those without diabetes or chronic kidney disease (15.2 mm Hg, 95% CI, 10.1–20.2 and 8.6 mm Hg, 95% CI, 5.0–12.3, respectively; P=.04). After multivariable adjustment, clinic SBP ≥150 mm Hg vs <150 mm Hg was associated with a larger white‐coat effect. Studies are needed to investigate the role of ABPM in guiding the initiation and titration of antihypertensive treatment, especially among older adults.

Diagnostic performance of matrix-assisted laser desorption ionisation time-of-flight mass spectrometry in blood bacterial infections: a systematic review and meta-analysis.

Abstract Background: Matrix-assisted laser desorption ionisation time-of-flight mass spectrometry (MALDI-TOF MS) has shown promise in decreasing time to identification of causative organisms compared to traditional methods; however, the utility of MALDI-TOF MS in a heterogeneous clinical setting is uncertain. Objectives: To perform a systematic review on the operational performance of the Bruker MALDI-TOF MS system and evaluate published cut-off values compared to traditional blood cultures. Methods: A comprehensive literature search was performed. Studies were included if they performed direct MALDI-TOF MS analysis of blood culture specimens in human patients with suspected bacterial infections using the Bruker Biotyper software. Sensitivities and specificities of the combined studies were estimated using a hierarchical random effects linear model (REML) incorporating cut-off scores of ≥1.7 and ≥2.0. Results: Fifty publications were identified, with 11 studies included after final review. The estimated sensitivity utilising a cut-off of ≥2.0 from the combined studies was 74.6% (95% CI = 67.9–89.3%), with an estimated specificity of 88.0% (95% CI = 74.8–94.7%). When assessing a cut-off of ≥1.7, the combined sensitivity increases to 92.8% (95% CI = 87.4–96.0%), but the estimated specificity decreased to 81.2% (95% CI = 61.9–96.6%). Conclusions: In this analysis, MALDI-TOF MS showed acceptable sensitivity and specificity in bacterial speciation with the current recommended cut-off point compared to blood cultures; however, lowering the cut-off point from ≥2.0 to ≥1.7 would increase the sensitivity of the test without significant detrimental effect on the specificity, which could improve clinician confidence in their results.

Physical Activity and Incident Hypertension in African Americans: The Jackson Heart Study

There is limited empirical evidence to support the protective effects of physical activity in the prevention of hypertension among African Americans. The purpose of this study was to examine the association of physical activity with incident hypertension among African Americans. We studied 1311 participants without hypertension at baseline enrolled in the Jackson Heart Study, a community-based study of African Americans residing in Jackson, Mississippi. Overall physical activity, moderate–vigorous physical activity, and domain-specific physical activity (work, active living, household, and sport/exercise) were assessed by self-report during the baseline examination (2000–2004). Incident hypertension, assessed at examination 2 (2005–2008) and examination 3 (2009–2013), was defined as the first visit with systolic/diastolic blood pressure ≥140/90 mm Hg or self-reported antihypertensive medication use. Over a median follow-up of 8.0 years, there were 650 (49.6%) incident hypertension cases. The multivariable-adjusted hazard ratios (95% confidence interval) for incident hypertension comparing participants with intermediate and ideal versus poor levels of moderate–vigorous physical activity were 0.84 (0.67–1.05) and 0.76 (0.58–0.99), respectively (P trend=0.038). A graded, dose–response association was also present for sport/exercise-related physical activity (Quartiles 2, 3, and 4 versus Quartile 1: 0.92 [0.68–1.25], 0.87 [0.67–1.13], 0.75 [0.58–0.97], respectively; P trend=0.032). There were no statistically significant associations observed for overall physical activity, or work, active living, and household-related physical activities. In conclusion, the results of the current study suggest that regular moderate–vigorous physical activity or sport/exercise-related physical activity may reduce the risk of developing hypertension in African Americans.

Relational Intimacy and Sexual Frequency: A Correlation or a Cause? A Clinical Study of Heterosexual Married Women

Researchers and practitioners have noted the importance of using clinical samples in sex therapy research. This study investigated the relationship between perceived levels of marital intimacy, sexual frequency, and sexual functioning among heterosexual married women. A clinical sample of 67 women completed the Couples Satisfaction Index (CSI), the Miller Social Intimacy Test (MSI), the Sexual Satisfaction Scale for Women (SSS-W), the Inclusion of the Other in the Self Scale (IOS), and the Female Sexual Functioning Index (FSFI-6). Data analyses revealed that marital intimacy acted as a predictor in univariate relationships on sexual frequency and sexual functioning but did not act as a mediator on sexual frequency and sexual functioning. Overall, these findings may further the discussion in the treatment of relational intimacy, sexual desire discrepancy, and female sexual dysfunction.

Clinic and ambulatory blood pressure in a population-based sample of African Americans: the Jackson Heart Study

Blood pressure (BP) can differ substantially when measured in the clinic versus outside of the clinic setting. Few population-based studies with ambulatory blood pressure monitoring (ABPM) include African Americans. We calculated the prevalence of clinic hypertension and ABPM phenotypes among 1016 participants in the population-based Jackson Heart Study, an exclusively African-American cohort. Mean daytime systolic BP was higher than mean clinic systolic BP among participants not taking antihypertensive medication (127.1[standard deviation 12.8] vs. 124.5[15.7] mm Hg, respectively) and taking antihypertensive medication (131.2[13.6] vs. 130.0[15.6] mm Hg, respectively). Mean daytime diastolic BP was higher than clinic diastolic BP among participants not taking antihypertensive medication (78.2[standard deviation 8.9] vs. 74.6[8.4] mm Hg, respectively) and taking antihypertensive medication (77.6[9.4] vs. 74.3[8.5] mm Hg, respectively). The prevalence of daytime hypertension was higher than clinic hypertension for participants not taking antihypertensive medication (31.8% vs. 14.3%) and taking antihypertensive medication (43.0% vs. 23.1%). A high percentage of participants not taking and taking antihypertensive medication had nocturnal hypertension (49.4% and 61.7%, respectively), white-coat hypertension (30.2% and 29.3%, respectively), masked hypertension (25.4% and 34.6%, respectively), and a nondipping BP pattern (62.4% and 69.6%, respectively). In conclusion, these data suggest hypertension may be misdiagnosed among African Americans without using ABPM.

The impact of the TelEmergency program on rural emergency care: An implementation study

Introduction Timely, appropriate intervention is key to improving outcomes in many emergent conditions. In rural areas, it is particularly challenging to assure quality, timely emergency care. The TelEmergency (TE) program, which utilizes a dual nurse practitioner and emergency medicine-trained, board-certified physician model, has the potential to improve access to quality emergency care in rural areas. The objective of this study was to examine how the implementation of the TE program impacts rural hospital Emergency Department (ED) operations. Methods Methods included a before and after study of the effect of the TE program on participating rural hospitals between January 2007 and December 2008. Data on ED and hospital operations were collected one year prior to and one year following the implementation of TE. Data from participating hospitals were combined and compared for the two time periods. Results Nine hospitals met criteria for inclusion and participated in the study. Total ED volumes did not significantly change with TE implementation, but ED admissions to the same rural hospital significantly increased following TE implementation (6.7% to 8.1%, p-value = 0.02). Likewise, discharge rates from the ED declined post-initiation (87.1% to 80.0%, p-value = 0.003). ED deaths and transfer rates showed no significant change, while the rate of patient discharge against medical advice significantly increased with TE use. Discussion In this analysis, we found a significant increase in the rate of ED admissions to rural hospitals with TE use. These findings may have important implications for the quality of emergency care in rural areas and the sustainability of rural hospitals’ EDs.