Perspectives on the Future of Epidemiology: A Framework for Training.

Over the past century, the field of epidemiology has evolved and adapted to changing public health needs. Challenges include newly emerging public health concerns across broad and diverse content areas, new methods, and vast data sources. We recognize the need to engage and educate the next generation of epidemiologists and prepare them to tackle these issues of the 21st century. In this commentary, we suggest a skeleton framework upon which departments of epidemiology should build their curriculum. We propose domains that include applied epidemiology, biological and social determinants of health, communication, creativity and ability to collaborate and lead, statistical methods, and study design. We believe all students should gain skills across these domains to tackle the challenges posed to us. The aim is to train smart thinkers, not technicians, to embrace challenges and move the expanding field of epidemiology forward.

Posttraumatic Stress Disorder 23 Years After the 1988 Spitak Earthquake in Armenia.

This population-based longitudinal study examined the rates and predictors of posttraumatic stress disorder (PTSD) among 725 differentially exposed survivors of the 1988 Spitak earthquake in Armenia, 23 years after the event. Participants had been previously evaluated in 1991. Evaluations included assessment of current PTSD (based on DSM-5 criteria), and a variety of potential risk and protective factors. For the whole sample, the rate of PTSD attributed to the earthquake decreased from 48.7% in 1991 to 11.6% in 2012 (p < .001). A "dose of exposure" pattern persisted, and 15.7% of participants who were in Spitak (high exposure) and 6.6% of participants who were in Kirovagan (low exposure) during the earthquake met the criteria for PTSD (p = .003). Additionally, in 2012, another 9.9% of participants met PTSD criteria due to post-earthquake traumas, which is a 5-fold increase from pre- to postearthquake (p < . 001). Factors positively associated with PTSD included earthquake-related job loss, exposure to post-earthquake traumas, depression at baseline, and chronic illness since the earthquake. Factors inversely associated with PTSD included housing assistance within two years after the earthquake, support of family and/or friends, and to a lesser degree, higher education and high living standard. These variables accounted for 23.1% of the variance in current PTSD severity scores. These findings indicate that PTSD rates subside significantly after a catastrophic disaster, although earthquake-related PTSD persists among a subgroup of exposed individuals. Predictors of PTSD identified in this study provide guidance for planning acute and longer-term postdisaster public mental health recovery programs.

A post-earthquake psychopathological investigation in Armenia: methodology, summary of findings, and follow-up.

The post-earthquake psychopathological investigation (PEPSI) was designed to probe the short-and long-term effects of the earthquake in northern Armenia on 7 December 1988 on survivors' mental and physical health. Four phases of this study have been conducted to date, and, overall, more than 80 per cent of a sub-sample of 1,773 drawn from an initial cohort of 32,743 was successfully followed during 2012. This paper describes the methodology employed in the evaluation, summarises previous findings, details the current objectives, and examines the general characteristics of the sample based on the most recent follow-up phase outcomes. Despite a significant decrease in psychopathology rates between 1990 and 2012, prevalence rates of post-traumatic stress disorder and depression among study participants in 2012 were greater than 15 and 26 per cent, respectively. The paper also notes the strengths and limitations of the study vis-à-vis future research and highlights the importance and potential practical implications of similar assessments and their outcomes.

Loss and psychosocial factors as determinants of quality of life in a cohort of earthquake survivors.

BACKGROUND: Despite the existing evidence of a long lasting effect of disaster related experiences on physical and psychological health, few studies have evaluated long-term quality of life (QOL) outcomes of disaster survivors and the factors associated with such outcomes. METHODS: 23 years after the 1988 Spitak earthquake in Armenia, the associations of demographic characteristics, trauma exposure and psychosocial variables on QOL were explored among a cohort of 725 exposed individuals. The EQ-5D-5 L instrument was applied to measure QOL of participants. Multivariate linear and ordinal logistic regressions were applied to evaluate the determinants of QOL and its underlying five domains (mobility, self-care, usual activity, pain/discomfort, and anxiety/depression). RESULTS: Older age, current depression, post-traumatic stress disorder and anxiety symptoms were negatively associated with QOL. Additionally, those with severe losses (who did not receive any financial/material aid) had significantly poorer QOL outcomes, with higher odds of mobility difficulties (OR = 1.86, p < 0.05), self-care difficulties (OR = 2.85, p < 0.05), and mood problems (OR = 2.69, p < 0.05). However, those with severe earthquake related losses who received financial/material aid reported less self-care difficulties (OR = 0.21, p < 0.05) usual activity difficulties (OR = 0.40, p < 0.05), and mood problems (OR = 0.44, p < 0.05). Finally, each unit increase in current social support score was found to be significantly associated with a better QOL outcome and better self-reported outcomes across all underlying domains of QOL. CONCLUSIONS: These findings suggest that earthquake related loss and concurrent psychopathology symptoms can have adverse impact on the QOL of survivors. They also indicate that well-targeted post-disaster financial/material aid and social support should be considered as means for improving the long-term QOL outcomes of disaster survivors.

Rate and predictors of postpartum depression in a 22-year follow-up of a cohort of earthquake survivors in Armenia.

Disasters have serious long-term impact on mental health for those exposed. The aim of this study was to identify predictors of postpartum depression among survivors of the 1988 devastating earthquake in Armenia. A nested case-control design was applied to investigate postpartum depression in a large-scale cohort of survivors followed between 1990 and 2012. From an original group of 725 adults who were assessed for psychopathology in 1990, 146 women reported having a delivery after the earthquake and were included in this study. Women with postpartum depression were identified using Edinburgh Postnatal Depression Scale. A logistic regression model was fitted to identify the predictors of postpartum depression. Of the 146 women, 19 (13.0%) had postpartum depression. Five independent predictors of postpartum depression were identified: number of woman's stressful life events (odds ratio (OR)=2.06), her prior history of postpartum depression (OR=16.98), delivering sick/dead neonate (OR=13.65), poor living standards during the post-earthquake decade (OR=5.77), and perceiving oneself reliable in 1990 (OR=0.24). Anxiety in 1990 was marginally significantly related to the outcome (OR=3.75). The rate of postpartum depression in this 22-year cohort was similar to that among the Armenian general population. Earthquake exposure was not related to postpartum depression, indicating that the impact of disaster-related trauma diminishes over time. The identified predictors provided evidence to develop interventions targeting groups of women most prone to postpartum depression under such circumstances.

Epidemiology of Hajj pilgrimage mortality: Analysis for potential intervention.

BACKGROUND: The Saudi Vision 2030 predicts the attendance of 30 million pilgrims each year by 2030. Cost-effective healthcare services during the Hajj are important to manage this increase in the number of pilgrims. Little is known about the impact of the existing chronic diseases on morbidity and mortality during the Hajj. Furthermore, the quality of services provided by Hajj hospitals warrants further study. This study aims to describe patterns of inpatient, all-cause mortality during the Hajj and the relationship between mortality and preexisting chronic diseases as well as the services provided in Hajj hospitals. METHODS: The population included pilgrims who were admitted to Hajj hospitals in Makkah and sacred sites between 2012 and 2017, excluding 2015. A retrospective, matched, case-control study design was utilized. 2237 cases of mortality were matched to 4474 control cases based on age and gender. The data were extracted from hospital admissions offices and medical records. Hierarchical, logistic regression models were used to examine the medical services. The effect measure modification of the copresence of more than one chronic disease was also examined. RESULTS: The rate of inpatient all-cause mortality was higher in Makkah hospitals compared to sacred site hospitals. Inpatient, all-cause mortality was significantly associated with diabetes, hypertension, and cardiovascular diseases. Effect measure modification was present between diabetes and cardiovascular diseases, hypertension, and cardiovascular diseases, but not between diabetes and hypertension. Patients who received medical services were more likely to die during their hospital stay compared to patients not receiving services. CONCLUSION: The current focus on public health issues during the Hajj should be equally distributed between communicable and non-communicable diseases. Although advanced services are provided by Hajj hospitals, interventions to address the increased risks, including mortality, faced by pilgrims with preexisting, chronic diseases should be further investigated and considered.

Short and long term determinants of incident multimorbidity in a cohort of 1988 earthquake survivors in Armenia.

BACKGROUND: Multimorbidity, presence of two or more health conditions, is a widespread phenomenon affecting populations' health all over the world. It becomes a serious public health concern due to its negative consequences on quality of life, mortality, and cost of healthcare services utilization. Studies exploring determinants of multimorbidity are limited, particularly those looking at vulnerable populations prospectively over time. This study aimed at identifying short and long term socioeconomic, psychosocial, and health behavioral determinants of incident multimorbidity among a cohort of the 1988 Armenian earthquake survivors. METHODS: The study included a representative subsample of 725 from a larger initial cohort of the earthquake survivors. Data on this subsample were collected via four phases of this cohort study during the period 1990-2012. The final logistic regression analysis eliminated all those cases with baseline multimorbidity to investigate short and long term determinants of incident multimorbidity; this subsample included 600 participants. RESULTS: More than 75% of the studied sample had multimorbidity. Perceived low affordability of healthcare services, poor living standards during the post-earthquake decade, and lower education were independent predictors of incident multimorbidity developed during the period 1990-2012. Stressful life events and poor social support were among psychosocial determinants of incident multimorbidity. Participants' baseline BMI reported in 1990 was independently associated with incident multimorbidity. CONCLUSIONS: Most of the identified determinants of incident multimorbidity in our study population were markers of social inequities, indicating that inequities pose a serious threat to both individual and public health-related outcomes. Strategies targeting to decrease such inequities along with promotion of healthy lifestyle and strengthening of social networks may considerably reduce multimorbidity among population groups with similar socioeconomic and cultural profiles.

Interbirth interval and history of previous preeclampsia: a case-control study among multiparous women.

BACKGROUND: Preeclampsia is a disorder with a reported incidence of 2%-8% among all pregnancies, accounting for more than 50,000 deaths worldwide each year. In low- and middle- income countries maternal/perinatal morbidity and mortality associated with preeclampsia are high due to the lack of proper prenatal and hospital care and limited access to neonatal intensive care. The objectives of our study were to determine the association of long interbirth interval (IBI) and preeclampsia and to investigate the interactions between long IBI and other risk factors among multiparous women in Yerevan, Armenia. METHODS: We conducted a hospital-based case-control study among 36 multiparous women with preeclampsia (cases) and 148 without preeclampsia (controls) during their last pregnancy, selected from the two largest maternity hospitals in Armenia. The data were collected through telephone-based structured interviews and analyzed using STATA software. The study applied univariate and multivariate logistic regression analyses. RESULTS: The study found a significant interaction between IBI and previous history of preeclampsia. Among women without a history of previous preeclampsia, the odds of having preeclampsia among women with long IBI (greater than or equal to five years) was 6.88 time higher compared to those with short IBI (CI: 1.75-27.05; p = 0.006) after adjusting for confounders; among women with a history of previous preeclampsia the odds ratio was 0.60 (CI: 0.07-4.99; p = 0.638). The final fitted model for preeclampsia among multiparous women who had planned their pregnancies included IBI, time to pregnancy, Body Mass Index, method of contraception and household monthly income. CONCLUSIONS: Long IBI appeared to be a strong risk factor for preeclampsia development only among women without a history of previous preeclampsia. This finding may contribute to a new approach in understanding the etiology of preeclampsia and may be useful for developing further recommendations for this particular subgroup of women that are at higher risk for preeclampsia development in subsequent pregnancies.

Triggers for attacks in familial Mediterranean fever: application of the case-crossover design.

The etiology of recurrent attacks of serositis in familial Mediterranean fever (FMF) is not completely understood. Uncontrolled clinical case series have reported that factors associated with emotional, physiological, or physical stress precede and might trigger the attacks. This case-crossover study, conducted between July 2007 and May 2008, aimed to estimate the role of precipitating factors in attacks in a sample of Armenian FMF patients in Yerevan, Armenia, where 104 patients contributed 55 case and 189 control time periods. The authors used conditional logistic regression to compare frequency of exposure to stressful events, strenuous physical activity, menstrual periods, and high-fat food consumption prior to FMF attacks and on attack-free random days. Multiple stressful life events predicted FMF attacks 2 days following the event. After adjustment for treatment, an additional stressful event was associated with an estimated 70% increase in the odds of having an FMF attack on the second day (95% confidence interval: 1.04, 2.79). High levels of perceived stress were also associated with FMF attacks. Physical exertion and high-fat diet did not increase the likelihood of FMF attacks. The possibility of prevention of attacks in FMF needs to be tested through stress-reduction interventions.

The course, trajectories and predictors of depression 23 years after the 1988 Spitak earthquake in Armenia: A prospective cohort study.

There is a scarcity of long-term studies of depression after natural disasters. This prospective epidemiologic study evaluated the course of depression and factors associated with it among survivors 23-years post-earthquake. A geographically stratified subsample of 725 adults exposed to the Spitak earthquake was assessed for depression using DSM-III-R criteria in 1991 and 2012. Additionally, an adapted CES-D-16 scale was used to assess depressive symptoms in 2012. A fitted multi-nominal logistic regression model identified predictors of different trajectories of depression. For the whole group, the rate of clinical depression dropped from 51.5% in 1991 to 31.7% in 2012. Earthquake intensity and experiencing strong fear at baseline were strongly associated with chronic (depressed at both assessments, 18.1%) and recovered (depressed only in 1991, 33.5%) depression trajectories compared with the healthy group, with odds ratios (OR) over 4.0 and 3.0, respectively. Predictors distinguishing chronic depression included earthquake-related nuclear-family deaths (OR=3.79), chronic illnesses at baseline (OR=1.24), and social support (OR=0.91). Predictors of the late-onset trajectory (depressed only in 2012, 13.5% of the cohort) included post-earthquake trauma (OR=1.42), socioeconomic status (OR=0.45), and social support (OR=0.86). The factors associated with the different trajectories of depression provide guidance for planning more effective mental health interventions after disasters.

Epidemiology: a problem-solving journey.

As a scientific discipline, epidemiology has helped liberate the practice of public health and medicine from dogmatic thinking over the past century. This commentary highlights some integrating principles to explain why epidemiology is a problem-solving discipline. The first of these is that epidemiology is an information science. Epidemiology generates information for decision-making at all levels of the health-care system, including information for both individuals and the general public. Although all scientific disciplines produce data that may be used for decision-making, there is more immediacy for the decisions in epidemiology. The second principle is that epidemiology operates within an environment of complex systems. Etiologic factors operate in complex systems, and the use of a systems analysis approach in investigating health problems must be considered. The third principle is that epidemiology is not just a scientific discipline but a professional practice area as well. Epidemiology has a solid disciplinary scientific base, and its practice requires well-grounded academic preparation. Its objectives are very much within the public-social domain, and a well-defined, outcome-oriented, prevention-based philosophy steers its practice. In a number of universities, epidemiology is taught today as a research discipline rather than as an operational, problem-solving one. In conclusion, this commentary emphasizes the need to accept the fact that epidemiology has as much of a social role as a scientific one. Public health action, problem solving, and a sense of mission are what brings many students to epidemiology. A problem-solving, action-oriented epidemiology is consistent with that sense of mission.

Prevention of amyloidosis in familial Mediterranean fever with colchicine: a case-control study in Armenia.

OBJECTIVE: To determine whether or not the use of colchicine decreases the risk of amyloidosis among Armenian patients with familial Mediterranean fever (FMF). SUBJECTS AND METHODS: The study included 99 Armenian patients from the Center of Medical Genetics database with genetically ascertained FMF; 33 had renal amyloidosis and 66 were randomly selected control patients without renal amyloidosis. Self- reported colchicine use was assessed by interviewer-based questionnaire. RESULTS: The patients with incident amyloidosis were more likely to be older men, but younger at the time of disease onset, and more likely to have had a family history of amyloidosis and M694F mutation in the MEFV gene compared to patients without amyloidosis. The risk of amyloidosis decreased with adequate colchicine use rather than nonadequate use (adjusted odds ratio, OR, 0.48, 95% confidence interval, CI, 0.16-1.43), continuous colchicine use rather than interrupted use (adjusted OR 0.15, 95% CI 0.04-0.53), earlier rather than later initiation age of colchicine treatment (adjusted OR 0.95, 95% CI 0.90-1.01), current colchicine rather than ever/never colchicine use (adjusted OR 0.20, 95% CI 0.05-0.89). CONCLUSION: The study demonstrated that colchicine treatment is effective in preventing amyloidosis among Armenian patients with FMF and that earlier initiation and continuous therapy at an adequate dose of 1.2-1.8 mg/day may be associated with a decreased amyloidosis risk among Armenian patients with FMF.

Low bone mineral density, coronary heart disease, and stroke mortality in men and women: the Third National Health and Nutrition Examination Survey.

PURPOSE: The aim of this study is to determine the long-term association of bone mineral density and cardiovascular disease mortality. METHODS: The data used are from the Third National Health and Nutrition Examination Survey (NHANES III), a nationally representative sample of noninstitutionalized civilians. A cohort of white, black, and Mexican-American persons ages 50 years and older at baseline (1988-1994) was followed through 2000 for coronary heart disease (CHD; n = 4690) and stroke mortality (n = 5272) using the NHANES III Linked Mortality File. RESULTS: Death certificates were used to identify 369 CHD and 166 stroke deaths. Results were evaluated to determine the relative risk of CHD or stroke per one standard deviation lower bone mineral density after adjusting for multiple risk factors. In Cox proportional hazards models, risk of CHD death and risk of stroke death were not associated with low bone mineral density among men. For women, no significant associations were found for stroke (relative risk, 1.34; 95% confidence interval, 0.86-2.07, p = 0.20) or CHD (relative risk, 1.26; 95% confidence interval, 0.88, 1.80; p = 0.21). CONCLUSIONS: Low bone mineral density was not associated with risk of cardiovascular disease in men. Among women with low bone mineral density, risk of CHD and stroke were elevated, but no significant associations were found.

Latency and exposure-health associations in Gulf War veterans with early fatigue onsets: a case-control study.

PURPOSE: To see if self-reported exposures were associated with health in early-onset Gulf War illnesses (GWIs) cases and healthy Gulf War veteran controls. METHODS: Forty-nine cases and 44 controls completed questionnaires about wartime exposures and symptoms experienced. Odds ratios were calculated using 2 x 2 tables and logistic regression. The incubation curve of fatigue onsets in cases was drawn to highlight exposure/health associations using Sartwell's method and tested with the Shapiro-Wilk test. The incubation period was defined as the time from arrival in the Persian Gulf to fatigue onset. RESULTS: The incubation curve was right skewed and lognormally distributed (p = 0.48; p > 0.05 indicates lognormality), suggesting an association between a wartime exposure and fatigue. Exposure to oil fire smoke, pesticides, contaminated food or water, dead animals, scud missile attacks, dead bodies, prisoners of war, artillery or small arms fire, and chemical suits was significantly associated with GWIs. Pyridostigmine bromide (PB) was the only continuous exposure significantly associated with GWIs. The odds of having GWIs increased by 1.3% for every PB pill taken (95% confidence interval 1.001-1.02). There were significant trends toward worse health with greater intake of PB. CONCLUSIONS: These analyses suggest that wartime exposures, including exposure to PB, are associated with fatigue.

A positive or a negative confounding variable? A simple teaching aid for clinicians and students.

Anticipating the direction of a confounding variable can be problematic especially to introductory students. Using elementary rules of mathematics, we describe below a simple instructional tool for deriving the direction of confounding bias. The tool is illustrated with examples and a heuristic mathematical justification is also described.

Application of case-crossover and case-time-control study designs in analyses of time-varying predictors of T-cell homeostasis failure.

PURPOSE: To evaluate the association of sexual behavior and recreational drug exposures with T-cell homeostasis failure (TCHF), which corresponds to the onset of a rapid decline in an individual's T lymphocyte count, which occurs on average approximately 1.75 years prior to an initial diagnosis of acquired immunodeficiency syndrome (AIDS). METHODS: A case-crossover design and a case-time-control design, both nested within the Multicenter AIDS Cohort Study of 4954 homosexual and bisexual men initiated in 1983. RESULTS: In the case-crossover analysis, use of both recreational drugs and hashish were found to be protective against TCHF (odds ratios < or = 0.41), based on comparisons with four earlier control periods. However, a significant decreasing trend in the prevalence of these exposures was observed over time, thus motivating the implementation of the case-time-control design. Using the latter approach, the associations of drug use (odds ratio=0.53; 95% confidence interval (CI): 0.22, 1.28) and hashish use (odds ratio=0.46; 95% CI: 0.20, 1.05) with TCHF were no longer statistically significant. CONCLUSIONS: The difference in inferences between these approaches demonstrates the importance of evaluating temporal trends in exposures when using a case-crossover design.

Prospective study of predictors of poor self-rated health in a 23-year cohort of earthquake survivors in Armenia.

Long-term prospective studies exploring general health outcomes among disaster survivors are rare. Self-rated health (SRH) - a proven correlate of morbidity and mortality prognosis - was used to investigate predictors of perceived health status among a 23-year cohort of survivors of 1988 Spitak earthquake in Armenia. A geographically-stratified subsample of 725 adults from a larger initial cohort was followed during the period of 1990-2012. A logistic regression model identified predictors of SRH. Adjusted relative risks for the long-term predictors of SRH were calculated. The rate of poor SRH among the survivors was 18.8%, fair 56.5%, and good/excellent 24.7%. In the fitted model, long-term risk factors of poor SRH included baseline body mass index, baseline multi-morbidity, number of experienced stressful life events, and perceived poor living standards during the post-earthquake decade, while participation in sports in the early 1990s was a protective factor. Short-term protective factors included socio-economic status score, social support, employment and dignity, while current household size was a risk factor for poor SRH. No association was found between earthquake exposure severity and SRH after 23 years. However, the identified predictors included a number of modifiable lifestyle, material and psychological factors. Thus, interventions targeting these factors could have a long-lasting impact on disaster victims' health status.

Urinary incontinence and psychological distress in community-dwelling older adults.

OBJECTIVES: To examine the relationship between urinary incontinence (UI) and psychological distress in older adults. We hypothesized that persons with UI associated with condition-specific functional loss would be most likely to report psychological distress. DESIGN: A population-based longitudinal survey. SETTING: Continuing participants in a study of community-dwelling adults who were initially living in East Baltimore in 1981. PARTICIPANTS: Persons aged 50 and older (n=781) at follow-up interviews conducted between 1993 and 1996 for whom complete data were available. MEASUREMENTS: Participants were classified as incontinent if they reported any uncontrolled urine loss within the 12 months preceding the 13-year follow-up interview. Condition-specific functional loss secondary to UI was further assessed based on a series of questions relating directly to participants' inability to engage in certain activities due to their UI. Psychological distress was assessed using the General Health Questionnaire (GHQ) at interviews in 1981 and at the 13-year follow-up. RESULTS: Persons with UI were more likely to experience psychological distress as measured by the GHQ than were persons without UI (unadjusted odds ratio (OR)=1.74, 95% confidence interval (CI)=1.13-2.68). Persons with condition-specific functional loss secondary to UI were substantially more likely to have psychological distress as measured by the GHQ than were persons without UI (unadjusted OR=4.02, 95% CI=1.86-8.70). In multivariate models that controlled for potentially influential characteristics such as age, gender, ethnicity, and chronic medical conditions the association between condition-specific functional loss secondary to UI and psychological distress remained statistically significant. Among people with UI, persons with persistently elevated GHQ scores were much more likely to report condition-specific functional impairment from UI (adjusted OR=6.55, 95% CI=1.94-22.12). CONCLUSION: Individuals with UI, especially when incontinence was associated with condition-specific functional loss, were more likely to have psychological distress than were other older adults. Our findings support a general conceptual model that condition-specific functional impairment mediates the relationship between a chronic medical condition and psychological distress.

Incident hypertension associated with depression in the Baltimore Epidemiologic Catchment area follow-up study.

BACKGROUND: This study evaluates the role of depression as a specific risk factor for hypertension. METHODS: This study analyzed the prospective data in the Baltimore Epidemiologic Catchment Area (ECA) Follow-up Study (n=1920), a longitudinal population-based study of mental illness in East Baltimore. Incident cases of hypertension as assessed by self-report (n=148) in 1993 were compared to the remaining cohort without hypertension (n=901) across three waves of ECA interviews (1981, 1982, 1993). Depression and related symptoms were measured at baseline (1981) by the Diagnostic Interview Schedule (DIS) and categorized as dysphoria, dysthymia, or major depressive episode (MDE) according to Diagnostic and Statistical Manual (DSM) III criteria. RESULTS: Individuals with a major depressive episode compared to those who reported never having dysphoria had a marginally significant increased risk for hypertension (Odds Ratio (OR)=2.16; 95% Confidence Interval (CI) (0.94,4.98)) after adjustment for age, gender, race, body mass index, Nam-Powers socioeconomic score, alcohol usage, smoking, exercise, diabetes status, and number of general medical visits. MDE reported to have begun more than a year before the baseline measurement was associated with an increased risk for incident hypertension (Adjusted OR=3.67, 95% CI (1.25,10.79). LIMITATIONS: Potential misclassification of self-reported hypertension outcome. CONCLUSIONS: Even though the data are based on self-report of hypertension, these findings suggest that depression may be an independent risk factor for hypertension particularly for those with recurrent episodes or a long term history of the disease.

Predictors of cause-specific hospital readmission in patients with heart failure.

BACKGROUND: Repeated hospital readmissions are frequent and increasing over time in patients with heart failure (HF). The predictors for readmission in patients with HF are not completely understood. HYPOTHESIS: The study was undertaken to investigate the time course of readmission by specific cause in patients with HF, and to examine the independent effects of HF etiology and left ventricular (LV) function on cause-specific readmissions. METHODS: A retrospective cohort of 493 consecutive patients with HF was followed for readmission for 16.5 +/- 12.3 months. Ischemic etiology of HF was defined as history of myocardial infarction (MI), coronary artery bypass graft (CABG), percutaneous transluminal coronary angioplasty (PTCA), or > or = 70% coronary stenosis. Left ventricular function was assessed echocardiographically. Cause-specific readmissions were classified as HF, cardiovascular disease (CVD) other than HF, and other non-CVD. RESULTS: The annual readmission rate was 56.6%. Median time to readmission was 91 days, with 18.3% patients readmitted within 1 month after discharge. Ischemic etiology independently predicted all-cause readmission: Cox hazard ratio (95% confidence interval): 1.40 (1.11-1.79). This relationship was significant in women (1.83 [1.31-2.55]), but not in men (1.15 [0.82-1.62]), while readmissions were equally frequent in both genders. Similarly, ischemic etiology significantly predicted readmission for CVD in women (4.18 [2.14-8.19]), but not in men (1.49 [0.83-2.67]). However, LV dysfunction independently predicted readmission for recurrent HF (2.44 [1.46-4.08]), while ischemic etiology was not predictive in either gender. CONCLUSIONS: Readmissions for recurrent HF comprise only one-third of total hospital readmissions in patients with HF. Ischemic etiology is a significant predictor of readmission, and most of this effect is mediated through a four-fold increased risk of readmission for CVD other than HF in women. Readmission for recurrent HF is predicted by LV dysfunction but not by ischemic etiology. Patients with HF can be accurately risk stratified for cause-specific readmission with available clinical data.