Depression and Clinical Outcomes in Heart Failure: An OPTIMIZE-HF Analysis

Article Outline

• Abstract
• Methods
  • Patient Eligibility
  • Statistical Analysis
• Results
  • In-hospital Treatments
    • Procedures
    • Discharge Instructions
    • Medications
  • Clinical Outcomes
• Discussion
  • Limitations
• References
• Copyright

Abstract 

Background

Depression is a risk factor of excessive morbidity and mortality in heart failure. We examined in-hospital treatment and postdischarge outcomes in hospitalized heart failure patients with a documented history of depression from the Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients with Heart Failure.

Methods

We identified patient factors associated with depression history and evaluated the association of depression with hospital treatments and mortality, and early postdischarge mortality, emergency care, and rehospitalization.

Results

In 48,612 patients from 259 hospitals, depression history was present in 10.6% and occurred more often in females, whites, and those with common heart failure comorbidities, including chronic pulmonary obstructive disease (36% vs 27%), anemia (27% vs 16.5%), insulin-dependent diabetes mellitus (20% vs 16%), and hyperlipidemia (38% vs 31%), all P <.001. Patients with depression history were less likely to receive coronary interventions and cardiac devices, all P <.01; or be referred to outpatient disease management programs, P <.001. Length of hospital stay was longer with depression history (7.0 vs 6.4 days, P <.001). In 5791 patients followed-up at 60-90 days postdischarge, those with depression history had higher mortality (8.8% vs 6.4%; P=.025). After multivariable modeling, depression history remained a predictor of length of hospital stay, P <.001 and postdischarge mortality, P=.02.

Conclusions

Depression history at heart failure hospitalization may be a predictor of prolonged length of hospital stay, less use of cardiac procedures and postdischarge disease management, and increased 60-90 day mortality. Patients with depression might represent a vulnerable group in which improved use of evidence-based treatment should be considered.

Keywords: Early clinical outcomes, Length of stay, Mortality, Rehospitalization, Treatment

Symptoms of depression are prominent in elderly hospitalized patients with heart failure.¹, ² Depressive symptoms on hospital admission were associated with decrease in functional status at 6 months,³ death at 3 months,⁴ 6 months,³ and 1 year;⁴ and hospital readmission at 3 months and 1 year.⁴ In a longitudinal design, and using regression analyses, physical symptoms were more closely related to depression than limitations in physical functioning.⁵ In patients with heart failure, predictors of development of depressive symptoms 1 year after initial assessment were living alone, alcohol abuse, believing medical care to be an economic burden, and worse quality of life.⁶ Other correlates of depression in elderly, hospitalized heart failure patients were poorer perceived emotional-informational support, higher fatigue, poorer perception of health and not living with family, explaining 49% of the total variance for psychological distress.⁷

In a meta-analysis on depression in heart failure, depression measures varied and included clinical interviews using many structured formats, unique depression symptom inventories, and medical record documentation of depression or antidepressant medication use.⁸ Hospitalization, care costs, and clinical events (6 months to many years after hospitalization) were studied;⁸ however, in-hospital treatment and early outcomes were not well described. Currently, little is known about the effect that depression presence has on characteristics, treatment, and early (60-90 day) outcomes in a nationally representative cohort of patients hospitalized for heart failure.

The Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients with Heart Failure (OPTIMIZE-HF) is a heart failure registry and performance-improvement program for patients hospitalized with heart failure exacerbation.⁸ The objectives of this analysis of OPTIMIZE-HF data are to assess prevalence of depression (symptomatic or syndromal) based on medical record documentation, examine the influence of depression on patient characteristics and quality of core heart failure medical care, and improve understanding of the impact of depression on in-hospital treatment and early outcomes.

Back to Article Outline

Methods 

Patient Eligibility 

OPTIMIZE-HF, a comprehensive hospital-based registry and process-of-care improvement program, has been previously described.⁹, ¹⁰, ¹¹ Patients were eligible for enrollment if they were hospitalized for worsening heart failure as the primary cause of admission or had significant heart failure symptoms that developed during hospitalization when heart failure was not the initial reason for admission. Patients had systolic dysfunction (ejection fraction <40%), heart failure symptoms in the setting of preserved systolic function (diastolic heart failure), or heart failure without documentation of ventricular function.

The OPTIMIZE-HF Registry gathered data from March 1, 2003 to December 31, 2004 on patient characteristics, hospital treatments, medication prescriptions, consultations for postdischarge services, and outcomes using a web-based information system. Data were obtained from all United States regions, and institutions represented community hospitals to large tertiary medical centers.⁹, ¹⁰, ¹¹ Automated electronic data checks were used to prevent out-of-range entry or duplicate patients. The registry coordinating center was Outcome Sciences, Inc. (Cambridge, Mass). History of depression was assessed by a review of patients' medical records. Analysis included 48,612 patients at 259 hospitals. A prespecified subgroup (10%) of OPTIMIZE-HF patients had 60-90 day follow-up data, representing 5791 patients and 91 hospitals. The follow-up cohort was demographically similar to patients in the overall registry.⁹, ¹⁰, ¹¹ Each participating center's institutional review board or a central institutional review board approved the protocol, and written informed consent was obtained before enrollment of patients who participated in follow-up data collection.

Statistical Analysis 

Continuous and categorical variables were described by depression status; significance was determined using Pearson chi-squared test and Wilcoxon rank-sum tests, as appropriate. Three regression modeling methodologies were used in the analyses. Generalized Estimating Equations (GEE) with binary family, link logit, and exchangeable correlation structure were used for all regression modeling, except for postdischarge mortality and when testing depression as predictor of length of stay. This analysis methodology accounted for potential intra-group correlation due to clustering within hospital site. The independent effect of depression on postdischarge mortality was assessed by Cox regression analysis. The potential clustering effect within hospital sites was accounted for by means of random effects (shared frailty). The proportionality assumption for the hazard function over time was tested by mean of the Schoenfeld residuals. Zero-truncated negative binomial regression was used when length of stay was modeled as an outcome, and association with covariates was reported as incidence rates ratios. Patients transferred to other institutions at discharge were excluded from the length-of-stay variable calculation.

Candidate covariates for all multivariable analyses were selected based on previous medical knowledge or on statistical significance (P=.15) in univariate analyses, then a parsimonious, highly predictive model was derived using a backward step-down selection.¹² The functional form of continuous variables was examined by means of fractional polynomials and transformed when appropriate.¹³ Predictive ability of the regression models was assessed by estimating the area under the receiver operating characteristic curve (GEE models) or Harrell's C-statistics (Cox model).¹⁴ A 2-sided P-value of <.05 was considered statistically significant for all analyses. All analyses were performed using Stata 10 (StataCorp LP, College Station, Tex).

Back to Article Outline

Results 

Of 48,612 patients hospitalized for heart failure, 10.6% had depression recorded in the medical record (depression was noted in 13.6% of 5719 patients in the follow-up cohort). The overall population was elderly, with a mean age of 73.2 years. Patients with history of depression were more often female, white, and those with common heart failure comorbidities. Depressed patients were more likely to have a medical history of coronary artery disease and myocardial infarction and less likely to have a treatment history of coronary artery bypass graft surgery or percutaneous coronary intervention (Table 1). Using GEE multivariable analyses, predictors of depression are presented in Figure 1.

Table 1. Patient Characteristics at Hospital Admission, N=48,612

Variable, % or Mean (SD)	History of Depression (n=5164)	No History of Depression (n=43,448)	⁎P Value
Age, years	72.74(14)	73.25(14)	.239
Women	59.7	50.6	<.001
Caucasian	82.3	73.2	<.001
LVSD (ejection fraction <40%)	46.6	52.4	<.001
Ejection fraction	40.9(17.9)	38.7(17.6)	<.001
Ischemic etiology	45.2	45.8	1.00
Acute renal failure	4.1	2.9	<.001
Anemia	26.9	16.5	<.001
Coronary artery disease	53.0	49.1	<.001
Cerebral vascular accident/TIA	20.4	15.0	<.001
Insulin dependent diabetes mellitus	20.4	16.2	<.001
Hyperlipidemia	38.0	31.4	<.001
Hypertension	73.5	70.6	<.001
Chronic obstructive pulmonary disease	36.0	26.6	<.001
Pulmonary hypertension	9.3	5.9	<.001
Peripheral vascular disease	18.3	13.1	<.001
Thyroid abnormality	20.6	14.4	<.001
History of CABG or PCI	27.0	28.4	.778
History of myocardial infarction	24.1	22.3	.085

CABG=coronary artery bypass graft surgery; LVSD=left ventricular systolic dysfunction; PCI=percutaneous coronary intervention; TIA=transient ischemic attack.

• 
  • View Large Image
  • Download to PowerPoint

• Figure 1. 

  Predictors of depression using Generalized Estimating Equations regression – exchangeable correlation. Admission systolic blood pressure, admission hemoglobin, and admission serum creatinine are not shown on this plot.

In-hospital Treatments 

There were no baseline differences between depressed and nondepressed heart failure patients in primary cause of admission. Patients with documented history of depression were more likely not to receive any procedures while hospitalized. About cardiac procedures, patients with a history of depression were less likely to receive cardiac angiography, percutaneous coronary intervention or coronary artery bypass graft surgery, electrophysiology study, implantable cardioverter-defibrillator placement, and cardiac resynchronization therapy (Figure 2). Additionally, patients with history of depression were equally likely to receive a cardiac rehabilitation referral (P=.49) but less likely to receive an order for outpatient heart failure disease management program referral; 12.6% vs 14.9%; P <.001.

• 
  • View Large Image
  • Download to PowerPoint

• Figure 2. 

  Percent of patients who received procedures while hospitalized by history of depression status. Striped bars=depression; solid bars=no depression.

While hospitalized, patients with history of depression were less likely to receive core heart failure education instructions, defined as documentation of instructions in the medical record, for activity (nondepressed, 87.5% vs depressed, 86.4%; P=.03), follow-up instructions (nondepressed, 93.4% vs depressed, 92.0%; P=.001), and what to do when symptoms worsen (nondepressed, 69.9% vs depressed, 68.1%; P=.02), but not for medications, diet, and daily weight monitoring. Although these differences were statistically significant, given the small percentage change, they are unlikely to be clinically meaningful.

There was no difference at hospital admission in use of beta-blocker, digoxin, aspirin, lipid-lowering agent (including statin use), or warfarin by history of depression. In the follow-up cohort, there was no difference in beta-blocker use in eligible patients with left ventricular systolic dysfunction, based on history of depression (84.8%) vs no history of depression (83.9%), P=.71. Likewise, there was no difference in angiotensin-converting enzyme inhibitor use in eligible patients, based on history of depression (63.4%) vs no history of depression (67.0%), P=.09.

Clinical Outcomes 

Hospital length of stay was significantly longer in patients with a medical history of depression (Table 2). However, as noted above, it is unlikely that these small differences are clinically meaningful. After multivariable modeling using the Huber/White/sandwich estimator of variance to adjust for potential cluster effect due to hospital site grouping, history of depression remained a significant predictor of hospital length of stay (incidence rate ratio 1.08; 95% confidence interval [CI], 1.04-1.11; P <.001); only medical history of acute renal failure and chronic kidney dysfunction had higher incidence rates ratios for longer hospital length of stay (incidence rates ratios 1.29 and 1.09, respectively).

Table 2. In-Hospital and 60-90 Day Postdischarge Outcomes in Hospitalized Heart Failure Patients by History of Depression Status

Variable	History of Depression (n=5164)	No History of Depression (n=43,448)	Unadjusted			Adjusted
			Regression Estimates	P Value	AUC	Regression Estimates	P Value	AUC
In hospital
All-cause mortality, %	4.20	3.7	1.11(0.95-1.29)a	.197	0.5068	1.10(0.90,1.34)a	.351	.7710
Length of stay, daysd	6.16	5.74	1.09(1.05-1.12)c	<.001	—	1.08(1.04,1.11)c	<.001	—
Postdischarge, %	(n=786)	(n=5005)
All-cause mortality	8.8	6.4	1.36(1.04-1.79)b	.027	0.5214	1.46(1.05,2.03)b	.025	.740
Mortality or rehospitalization	37.1	34.6	1.15(0.98-1.35)a	.098	0.5065	1.21(0.99,1.47)a	.064	.6492
Emergency department visits without hospitalization	15	12.2	1.15(0.93-1.43)a	.205	0.5148	1.15(0.92,1.43)a	.235	.5964
Hospitalization	29.9	29.6	1.08(0.91-1.27)a	.377	0.5009	1.09(0.90,1.33)a	.386	.6213
Cardiovascular hospitalization	22.5	21.9	1.06(0.88-1.27)a	.56	0.5021	1.01(0.83,1.25)a	.89	.6262

The Generalized Estimating Equations (GEE) multivariable logistic model for in-hospital mortality included the following variables: age, race; prior history of: depression, smoking, acute renal failure, cerebrovascular accident, dialysis, hyperlipidemia, liver disease, chronic obstructive pulmonary disease, and peripheral vascular disease; admission medications: beta-blocker, angiotensin converting enzyme (ACE) inhibitor, aldosterone antagonists, and lipid-lowering agent; admission vital signs: weight, heart rate, systolic blood pressure, and diastolic blood pressure; admission laboratory: sodium, hemoglobin and creatinine; no known heart failure before this admission, and left ventricular systolic dysfunction (LVSD) status.

The multivariable Cox regression model for postdischarge mortality included the following variables: age, race; prior history of: depression, ischemic heart disease, hypertension, liver disease, and diabetes; procedures: any revascularization procedure (either coronary artery bypass graft surgery or percutaneous coronary intervention), and mechanical ventilation; discharge medications: ACE inhibitor, aldosterone antagonists, digoxin, and lipid-lowering agent; discharge vital signs: heart rate, systolic blood pressure, diastolic blood pressure; admission laboratory: serum sodium; discharge laboratory: serum creatinine.

The GEE multivariable logistic model for postdischarge mortality or rehospitalization included the following variables: length of stay; prior history of: depression, chronic obstructive pulmonary disease, and diabetes; procedures: coronary angiography, mechanical ventilation, and cardiac resynchronization therapy (CRT); discharge medications: ACE inhibitor, angiotensin receptor blocker, hydralazine, and lipid-lowering agent; discharge vital signs: heart rate, and systolic blood pressure; admission laboratory: serum sodium and hemoglobin; discharge laboratory: serum creatinine; no known heart failure before this admission, LVSD status and ischemic etiology.

The GEE multivariable logistic model for postdischarge rehospitalization included the following variables: length of stay; prior history of: depression, implantable cardioverter-defibrillator (ICD), atrial arrhythmia, chronic obstructive pulmonary disease, and renal insufficiency; procedures: mechanical ventilation, and CRT; discharge medications: hydralazine, and nitrates; admission laboratory: hemoglobin; discharge laboratory: serum creatinine; no known heart failure before this admission, and LVSD status.

The GEE multivariable logistic model for postdischarge emergency department visits without hospitalization included the following variables: length of stay; prior history of: depression, acute renal failure, cerebrovascular accident, and chronic obstructive pulmonary disease; procedures: mechanical ventilation; admission laboratory: hemoglobin; ischemic etiology.

The GEE multivariable logistic model for postdischarge cardiovascular hospitalization included the following variables: age, length of stay; prior history of: depression, acute renal failure, cerebrovascular accident, chronic renal insufficiency (serum creatinine >2.0 mg/dL), ICD, pulmonary reactive airway disease, smoking in the past year, and chronic obstructive pulmonary disease; procedures: mechanical ventilation, and CRT; discharge medications: nitrates; admission laboratory: hemoglobin; LVSD status.

There was no difference in in-hospital mortality based on history of depression; (odds ratio [OR] 1.10, 95% CI, 0.90-1.34; P=.351). Postdischarge mortality (OR 1.46, 95% CI, 1.05-2.03; P=.025; Figure 3, Table 2) was increased in patients hospitalized with heart failure and with history of depression, but other 60-90 day postdischarge outcomes did not differ by depression status (Table 2).

• 
  • View Large Image
  • Download to PowerPoint

• Figure 3. 

  Kaplan-Meier curve for postdischarge mortality.

Back to Article Outline

Discussion 

Among a large, representative population of patients admitted to the hospital for heart failure decompensation, OPTIMIZE-HF demonstrated that patients with a medical record history of depression are more likely to be women, white, and have multiple comorbidities, and less likely to receive cardiac procedures during hospitalization, be referred to an outpatient heart failure disease management program, and receive some components of heart failure education as recommended by the Joint Commission and consensus-developed chronic heart failure guidelines. There was a significant increase in hospital length of stay and postdischarge 60-90 day all-cause mortality in patients with depression history.

In OPTIMIZE-HF, 10.6% of patients had a documented medical history of depression. In a meta-analysis of depression prevalence in patients with heart failure, the range was 9% to 60%.⁸ Of 8 studies in hospitalized patients using medical record diagnosis or other formal assessment, depression frequency was 16%. In contrast, 7 studies that used patient questionnaires and allowed more liberal depression definitions reported a mean depression rate of 38%.⁸ Despite lower prevalence of depression using medical record documentation, our results were similar to prior studies in that women and whites were more likely to have a history of depression compared with men and minority groups,⁸ and heart failure medications were not used based on depression status.¹, ¹⁵

No previous investigations examined depression and in-hospital procedures, referral for postdischarge programs, or discharge education. Our results support that depression history might be a predictor for disparate hospital or posthospital heart failure care, even though national consensus guidelines for heart failure¹⁶, ¹⁷ do not distinguish care recommendations by depression history. The degree that medical comorbidities played a role in decisions related to cardiac procedures is unknown. The higher rate of comorbidities in depressed patients could have increased procedural risk and limited treatment options. Heart failure treatment guidelines, however, do not exclude patients with comorbidities or poor clinical prognosis from receiving evidence-based medical care; the exception being the contraindication for use of an implantable cardioverter-defibrillator in patients with a life expectancy of <1 year.¹⁷ Prospective studies are needed that focus on the effects of depression on clinical outcomes and uniformity of hospital services.

Hospital length of stay is an important aspect of the economic burden of heart failure. In our study, length of stay was slightly longer in patients with depression history. Koenig found a nonsignificantly lower length of stay in nondepressed patients (11.3 days) compared with those with minor or major depression (13.4 days and 13.0 days, respectively).² It is unknown if identification of depression symptoms early after hospital admission would prompt pharmacologic and cardiac treatment strategies that could reduce length of stay to that of nondepressed patients.

Only 2 studies on depression in hospitalized patients with heart failure included reports of early follow-up, both at 12 weeks after discharge. Similar to our findings, readmission rates were comparable in depressed and nondepressed patients,², ⁴ even after multivariable analyses.⁴ Whereas depression was associated with excess mid (3-9 months) and late (1 year) rehospitalization after discharge,², ⁴ early rehospitalization for decompensated heart failure is most likely due to other factors. Of note, in OPTIMIZE-HF, 60-90 day emergency care without hospitalization also was not affected by depression history. Although depression may be a risk factor for nonadherence to medical treatment¹⁸ and poorer physical functioning,⁵ our results on re-hospitalization and postdischarge emergency care provide further evidence that depression is probably not the factor prompting cardiac instability.

In OPTIMIZE-HF, in-hospital mortality was not associated with history of depression (P=.77). These results differed from in-hospital mortality reported in one small single-site subanalysis of ADHERE (Acute Decompensated Heart Failure Registry) data,¹⁹ in which researchers found significantly more in-hospital deaths in patients with history of depression (n=34) than those without (n=137; 17.7% vs 6.6%, P <.05). However, on multivariable analysis, only the combined endpoint, in-hospital mortality or cardiopulmonary resuscitation, was associated with history of depression.¹⁹ With a substantially larger cohort, our findings make it unlikely that depression contributes substantially to in-hospital mortality.

OPTIMIZE-HF investigators also found a significant association between history of depression and 60-90 day mortality. Of cohort studies included in a mortality meta-analysis,⁸ mortality follow-up time ranged from 6 months to more than 4 years after hospitalization, with 1 exception. In a single-center hospital study of 374 patients, 3-month mortality in depressed patients was nonsignificantly higher than in nondepressed patients.⁴ However, when mortality was assessed by Beck Depression Inventory scores, depressed patients had a significantly higher mortality (11.9% vs 5.7%, OR 2.26; 95% CI, 1.04-4.91; P=.04) than nondepressed patients.⁴ Research is needed to learn the optimal intensity of depression assessment and treatment required to decrease postdischarge mortality after a heart failure hospitalization. The hospital experience provides an opportunity for interdisciplinary collaboration that might benefit these vulnerable patients. The randomized, controlled multicenter MOOD-HF (MOrbidity, mOrtality and mood in Depressed Heart Failure patients) trial investigates use of a serotonin re-uptake inhibitor in depressed patients.²⁰ Study results will provide evidence-based recommendations for managing depression with drug therapy.

Limitations 

This study was not a prospective randomized trial, thus, interpretations of these data are limited; definitive cause-and-effect relationships cannot be established. Depression was ascertained by medical record review alone, which may have affected the reported depression prevalence. This method of data collection also did not allow for the distinction of depression as a symptom versus a syndrome. Information about depression severity, current depression status, and depression medications were not captured. Consequently, actual depression rate at hospitalization may have been higher than reported. Heart failure medication use was as reported by patients and as documented in medical records; adherence rates were not studied. Effect of depression on clinical outcomes may have been influenced by unmeasured confounders and variables that were measured but not documented or dropped from consideration in the model due to missing variables. All-cause, but not cardiovascular mortality, was captured in OPTIMIZE-HF. Findings reported in this study may not apply to hospitals that differ in patient characteristics or care patterns from OPTIMIZE-HF hospitals.

In conclusion, OPTIMIZE-HF reveals that patients with heart failure and depression history differ in a number of important characteristics from nondepressed patients and suggests that depressed patients are less likely to receive cardiac procedures and heart failure disease management program care than their nondepressed counterparts. Depression may be a predictor of longer hospital stay and early postdischarge mortality, but not in-hospital mortality or early rehospitalization. Patients with depression may represent a vulnerable group in which improved use of evidence-based treatment should be considered.

Back to Article Outline

References 

1. Lesman-Leegte I, Jaarsma T, Sanderman R, et al. Depressive symptoms are prominent among elderly hospitalised heart failure patients. Eur J Heart Fail. 2006;8:634–640
   • View In Article
   • MEDLINE
   • CrossRef
2. Koenig HG. Depression in hospitalized older patients with congestive heart failure. Gen Hosp Psychiatry. 1998;20:29–43
   • View In Article
   • Abstract
   • Full-Text PDF (1689 KB)
   • CrossRef
3. Vaccarino V, Kasl SV, Abramson J, Krumholz HM. Depressive symptoms and risk of functional decline and death in patients with heart failure. J Am Coll Cardiol. 2001;38:199–205
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (156 KB)
   • CrossRef
4. Jiang W, Alexander J, Christopher E, et al. Relationship of depression to increased risk of mortality and rehospitalization in patients with congestive heart failure. Arch Intern Med. 2001;161:1849–1856
   • View In Article
   • MEDLINE
   • CrossRef
5. Friedman MM, Griffin JA. Relationship of physical symptoms and physical functioning to depression in patients with heart failure. Heart Lung. 2001;30:98–104
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (99 KB)
   • CrossRef
6. Havranek EP, Spertus JA, Masoudi FA, et al. Predictors of the onset of depressive symptoms in patients with heart failure. J Am Coll Cardiol. 2004;44:2333–2338
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (108 KB)
   • CrossRef
7. Yu DS, Lee DT, Woo J, Thompson DR. Correlates of psychological distress in elderly patients with congestive heart failure. J Psychosom Res. 2004;57:573–581
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (134 KB)
   • CrossRef
8. Rutledge T, Reis VA, Linke SE, et al. Depression in heart failure (A meta-analytic review of prevalence, interaction effects, and associations with clinical outcomes). J Am Coll Cardiol. 2006;48:1527–1537
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (210 KB)
   • CrossRef
9. Fonarow GC, Abraham WT, Albert NM, et al. Association between performance measures and clinical outcomes for patients hospitalized with heart failure. JAMA. 2007;297:61–70
   • View In Article
   • CrossRef
10. Fonarow GC, Abraham WT, Albert NM, et al. Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients with Heart Failure (OPTIMIZE-HF): rationale and design. Am Heart J. 2004;148:43–51
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (275 KB)
    • CrossRef
11. Fonarow GC, Abraham WT, Albert NM, et al. Influence of a performance-improvement initiative on quality of care for patients hospitalized with heart failure: results of the Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients with Heart Failure (OPTIMIZE-HF). Arch Intern Med. 2007;167:1493–1502
    • View In Article
    • CrossRef
12. Ambler G, Brady AR, Royston P. Simplifying a prognostic model: a simulation study based on clinical data. Stat Med. 2002;21(24):3803–3822
    • View In Article
    • MEDLINE
    • CrossRef
13. Royston P, Ambler G, Sauerbrei W. The use of fractional polynomials to model continuous risk variables in epidemiology. Int J Epidemiol. 1999;28(5):964–974
    • View In Article
    • MEDLINE
    • CrossRef
14. Harrell FE, Margolis PA, Gove S, et al. Development of a clinical prediction model for an ordinal outcome: the World Health Organization Multicentre Study of Clinical Signs and Etiological agents of Pneumonia, Sepsis and Meningitis in Young Infants. Stat Med. 1998;17(8):909–944
    • View In Article
    • MEDLINE
15. Rumsfeld JS, Jones PG, Whooley MA, et al. Depression predicts mortality and hospitalization in patients with myocardial infarction complicated by heart failure. Am Heart J. 2005;150:961–967
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (145 KB)
    • CrossRef
16. Adams KF, Lindenfeld J, Arnold JMO, et al. HFSA 2006 comprehensive heart failure practice guideline. J Card Fail. 2006;12:e1–e122
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (37 KB)
    • CrossRef
17. Hunt SA, Abraham WT, Chin MH, et al. ACC/AHA 2005 Guideline Update for the Diagnosis and Management of Chronic Heart Failure in the Adult: A Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Update the 2001 Guidelines for the Evaluation and Management of Heart Failure) (American College of Cardiology Website). http://www.acc.org/clinical/guidelines/failure/index.pdfAccessed August 18, 2005
    • View In Article
18. DiMatteo MR, Lepper HS, Croghan TW. Depression is a risk factor for noncompliance with medical treatment. Arch Intern Med. 2000;160:2101–2107
    • View In Article
    • MEDLINE
    • CrossRef
19. de Denus S, Spinler SA, Jessup M, Kao A. History of depression as a predictor of adverse outcomes in patients hospitalized for decompensated heart failure. Pharmacotherapy. 2004;24:1306–1310
    • View In Article
    • MEDLINE
    • CrossRef
20. Angermann CE, Gelbrich G, Störk S, et al. Rationale and design of a randomized, controlled, multicenter trial investigating the effects of selective serotonin re-uptake inhibition on morbidity, mortality, and mood in depressed heart failure patients (MOOD-HF). Eur J Heart Fail. 2007;9:1212–1222
    • View In Article
    • CrossRef