The primary aim of this project is to test the efficacy of an inpatient congestive heart failure (CHF) educational intervention compared with usual care among inpatients at Griffin Hospital, who are largely drawn from the population of the Naugatuck Valley in Connecticut.

The educational intervention will utilize:

• written educational materials suitable for patients with low health literacy - alternatives to written materials (e.g., video- and audiotapes) that may more effectively communicate health information to elderly patients and those with low health literacy
• a one-on-one educational session with a nurse patient educator. The educational session will use as its framework guidelines provided by the America Medical Association (AMA) to improve communication between healthcare providers and low health literacy patients. The investigators hypothesize that CHF patients who receive this educational intervention will have fewer hospital readmissions or deaths than the usual care group. The investigators further hypothesize that patients with low health literacy will derive more benefit from the intervention than patients with higher literacy.

The secondary aims of the project are to:

• assess whether patients in the education and usual care groups differ on post-discharge CHF knowledge and on satisfaction with hospital care. Compared with usual care, the investigators hypothesize that CHF patients who receive the educational intervention will have better knowledge of CHF and will be more satisfied with the care they received in the hospital.

The potential impact of the proposed project may be to increase disease knowledge and health literacy, and improve adherence to CHF treatments. This, in turn, may contribute to improved medical outcomes and reduced hospital readmissions for CHF patients. In addition, if this preliminary study provides evidence of a promising educational intervention suitable for patients with low health literacy, th investigators will endeavor to test the intervention in ethnically diverse populations throughout Connecticut.

Variables and Measures

We will assess health literacy with the Short Test of Functional Health Literacy in Adults (STOFHLA). The STOFHLA is a brief version of the widely used instrument, the Test of Functional Health Literacy in Adults (TOFHLA) that measures patients' ability to comprehend situations often encountered in the health care setting; e.g., reading registration forms and instructions for diagnostic tests. It is a 36-item reading comprehension test written in 14-point font that requires 7 minutes to complete.

Primary Outcome The primary outcome is readmission or death from any cause within 90 days of discharge. This outcome will be ascertained by: 1) reviewing electronic and written medical records to identify patients re-admitted to Griffin Hospital; 2) monitoring obituaries; and 3) conducting three-month post-discharge follow-up phone calls to determine if the patient was admitted elsewhere since discharge from Griffin.

Secondary outcomes

• Change in knowledge of CHF from study enrollment to two weeks post-discharge will be measured by the Knowledge of CHF Questionnaire.
• Patient satisfaction with hospital care will be examined two weeks post-discharge with the 27-item Hospital Consumer Assessment of Health Plans Survey (HCAHPS) Telephone Survey Instrument.
• Feedback on educational materials will be obtained in narrative from patients.

Other variables of interest Through review of electronic and written medical records, we will ascertain subject characteristics including demographics and socioeconomic factors and disease-related variables

Procedures Following IRB review and approval, Dr. Dorothea Wild, a study co-investigator, will review and screen new admissions to identify potentially eligible patients. Within 24 hours of admission, she will contact the attending of record to request permission to enroll the patient in the study; if permission is granted, Dr. Wild will approach the patient to describe the study and obtain consent to participate.

Obtaining consent Dr. Wild will obtain written informed consent from each patient who agrees to participate. She will inform patients that they will: 1) receive either the standard educational material provided by the Hospital for CHF patients or a more intensive educational intervention developed for this study; 2) have a 50% chance of being assigned to each group; and 3) not know whether they are receiving standard or intensive education.

Randomization, allocation concealment, and blinding Dr. Haq Nawaz will use software developed by the Johns Hopkins Division of Oncology Biostatistics to randomize patients to the usual care and educational intervention groups. Dr. Nawaz will keep group assignments in consecutively numbered opaque envelops and will reveal the assignment only to the nurse patient educator when the intervention is assigned. All other study personnel who collect and manage data will remain blind to study assignment. Patients will also be blind to group assignment throughout the trial.

Each patient in the group receiving the educational intervention will:

• view the HFSA videotape;
• receive a copy of "Managing Your Health with Heart Failure" to review in hospital and to take home;
• receive pre-printed charts on which to record daily weigh, evidence of swelling, and whether or not s/he took prescribed diuretics, with instructions on when to inform the doctor about weight or swelling problems;
• receive a copy of "Following a Low-Salt Diet" to review and take home;
• participate in a 45 - 60 minute one-on-one session with a nurse patient educator that reviews the above information using as a framework the 6-step strategy recommended by the AMA; and
• be encouraged to take written/pictorial notes to augment the above materials.

The nurse educator will elicit feedback from patients on the quality and accessibility of the written materials.

Each patient in the educational intervention and usual care groups will receive and review with a staff nurse:

• written information on diagnosis and treatment of CHF that is not designed for low literacy patients; and
• a pre-printed discharge instruction sheet that explains the patient's medications and dietary instructions, identifies symptoms that should be reported to the primary care physicians, and provides a physician follow-up appointment.

Post-discharge assessments The data manager will administer the Knowledge of CHF Questionnaire and the measure of patient satisfaction by telephone two weeks post-discharge. She will also regularly review obituaries and place a follow-up phone call three months post-discharge to ask about any subsequent hospitalizations, including to facilities other than Griffin.

Data safety and monitoring Because the proposed study involves an educational intervention, potential risks of participation are minimal. Nonetheless, the PI, Dr. Calvocoressi, and one co-investigator, Dr. Nawaz, will conduct monthly data and safety reviews to assess risk and monitor adverse events. An additional review will be conducted following the report of any adverse event. We will also assess whether the event can be attributed to the research (i.e., definite, possible, unrelated). Any adverse events graded 2 or higher will immediately be reported to the Griffin Hospital IRB and to the funding agency.

Data Analysis Plan The analysis plan will include univariate, bivariate, and multivariate analyses conducted with SAS software. We will conduct all analyses according to the intent-to-treat principle and will use a 0.05 (two-tailed) level of significance.

Assessing subject characteristics and the adequacy of the randomization procedure We will examine whether the control and intervention groups significantly differ on subject characteristics (e.g., socio-demographic factors, features of the disease, involvement of caregivers) that may impact the relationship between treatment group and study outcomes. For categorical variables, we will use the Chi square test for this purpose; for continuous variables, we will use the Student's t-test.

Primary outcome We will use the t-test for proportions to examine the difference in the proportions of patients in the educational and usual care groups who are re-hospitalized or die within 90 days of discharge. In addition, we will employ logistic regression to examine whether the educational and usual care groups differ on re-hospitalization/death (yes/no) at 90 days adjusted for pertinent covariates. We will conduct a secondary analysis to test the robustness of these findings. That analysis will examine whether time to re-hospitalization/death differs between the educational and usual care groups using the log rank test and the proportional hazard (Cox) model.

We will examine the interaction between treatment group (educational versus usual care) and health literacy (i.e., low versus adequate) in relation to readmission/death at 90 days post-discharge to assess whether the impact of the educational intervention differed by health literacy level.

Secondary outcomes We will examine the difference between the educational and usual care groups in knowledge of CHF at study enrollment and following discharge by calculating a continuous change score based on the two administrations of the knowledge of CHF questionnaire. We will use the paired t-test for the bivariate analysis. To control for covariates, including knowledge score at the time of enrollment, we will conduct a linear regression analysis.

The difference between the educational and usual care groups in satisfaction with treatment, measured post-discharge on a continuous scale, will be examined using the t-test and linear regression to adjust for covariates.

Power and Sample Size Calculations We have based these calculations on our primary outcome, the proportions of CHF patients in the educational and usual care groups who either die or are re-admitted to the hospital for any cause within 90 days of discharge.

A review of the Griffin Hospital database over the past three years indicates that 33.5% of CHF patients meeting our inclusion criteria were re-admitted to Griffin Hospital within 90 days. Though the patient population is quite stable and most CHF patients needing readmission probably return to this facility, we assume that an additional 5% were admitted elsewhere during the 90 days following discharge from Griffin Hospital. In addition, based on previously conducted studies of educational interventions in CHF (19), we expect that an additional 6% will die within this period. Thus, we base our calculations on 45% of the usual care group experiencing all-cause readmission or death within 90 days of discharge. For the educational group we based our calculations on a 25% absolute decrease in these events during the same time period (i.e., 20% will be re-admitted or die). This is a clinically meaningful decrease and is consistent with the difference in readmissions/deaths between control and educational intervention groups in a recent CHF outpatient study.

We used the following formula for our calculations:

n = {[p1(q1) + p2(q2)]/ (p2-p1)}(f(alpha,beta))

where n = number of subjects needed in each group p1 = percentage of patients in the control group who were not re-hospitalized q1 = percentage of patients in the control group who were re-hospitalized p2 = percentage of patients in the intervention group who were not re-hospitalized q2 = percentage of patients in the control group who were re-hospitalized f(alpha,beta) = 7.9 when alpha = .05 (two-sided) and beta = .20

The number of patients per group needed to detect a 25% decrease in 90 day all-cause readmission/death is 52; the total sample needed for this study is 104. Assuming that 10% of patients will refuse to participate or will be lost to follow-up, we will need to enroll 116 eligible patients.

Over the past three years, there were, on average 174 patients admitted to Griffin annually with a primary diagnosis of CHF. Seventy-three of these patients per annum would have been excluded from participation (e.g., discharged to a skilled nursing facility, too ill to participate), leaving an average of 101 patients per year meeting inclusion criteria. We should, therefore, be able to recruit the 116 subjects need for this study in approximately 14 months.

Inclusion Criteria:

• Eligible consecutive patients aged 50 years and over admitted with a primary diagnosis of CHF at Griffin Hospital will be invited to participate regardless of age, gender, race, or education level.
• Because dietary and medication non-compliance affect patients with both systolic and diastolic heart failure, the investigators will include patients with both of these conditions, regardless of their ejection fraction.

Exclusion Criteria:

• The investigators will exclude patients with clinical conditions and communication barriers that would limit their ability to participate in and/or benefit from this educational intervention.
• The investigators will also exclude patients whose planned discharge is to another hospital or to a structured setting in which medical personnel are responsible for their care (e.g., a skilled nursing facility), thus limiting their ability to implement a largely self-directed self-care regimen upon leaving the hospital.
• In addition, the investigators will exclude any patient who does not have a telephone and cannot, therefore, be contacted to obtain post-discharge follow-up data.

Specific exclusion criteria include:

• A diagnosis of dementia or other severe mental disorder (e.g., acute delirium, psychosis)
• Clinical instability or need for transfer to another hospital for acute intervention (e.g., experiencing cardiogenic shock, or needing valve surgery or acute coronary intervention)
• Terminal illness or intubation
• Moderate to severe uncorrected vision or hearing problems
• Inability to speak English or to provide informed consent
• Lack of access to a telephone
• Planned discharge to a structured facility (e.g., skilled nursing facility, intermediate care facility)

18 months