We conducted a secondary analysis to evaluate the relationships between levels of patient activation and heart failure self-management knowledge, self-efficacy, and self-management behaviors in heart failure patients discharged from rural critical access hospitals to home. The data for this paper formed the baseline data from a NIH-funded randomized controlled trial titled “Patient AcTivated Care at Home (PATCH)” which aims to examine the feasibility of a 12-week home-based intervention to improve heart failure self-management adherence. This trial can be found on https://www.clinicaltrials.gov/ct2/show/NCT01964053.

Participants were recruited from October 2013 to December 2014 from two rural critical access hospitals in southeast Nebraska. Patients were eligible for the study if they: 1) were age 21 or older; 2) had heart failure as one of their discharge diagnoses; 3) had New York Heart Association (NYHA) class II to IV heart failure or had NYHA class I heart failure and had at least one heart failure-related hospitalization or emergency department visit in the previous year; 4) were discharged to home; 5) passed the Mini-Cog screen test screening for dementia ²² ; 6) understood English; and 7) had access to a phone.

We excluded patients who: 1) had depressive symptoms (received a score of 3 or above on the Patient Health Questionnaire-2 (PHQ-2) ²³ ; 2) were diagnosed with liver cirrhosis; 3) were diagnosed with chronic renal failure; and 4) were diagnosed with other end stage and/or terminal illness (e.g. cancer) which limited the patient’s ability to perform self-management behaviors. The study setting is described in more detail in the study protocol ²⁴ .

We collected socio-demographic characteristics including age, gender, educational attainment, race/ethnicity, annual household income, marital status, and smoking status with a structured questionnaire. Clinical characteristics included comorbidities, echocardiographic ejection fraction (EF), and New York Heart Association (NYHA) Functional Classification.

We measured patient activation using the Short Form of the Patient Activation Measure (13-item version, available at http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1361231/table/tbl1/) which has similar reliability and validity to the long form (22 item version) across different ages, genders and health condition status ¹⁵ . Each item of the form was scored on the 5-point Likert response scale. For the ease of interpretation, the raw scores were transformed from the original metric to a 0–100 metric with higher scores indicating higher activation levels. Based on the patient activation score, patients were categorized into four levels: level 1 (score <47.0), level 2 (score 47.1–55.1), level 3 (score 55.2–67.0), and level 4 (score >67.0) ^{12, 15} .

We measured self-efficacy using the Self-Care of Heart Failure Index (SCHFI) which has six questions on a 4 point Likert scale ¹⁴ . The SCHFI was used to access the degree of achievement in self-care maintenance and management ²⁵ . The raw scores were standardized to a 100 point scale, with higher scores indicating higher efficacy. This measurement tool is reliable to assess self-efficacy with Cronbach’s alpha coefficient of .83 ¹⁴ .

We measured patients’ heart failure knowledge using the 27 multiple-choice questions of the Atlanta Heart Failure Knowledge Test (AHFKT) ²⁶ . Each correct answer scored 1 point and the total score ranged from 0 to 27. This questionnaire established high reliability to evaluate heart failure knowledge with a Cronbach alpha coefficient of .84 ²⁶ .

We used the 29-item Revised Heart Failure Self-Care Behavior Scale (RHFSCS) to assess patients’ behavior in six heart failure care domains: (1) seeking medical help, (2) being aware of the effects of heart failure, (3) prevention of complications, (4) awareness of deleterious effects of medical care, (5) accepting heart failure, and (6) learning to live with heart failure ²⁷ . Each response is granted a score from 0 (none of the time) to 5 (all of the time) ²⁷ . The internal reliability of this questionnaire is consistent with a Cronbach alpha coefficient of .84 ²⁷ .

Statistical analyses were conducted using SPSS version 20, with a p value less than .05 considered significant. We reported mean and standard deviation for continuous variables. Frequency and percentage were used to report categorical variables. Since patient activation level was an ordered variable, the Chi-square test for trend was used to assess correlations between four activation levels and the categorical variables (e.g. patient’s demographic and clinical characteristics). To compare continuous variables (e.g. behaviors, heart failure knowledge, Self-efficacy) across four activation levels, we used the ANOVA test for normally-distributed variables and the Kruskal-Wallis test for the non-normally distributed variables.

The study protocol was approved by the University of Nebraska Medical Center Institutional Review Board (IRB) and hospital ethical committees (IRB PROTOCOL # 228-13-EP).

There were 101 patients enrolled in this study. Table 1 describes socio-demographic and clinical characteristics. The mean age was 70±12.1 years. The majority of participants were women (63%). Two-thirds of participants had an annual household income of less than $50,000 or an educational attainment of high school graduate or lower. All the patients had multiple comorbidities including more than 2 of these chronic conditions: Hypertension (98.0%), Coronary artery disease (94.1%), Dyslipidemia (83.2%), Diabetes mellitus (41.6%), COPD/asthma (39.4%), CVA/stroke (17.8%), and cancer (6.9%). The majority of participants are at NYHA class II and III with the average EF of 55.8±11.1.

The mean patient activation score was 57 and the median was 52 (95% CI 53–61). The patient activation score did not distribute normally with a positively skewed pattern. About 40% of participants believed that they were responsible for caring for their health and illness but failed to take action to manage their heart failure condition (level 1). A quarter of participants felt confident and knowledgeable enough to manage their health but failed to take action (level 2). Less than 40% of participants actually took some actions to manage their condition (level 3 and 4).

Patient activation levels were not significantly different across age group, gender, marital status, household income or smoking status. There was an association between activation level and educational attainment. Participants with educational attainment beyond high school had higher patient activation scores on the PAM compared to patients who only graduated high school or lower (62 vs. 54, p=.03).

Using the Chi-square test for trend, we found an increasing trend between knowledge and self-efficacy across the four levels of patient activation. Patients with higher heart failure knowledge or greater self-efficacy were at a higher level of patient activation (p=.005 and p<.001, respectively). Patients who were at a higher level of patient activation also had higher scores of self-management behavior (p<.001). Table 2 describes the average score of self-management knowledge, efficacy, and behavior among 4 levels of patient activation level.

Heart failure (HF) patients discharged from rural hospitals: demographic, clinical, patient activation level, HF knowledge and self-management behavior characteristics

Data was collected by questionnaires through patient interview and medical record review.

Details of labelling and abbreviation of the variables were included in the dataset ⁴⁰ .

Our study had some limitations. We used only baseline data; therefore, we are unable to demonstrate the changes of patient activation scores, knowledge scores, self-efficacy and self-management behaviors over time. As a result, we could not establish the temporal relationship between knowledge, efficacy, patient activation level and their impact on behavior in this article, which will be reported in future manuscripts. Secondly, the patients who agreed to participate in the original randomized controlled trial might be more motivated than patients who refused to participate, therefore creating a pool of heart failure patients with higher activation levels compared to the general rural heart failure population. If this assumption is valid, the overall activation level in general rural heart failure population might be even lower than what is reported in this article, which indicates a great need to conduct interventions to enhance rural heart failure activation to engage in self-management behaviors.

Using the short version of PAM (13 items) is a feasible way to assess a patient’s activation level and identify those with low levels in both inpatient and outpatient settings. The assessment results allow clinicians to develop tailored interventions to support those high risk patients. Additionally, the patients with high activation levels could be encouraged and recruited as behavior coaches and/or peer supporters for those with low activation scores.