Study Area and Period
The study was conducted in tertiary hospitals of the Tigray region from March 01 – April 30, 2020. ACSH and AUCSH are university hospitals in Mekelle and Axum, Tigray region respectively commenced rendering all the specialized and non-specialized services including special clinic services (19, 20). The hospital provided follow-up care for 950 heart failure patients according to data registered in 2019 before the data collection period.
Study design
An institutional-based cross-sectional study was employed.
Source population
All adult patients with heart failure on follow-up at the cardiology unit of ACSH and AUCSH during the data collection period.
Study population
All selected adult patients with heart failure on follow-up at the cardiology unit of ACSH and AUCSH.
Inclusion and Exclusion criteria
All adult patients with heart failure who were on follow-up for three months and above were included in the study. However, patients who were critically ill during the data collection period were excluded from the study.
Sample size determination
The sample size for the study was determined by using single population proportion formula assuming 5% marginal error (d), 95% CI (alpha = 0.05), and 50% proportion (p) of poor HRQoL in HF patients as there is no previously conducted study in a similar setting to or in Ethiopia.
Where
n1 = required initial sample size
z = the desired level of confidence interval 95% (Z = 1.96)
P = proportion of poor HRQoL in HF patients
q = proportion of good HRQoL in HF patients (1˗ 0.5 = 0.5)
d = Marginal error (0.05)
A correction formula was introduced as
,
which gives a sample size of 274 and considering 10% for the non-response rate, the final sample size was 301.
Sampling techniques and procedures
The sample size of each tertiary hospital was determined using proportional sample allocation. A systematic random sampling method was used to enroll study participants. Sampling interval (K) was calculated then by selecting a random number from one up to “k” using the lottery method, study samples were selected.
Study variables
Independent variables
The independent variables were socio-demographic variables (age, sex, marital status, educational level, income, occupation, number of household members, religion, ethnicity, and place of residence), clinical characteristics of patients (NYHA class, duration of HF, previous hospitalization, and presence of comorbidity), behavioral practices of patients (smoking and salt consumption) and presence of social support.
Dependent variable
Health-related quality of life was the dependent variable.
Data collection tools and procedures
Data were collected through interviews using pretested structured questionnaires and documentary reviews from patient medical charts. Two trained BSc nurses were employed as data collectors and two experienced BSc nurses were assigned as supervisors. Study participants were informed about the length of time the questionnaire needed, anonymity & confidentiality. The questionnaire had five parts, socio-demographic-related questions, clinical related questions, behavioral-related questions, questions related to the presence of social support, questions related to assessing HRQoL adapted from medical outcomes study SF-36 health survey (21).
The SF-36 health survey is one of the most widely used measures of HRQoL consisting of 36 items and covering eight dimensions: physical functioning, role limitations due to physical health problems, bodily pain, general health perception, vitality, social functioning, role limitations due to emotional health problems, and emotional well-being/mental health. Scores on all the subscales are transformed linearly to a possible range of 0-100 where higher scores represent better HRQoL (21, 22).
The Cronbach's alpha value of the scales of the instrument was 0.88 which indicates that the instrument has acceptable internal consistency reliability. Furthermore, external validity was addressed through the use of probability sampling technique, strong inclusion and exclusion criteria. The content validity was also assured through conducting a relevant literature review before the development of the instrument, adapting a validated questionnaire from Medical care, evaluating the tool by senior researchers, and translating the questionnaire into local language to suit the language ability.
Operational definitions
Good HRQoL is defined as when the patient scored a mean score of SF-36 ≥ 60 (22).
Poor HRQoL is defined as when the patient scored a mean score of SF-36 ˂ 60 (22).
Comorbidity is defined as the presence of one or more disorders in addition to the index disease (23).
Smoking is defined as the experiences of ever cigarette smoking status (24).
Salt-free is defined as the use of spices other than salt when cooking and avoid eating foods prepared outside the home (24).
Social support is defined as the real resources provided by others that enable a person to feel cared for or when family members and friends are involved in the care process (25, 26).
Data quality assurance
Data quality was ensured by giving training for data collectors and supervisors and by providing day-to-day supervision. The questionnaire was translated into the local language (Tigrigna) and retranslated into English by experts to ensure its consistency. Each questionnaire was checked for completeness, missed values, and unlikely responses at the spot. A pre-test was conducted on 5% of the sample size in Suhul general hospital to see the applicability of the instruments and necessary amendments were done. Every questionnaire was checked by the principal investigator at the spot (27).
Data processing and analysis
Data were checked visually for its completeness and the questionnaire was coded, entered, and cleaned using Epi-Data manager version 4.4.2.1 for windows and exported to SPSS version 22.0 for statistical analysis. Multi-collinearity was checked among predictor variables and a Variance Inflation Factor (VIF) value greater than 1.42 did not appear in this study, indicating that serious multi-collinearity did not exist.
Descriptive statistics were computed through percentage, frequency, mean and standard deviation, and results were summarized and presented by texts, tables, and figures. Binary logistic regression was used to determine the statistical association between a set of independent variables and the outcome variable. Variables with P-values < 0.25 in the bivariate regression analysis were included in the multivariable analysis. Adjusted odds ratio with a 95% CI was computed and statistical significance was declared at p-value < 0.05 (27).
Ethical considerations
Ethical clearance was obtained from the Mekelle University College of Health Sciences Institutional Review Committee (IRC). The purpose and objective of the study were described to the study participants and written consent was obtained. Respondents have been informed of all the necessary information and they decided freely by themselves to involve, refuse, or discontinue participation at any time they want. Information was collected anonymously; confidentiality and privacy were assured and maintained throughout the study period.