Study setting
This study was conducted in Addis Ababa, the capital city of Ethiopia. As per the 2007 national census [14], its total population was 3,384,569 with annual growth rate of 3.8%. There are 12 public hospitals in Addis Ababa [15] six of which and 103 health centers are under the city administration [16]. This study was conducted in two hospitals namely; Saint Paul’s Hospital Millennium Medical College (SPHMMC); a federal referral hospital and Ras Desta Damtew Memorial Hospital (RDDMH)-one of the city administration’s hospital and, five health centers namely; Arada, Lideta, Nifas Silk Lafto Wereda 9, Akaki Kality and Bulbula health centers.
Design and period
A multi-facility based cross sectional study design was conducted throughout August 2020.
Source and study population
The source population was all adult outpatients with T2DM and chronic hypertension in the study area. The study population is all outpatients with T2DM and hypertension visiting public hospitals and health centers and fulfilled the inclusion criteria.
Inclusion and exclusion criteria
Inclusion criteria
Adult outpatients with at least one chronic illness and started taking medications for at least six months, visiting the facility during the study period, willing to participate in the study, and those who could speak Amharic language were included in the study.
Exclusion criteria
Patients with any long term or temporary psychiatric problems, patients admitted to the emergency/inpatient department for any reason, and patients aged below 14 were not included.
Sample size and Sampling technique
The single population proportion formula was employed to estimate sample size required for the study. To attain maximum size, the proportion of patients with affected HRQoL was considered as 50%. A 95% confidence level and a 5% tolerable error assumption were considered to estimate the sample size. Taking in to account of 10% contingency for nonresponse, the required number of respondents was 423. Both hospitals and health centres were selected intentionally (purposively) grounded on patient flow and socio-economic aspects. Participants from the respective facilities were included based on availability during the framed data collection period. In the same vein, allocation per T2DM and hypertension cases in each facility was determined based on follow up outpatient loads reported in similar settings [17-20].
Study variables
Dependent variables:
The dependent variable of this study were HRQoL and GQoL
Independent variables:
Socio-demographic: age, sex, marital status, education, occupation, religion, substance use history, income level, type of relation with others, number of people around.
Clinical related: type of chronic illness, number of comorbidities, time since diagnosis, time since initiation of treatment.
Facility related: type of follow up facility.
Data collection instrument and procedure
Data was collected using a structured interviewer-administered questionnaire prepared from the literature. Patients’ HRQoL was measured using the standardized EQ-5D-3L generic tool [6]. The tool has five dimensions namely: mobility, self-care, usual activities, pain/discomfort and depression/anxiety on three levels as; no problem, some or moderate problem and severe problems. Likely, patients’ subjective judgement about their current state of overall health was measured using the Euroqol visual analogue scale (EQ VAS) [6] anchored with adapted patient aiding terms [21]. Collected quantitative data included; socio-demographic profiles, clinical and social factors, substance use history and quality of life scores.
Data quality management
The quality of data collected was ensured through the use of trained data collectors, continuous supervision, and application of a pre-tested instrument in the data collection. Furthermore, the instrument was translated to Amharic, a national working language in the country, and back translated for analysis and reporting. Content validity of the instrument, on selected variables of interest to the research question, was checked by experts in the team. A Cronbach’s alpha test was done to measure reliability of the EQ-5D-3L scale on similar populations to health centres and hospitals (10% for each) and score showed high internal consistency (α= 0.85 and 0.91 respectively).
Operational definitions
Patients with chronic non-communicable diseases: patients diagnosed with T2DM or/and hypertension visiting the selected health facilities and commenced with medications at least six months earlier to the study period. Additional follow up comorbidities may or may not present.
Affected HRQoL: A cumulative score from mobility, self-care, usual activities, pain/discomfort and depression/anxiety aspects of a patient and summing up to 6 or more (reporting at least one moderate or severe problem) in the EQ-5D-3L instrument. Score sums of 5 (reporting no problems to all dimensions) were considered not affected.
Global quality of life (GQoL): Patients’ self-rating of their own current overall health status on a numerical simplified EQ VAS scale ranging from 0 (worst imaginable health state) to 100 (best imaginable health state).
Any problems: refer a patient’s experience of least one or a combination of problems at moderate to severe degree on the EQ-5D-3L tool.
Ethical approval and consent
Ethical approval has been obtained from Saint Paul’s hospital millennium medical college (SPHMMC). Additional approval was sought from the Addis Ababa Health Bureau (AAHB) research ethics committee. A support letter was written from SPMMC and AAHB to the respective facilities. Data and information acquired through the study were kept confidential. No individual identifiers were collected and analysis was made in aggregate. Inclusion in the study was solely on voluntarily basis.
Data analysis
Collected data was coded, cleaned manually and entered in SPSS version 26.0 for windows. We used descriptive statistics to present univariate and bivariate analysis. Binary logistics regression was used to identify associated factors with affected HRQoL. A 95% confidence level was considered with statistical significance of (p<=0.2) for the bi-variable and (p<=0.05) the multivariable outputs. A Shapiro-Wilk test was considered for checking normality of GQoL scores. Because the data showed a non-parametric distribution (W(409)=0.94, p=0.000), mean difference by sex, comorbid condition, and facility types was tested using the Mann Whitney U test. Potential difference by diagnosis type was tested using the Kruskal Wallis test. The ranked median difference test for Age category vs. GQoL was tested using the Spearman’s rank correlation coefficient. A p-value of 0.05 was considered as a cut point of statistical significance in all difference tests.