Measurement and data collection
Version 1.0 of the
SF-3623 was translated
into Vietnamese by a professional translator and then translated back
into English by another. No significant difference between the two
translations was detected. The survey consisted of 36 questions related
to eight health notions including physical functioning, role limitations
due to physical health problems, role limitations due to emotional
problems, energy/fatigue, emotional well-being, social functioning,
bodily pain and general health perceptions, divided into two groups of
physical and mental
health.15 The responses
were scored following the RAND scoring
method.24 The 36-Item
Health Survey 1.0 included the same set of items as the SF-36. In this
research, we used the RAND 36-Item Health Survey scoring process
consisting of 2 steps. Firstly, pre-coded numeric values were recoded
according to the scoring scale described hereinafter. Responses 1 to 5
for items number 1, 2, 20, 22, 34, and 36 were recoded to values of 100,
75, 50, 25, and 0, respectively. Responses 1 to 3 for items number 3 to
12 were recoded to values of 0, 50, and 100, respectively. Responses 1
and 2 for items from 13 to 19 would be recoded to values of 0 and 100,
respectively. Responses 1 to 6 for items number 21, 23, 26, 27, 30 were
recoded to values of 100, 80, 60, 40, 20, and 0, respectively. Responses
1 to 6 for items number 24, 25, 28, 29, 31 were recoded to values of 0,
20, 40, 60, 80, and 100, respectively. Responses 1 to 5 for items number
32, 33, 35 were recoded to values of 0, 25, 50, 75, and 100,
respectively. All the items were scored so that that a higher score
indicates a better state of health-related QoL. In addition, each score
of items ranged from 0 to 100, with 0 indicating the lowest and 100 the
highest possible score. Scores are represented as percentage of the
total possible score
achieved.24 Secondly,
items in the same scale were averaged together to create 8 sub-scale
scores that yield eight health concepts. If the respondent missed one of
the two items, the person’s score was set to be equal to that of the
non-missing item.24
The high reliability of this scoring method for each of eight scales has
been identified, with Cronbach’s alpha ranging from 0.78 to
0.93.24 The
International QoL Assessment project has translated and validated the
SF-36 for use in 45
countries.25 In this
study, we also verified the suitability of the SF-36 based on our pilot
study after translation and adaptation to the Vietnamese language and
culture. The pilot research was conducted on 10 patients, and the
results showed that all the scales were technically reliable and
feasible for the main survey. In the main survey, the selected patients
were interviewed face-to-face when they visited Vietnam Heart Institute
(about 80% of the patients) or via telephone when any of them did not
come at repeated surveys (20%).
The main predictor of interest is the QoL of patients across rounds
(i.e., baseline, 1st and 3rd months
of follow up)
Key selected characteristics. The demographic factors covered age
(years), gender (male or female), educational level (lower level
representing under or equal to high school or higher level including
college or university and post-graduate level), marital status (single
or married), occupation (retired or not retired/farmers), region (rural
or urban), past coronary intervention before the current intervention
(yes or no), hypertension (yes or not), hyperlipidaemia (yes or no),
diabetes mellitus (yes or no), the number of stents inserted (1, 2, and
3 stents), smoking history (yes or no before the onset of UA), alcohol
intake history (yes vs. no before the onset of UA), and troponin T
(under or over 0,01 ng/ml).
Statistical analysis . The frequencies and percentages
for categorical variables, and means and standard deviations (SD) for
quantitative variables were used to describe the data. We calculated
mean scores and standard deviations for eight subscales of the SF-36 at
pre-intervention, one month, and three months after the intervention. In
each of the eight subscales, higher scores indicated a better state of
health-related QoL. As a repeated measures design, we used a
mixed-effects model with random intercept to detect factors associated
with QoL. Each individual had three observations, one at baseline, one
at 1 month and the other at 3 months, with correlation assumed between
observations within the same individual. All analyses were conducted
using Stata version 16 (Stata Corporation, College Station, TX)
Research ethics . All subjects received verbal
descriptions of the purpose and methods and their right to refuse
participation at any time. They were reassured that their participation
was voluntary and anonymous, and the refusal would not affect their
treatment or nursing care. The protocol was approved by the Research
Assessment Board from the Institute for Preventive Medicine and Public
Health at Hanoi Medical University according to Decision No.
5403/QD-DHYHN dated 06/12/2016.