Filetype PDF | Posted on 23 Sep 2022 | 4 years ago
Authentication
digital resources
244x Filetype PDF File size 0.20 MB Source: ijshr.com
International Journal of Science and Healthcare Research
DOI: https://doi.org/10.52403/ijshr.20211008
Vol.6; Issue: 4; Oct.-Dec. 2021
Website: ijshr.com
Original Research Article ISSN: 2455-7587
The RAND-36 Health Survey 1.0: Translation,
Reliability, Cross-Cultural Adaptation and
Validation of the Gujarati Version
Hemang Jani1, Gauravi Dhruva2, Dinesh Sorani3, Yagnik Dave4
1Inch. Principal/Associate Professor, Harivandana Physiotherapy College, Rajkot, Gujarat, India.
2
HOD, Pathology Department, PDU Medical College, Rajkot.
3
Principal, Govt. Physiotherapy College, Jamnagar.
4Assistance Professor, Faculty of Physiotherapy, Marwadi University, Rajkot.
Corresponding Author: Hemang Jani
ABSTRACT Product-moment correlations to test the
equivalence of the corresponding Gujarati and
Background: The Short Form 36 Item Survey is English versions of the RAND-36 ranged from
the most typically used instrument for assessing 0.73 to 0.92. The Gujarati version of the RAND-
health-related quality of life.1 Two identical 36 has high internal consistency (Cronbach’s
versions of the initial instrument are currently α=0.809) and test-retest reliability (Intra-class
available: the general public domain, license- correlation coefficient=0.746, 95% CI: 0.58,
free RAND-36, and also the commercial SF 36.2 0.94).
RAND 36 don't seem to be available within the Conclusions: The Gujarati version of the
Gujarati language. The aim of this study was to RAND-36 performed well and the findings
translate and culturally adapt the RAND 36 into suggest that it is a reliable and valid measure of
the Gujarati language and measure its reliability health-related quality of life among the general
and validity. Gujarati population.
Methods: According to the guidelines by the
International Quality of Life Assessment Keywords: RAND-36, cross-cultural translation,
project, a test of item-scale correlation, a quality of life, health status assessment,
sequence of translation, and validation were Gujarati.
implemented for the translation of the Gujarati
version of the RAND-36. Following pilot INTRODUCTION
testing, the English and the Gujarati versions of Health is defined as complete
the RAND-36 were administered to a random mental, physical and social well-being with
sample of 120 apparently healthy individuals to the absence of disease. The RAND 36
test validity and 96 respondents completed the Health Survey is a quality of life
Gujarati RAND-36 again after two weeks to test Questionnaire. It consisted of the same
reliability. Data were analyzed using one-way items included in the Medical Outcomes
analysis of variance, multi-trait scaling analysis, Study 36 Items Short-Form Health Survey
one-way analysis of variance, Pearson’s (SF-36). Both tools came out from work
product-moment correlation analysis, and Intra-
Class Correlation (ICC) at p < 0.05 begun at RAND in 1984 as part of the
Results: The median Cronbach's alphas for the Medical Outcomes Study (MOS) and
Gujarati RAND-36 in multiple subgroups include multi-item scales that assess eight
exceeded 0.70 for every scale except one. Two health concepts: physical functioning,
of the English RAND-36 scales had median bodily pain, role limitations due to physical
Cronbach's alphas that exceeded 0.70; the rest health problems, role limitations due to
exceeded 0.50. Test-retest correlations were emotional problems, mental health
found statistically significant for both versions. (emotional well-being), social functioning,
International Journal of Science and Healthcare Research (www.ijshr.com) 52
Vol.6; Issue: 4; October-December 2021
Hemang Jani et.al. The RAND-36 health survey 1.0: translation, reliability, cross-cultural adaptation and
validation of the Gujarati version.
vitality (energy/fatigue), and general health. gather demographic information (i.e,
The only difference between the measures is gender, age, and education), and the
the scoring algorithm applied when questionnaires were coded so that non-
calculating two of the scale scores (bodily respondents could be contacted. The list that
pain, general health), which is independent linked the questionnaire codes to specific
of the evaluation process. Hence, for the individuals can be accessed by only the
comparison with published general second author. These questionnaires were
population norms, SF-36 scoring was used. administered twice (time 1 = initial; time 2
There is a growing body of evidence that the = retest) as follows:
SF-36 can be successfully translated, Group A. subjects completed the English
validated, and normalized for use in other version first then the Gujarati version. A
countries.3 This article describes the sealed envelope containing both versions to
administration and evaluation of a new the subjects was distributed by the assigned
Gujarati translation of the RAND-36 in a coordinator.
sample of Gujarati People. The translation The subjects were given instruction
process that initiated this work was to complete both questionnaires as follows:
conducted in agreement with a commonly (1) Complete the English questionnaires put
accepted cross-cultural adaptation it in the envelope provided and close the
4,6
methodology. envelope,
(2) Complete the Gujarati questionnaire, put
METHODS it in the other envelope and close the
Sample envelope, and
This study was conducted at (3) Return both envelopes to the assigned
Harivandana Physiotherapy clinic Rajkot, coordinator.
Gujarat. India. A convenience sample was Aside from the order of
selected from several departments at administration, instructions for Groups B
Harivandana Physiotherapy clinic Rajkot. through F were as discussed for Group A.
This sample was selected by asking each of Group B. Participants completed the
the departments to generate a list of Gujarati version first, immediately followed
bilingual (Gujarati and English) Gujarati by the English version.
employees who were willing to participate Group C. Participants completed the English
in the study. Participants were selected from version followed by the Gujarati version
each department to work as coordinators after a 2-week interval.
with the second author. After obtaining Group D. Participants completed the
complete lists of participants from the Gujarati version followed by the English
departments, each individual was assigned version after a 2-week interval.
randomly to one of six groups. 60 subjects Group E. Participants completed the
were assigned to each of Groups A, B, and Gujarati version two times (2 weeks apart).
C and 60 subjects were assigned to each of Group F. Participants completed the English
Groups D, E, and F. After taking written version two times (2 weeks apart).
informed concern all the participants were
enrolled in this study. Scoring:
All scales were converted linearly to
Data Collection: a 0 to 100 possible variation of scores, with
Data were collected by distributing 0 and 100 representing the least and most
the English and/or Gujarati versions of the favorable health state, respectively. All
RAND-36 to the participants through the scores indicate the percent of the total
coordinators. Cover letters that explain the possible score for that scale. The RAND 36
purpose of the study were provided. was scored according to the methodology
Additional questions were included to used for the published norms for the SF 36.7
International Journal of Science and Healthcare Research (www.ijshr.com) 53
Vol.6; Issue: 4; October-December 2021
Hemang Jani et.al. The RAND-36 health survey 1.0: translation, reliability, cross-cultural adaptation and
validation of the Gujarati version.
in order to compare scores from this sample ANOVA test was applied for comparing the
to those from the Indian general population. age of the six groups found no statistically
For other detail examination, the pain and significant difference [F (5, 358) = 0.929, P
therefore scoring of general health > 0.05].
perception scales followed the RAND 36
recommendations.1 Internal Consistency Reliability:
Gujarati Version Initial
Data Analysis: Administration. Alpha reliability
Statistical analyses were conducted coefficients for the initial administration of
using SPSS 20. the Gujarati version (Groups B, D, and E, n
Descriptive Statistics. The average ages, = 48) ranged from 0.60 (general health) to
percentage of men and women, and degree 0.87 (physical functioning). Reliability
of education of respondents in each category coefficients exceeded 0.70, except for the
were determined. general health scale which follows
Internal Consistency Reliability. Nunnally's standard of acceptable reliability
9
Cronbach's coefficient alpha was for group-level studies.
used to estimate internal consistency
reliability for each of the administrations of Retest Administration:
8
Gujarati and English versions. Alpha coefficients for retest
administrations (Groups A, C, and E, n =
Test-Retest Reliability. 48) ranged from 0.57 (general health) to
To assess test-retest reliability 0.88 (physical functioning) and were 0.70 or
during a 2-week interval, Pearson product- higher for all scales except for general
moment correlations were computed health.
between initial and retest administrations for
both Group E and Group F. Internal Consistency Reliability:
English Version Initial Administration.
Equivalent Forms Reliability. Alpha coefficients for the initial
Pearson product-moment correlation administrations of the English version
coefficients were computed to evaluate (Groups A, C, and F, n = 48) ranged from
equivalent forms reliability between the 0.56 (role limitations-physical, social
Gujarati and English versions administered functioning) to 0.89 (physical functioning).
on the same day (Groups A and B). Scale Alpha coefficients for four other scales were
Score Means. Scale score means were less than 0.70 [Role limitations-emotional
computed for the Gujarati and English (0.66), Mental health (0.68), Vitality (0.58),
versions to determine if central tendency General health (0.61)].
varied by version. Retest Administration. Alpha coefficients
for the retest administrations (Groups B, D
RESULTS and F, n = 48) ranged from 0.60 (social
Subjects of the 120 subjects functioning) to 0.90 (physical functioning).
randomly assigned to the six study groups, Alpha coefficients for two other scales were
96 (78%) individuals returned their less than 0.70 (general health (0.68), vitality
questionnaires. Of these, 24 individuals (0.61)). Test-Retest Reliability Two-week
were excluded from the study because they test-retest product-moment correlations
had missing data for one or more items. were all statistically significant and ranged
Thus, the analytic sample size was 96. The from 0.29 (bodily pain) to 0.80 (mental
mean age of the subjects was 34.9 + 7.0 health) for the Gujarati version (n = 48) and
years, with a range of 25 to 55 years. from 0.46 (role limitations-physical) to 0.77
Seventy-eight percent of the subjects were (general health) for the English version (n =
male, and 52% were college graduates. 48). Equivalent-Forms Reliability
International Journal of Science and Healthcare Research (www.ijshr.com) 54
Vol.6; Issue: 4; October-December 2021
Hemang Jani et.al. The RAND-36 health survey 1.0: translation, reliability, cross-cultural adaptation and
validation of the Gujarati version.
Equivalent-forms product-moment functioning) to 0.91 (mental health).
correlations between corresponding scales Correlations for those administered the
for those administered the English version Gujarati version immediately followed by
immediately followed by the Gujarati the English version (n = 48) ranged from
version (n=48) ranged from 0.78 (social 0.73 (vitality) to 0.92 (mental health).
Table-1: Mean RAND-36 Scale Scores for the Initial Administrations of the Gujarati and English Versions of the RAND-36
Scale Number of Initial Gujarati Administration Initial English Administration
Items (n = 48) (n = 48)
Physical functioning 10 87.3 (17.1) 82.4 (20.6)
Role limitations due to physical health 4 84.3 (30.3) 90.3 (19.4)
Role limitations due to emotional 3 76.3 (35.2) 82.3 (29.3)
problems
Energy/fatigue 4 67.6 (16.8) 68.1 (16.8
Emotional well-being 5 75.2 (16.1) 75.6 (15.6)
Social functioning 2 82.0 (19.3) 79.0 (20.7)
Pain 2 80.3 (21.9) 81.1 (19.0)
General Health 5 71.1 (16.1 74.5 (15.1)
+P < 0.01. ++P < 0.05. P < 0.0001. P< 0.001.
"Scale scores were calculated using the RAND-36 scoring system.”
DISCUSSION and Gujarati versions. The size of the
The median internal consistency correlations indicates a noteworthy degree
reliability coefficients for all administrations of variation between initial and retest
(Groups A, C, and E) of the Gujarati version administrations. The study timing may have
exceeded 0.70 for every scale except for contributed to the fluctuation in scale
general health (median alpha = 0.59). The scores.
median internal consistency reliability Many Gujarati citizens suffer from
coefficients for all administrations (Groups Covid-19 and related symptoms during the
B, D, and F) of the English version time of year the study were conducted.
exceeded 0.50. Therefore, the results of this Because of attainable actual changes, test-
study provide support for the reliability of retest reliability estimates need to be
the Gujarati version and are consistent with evaluated with caution, especially when
previous reliability estimates reported for studying a dynamic process such as health
the English version. status.11 The equivalence of the
Both the Gujarati and English corresponding English and Gujarati versions
versions tended to have internal consistency of the RAND 36 scale was assessed using
reliability coefficients equal to or above the correlations.
acceptable standards for group Product moment correlations were
comparisons;9'10 however, the Gujarati ranging from 0.73 to 0.92 between
version had higher median values than the corresponding scales. These results provide
English version. Although all participants strong support for the equivalence of the
were bilingual, their mother language was Gujarati and English versions. The results
Gujarati. found by this study indicate the good
As a result, it is likely that the reliability of a Gujarati version of the
participants had a better understanding of RAND-36 and its equivalence with the
the Gujarati version than the English English version.
version, leading to more internally
consistent responses. In addition, the CONCLUSIONS
Gujarati version was adapted to the Gujarati The Gujarati version of the RAND-
culture, whereas the English version was 36 performed well and the findings suggest
developed for the dominant US culture. that it is a reliable and valid measure of
Test-retest correlations during a 2-week health-related quality of life among the
time interval were similar for the English general Gujarati population.
International Journal of Science and Healthcare Research (www.ijshr.com) 55
Vol.6; Issue: 4; October-December 2021
no reviews yet
Please Login to review.
