·Investigation·
Validating
a tool to assess eye health knowledge, attitude and practice in Cambodia and
Vietnam
Noela Prasad1,2, Gail M Ormsby3,4,
Robert P Finger5
1Institute
for Health and Sport, Victoria University, Melbourne, VIC 3011, Australia
2VACCHO,
Collingwood, VIC 3066, Australia
3Professional
Studies, Faculty of Business, Education, Law and Arts, University of Southern
Queensland, West Street, Toowoomba, QLD 4350, Australia
4Lifestyle
Research Centre, Avondale College of Higher Education, Cooranbong, NSW 2265,
Australia
5Department
of Ophthalmology, University of Bonn, Regina-Pacis-Weg 3, Bonn 53113, Germany
Correspondence
to: Gail M Ormsby. Professional Studies, Faculty of
Business, Education, Law and Arts, University of Southern Queensland, West
Street, Toowoomba, QLD 4350, Australia. go.gailormsby@gmail.com
Received:
Abstract
AIM: To develop an eye health knowledge, attitude and practice (EH-KAP)
field-based assessment tool for use in implementing effective eye health care
services.
METHODS: An instrument development and validation study. A
Vietnam EH-KAP dataset were used to identify and eliminate redundant questions
to develop a standardized tool. Face validity was assessed by the KAP survey
team. Internal validity (congruency/criterion) was assessed by comparing
descriptive analysis of two datasets (n=531; n=38) collected from
the same sampling frame at different time points. Weighted scores were
calculated for each construct. Kappa values for test-retest and inter-observer
agreement were calculated to check the reliability of responses. The modified
version was assessed by analysing the raw and ungrouped data. Responses were weighted
and agreement was tested by comparing construct scores.
RESULTS: Totally 38 respondents were included in this
validation process (mean age 58.5y). Mean scores for knowledge were 9.15 (old
questionnaire n=531) and 5.05 (modified version). For attitude, the
scores were 2.23 and 2.42, and for practice the scores were 3.33 and 2.21.
Test-retest agreement was between 62% to 93% (Kappa 0.24 to 0.86) for the
ungrouped raw data, and 55% to 72% (Kappa 0.42 to 0.65) for KAP domain.
Inter-observer Kappa value for ungrouped data was 0.37 and 0.45 for the
weighted scores.
CONCLUSION: This standardized tool applied at critical time
points can assess trends in KAP within the same population and for comparison
across groups. If used alongside a Rapid Assessment of Avoidable Blindness
(RAAB), this tool provides a comprehensive perspective on eye-health of a
population.
KEYWORDS: validation;
knowledge; attitude; practice; eye health; knowledge, attitude and practice;
questionnaire
DOI:10.18240/ijo.2019.11.16
Citation: Prasad N,
Ormsby GM, Finger RP. Validating a tool to assess eye health knowledge,
attitude and practice in Cambodia and Vietnam. Int J Ophthalmol 2019;12(11):1767-1774
INTRODUCTION
Knowledge,
attitude and practice (KAP) surveys have been utilised to gain a
cross-sectional snap-shot of behavioural patterns within populations[1-2]. When administered at baseline,
and again at the end of the project, this tool can evaluate change in KAP in
response to specific interventions or programs[1-2]. Therefore, applying information from prior experience,
we developed a tool to assess KAP related to eye health. The intention was to
use a KAP assessment not only to inform health promotion and policy, but also
to assess effectiveness as a part of the programme planning cycle. This paper
describes the process of standardizing a valid tool that reliably measures KAP
concerning eye health, since there was no standardized tool identified in the
literature.
SUBJECTS AND METHODS
Ethical
Approval Ethics
approval was obtained from the Royal Victorian Eye and Ear Hospital, the Takeo,
Thanh Hoa and Nghe An Provincial Health Departments. The study followed the
tenets of the Declaration of Helsinki. Permissions were granted by local
authorities to conduct field-work. Oral consent was obtained from each
participant after they received information regarding the project and their
queries were clarified in the local dialect. No incentive or monetary
compensation was provided for participation.
Definitions Three
attributes to be measured using this tool in the target community were
knowledge, attitudes and intended practice, which are defined as follows.
Knowledge possessed by a community is their understanding of that topic or the
level of correct factual knowledge. Attitude refers to people’s feelings and
perceptions toward eye care and services as well as any preconceived ideas they
may have towards it[3].
Practice is defined as previous and intended future use of eye care services.
Exploratory
Surveys Despite the
importance of assessing KAP as a part of the planning cycle for eye health
programs, a literature search failed to identify any standardized tool for
measuring these constructs. Interviewer administered surveys were conducted in
Takeo Province in Cambodia (2010; n=599)[2], and in Son La (2010; n=300),
Thanh Hoa (2009; n=274) and Nghe An (2009; n=257) provinces of
Vietnam (unpublished) prior to commencing programmes for the delivery of eye
care services funded through the Australian Avoidable Blindness Initiative
Program. Different methodology and data collection instruments had been used at
each location, so these data were used to inform the development of a generic
validated KAP survey tool.
Tool
Development and Validation This
validation study was conducted during August 2013. Initially, face validity and
content validity of the “original” KAP questionnaire (Thanh Hoa and Nghe An)
were reviewed. Criteria assessed 1) adequacy of response options; that is,
whether at face value, the questions appear to elicit valid responses, 2)
construct and content validity, by examining whether all relevant aspects of
the construct were covered. The tool was then revised to eliminate redundant
questions and to collapse response options. It was translated into Kinh
(Vietnamese), assessed again to ensure face and content validity in the Vietnamese
context. Finally, it was back-translated into English, and reviewed by
bilingual eye-health workers.
Procedure Phase 1:
Identified which constructs a particular item sought to answer. For example,
the first construct knowledge, comprising factually correct information
obtained from a trusted and accessible source, related to three “items of
interest” - specific disease conditions, treatment options, and costs incurred.
Phase 2: Clarified the questions. Two criteria were applied to clarify each question:
a) ensure the question measured a discrete domain (attribute/trait/variable)
that informs a single construct; b) check for redundancy of
attribute/trait/variable questions. Phase 3: Eliminated redundant questions.
Phase 4: Checked the appropriateness of context, ensuring the question was
framed to measure what it intended to, within both the English and Kinh
questionnaires.
To ensure a
rigorous assessment, guidelines were provided as a manual of operation (in
English and Kinh) to be used as a reference point during data collection. The
questionnaire was then pre-tested in Thanh Hoa Province by a health centre
nurse, a community representative, two district hospital nurses, one NGO
development worker, an individual with low vision, and an interpreter. For
example, the process involved a role-play; where a nurse administered the
questionnaire to a community representative, while another nurse observed and
followed the English translation. All responses and limitations were noted and
rectified according to the socio-cultural context.
Due to the
absence of a gold standard against which knowledge, attitude or practice could
be assessed, the former (2009/2010) versions of the eye health
knowledge, attitude and practice (EH-KAP) tool were
assumed to be the “criterion” against which the revised tool was compared.
Reliability The final
questionnaire (Vietnamese version) was used for reliability testing among
sixteen adults chosen to represent the range of possible respondents. This
involved the intentional inclusion of people with a disability and ensured
gender equity.
Gauging
reliability was especially important because this tool was intended to measure
change over time. Test-retest and inter-observer reliability were assessed,
with the results reported in this paper. Ten volunteers (interviewers) were
trained to explain the purpose of this survey to potential respondents and to
administer the survey tool. They were informed of their responsibilities on the
research team, and were asked to read through the questionnaires (item-by-item
review of the questions, including skip patterns and special instructions), and
data entry screens to ensure that they understood the questions, definitions
and process of performing quality checks. The sampling strategy was explained,
with emphasis on the respondent selection procedures. The interviewers were
taught interview techniques, listening skills, confidentiality procedures, and
how to obtain informed consent from respondents. An “effective interview”
process was demonstrated to reinforce this knowledge. Finally, inter-observer
reliability among the ten trained interviewers and test-retest reliability
using different groups of respondents were conducted.
Field
Testing The manual
of operations and data collection tool were field-tested in a village setting.
Two villages were randomly selected from a commune in the district of Thanh
Hoa, using the same methodology and inclusion criteria as had been used for the
previous KAP survey conducted in 2009. In each village, a random start was
identified and houses on the left side of the street were consecutively visited until a
total of 30 individuals aged 30y or more were surveyed. Back-calculating from
the previous KAP survey conducted in Thanh Hoa and Nghe An Provinces, it was
evident that by surveying a sample of 530 people from a sampling frame that
comprised the general population of 4 000 000 (or, if stratified sampling was
done, a subgroup of 1 400 000 people above 30 years of age), we can measure
“adequate knowledge related to eye health” with 95% confidence, if 10% of the
sample has “adequate knowledge”.
Weighting
and Scores We applied
weighting arbitrarily based on the richness of information obtained through the
baseline surveys, as well as the “value” of accurate knowledge for each item,
with the latter based on clinical evidence. For example, “People with diabetes
should have their eyes examined once a year”, was assigned a maximum weight of
1.0, as it aligned with clinical guidelines, and “the treatment for diabetic eye
disease includes surgical treatments”, was assigned a relatively lower weight
of 0.25. The rationale was that it is of greater value to preserve the eye
health of an individual. That is, it is more important that a person with
diabetes knows they need an annual eye examination, relative to the value of
knowing what treatment options exist.
The scoring
system was applied to the responses to obtain the weighted summary scores for
each domain. Based on the questions and response options obtained from both
questionnaires, the maximum obtainable score in the “old” version of the
questionnaire for Knowledge was 7.50, and 3.00 for attitude, while the maximum
practice obtainable score with the “old” questionnaire was 9.00. The scoring
system and weighting of response items are depicted in Table 1.
Table 1
Scoring system for responses elicited by the EH-KAP tool
|
Knowledge |
|
|
Poor nutrition can cause decreased vision or
blindness |
0.25 |
|
Vitamin A deficiency can cause decreased vision
or blindness |
0.25 |
|
Sunlight can cause decreased vision or blindness |
0.25 |
|
Complications of other diseases can cause
decreased vision or blindness |
0.25 |
|
Decreased vision or blindness can be genetic |
0.25 |
|
Eye injury can cause decreased vision or
blindness |
0.25 |
|
Old age can cause decreased vision or blindness |
0.25 |
|
Cataract can cause decreased vision or blindness |
0.25 |
|
Trachoma/Trichiasis can cause decreased vision or
blindness |
0.25 |
|
Corneal ulcers can cause decreased vision or
blindness |
0.25 |
|
Glaucoma can cause decreased vision or blindness |
0.25 |
|
Diabetic retinopathy can cause decreased vision
or blindness |
0.25 |
|
Age-related macular degeneration can cause
decreased vision or blindness |
0.25 |
|
Other variations describing a known cause of
decreased vision or blindness |
0.25 |
|
Poor vision or blindness can be prevented |
1 |
|
Red-eye can be prevented by avoiding sharing
towels and other personal items |
0.25 |
|
Red-eye can be prevented by washing hands |
0.25 |
|
Red-eye can be prevented by avoiding touching or
rubbing infected eyes |
0.25 |
|
Other variations describing a known method of
preventing “red-eye” |
0.25 |
|
Red-eye can be treated by washing eyes |
0.25 |
|
Red-eye can be treated by eye drops |
1 |
|
Other variations describing a known method of
treating “red-eye” |
0.25 |
|
The best way to prevent or treat blurred vision
or refractive error is avoiding reading or watching television |
0.25 |
|
The best way to prevent or treat blurred vision
or refractive error is to use spectacles or contact lenses |
1 |
|
The best way to treat cataract is to protect eyes
from sunlight when outdoors |
0.25 |
|
The best way to treat cataract is to stop smoking |
0.25 |
|
The best way to treat cataract is surgical
removal |
1 |
|
The best way to treat cataract is to use
spectacles |
0.25 |
|
Other variations describing a known method of
treating cataract |
0.25 |
|
The best way to prevent trachoma is facial
cleanliness |
0.25 |
|
The best way to prevent trachoma is environmental
hygiene |
0.25 |
|
Other variations describing a known method of
preventing trachoma |
0.25 |
|
Ways to treat trachoma/Trichiasis are antibiotics |
0.25 |
|
Ways to treat trachoma/Trichiasis are surgery to
correct trichiasis |
0.25 |
|
Other variations describing a known method of
treating trachoma / trichiasis |
0.25 |
|
Ways to prevent corneal ulcers are by using eye
protection in high risk situations |
1 |
|
Ways to prevent corneal ulcers are by hygienic
use of contact lenses |
0.25 |
|
Other variations describing a known method of
preventing corneal ulcers |
0.25 |
|
Ways to treat corneal injury and ulcers are by
instilling eye drops |
0.25 |
|
Ways to treat corneal injury and ulcers are by
getting immediate medical attention |
1 |
|
Other variations describing a known method of
treating corneal ulcers |
0.25 |
|
Attitude |
|
|
We seek treatment when a member of my family or I
have an eye problem |
1 |
|
If we did not go for treatment, this could be
because doctor advised that immediate treatment is not required |
0.25 |
|
If we did not go for treatment, this could be
because we are waitlisted for scheduled treatment |
0.25 |
|
If we did not go for treatment, this could be
because of being medically unfit for surgery |
0.25 |
|
Other variations describing a valid reason for
not undergoing treatment |
0.25 |
|
People with a disability get appropriate care |
1 |
Table 1
(Continued)
|
Knowledge |
|
|
A child who has a disability can go to a normal school |
1 |
|
A child who has a disability can go to a special school |
0.25 |
|
Practice |
|
|
I get my eyes checked more than once a year |
0.5 |
|
I get my eyes checked at least once a year |
1 |
|
I get my eyes checked whenever I have a problem (less than once a
year) |
0.25 |
|
I sought treatment for my eye problem |
1 |
|
To treat eye problems, I go to the commune health station |
0.25 |
|
To treat eye problems, I go to the village health worker |
0.25 |
|
To treat eye problems, I go to the district hospital eye unit |
0.25 |
|
To treat eye problems, I go to the provincial eye hospital |
0.25 |
|
To treat eye problems, I go to a private doctor/hospital |
0.25 |
|
To treat eye problems, I go to the opticals/glasses shop |
0.25 |
|
Other variations describing a valid eye care service provider |
0.25 |
|
I have a health insurance card |
0.5 |
|
I use my health Insurance to obtain subsidies for treatment of eye
problems |
1 |
|
To protect my eyes, I use clean water to wash my face frequently |
0.25 |
|
To protect my eyes, I avoid sharing personal items like towels |
0.25 |
|
To protect my eyes, I wear eye protection when working or driving |
0.25 |
|
To protect my eyes, I wear sunglasses when outdoors |
0.25 |
|
To protect my eyes, I wear a hat when outdoors |
0.25 |
|
To protect my eyes, I get my eyes checked periodically |
0.25 |
|
To protect my eyes, I eat health and vitamin A-rich foods |
0.25 |
|
Other variations describing a valid practice that protects the eyes |
0.25 |
|
Modified practice |
|
|
I have had the inside of my eyes examined for diabetic eye disease |
1 |
|
Diabetes Mellitus knowledge |
|
|
People with diabetes should have their eyes examined once a year |
1 |
|
People with diabetes should have their eyes examined once in six
months |
0.5 |
|
Modified attitude |
|
|
A person with diabetes needs an eye examination even if his/her blood
sugar levels are stable |
1 |
|
The treatment for diabetic eye disease includes controlling blood
sugars |
0.25 |
|
The treatment for diabetic eye disease includes laser |
0.25 |
|
The treatment for diabetic eye disease includes surgical treatments |
0.25 |
|
Modified knowledge |
|
|
I use my glasses (or contact lenses) for near vision only |
0.25 |
|
I use my glasses (or contact lenses) only for viewing distant objects |
0.25 |
|
I use my glasses (or contact lenses) only for viewing both near and
distance |
0.25 |
|
Modified practice |
|
|
My glasses were prescribed at the government hospital |
0.25 |
|
My glasses were prescribed at a private hospital/clinic |
0.25 |
|
My glasses were prescribed at the opticals/glasses shop |
0.25 |
|
My glasses were purchased at the eye hospital |
0.25 |
|
My glasses were purchased at the district eye unit |
0.25 |
|
My glasses were purchased at the opticals/glasses shop |
0.25 |
|
My glasses were purchased at the market (or at another shop except a
specific spectacles shop) |
0.25 |
|
My glasses were not purchased by me; someone donated the glasses or
purchased them for me |
0.25 |
|
Other variations describing a valid person/place from where glasses
were purchased |
0.25 |
|
When I wear my glasses (or contact lenses) I can see more clearly |
1 |
Data
Management and Statistical Analysis Two Vietnamese health personnel were
trained to enter data into the customised form developed in Microsoft Access.
Data were cleaned, and where possible, missing data were traced from the
original questionnaires. Analysis was performed in STATA 8.2 (StataCorp.
College Station, Texas, USA).
Criterion
validity was assessed by comparing descriptive analyses of the two datasets
obtained from Thanh Hoa. The first dataset had been elicited using the “old”
questionnaire, and the second using the tool that was being tested
(congruency). Kappa values were calculated from the results of the test-retest
and inter-observer reliability testing using SPSS (version number 10, IBM,
Chicago, USA).
RESULTS
A total of
38 respondents were included in this validation process. Of the sixteen
respondents that could accurately provide their age, eight were over 50 years of age
(four men and four women), and eight between 30 to 50 years of age. The mean
age was 58.5 years (95%CI 52.9-64.2). The four representatives of each gender
from each age stratum included individuals with a
seeing/hearing/mobility/understanding disability and at least one person who
had type 2 diabetes mellitus, as well as individuals who had no formal
education and were illiterate, individuals who had completed some education but
not beyond primary school, and individuals who had completed high school (at
least 12y of formal education).
Following a
construct and face validity assessment, the questions were grouped together and
arranged in a logical sequence. Any potentially redundant questions or
responses were identified, and questions were modified based on the data and
where data was lacking, modification were based on a review of literature. The
“old” version of the questionnaire had 100 questions for “service users” and
collected information on demographics, knowledge of common eye diseases,
attitude of the interviewee regarding eye diseases, attitude of the
interviewees towards people with disabilities, practice related to management
and prevention of eye problems, knowledge and practice of prevention and
treatment of children’s eye diseases, and accessibility and affordability of
eye care services. Following the revision process, the tool contained 50
questions, collecting: “general information” including demographic details (10
questions), 13 questions pertaining to the domain that measured the construct
knowledge, 7 questions for attitude, and 8 for practice.
Two sets of
“additional questions” were included: 8 questions pertaining to individuals who
self-reported having diabetes mellitus, and 4 questions for individuals who
reported having required correction of refractive error. The revised tool
targeted only service users, and
took approximately 30min to administer. A descriptive analysis of the three
attributes measured in the two datasets, collected from the same sampling frame
at different points in time is presented for comparison. The first dataset was
acquired using the “old” unmodified version of the questionnaire and the second
dataset of the modified tool. The constructs (KAP) are referred to as “domains”
once they have been measured through assigning scores to responses for
questions relating to each construct.
Based on the
same scoring system (Table 1), using the responses from the revised tool, the
maximum score obtainable for knowledge was 14.75, and the minimum score was
0.25. For attitude, the maximum and minimum scores were 4.25 and 0.25
respectively, and for practice the scores obtainable were 8 and 0.25
respectively. The section titled “additional questions”, comprised of questions
directed to individuals who self-reported having diabetes mellitus, and to
people who reported having been advised spectacle correction. The knowledge
domain for persons with diabetes had a maximum obtainable score of 1.0 for
knowledge; attitude had 1.5 and practice had 1.75. All three domains had a
minimum obtainable score of 0.25. The sub-group that had been advised spectacle
correction responded to questions related to knowledge (maximum obtainable
score 0.75) and practice (maximum obtainable score 3.25).
To assess
reliability, “raw data” were analysed across the entire range of 212 response
options that comprised the revised tool. Subsequently, individual responses were
weighted by applying scores, and agreement was tested across the eight domain
scores. Test-retest and inter-observer agreement results are depicted in Table
2.
Table 2
Results of test-retest and inter-observer agreement
|
Group |
Raw data |
Domain scores |
||
|
Agreement (%) |
Kappa value (SE) |
Agreement (%) |
Kappa value (SE) |
|
|
Interviewer
test-retest agreement |
|
|
|
|
|
Interviewer 1 |
87.03 |
0.71 (0.07) |
54.55 |
0.42 (0.10) |
|
Interviewer 2 |
90.43 |
0.73 (0.06) |
54.55 |
0.46 (0.10) |
|
Interviewer 3 |
88.59 |
0.74 (0.07) |
54.55 |
0.42 (0.10) |
|
Interviewer 4 |
89.73 |
0.80 (0.07) |
45.45 |
0.37 (0.09) |
|
Interviewer 5 |
85.56 |
0.72 (0.06) |
54.55 |
0.50 (0.09) |
|
Interviewer 6 |
87.50 |
0.74 (0.07) |
63.64 |
0.53 (0.11) |
|
Interviewer 7 |
92.97 |
0.86 (0.07) |
63.64 |
0.53 (0.11) |
|
Interviewer 8 |
93.48 |
0.86 (0.07) |
63.64 |
0.53 (0.11) |
|
Interviewer 9 |
91.44 |
0.84 (0.07) |
63.64 |
0.59 (0.10) |
|
Interviewer 10 |
88.83 |
0.78 (0.06) |
63.64 |
0.59 (0.10) |
|
Respondent
inter-observer agreement |
|
|
|
|
|
Respondent 1 |
84.32 |
0.66 (0.06) |
72.73 |
0.65 (0.12) |
|
Respondent 2 |
61.70 |
0.24 (0.06) |
63.64 |
0.59 (0.10) |
|
Respondent 3 |
73.51 |
0.39 (0.06) |
72.73 |
0.65 (0.12) |
|
Respondent 4 |
77.54 |
0.52 (0.06) |
54.55 |
0.47 (0.10) |
|
Respondent 5 |
75.68 |
0.52 (0.06) |
63.64 |
0.53 (0.11) |
|
Respondent 6 |
85.33 |
0.70 (0.07) |
54.55 |
0.42 (0.11) |
Validity was
further assessed to determine the extent to which the tool measured constructs
it intended to measure. This was achieved through comparing the results
obtained using the “old” questionnaire on the larger group, with the results of
administering the “revised” tool to a small sub-group within the sampling frame
of the larger KAP study conducted in Thanh Hoa/Nghe An Provinces (2010, n=531).
Internal
validity was assessed by comparing scores obtained by the two groups for each
of the three constructs–knowledge, attitude and intended practice. The mean
scores for knowledge were 9.15 (n=38, 95%CI 8.47 to 9.83) and 5.05 (n=531,
95%CI 4.92 to 5.17) in the dataset elicited by the smaller and larger groups
respectively. For attitude, the scores were 2.23 (n=38, 95%CI 2.03 to
2.43) and 2.42 (n=531, 95%CI 2.37 to 2.47), and for practice, the scores
were 3.33 (n=38, 95%CI 2.95 to 3.71) and 2.21 (n=531, 95%CI 2.10
to 2.31).
The overall
total scores for the KAP survey were 15.23 (n=38, 95%CI 14.18 to 16.29)
for the validation sub-sample and 9.68 (n=531, 95%CI 9.48 to 9.87) for
the larger sample.
Power and
Effect Size Back-calculating
from the KAP survey previously conducted in Than Hoa and Nghe An Provinces, it
was evident that by surveying a sample of 530 people from a sampling frame comprising
a general population of 4 000 000 (or, if stratified sampling was done, a
sub-group of 1 400 000 people above 30 years of age), “adequate knowledge” may be measured
with 95% confidence.
DISCUSSION
This
research followed a rigorous scientific process to validate a standardised KAP
tool as a component of a comprehensive eye-health field assessment. The
approach addressed inconsistencies in methodology and weaknesses in previous
KAP surveys. The EH-KAP tool proved to be a reliable and valid instrument to
assess KAP related to eye health at community level in Vietnam. Using this tool
alongside other cross-sectional survey such as the Rapid Assessment for
Avoidable Blindness (RAAB) surveys may provide valuable additional information
necessary for successful healthcare planning and implementation[4].
Health-related
behaviour is most often assessed through KAP surveys, though other research
methods may be used either alongside a social survey or as an alternative to it[5-6]. KAP surveys can provide
information useful to a range of stakeholders and informs health promotion
campaigns and strategies, the development of health promotion materials. KAP
surveys are also a method of assessing effectiveness of interventions to
improve health and can hence inform health policy[7]. When used to understand health
behaviour, data collected from a KAP survey complements, and in fact often
leads to deeper investigation through focus group discussions or in-depth
interviews[1].
One of the key characteristics of research into KAP is to employ appropriate
methodology: though a KAP survey provides a standardized ecological overview,
it may only uncover the first layer of “truth” and deeper investigation through
qualitative research is often essential. This is possibly one reason for the
paucity of validated tools to study KAP, and why psychometric testing is rarely
performed or reported[1,7-8]. Further, being population- or
site-specific by nature, generalizability of data obtained through a KAP survey
is inevitably limited. The questions in this tool are intentionally open-ended
to ensure richness of data, though future researchers may modify response
options for ease of administration by the interviewers, and to the local
context.
In the field
of eye care, KAP studies have primarily been used to explore behaviour in well-defined
groups of individuals, often in relation to a specific disease condition or
intervention[9].
At the time of commencing this validation, a review of published literature
revealed no report of a KAP survey having been conducted at baseline to inform
the delivery of eye care interventions in the target population. Consequent to
the global initiative to eliminate avoidable blindness (VISION 2020: The Right
to Sight), systematic assessments of eye and vision related morbidity have
formed the basis of planning national programmes for elimination of avoidable
blindness.
Though there
is evidence that barriers exist to the uptake of eye-care services among
specific groups[10-18], yet no
systematic tool is available to assess these barriers at baseline to inform
programme planning, and at end-line to assess effectiveness of the programme in
overcoming these barriers. Applying a standard measure throughout the planning
cycle will result in more efficient and effective delivery of eye-care services
to a target population.
Development
of a questionnaire, instrument or tool for data collection de novo involves a
series of processes to ensure that the resulting tool is valid, reliable and responsive
to change[19-20]. To revise this
KAP questionnaire, prior experience from both the Cambodia KAP survey (n=599)
and from the surveys conducted in Son La (n=300) Thanh Hoa and Nghe An (n=531)
Provinces of Vietnam were built upon by adding/rephrasing questions and
expanding response options through a review of relevant literature. In an
effort to ensure that the phrasing of questions was acceptable, the generic
English tool was translated and tested repeatedly. This was done to maximize
validity and to eliminate potential response bias resulting from racial or
ethnic cultural experience or the lack of it[21-23]. The result is a simple, standardized and focused tool
that can be applied repeatedly over time (3-5y) to assess trends in KAP within
the same population and for comparison across groups. Though the scoring system
applied was the same, the total scores obtainable are different for each of the
two questionnaires. This is partly because in the old questionnaire, though the
number of individual items scored was more numerous, several questions measured
a single trait (redundancy), and some questions that were included have in fact
been proven to be inappropriate[24]. However, narrow confidence intervals around total
scores suggest that precision has not been compromised, and the larger scores
obtainable with the revised questionnaire possibly leaves room for greater responsiveness
to change.
A good tool
would be valid, reliable and sensitive to change. Construct validity
(congruent) and responsiveness to change can be assessed from data collected at
sequential time-points during the project cycle[25]. It is also essential that for
construct validity of a tool to be demonstrated, the domain scores or more
realistically in this context, scores obtained to individual questions (item
scores) must correlate with related verifiable variables. For example, reported
utilization of health care facilities may be verified using hospital records[26]. From the survey
reported here and the previous (2009) KAP survey in the same Province, it may
be surmised to a certain extent that the data obtained adequately reflects
actual KAP within the local community. This assumption was based on informal
discussions and feedback obtained from both the respondents and the
interviewers during the process of pre-testing and revising the tool. However,
identifying and measuring variables such as these concurrently is not feasible
in the present context. What remains unclear and needs consideration when using
tools such as this one, is the possible impact of non-sampling errors on
response reliability[27].
Though such errors have may be insignificant, in certain cultures such as in
Vietnam, contextual issues such as household-level effects may influence
health-seeking behaviour[28].
For fully informed planning of an eye care programme to serve a given
population, rapid assessments such as the RAAB and EH-KAP surveys can be
supplemented with audits and inventories of available manpower and materials
(human resource capability, infrastructure and equipment) from the perspective
of service providers[29].
This survey instrument used alongside a RAAB,
provides a comprehensive perspective on the eye-health in a population, an
inference deduced also in Cambodia[3,30-31].
ACKNOWLEDGEMENTS
The authors
thank the following individuals for their contribution and support in Cambodia:
Anna-Lena Arnold and Professor Jill Keeffe (CERA) who coordinated the overall
survey, Te Serey Bonn (Caritas Takeo Eye Hospital) who helped with data
collection and translation into Khmer and into English, and Dr Sarin Sun,
Evangeline Dunton, Chim Thoeun, Som Putvibol, Vong Chrean, Ken Punlork, Chrek
Vannak, Chan Sam Nang, Koy Chanthoeun and Seng Puthisey who assisted with data
collection. The authors acknowledge the valuable contribution made by the
leaders and staff of Thanh Hoa Eye Hospital and Nghe An Eye Centre and Nathalie
Maggay, Nguyen Hong Nga, Nguyam Ngoc Anh from CBM Vietnam; and the services of
PeaPROs which conducted the initial KAP survey.
The authors
acknowledge the support from the Centre for Eye Research Australia (CERA),
Royal Victorian Eye & Ear Hospital, Melbourne.
Foundations:
Supported by the Australian Government Avoidable
Blindness Initiative granted through CBM Australia.
Conflicts of
Interest: Prasad N, None; Ormsby
GM, None; Finger RP, None.
REFERENCES