The Extent of Chronic Wounds in Canada: What We Know and What We Don't
Know
by M. Gail Woodbury and Pamela E. Houghton
Suggestions for Minimizing Bias in Prevalence Studies
SUBJECTS
Whole Population: For example, all patients within a health care
setting. Avoid clinic audits with specific referral patterns or exclusion
of facility units.
Random Sample: For example, the population of a region or community—this
must be done using accepted methods, which are described. Avoid convenience
samples, i.e., sampling of only patients on health service in community.
METHOLOLOGY
Study Design: Use a cross-sectional design, a one-time "snap-shot."
Data Source: Direct clinical examination
Collection Period: Allow sufficient time to assess whole or sample
population, but not too long to miss individuals due to discharge from
the facility or service—may take longer in community: acute care
— 1 day, LTC — 1 day to 1 week, community agency —1 week
to 1 month
Measurement of Outcome: Use standard measures to characterize ulcer
severity or ulcer etiology
• Pressure ulcers — NPUAP staging. Stage 1 = persistent non-blanchable
erythema
• Venous ulcers — use ankle brachial pressure index (ABPI)
• Neuropathic ulcers — use monofilaments sensory tests
Use two observers working independently; one should not be a staff
member.
Check reliability of outcome measurement and report agreement.
Sample Size: The following sample sizes are recommended: Acute
care – 300, LTC – 330, Community agencies – 200. These
sample sizes are large enough to detect prevalence of 25%, 30% and 15%
respectively, with 95% CI of 10%.
PREVALENCE STUDY REPORTING
Design and Methodology: Describe the study design and methodology
using the terms subject inclusion/exclusion, data collection procedures
(who, what, where, when, how), and measurement of outcome as outlined
above.
Response Rate: Indicate the number of people who participate in
the study (potential number minus those who do not respond or consent)
divided by the number in the total potential population — should
be at least 70%.
Describe the group of people who refuse to participate to indicate that
those who participated were representative of the whole population.
Prevalence Estimate and 95% Confidence Interval (CI): Report prevalence
as %. 95% CI is a measure of sample variability. The larger the sample
size, the narrower the CI. 95% CI = the range of values within which one
can be 95% confident that the true prevalence estimate lies.
Subgroup Analysis: Provide prevalence estimates (% people in population)
for subgroups (age, stage, setting, region, etc.). Avoid descriptions
expressed as proportion of total number of wounds examined.
Population Demographics: Describe the population that has been
investigated (mean age, sex, medical status) so that others can compare
their population with yours.
Suggestions for Minimizing Bias in Incidence Studies
SUBJECT INCLUSION/EXCLUSION
Whole Population: For example, all patients within a health-care
setting. Avoid clinic audits with specific referral patterns or exclusion
of facility units.
Random Sample: For example, the population of a region or community
—must be done using accepted methods, which are described. Avoid
convenience samples, sampling of only patients on health service in community.
Consider whether incidence should be measured in people who are wound-free
at the beginning of the study or should people who already have a wound
(who are at risk of developing another wound) be included.
METHODOLOGY
Study Design: Incidence is measured by identifying a group of people
who are at risk of developing chronic ulcers and examining them periodically
to discover and count new cases that develop during that interval.
A prospective cohort design should be used.
Data Sources: Direct observation
Collection Period: Often the collection period is related to length
of stay in a health-care setting: acute care – 5 days, LTC –
1 month, community agency – 2 months; conversely, the collection
period could be a year to obtain the annual incidence.
Measurement of Outcome: Count the number of persons who develop
new wounds. Persons with more than one wound should be counted only once.
Use standard measures to characterize ulcer severity or ulcer etiology
(see Prevalence). Use two observers working independently; one should
not be a staff member. Check reliability of outcome measurement and report
agreement.
Sample size: A sample size of 250 is sufficient to detect incidence
of less than 25% with 95% CI of 10%.
INCIDENCE STUDY REPORTING
Design and Methodology: Describe the study design and methodology
in terms of subject inclusion/exclusion, data collection procedures, and
measurement of outcome as outlined above.
Response Rate: Indicate the number of people who participate in
the study (potential number minus those who do not respond or consent)
divided by the number in the total potential population — should
be at least 70%.
Describe the group of people who refuse to participate to indicate that
those who participated were representative of the whole population.
Incidence Estimate and 95% Confidence Interval: A measure of sample variability.
The larger the sample size, the narrower the CI. 95% CI = the range of
values within which one can be 95% confident that the true prevalence
estimate lies.
Subgroup Analysis: Provide prevalence estimates (% people in population)
for subgroups (age, stage, setting, region, etc.). Avoid descriptions
expressed as proportion of total number of wounds examined.
Population Demographics: Describe the population that has been
investigated (mean age, sex, medical status) so that others can compare
their population with yours.
|
|
|
