394 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2010 vol. 34 no. 4© 2010 The Authors. Journal Compilation © 2010 Public Health Association of Australia
Comparing self-reported and measured high blood
pressure and high cholesterol status using data from
a large representative cohort study
Abstract
Objective: To examine the relationship
between self-reported and clinical
measurements for high blood pressure
(HBP) and high cholesterol (HC) in a
random population sample.
Method: A representative population sample
of adults aged 18 years and over living in
the north-west region of Adelaide (n=1537)
were recruited to the biomedical cohort
study in 2002/03. In the initial cross-sectional
component of the study, self-reported HBP
status and HC status were collected over
the telephone. Clinical measures of blood
pressure were obtained and fasting blood
taken to determine cholesterol levels. In
addition, data from a continuous chronic
disease and risk factor surveillance system
were used to assess the consistency of self-
reported measures over time.
Result: Self-report of current HBP and
HC showed >98% specificity for both, but
sensitivity was low for HC (27.8%) and
moderate for HBP (49.0%). Agreement
between current self-report and clinical
measures was moderate (kappa 0.55)
for HBP and low (kappa 0.30) for HC.
Demographic differences were found with
younger people more likely to have lower
sensitivity rates. Self-reported estimates
for the surveillance system had not varied
significantly over time.
Conclusion: Although self-reported
measures are consistent over time
there are major differences between the
self-reported measures and the actual
clinical measurements. Technical aspects
associated with clinic measurements could
explain some of the difference.
Implications: Monitoring of these broad
population measures requires knowledge
of the differences and limitations in
population settings.
Key words: hypertension, cholesterol,
data collection, validity, cross sectional
survey.
Aust NZ J Public Health. 2010; 34:394-400
doi: 10.1111/j.1753-6405.2010.00572.x
Anne Taylor, Eleonora Dal Grande, Tiffany Gill Population Research and Outcomes Studies Unit, South Australia Health
Sandra PickeringHealth Observatory , Department of Medicine, University of Adelaide, The Queen Victoria Hospital, South Australia
Janet GrantPopulation Research and Outcomes Studies Unit, South Australia Health
Robert AdamsHealth Observatory , Department of Medicine, University of Adelaide, The Queen Victoria Hospital, South Australia
Patrick PhillipsEndocrine and Diabetes Service, University of Adelaide, The Queen Victoria Hospital, South Australia
Hy p e r t e n s i o n a n d h y p e r -
cholesterolemia are two of the most
important factors in the prevention
and control of diabetes, cardiovascular disease
and metabolic syndrome. Determining
population measures of high blood pressure
(HBP) and total high cholesterol (HC)
using clinical measurements is an expensive
and time-consuming undertaking. Most
commonly these measurements are gained,
at the population level, by surveys, with
respondents self-reporting their HBP/HC
status based on recall as to whether they
had been told by a doctor or nurse they had
high readings.
Many chronic disease and risk factor
surveillance systems rely on self-report as the
basis of the data collection.1-3 These systems
provide an important service in highlighting
prevalence estimates for health planners,
policy makers and health promotion experts.
Although acknowledgement is made that
Submitted: July 2009 Revision requested: November 2009 Accepted: February 2010Correspondence to:Dr Anne Taylor, Population Research and Outcomes Studies Unit, SA Dept of Health, Level 8, CitiCentre Building, 11 Hindmarsh Sq, Adelaide, South Australia 5000. Fax: (08) 8226 6244; e-mail: [email protected]
the values are only ‘estimates’ of the true
prevalence, they are important as monitoring
indicators when population studies based
on actual clinically controlled studies
are undertaken with considerable gaps
between years of collection. It is therefore
important that the self-reported estimates
are as accurate as possible, of the highest
standard, and the differences are known. One
way of assessing the appropriateness of the
self-reported data is by validity studies with
criterion validity studies the most highly
relevant.
Studies overseas have determined
the agreement between self-reported
hypertension and/or cholesterol against actual
measurements using three methodologies.
The first group have used administrative data
(use of health service or personal medical
records) to assess the presence of HBP or
HC against self-reported measures.1-5,9,10
Sensitivity (true positives), specificity (true
Chronic illnesses Article
2010 vol. 34 no. 4 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 395© 2010 The Authors. Journal Compilation © 2010 Public Health Association of Australia
negatives) and the predictive value of a positive or negative answer
were calculated to evaluate the utility of these data. Studies have
reported sensitivity rates for HBP varying between 73% and 83%
and specificity between 81% to 95%.5-7 For HC, sensitivity ranges
were lower and varied from 59% to 64% with specificity ranging
from 84% to 93%.5,9,10
Secondly, studies have assessed the overall agreement of these
two variables/indicators using self-completed questionnaires
followed by clinical examinations using non-random, convenience
samples or targeted racial/ethnic groups.11-14 The only study to
assess sensitivity reported 87% for persons aged between 18 and
74 years for HBP.14 Specificity was also high (96%).
There are limited contemporary studies that have assessed
clinically measured HBP and/or HC against self-reported from
large random population samples. Natarajan et al.15 using
NHANES III (1988-1991) data, has reported sensitivity of 51%,
specificity of 89% (and Predictive Value Positive (PVP) 87% and
Predictive Value Negative (PVN) 55%) for HC. Vargas et al.16 used
the NHANES III (1988-1994) data set reported a sensitivity of
71%, specificity of 90%, for HBP with a PVP of 72% and PVN
of 89%. Giles et al.17 in a major regional US study, had similar
findings for HBP although there was some variation between
white and black respondents. In a 1993 assessment of BRFSS
respondents, values were overall lower for both HBP and HC
(sensitivity of 43% for HBP and 44% for HC, and specificity >
85% for both).18
To ascertain the agreement between these self-reported measures
in an Australian setting, data from the North West Adelaide
Health Study (NWAHS) were utilised. The data were also used
to undertake comparisons between different demographic groups.
The analysis was based on previous research assessing the
difference between self-reported and actual height and weight.19
In addition, data from the South Australian Monitoring and
Surveillance System (SAMSS), a continuous chronic disease and
risk factor system based on telephone interviews, were used to
assess the consistency of these self-reported measures over time.
MethodologySurvey samples
NWAHS is a representative biomedical cohort study of adults,
aged 18 years and over at the time of recruitment, living in the
north and west regions of Adelaide.20,21 The study is a collaboration
between SA Health, The Queen Elizabeth Hospital, the Lyell
McEwin Hospital, The University of Adelaide, the University
of South Australia and SA Pathology (formally the Institute
of Medical & Veterinary Science). Phase 1A of this study was
conducted primarily in 2000. Phase 1B of the study was conducted
between August 2002 and July 2003. Data used in these analyses
were obtained from Phase 1B. Details of the NWAHS design,
procedures and participants have been published elsewhere.20-22
Initially participants were recruited via a Computer Assisted
Telephone Interviewing (CATI) system based on a random
selection of telephone numbers in the Electronic White Pages
(EWP). The adult in the household with the last birthday was
selected for inclusion in the study. Interviews were conducted
in English. Prior to the recruitment telephone call, households
were sent an approach letter detailing the study. Respondents
successfully recruited into the study were sent a detailed
information pack and were asked to attend one of the two clinics in
the region at a time suitable to them. All protocols and procedures
were approved by the North Western Adelaide Health Service –
Ethics of Human Research Committee and the Aboriginal Health
Research Ethics Committee of South Australia.
Biomedical data were collected at the clinic. Participants were
asked if they were currently on medication for hypertension,
cholesterol or lipid lowering medication. These questions were
asked before blood pressure was measured and blood samples
drawn. Blood pressure was assessed by the average of two
measurements, recorded with a calibrated sphygmomanometer and
taken five to ten minutes apart while the participant was seated and
relaxed. HBP was defined as systolic blood pressure ≥140 mmHg
and/or diastolic blood pressure ≥90 mmHg.23,24
A fasting blood sample was drawn for glucose and lipid profile to
determine diabetes and cholesterol. HC was defined as total serum
plasma cholesterol ≥5.5 mmol/L.25 Other biomedical measures
and key findings are detailed elsewhere.26-29
Self-report data were collected in the initial recruitment
telephone call (including if they had ‘ever’ been told by a doctor
or nurse that they have HC or HBP, and if they still have HC or
HBP), and selected demographic information. Primary assessment
was made between clinical measurements and those reporting still
having HBP or HC. Additional comparisons were made between
those who self reported ‘ever’ having HBP or HC and the clinical
measurements.
To assess the consistency of the self-reported questions over
time, data from SAMSS (where the identical self-reported
questions on HBP and HC had been used since July 2002)
were used.30 All households in SA, with a number listed in the
EWP were eligible for selection in the SAMSS sample. A letter
introducing the surveillance system was sent to the household of
each selected telephone number. The letter informed people of the
purpose of the interview and indicated that they could expect a
telephone call within a certain time frame. Within each household,
the person who had their birthday last, aged zero years and over,
was selected for interview. Data were collected every month by
a contracted agency and interviews were conducted in English.
Proxy interviews were undertaken for respondents aged under 16
years. The response rate for SAMSS from July 2002 until June
2008 was between 65% and 70% each month. The Computer
Assisted Telephone Interview (CATI) system was used to conduct
the interviews. At least 10 call backs were made to the telephone
number selected to interview household members. Replacement
interviews for persons who could not be contacted or interviewed
was not permitted. Analysis for the comparison was limited to
SAMSS respondents aged 18+ years (n = 31,930).
Chronic illnesses Comparing self-reported and measured blood pressure and cholesterol
396 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2010 vol. 34 no. 4© 2010 The Authors. Journal Compilation © 2010 Public Health Association of Australia
Statistical analysesThe NWAHS data were weighted to reflect region, age groups,
sex and probability of selection in the household to the Australian
Bureau of Statistic’s 2001 Census population.31 SAMSS data were
weighted by age, sex and area of residence to reflect the structure of
the population in South Australia to the latest Census or Estimated
Residential Population. Probability of selection in the household
is calculated based on the number of people in the household and
the number of listings in the White Pages. Weighting is used to
correct for disproportionality of the sample with respect to the
population of interest.
Data were analysed using SPSS version 15.0. The conventional
p value of 5% was used as the criteria for statistical significance.
Comparisons between self-reported current HBP and HC and
biomedically measured HBP or HC were made in terms of
sensitivity, specificity, positive predictive value (PPV), and
negative predictive value (NPV). These comparisons were made
for age groups and gender and other relevant variables.
For the SAMSS data, graphic interpretation was presented
by two-way fractional-polynomial prediction plots using time
in months as the X variable. In addition, yearly estimates were
obtained by combining the relevant 12 months of data.
ResultsThe overall response rate for Phase 1B of the NWAHS study
was 47.1% with n=1,537 people participating in both the telephone
questionnaire and attending the clinic for biomedical assessments.
Clinic data collected were missing for n=12 (0.8%) on blood
pressure and cholesterol. Analysis was undertaken only on those
in which both self-reported and clinical measured details were
available (n=1525; 99% of total sample).
Table 1 shows the demographic profile of the NWAHS
respondents used in the analyses. Previous analyses have compared
NWAHS study with ABS census data to assess bias.22
From the NWAHS CATI interview, 23.9% (95% CI 21.8 – 26.1)
of the respondents reported having ‘ever’ been told they had HBP
and 15.8% (95% CI 14.1 - 17.8) reported they still had HBP. In
addition, 1% of respondents reported not knowing if they ever had
HBP or had never had their blood pressure measured. Of those
who reported previously being told they had HBP, 6.1% reported
not knowing if they still had HBP. At the clinic, 30.6% (95% CI
28.3 - 32.9) had HBP and/or were currently taking medication for
their blood pressure.
Overall 23.8% (95% CI 21.8 – 26.0) reported having ‘ever’
been told they had HC and 12.3% (95% CI 10.8 – 14.1) reported
they still had HC. In addition, 3.3% of respondents reported not
knowing if they ever had HC or had never had their cholesterol
measured. Of those who reported previously being told they had
HC, 27.2% reported not knowing if they still had HC. At the clinic
42.8% (95% CI 40.4 – 45.3) had HC and/or were currently taking
medication for their cholesterol.
Table 2 shows the prevalence of self-reported (current) and
clinical measured HBP and HC by selected demographic and
health risk factors and detail the sensitivity, specificity, PPV and
NPV for these measures. The trend analysis highlighted in Figures
1 and 2 using SAMSS data shows the consistency of the self-report
current HBP and HC with little variation over the years and with no
significant increase or decrease over time. The comparable figures
for HBP and HC in 2003 (12 months of SAMSS data combined)
were 16.3% and 12.8%, a non-significant difference from the
NWAHS self-reported figures of 15.8% and 12.3% (2002/03).
When analysis was undertaken comparing the self-reported
‘ever’ diagnosed with the clinical measurements, sensitivity
and specificity were 49.3% and 91.6% for HBP and 48.5% and
94.6% for HC. Kappas were 0.54 (p<0.001) and 0.46 (p<0.001)
respectively.
Table 1: Demographic profile of NWAHS respondents.
n %
Gender
Male 749 49.1
Female 776 50.9
Age groups
18 to 24 185 12.1
25 to 44 603 39.5
45 to 64 447 29.3
65 years and over 290 19.0
Country of birth
Australia 1,103 72.3
UK or Ireland 209 13.7
Europe 119 7.9
Asia 54 3.5
Other 29 1.9
Not stated 11 0.7
Marital status
Married or living with partner 958 62.8
Separated or divorced 128 8.4
Widowed 89 5.8
Never married 336 22.0
Not stated 14 0.9
Employment status
Full time employed 596 39.1
Part time / casual employment 306 20.0
Unemployed 48 3.1
Home duties 182 11.9
Retired 267 17.5
Student 71 4.6
Other 39 2.6
Not stated 16 1.1
Educational attainment
Secondary 633 41.5
Trade, Apprenticeship, Certificate, Diploma 621 40.7
Bachelor degree or higher 178 11.7
Other/Not stated 93 6.1
Gross household annual income
up to $20,000 309 20.2
$20,001-$40,000 352 23.1
$40,001-$60,000 322 21.1
$60,001+ 437 28.7
Not stated 105 6.9Total 1,525 100.0
Taylor et al. Article
2010 vol. 34 no. 4 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 397© 2010 The Authors. Journal Compilation © 2010 Public Health Association of Australia
DiscussionThe variation between current self-reported and clinically
measured HBP and HC seems considerable with the overall
difference between prevalence estimates of 14.8% for HBP and
30.5% for HC. There were some differences by sex, age group
and other key descriptors between those reporting current HBP/
HC and measured HBP/HC. Kappa values were generally in
the medium range. In terms of sensitivity, rates varied but were
generally about 50% (indicating that approximately 50% of true
cases were detected by self-report). Specificity was high (98+%)
indicating that most people who self-reported they did not have
current HBP or HC were found to not have a high reading in the
clinic environment. These analyses have shown that in a clinic
setting many more people are determined to have HBP or HC
than those reporting a current high measurement. These results are
similar to those reported for BRFSS18 but less than other similar
studies. Methodological, cultural, timing and guideline differences
make direct comparisons difficult. The SAMSS data show that the
self-report prevalence estimates are consistent and have not varied
considerably over time. Previous CATI field testing has shown
excellent reliability for both HBP and HC.32
These variations are less dramatic if the ‘ever’ estimates are
used with differences of 6.7% for HBP and 19.0% for HC. People
who self-report ‘no’ to current HBP/HC could be on medication
and rightly report that they currently did not have high readings.
In the clinic setting respondents who were on medication were
classified as having the condition.
There are many reasons for the differences between current
self-reported and clinic measurements besides the obvious
‘condition under-control’ scenario. Technical aspects associated
with measuring HBP could be a factor. Although the measurement
of BP within the NWAHS clinic was undertaken to best practice
Table 2: Proportion of participant with current high blood pressure from self-reported CATI recruitment interview and clinical measurements and sensitivity, specificity, positive, negative predictive values.
Prevalence Self-reported Measured Percentage Kappa P-value Sensitivity Specificity Positive Negative difference predictive predictive value value
Current high blood pressure Overall 15.8 30.6 -14.8 0.55 <0.001 49.0 98.8 94.6 81.5
Gender
Male 14.6 32.4 -17.8 0.47 <0.001 41.8 98.5 92.9 77.9
Female 17.0 28.8 -11.8 0.64 <0.001 56.9 99.1 96.1 85.0
Age groups
18 to 44 2.3 8.4 -6.1 0.26 <0.001 18.1 99.1 66.0 93.0
45 to 64 20.3 39.9 -19.6 0.49 <0.001 47.5 97.8 93.4 73.8
65 years and over 46.0 76.9 -30.9 0.40 <0.001 59.5 98.7 99.4 42.2
BMI categories
Normal/Underweight 8.3 15.7 -7.4 0.63 <0.001 51.2 99.6 96.2 91.7
Overweight 16.5 33.2 -16.7 0.54 <0.001 47.9 99.3 97.2 79.4
Obese 24.5 45.5 -21.0 0.47 <0.001 49.3 96.2 91.5 69.4
Smoking status
Non smoker 17.5 32.4 -14.9 0.58 <0.001 51.8 99.0 96.0 81.1
Ex-smoker 19.2 38.1 -18.9 0.54 <0.001 49.4 99.3 97.7 76.1 Current smoker 6.2 13.7 -7.5 0.38 <0.001 31.8 97.9 70.1 90.1
0
1020
3040
Pre
vale
nce
(%)
July
2003
Janu
ary 20
04
July
2004
Janu
ary 20
05
July
2005
Janu
ary 20
06
July
2006
Janu
ary 20
07
July
2007
Janu
ary 20
08
July
2008
Month
Actual Predicted 95% CI
010
2030
40
July
2003
Janu
ary 20
04
July
2004
Janu
ary 20
05
July
2005
Janu
ary 20
06
July
2006
Janu
ary 20
07
July
2007
Janu
ary 20
08
July
2008
July
2003
Janu
ary 20
04
July
2004
Janu
ary 20
05
July
2005
Janu
ary 20
06
July
2006
Janu
ary 20
07
July
2007
Janu
ary 20
08
July
2008
Male Female
Actual Predicted 95% CI
Pre
vale
nce
(%)
Month
Graphs by sex
Figure 1: Overall and gender trends in proportion of adults (18+ years) with high blood pressure, 2003 to 2008
5
Figure 1: Overall and gender trends in proportion of adults (18+ years) with high blood pressure, 2003 to 2008.
Chronic illnesses Comparing self-reported and measured blood pressure and cholesterol
398 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2010 vol. 34 no. 4© 2010 The Authors. Journal Compilation © 2010 Public Health Association of Australia
standards, issues outside the control of the clinic staff could
affect measurement readings. These include the ‘white coat
syndrome’33,34 and the fact that an individual’s blood pressure
varies considerably across the day and from day to day.35 To
diagnose HBP, readings should be taken over more than one
visit; even two measurements taken five to 10 minutes apart does
not ensure a correct diagnosis with blood pressure measurement
obtained at a single point of time manifestly overestimating the true
prevalence.36 As argued by Marshall7 ‘measured blood pressure
is intrinsically variable because every cardiac cycle produces a
different blood pressure’. It is also argued that the diagnosis of
hypertension should be deferred for two years37; that 24-hour blood
pressure monitoring is required to properly diagnose HBP38; and
that there is difference depending upon which arm is used for the
measurement.39 For clinically meaningful diagnosis of HBP it
has also been argued that the use of a unified cuff size rather than
one tailored to BMI could produce differences in readings.14,40
Also compounding the self-report recall is the ‘arbitrary’ and
changeable cut-off points (e.g. 140/90) that has seen best practice
guidelines change over the years.38
For HC, confusion exists between the different components of
total cholesterol. Identification of cardiac risk factors including
HC and HBP, and subsequent effective management is relatively
infrequent in general practice.41 Identifying HC requires blood
testing, which may contribute to a relative under-diagnosis,
particularly in younger people where the level of cardio-metabolic
risk may not be as well appreciated by patients or clinicians.
Notwithstanding, the large differences between current self-
report and clinical measures, there are some good outcomes of this
study. The fact that both HBP and HC specificity levels were very
high (98%+) indicates that recall was not a problem. Self-report
failed to identify 18.5% of people found to have HBP in the clinic
010
2030
Pre
vale
nce
(%)
July
2003
Janu
ary 20
04
July
2004
Janu
ary 20
05
July
2005
Janu
ary 20
06
July
2006
Janu
ary 20
07
July
2007
Janu
ary 20
08
July
2008
Month
Actual Predicted 95% CI
010
2030
July
2003
Janu
ary 20
04
July
2004
Janu
ary 20
05
July
2005
Janu
ary 20
06
July
2006
Janu
ary 20
07
July
2007
Janu
ary 20
08
July
2008
July
2003
Janu
ary 20
04
July
2004
Janu
ary 20
05
July
2005
Janu
ary 20
06
July
2006
Janu
ary 20
07
July
2007
Janu
ary 20
08
July
2008
Male Female
Actual Predicted 95% CI
Pre
vale
nce
(%)
Month
Graphs by sex
Figure 2: Overall and gender trends in proportion of adults (18+ years) with high cholesterol, 2003 to 2008
6
Figure 2: Overall and gender trends in proportion of adults (18+ years) with high cholesterol, 2003 to 2008.
Table 3: Proportion of participant with current high cholesterol from self-reported CATI recruitment interview and clinical measurements and sensitivity, specificity, positive, negative predictive values.
Prevalence Self-reported Measured Percentage Kappa P-value Sensitivity Specificity Positive Negative difference predictive predictive value value
Current high cholesterol
Overall 12.3 42.8 -30.5 0.30 <0.001 27.8 99.2 96.4 64.7
Gender
Male 13.7 43.5 -29.8 0.32 <0.001 30.5 99.2 96.7 65.0
Female 11.0 42.2 -31.2 0.27 <0.001 25.0 99.3 96.1 64.5
Age groups
18 to 44 2.8 27.8 -25.0 0.12 <0.001 8.8 99.6 88.4 73.9
45 to 64 16.9 55.6 -38.7 0.24 <0.001 28.7 97.9 94.4 52.2
65 years and over 31.5 64.3 -32.8 0.41 <0.001 49.0 100.0 100.0 52.1
BMI categories
Normal/Underweight 7.6 30.0 -22.4 0.30 <0.001 23.8 99.4 94.5 75.3
Overweight 13.4 46.2 -32.8 0.30 <0.001 28.8 99.9 99.5 62.1
Obese 16.8 54.1 -37.3 0.25 <0.001 29.2 97.8 94.0 53.9
Smoking status
Non smoker 11.5 41.3 -29.8 0.28 <0.001 26.3 98.9 94.2 65.6
Ex-smoker 17.6 50.5 -32.9 0.34 <0.001 34.4 99.4 98.4 59.8 Current smoker 6.7 35.6 -28.9 0.22 <0.001 18.3 99.8 98.0 68.8
Taylor et al. Article
2010 vol. 34 no. 4 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 399© 2010 The Authors. Journal Compilation © 2010 Public Health Association of Australia
and 35.3% of those found to have HC as a result of their blood
test. These clinic measurements could be true under-diagnosis
where the respondents were unaware of his/her condition and, in
the case of blood pressure, the difference could be classified as
pre-hypertension.42,43 There is also the consideration that under
reporting could be related to misinterpretation, inadequate or
inappropriate communication between patient and doctor or
the changing of guidelines.44 In common with other Western
countries, around 60% of Australians have health literacy levels
that are not at the ‘minimum required for individuals to meet
the complex demands of everyday life and work in the emerging
knowledge-based economy’.45 Studies predominantly from the US
have shown that limited functional health literacy is associated
with less frequent preventive health behaviours and less active
self-management of chronic conditions.46,47 In addition, it should
be considered that older persons are more aware of their overall
health status and medical problems and are more likely to have
regular medical check-ups.
The strength of this study lies in its representative nature,
the large random sample, the high response rate, the clinical
operations and stringent measurement standards. Although the
response rate associated with the complete study involvement,
including obtaining blood and other bio-medical measurements,
was 49.6% (60% of people interviewed by telephone), this is
relatively high when compared to other recent, comparable
Australian studies.27 There is a trend towards lower response rates
in all types of population surveys as people protect their privacy,
or are overwhelmed by marketing telephone calls or mail outs.
The additional commitments associated with involvement in a
cohort study (including invasive tests such as the taking of blood
samples) add to respondent burden.
Weaknesses of the clinic measurements include the arbitrary
cut off point undertaken in our analysis, the dichotomy of the
diagnosis being made based on the average of the two readings and
the problems associated with determining clinical high readings
of these two measurements as detailed previously. In addition,
this analysis only assessed the total cholesterol measurement
and did not attempt to break down the levels of different types
of cholesterol. Questionnaire construction in self-report studies
could also have been expanded so that the self-report and clinical
measurements were exactly comparable especially in terms of
medication and control. In addition, although relatively high in
comparison to some other major studies, the low response rate
adds some responder bias. Extensive analysis has been undertaken
on the bias associated with this study19 and although the data are
weighted by age, sex, area and probability of selection, some
bias, as a result of under or over representation of some groups,
would be expected.
The study has shown major differences between estimates of
HBP and HC based on self-report estimates and those obtained
from clinic measures. However, people without high clinical
readings of blood pressure and cholesterol are not likely to report
they have HBP or HC. This suggests that self-report is more
likely to under- than to over-estimate the true prevalence of these
conditions. We conclude that self-reports of HBP and HC are
suitable for population monitoring, provided the limitations of
this information are acknowledged.
AcknowledgementsThe authors would like to thank the NWAHS participants and
the clinic staff.
References1. Department of Health [publication and report page on the Internet]. Perth
(AUST): Government of Western Australia; 2006 [cited 2007 Oct]. Population Surveys: Western Australian Health and Wellbeing Surveillance System. Available from: http://www.health.wa.gov.au/publications/pop_surveys.cfm
2. NSW Health [public health page on the Internet]. Sydney (AUST): State Government of New South Wales; 2006 [cited 2007 Oct]. New South Wales Health Survey Program. Available from: http://www.health.nsw.gov.au/public-health/survey/hsurvey.html
3. Taylor A. (2006) Chronic Disease Surveillance in South Australia. South Australia Public Health Bulletin [serial on the Internet]. 2006 [cited 2007 Oct 17]; 3:6-8. Available from: www.dh.sa.gov.au/pehs/publications/PHB-chron-disease-ed3-06.pdf
4. Muhajarine N, Mustard C, Roos LL, Young TK , Gelskey DE. Comparison of survey and physician claims data for detecting hypertension. J Clin Epidemiol. 1997;50(6):711-8.
5. Robinson JR, Young TK, Roos LL, Gelskey DE. Estimating the burden of disease: comparing administrative data and self-report. Med Care. 1997;35(9):932-47.
6. Klungel OH, de Boer A, Paes AH, Seidell JC, Bakker A. Cardiovascular diseases and risk factors in a population-based study in The Netherlands: agreement between questionnaire information and medical records. Neth J Med. 1999;55(4):177-83.
7. Okura Y, Urban LH, Mahoney DW, Jacobsen SJ, Rodeheffer RJ. Agreement between self-report questionnaires and medical records data was substantial for diabetes, hypertension, myocardial infarction and stroke but not for heart failure. J Clin Epidemiol. 2004;57:1096-103.
8. Tormo MJ, Navarro C, Chirlaque MD, Barber X, and the EPIC group of Spain. Validation of self diagnosis of high blood pressure in a sample of the Spanish EPIC cohort: overall agreement and predictive values. J Epidemiol Community Health. 2002;54:221-6.
9. St Sauver JL, Hagen PT, Cha SS, Bagniewski SM, et al. Agreement between patient reports of cardiovascular disease and patient medical records. Mayo Clin Proc. 2005;80(2):203-10.
10. Martin LM, Leff M, Calonge N, Garrett C, Nelson DE. Validation of self-reported chronic conditions and health services in a managed care population. Am J Prev Med. 2000;18(3):215-18.
11. Fakiri FE, Bruijnzeels MA, Hoes AW. No evidence for marked ethnic differences in accuracy of self-reported diabetes, hypertension and hypercholesterolemia. J Clin Epidemiol. 2007;60:1271-9.
12. Alonso A, Beunza JJ, Delgado-Rodriguez M, Martinez-Gonzalez MA. Validation of self reported diagnosis of hypertension in a cohort of university graduates in Spain. BMC Public Health. 2005;5:94.
13. Perryman S, Beerman KA. Know your numbers: comparing participants and non-participants in a university health screening program. J Am Coll Health. 1997;46(2):87-91
14. Chen Y, Rennie DC, Lockinger LA, Dosman JA. Association between obesity and high blood pressure: reporting bias related to gender and age. Int J Obes [Lond]. 1998;22:771-7.
15. Natarajan S, Lipsitz SR, Nietert PJ. Self-report of high cholesterol – determinates of validly in US adults. Am J Prev Med. 2002;23(1):13-21.
16. Vargas CM, Burt VL, Gillum RF, Pamuk ER. Validity of self-reported hypertension in the National Health and Nutrition Examination Survey III, 1988-1991. Prev Med. 1997(26):678-85.
17. Giles WH, Croft JB, Keenan NL, Lane MJ, Wheeler FC. The validity of self-reported hypertension and correlates of hypertension awareness among blacks and whites within the stoke belt. Am J Prev Med. 1995;11(3):163-9.
18. Bowlin SJ, Morrill BD, Nafziger AN, Jenkins PL, Lewis C, Pearson TA. Validity of cardiovascular disease risk factors assessed by telephone survey: the behavioural Risk Factor Survey. J Clin Epidemiol. 1993;46(6):561-71.
19. Taylor AW, Dal Grande E, Gill TK, Chittleborough CR, Wilson DH, Adams RA, et al. How valid are self-reported height and weight? A comparison between CATI self-report and clinic measurements using a large representative cohort study. Aust NZ J Public Health. 2006;30:238-46.
Chronic illnesses Comparing self-reported and measured blood pressure and cholesterol
400 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2010 vol. 34 no. 4© 2010 The Authors. Journal Compilation © 2010 Public Health Association of Australia
20. Grant JF, Chittleborough CR, Taylor AW, Dal Grande E, Wilson DH, Phillips PJ, et al. The North West Adelaide Health Study: methodology and baseline segmentation of a cohort along a chronic disease continuum. Epidemiologic Perspectives & Innovations. 2006;3:4.
21. Taylor A, Dal Grande E, Chittleborough C, Cheek J, Wilson D, Phillips P, et al. The North West Adelaide Health Study - Key Biomedical Findings, Policy Implications and Research Recommendations. Adelaide (AUST): South Australian Department of Human Services; 2002.
22. Taylor AW, Dal Grande E, Gill T, Chittleborough CR, Wilson DH, Adams RJ, et al. Do people with risky behaviours participate in biomedical cohort studies? BMC Public Health. 2006;6:11.
23. Australian Institute of Health and Welfare [data online page on the Internet]. Canberra (AUST): AIHW; 2009 [cited Author: please supply year, month and day web page viewed]. Risk Factor Prevalence Study Survey 1989, Survey No. 3. Available from: http://www.aihw.gov.au/dataonline/riskfactors/index.cfm#RFPS
24. Chalmers J, et al. Guidelines for the Management of Hypertension. Geneva (CHE): WHO/ISH Hypertension Guidelines Committee, World Health Organisation; 1999.
25. National Stroke Foundation of Australia. Heart, Stroke and Vascular Diseases Australian Facts. Canberra (AUST): Australian Institute of Health and Welfare; 2001.
26. Adams RJ, Wilson DH, Appleton S, Taylor A, Dal Grande E, Chittleborough CR, et al. Under-Diagnosed Asthma in South Australia. Thorax. 2003;58(10):846-50.
27. Gill T, Chittleborough C, Taylor A, Ruffin R, Wilson D, Phillips P. Body mass index, waist hip ratio, and waist circumference: which measure to classify obesity? Soz Praventivmed. 2003;48(3):191-200.
28. Appleton SL, Ruffin RE, Wilson DH, Taylor AW, Adams RJ. Asthma is associated with cardiovascular disease in a representative population sample. Obesity Research and Clinical Practice. 2008;2:91-9.
29. Adams RJ, Appleton S, Hill CL, Wilson DH, Taylor AW, Chittleborough C, et al. Independent association of HbA1c and incident cardiovascular disease in people without diabetes. Obesity. 2009;17(3):559-63.
30. Population Research and Outcome Studies Unit. The South Australian Monitoring and Surveillance System (SAMSS) 2002 - 20 [brief report]. Adelaide (AUST): South Australian Department of Health; 2002.
31. Australian Bureau of Statistics. 2015.4 - Census of Population and Housing Selected Social and Housing Characteristics for Statistical Local Areas, South Australia, 2001. Canberra (AUST): ABS; 2002.
32. Daly A, Taylor A. Field Testing 2 Report - Alcohol consumption, Cardiovascular Disease and Tobacco Consumption - January 2004. In: Population Health Monitoring and Surveillance, Question Development Field Testing Reports [report page on the Internet]. Melbourne (AUST): National Public Health Partnership; 2004 [cited Author: please supply year, month and day web page viewed]. Available from: http://www.nphp.gov.au/catitrg/
33. McGrath BP. Ambulatory blood pressure monitoring and ‘white coat’ hypertension: saving costs. Med J Aust. 2002;176:571-2.
34. Pierdomenico SD, Lapenna D, Di Mascio R, Cuccurullo F. Short- and long-term risk of cardiovascular events in white-coat hypertension. J Hum Hypertens. 2008;22(6):408-14.
35. Turner MJ, van Schalkwyk JM. Blood pressure variability causes identification of hypertension in clinical studies: a computer simulation study. Am J Hypertens. 2008;21(1):85-91.
36. Hajjar I, Kotchen JM, Kotchen TA. Hypertension: trends in prevalence, incidence and control. Annu Rev Public Health. 2006;27:465-90.
37. Marshall TP. Blood pressure Variability: the challenge of variation. Am J Hypertens. 2008;21(1):3-4.
38. Krum H, Jelinek MV, Stewart S, Sindone A, Atherton JJ, Hawkes AL, et al. Guidelines for the prevention, detection and management of people with chronic heart failure in Australia 2006. Med J Aust. 2006;185(10):549-56.
39. Agarwal R, Bunaye Z, Bekele DM. Prognostic significance of between-arm blood pressure differences. Hypertension. 2008;51:657-62.
40. Fiebach NH, Hebert PR, Stampfer MJ, Colditz GAS, Willett WC, Rosner B, et al. A prospective study of high blood pressure and cardiovascular disease in women. Am J Epidemiol. 1989;130(4):646-54.
41. Wan Q, Harris MF, Davies GP, et al. Cardiovascular risk management and its impact in Australian general practice patients with type 2 diabetes in urban and rural areas. Int J Clin Pract. 2008;62:53-8.
42. Kaplan NM. Treating pre-hypertension: a review of the evidence. Curr Hypertens Rep. 2008;10(4);326-9.
43. Pickering TG. The natural history of hypertension: pre-hypertension or masked hypertension? J Clin Hypertens (Greenwich). 2007;9(10):807-10.
44. Bays HE, Chapman RH, Fox KM, Grandy S, Shield Study Group. Comparison of self-reported survey (SHIELD) versus NHANES data in estimating prevalence of dyslipidemia. Current Medical Research and Opinions. 2008;24(4):1179-86.
45. Australian Bureau of Statistics. Adult Literacy and Life Skills Survey, Summary Results, Australia 2006. Canberra (AUST): ABS; 2007. Catalogue No.: 4228.0.
46. Berkman ND, et al. Literacy and Health Outcomes. Rockville (MD): Agency for Healthcare Research and Quality; 2004. AHRQ Publication No.: 04-E007-1.
47. De Walt DA, Boone RS, Pignone MP. Literacy and health outcomes: a systematic review of the literature. J Gen Intern Med. 2005;19:128-39.
Taylor et al. Article