Get documents about "BEACH
Shared by: qingyunliuliu
-
Stats
- views:
- 12
- posted:
- 10/9/2011
- language:
- English
- pages:
- 54
Document Sample
BEACH
A BEACH SAND study
Estimated prevalence of chronic illnesses
(identified as Health Priority Areas)
among general practice patients
Stephanie Knox (now at CHERE)
Helena Britt
Christopher Harrison
Australian GP Statistics & Classification Centre
A collaborating Unit of the Australian Institute of Health and Welfare
& the University of Sydney
A collaborating unit of the Australian Institute of Health and Welfare and the University of Sydney
BEACH
Today’s aims
About us
BEACH methods
Representativeness
Some changes in morbidity managed by GPs
SAND sub-studies – methods
Results of the prevalance study
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
The Family Medicine Research Centre
A University of Sydney Research Centre academically attached to the
Department of General Practice & School of Public Health
A Collaborating Centre of the World Organization of Family Doctors (Wonca)
All moneys are self-generated through competitive grants,
consultancies and contracts.
Our work is health services research and development.
The Australian GP Statistics & Classification Centre
A collaborating unit of the Australian Institute of Health and Welfare and the
University of Sydney, within the FMRC.
Responsible for the BEACH program which is conducted under the AIHW
Act, with Ethics approval from the University and AIHW Ethics Committees
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
BEACH
Began April 1, 1998
Now in its 10th year (April 1, 2007)
~9,000 participants
Representing 7,500+ GPs
>40% of recognised GPs)
Data available ~ 900,000 encounters
Data being used by the profession,
researchers, governments & industry
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
BEACH Aims
to provide a reliable and valid continuous national data
source of timely GP–patient encounter information
to inform the community of the care provided to the vast
majority of the population by GPs - the gate keepers of our
medical system
to assess patient health risk factors, prevalence of disease,
or longer term management--on selected subjects in sub-
samples of patients- up to 20 sub-studies per year
(SAND- Supplementary Analysis of Nominated Data)
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
BEACH methods
A cross sectional encounter based study
Paper based data collection
National GP random sample (drawn by DoHA)
1,000 GPs per year
20 per week x 50 weeks a year - ever changing
100 consecutive encounters per GP
All types of encounters included
National data for 100,000 encounters p.a.
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Post-stratification weights
To ensure national representation of GP
encounters, each year the data are weighted
1. To correct for the under-representation of young
GPs <35 years (small but important)
2. To give more weighting to the encounters of
busy GPs and less to (e.g.) part-time GPs data
are weighted by HIC A1 Medicare claims
previous quarter.
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Observed vs expected age distribution
by state
NSW
30
25
20 Expected
% encs
encounters
15
observed
10
5
0
0-4 5- 15- 25- 45- 65- 75+
14 24 44 64 74
Expected = MBS (A1) claims, observed = BEACH sample
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Observed vs expected age distribution
by state
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Observed vs expected age distribution
by state
WA
30
Expected encounters
25 observed
20
% encs
15
10
5
0
0-4 5-14 15-24 25-44 45-64 65-74 75+
Age gp
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
BEACH variables (see recording form)
G.P characteristics (GP completed questionnaire)
Practice characteristics (GP completed questionnaire)
Patient characteristics
Encounter details
Patient reasons for encounter (up to 3)
Problems managed (up to 4)
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
BEACH variables
Management (of each problem)
Medications prescribed, supplied, advised (up to 4/problem)
Non-pharmacological treatments (up to 2/problem)
• Clinical treatments
• Therapeutic procedures
New referrals & admissions (up to 2 – linked to problem(s))
Pathology tests ordered (up to 5–each linked to problem(s))
Imaging & other tests ordered (up to 2 - linked)
(view form)
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
G.P Characteristics
age
sex
country of graduation
years in general practice
FRACGP
currently in training program
practice location (State,RRMA, ARIA, SEIFA)
practice size
use of computers
+ other variables over the years
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Patient characteristics
Age (from DoB)
Sex
Status to the practice (new/seen before)
NESB (Yes/no)
Aboriginal (self ID – Yes/no)
Torres Straight Islander (self ID – Yes/no)
Health care card holder (Yes/no)
VA card holder (yes/no)
Reasons for encounter (up to 3)
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Some changes over time
A quick look at:
patients seen
problems managed
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Age of patients at encounter
changing over time (% of workload)
30
1990-91
25 1998-99
2005-06
20
15
10
5
0
<15 yrs* 15-24 yrs 25-44* 45-64 yrs* 65-74 75 + *
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Summary of problems over time
160
1990-91
140 1998-99
2005-06
120
100
80
60
40
20
0
RFEs Problems New problems Chronic
managed problems*
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Selected morbidity management rates over
time - rate per 100 encounters
6
1990–91
5 1998–99
2000–01
4
2005–06
3
2
1
0
sion Lipid
s r iti s et es
pres oart h Diab
De Ost e
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Other problems managed over time rate per
100 encounters
10
1990-91
9
8 1998-99
7 2005-06
6
5
4
3
2
1
0
i s* * I* * * *
ion int RT t is itis OM ma
s ld pla U hi ill te th
ert en ea m nc ns cu As
yp ag co Br
o To A
H ph c k e
so Ba ut
Oe Ac
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Encounter data
Tells us about:
patient demand for care
the problems managed
How specific problems are managed by GPs
It does NOT provide prevalence estimates of
disease:
among the attending population
(attendance gives chance of selection)
among the population at large
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
SAND:
Supplementary Analysis of Nominated Data
“The bit on the bottom of the form”
Questions usually asked of the patient
Each GP pack divided into three: 40 + 30 + 30
• 40 BMI, smoking status, alcohol intake
• 30 nominated subject 1
• 30 nominated subject 2
Nominated subject 5 weeks -- 3000 random sample
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Available through
http://www.fmrc.org.au
(go to ‘BEACH’ and select
‘publications’)
Also see ‘Abstracts’ in the
publication section of our
website
Hard copies of reports from
http://www.aihw.gov.au
(cost $15-27 each book in GEP
series)
AGPSCC
Phone +61 2 9845 8151
email: gpstats@fmrc.org.au
A collaborating unit of the Australian Institute of Health and Welfare and the University of Sydney
BEACH
Population prevalence of disease
Important for health policy and health service
planning
In Australia usually rely on results from the
National Health Survey (NHS) (conducted by the
ABS)
NHS estimates population prevalence based on
patient self-reported morbidity
Using structured interview, with trained
interviewer, in respondent’s home to elicit
health-related information
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
National Health Survey
Advantages:
includes a representative sample of the total
population (n = about 27,000 in 2004)
Shortcomings:
Only conducted every 3 years ( 4 in the past)
Self-report likely to be unreliable (lack of
clinical knowledge, recall issues etc)
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
The international scene
Most countries rely on similar national surveys
Many people question the reliability of such estimates
Some studies suggest patient recall of morbidity better
for well defined conditions (e.g. Diabetes, heart
disease) than for ill-defined conditions (e.g. back pain)
Accuracy of household health data tested by Moore
(US) in 1972, on chronic disease prevalence
One third reported 1+ chronic disease (n=333)
106 medical records checked
False positive chronic disease identification by
patient self report = 25%; false negative 38%
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Prevalence data from EHRs?
Some have tried to produce prevalence
estimates from GP EHRs
This works better in countries with patient
registers than in fee for service system.
Major issue is incompleteness of record
Some question why we need to sample
people who do not utilise the health care
system (i.e. currently not a cost)
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Our study
Aims:
to use an ongoing national survey of GP
activity to provide estimates of the
prevalence of diagnosed morbidity among
the population of general practice patients
in Australia.
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Methods
A SAND (Supplementary Analysis of Nominated
data) substudy in the BEACH program
‘the bit on the bottom of the form‟
BEACH GPs 12/07/05-19/08/05 & 25/10/05-28/11/05
Each GP asked, for 30 consecutive patients
„Does this patient have any of the following
conditions which require ongoing
management?‟
Final sample 9156 from 305 GPs
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Conditions listed on the form
Tick boxes provided, multiple response allowed
Cardiovascular disease Psychological problems
--Ischaemic heart disease --Depression
--Cerebrovascular disease --Anxiety
--Peripheral vascular disease --Insomnia
--Congestive heart failure --Other psychological problem
--Hypertension – complicated Other problems
--Hypertension – uncomplicated --Hyperlipidaemia
--Other --Chronic back pain
Arthritis --Malignant neoplasms
--Osteoarthritis --Gastro-oesophageal reflux disease
--Rheumatoid arthritis Respiratory problems
--Other arthritis --Asthma
-- mild
Diabetes --moderate
--Type 1 --severe
--Type 2 --Chronic obstructive airways disease
--Other None of these conditions.
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Prevalence among whom?
Patient sample drawn at point of GP-patient
encounters is biased towards frequent
attenders, who are older and have more health
problems than population as a whole.
Sample will over-estimate prevalence of health
conditions in the total patient population
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Adjust to patient population
1. Enumerate the proportion of the sampled
general practice patients who have a
chronic condition requiring ongoing
management
2. Use these crude sample rates to estimate
prevalence in the population of patients
attending GPs at least once, by adjusting
for annual visit rates.
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Methods (cont.)
Crude rates calculated (% with the disease)
Crude rate weighted by visit frequency
related to age and sex, by weighting SAND
sample against age-sex distribution of
population who visited a GP (1+ times) in
12 months (April 2004-March 2005)
Used SAS V9.13 to adjust for the cluster
design of the study.
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
The underlying theory
Recording of morbidity present in the
patient will be more reliable with the
combined input of:
the patient,
the GP, and
the medical record (where available)
than patient self-report alone.
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Results – Age-sex distribution c/f
population attending GP
N = 9,156 patients from 305 participating GPs.
50
National GP patient population
45
SAND sample
40
35
30
% 25
20
15
10
5
0
Male 0-14 yrs 15-24 yrs 25-44 yrs 45-64 yrs 65-74 yrs 75+ yrs
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Prevalence – crude rates
(what’s in the waiting room)
40
35
30
25
20
15
10
5
0
ed VD obl'm thritis rl ipid ORD thm a betes OAD ancy
lis t ardi o r r G As C n
ne of C ch p 1+ A H ype +D
ia
M alig
No 1+ sy 1
1 +P
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Prevalence – crude rates
25
Per cent of attending patients
20
15
10
5
0
n rt rt VD
nsio H ype H y pe IHD reb
e rte plic p lic Ce
Hyp nco
m
Co
m
U
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Prevalence – crude rates
25
Per cent of attending patients
20
15
10
5
0
on e ty ia er
si nxi mn Ot h
res A In so
Dep
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Prevalence – crude rates
25
Per cent of attending patients
20
15
10
5
0
M id n
IDD
M
IDD OA ato k p ai
N eum bac
Rh o ni
c
r
Ch
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Prevalence – crude rates
25
Per cent of attending patients
20
15
10
5
0
a d te e
t hm Mi l ra ver
As Mode Se
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Prevalence – crude rates + adjusted
30
Per cent of attending patients
Per cent of patient population
25
20
15
10
5
0
t rt
io V
D
nsio n ype r y pe IHD rebVD rl ip
id
C ard e rte mplic
H
lic
H Ce Hyp
e
Hyp nco mp
U Co
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Prevalence – crude rates + adjusted
25
Per cent of attending patients
Per cent of patient population
20
15
10
5
0
l'm on ty ia the
r
ro b s si xi e mn
p r e An Ins
o O
P sy c
h Dep
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Prevalence – crude rates + adjusted
25 Per cent of attending patients
Per cent of patient population
20
15
10
5
0
ll) M M is OA id a in
s (a IDD NID
D
art hrit ato kp
ete y heu
m ac
Dia
b An R nicb
C hro
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Prevalence – crude rates + adjusted
25
Per cent of attending patients
Per cent of patient population
20
15
10
5
0
D a d te re AD cy
OR thm M il era ev e CO n an
G As Mo
d S alig
M
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Est prevalence in GP patient pop’n c.f.
population prevalence (NHS 2004)
25
Prevalence in patient population
Population prevalence NHS
20
15
10
5
0
n d
sio IHD VD ipi DM DM
n reb e rl ID NID
e rte Ce Hy
p
H yp
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Est prevalence in GP patient pop’n c.f.
population prevalence (NHS 2004)
25 Patient population
NHS
20
15
10
5
0
l'm on ty
h prob pres
si A nxi e
syc D e
A ny p
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Est prevalence in GP patient pop’n c.f.
population prevalence (NHS 2004)
20
Patient population
NHS
15
10
5
0
tis OA toi
d P a cy
hri a in/B st hm an
art um pa A gn
y e ck ali
An Rh ba M
ic
r on
Ch
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
What about population prevalence?
If we assume that all of the 15% who did
not attend a GP in 2005 have none of the
listed diseases under management,
we have to add these into the
denominator, to gain an estimate of
prevalence among the total population
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH Prevalence in patient population –
extrapolated to total population
c.f. NHS
20
Extrapolated patient prevalence
NHS 2004
15
10
5
0
n IHD lem ssio
n all OA ids P a pl
si o ob itis
rlip in/B t hm n eo
en pr pr
e thr pe pa As g't
ert h De Ar Hy ck li
Hy
p
s yc ba Ma
P ni c
ro
Ch
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Discussion
Differences in question
SAND ‘ongoing management’ vs NHS present/absent
SAND Chronic back pain vs back pain
Methods seem appropriate – supported by
decrease population estimate for prevalence of
Asthma to LESS than the NHS – this suggests that
the assumption of no attendance = no disease
does not apply to asthma.
Certainly the costs of collection are marginal
c.f. NHS
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Conclusion
At marginal cost to the BEACH program we can gain
reliable estimates of the prevalence of diseases classed
as National Health Priority Areas, among the population
who attend a general practitioner in Australia.
These estimates would assist in health care planning by
allowing estimates of future health service needs at a
national and local level.
Extrapolation to the total population may raise
methodological questions, but the results may well be
more reliable that the NHS.
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Many thanks
to the GPs
BEACH 06-07
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Beach 05-06
Estimated prevalence of NHPA morbidities. SPH Seminar August 2007
BEACH
Available through
http://www.fmrc.org.au
(go to ‘BEACH’ and select
‘publications’)
Also see ‘Abstracts’ in the
publication section of our
website
Hard copies of reports from
http://www.aihw.gov.au
(cost $15-27 each book in GEP
series)
AGPSCC
Phone +61 2 9845 8151
email: gpstats@fmrc.org.au
A collaborating unit of the Australian Institute of Health and Welfare and the University of Sydney
