Australian General Practice Statistics and Classification Centre

1
BEACH
Bettering the Evaluation And Care of Health
A continuous study of general practice activity
in Australia Professor Helena Britt
Professor Graeme Miller
Visit sponsored by Centre for Health Services and
Policy Research at UBC Canadian Institute for
Health Information Canadian Institutes of Health
Research Canadian Health Services Research
Foundation
2
Australian GP Statistics Classification Centre
(GPStats)

• A collaborating unit of the Australian Institute
  of Health and Welfare and the University of
  Sydney.
• Responsible for the BEACH program which is
  conducted under the AIHW Act, with Ethics
  approval from the University and AIHW Ethics
  Committees
• We are researchers, so data quality is essential

3
BEACH aims

• to provide a reliable and valid continuous
  national data source of timely GPpatient
  encounter information
• to assess patient risk factors and health states
  and their relationship to morbidity and health
  service activity.

4
Current status of BEACH

• Began April 1, 1998
• Now in its 8th year (began April 1, 2005)
• Over 7,500 participating GPs to date
• Data available about 725,000 encounters
• Data being used by, Government, researchers,
  industry
• Continuity allows measures of change

5
BEACH methods

• Paper based data collection
• National GP random sample (drawn by DHAC)
• 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.
• All questions approved by the AIHW Ethics
  Committee

6
BEACH variables

• G.P characteristics (GP completed
  questionnaire)
• Patient characteristics
• Encounter details
• Patient reasons for encounter (up to 3)
• Problems managed (up to 4)
• Management (of each problem)

7
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

8
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)

9
Encounter data

• Date, place/item no/direct or indirect, paid
  by?
• Problems managed (up to 4)
• and their status (new/old to patient)
• Management of each problem
• medications (up to 4 per problem),
  (new/continued)
• procedures/ clinical treatments (up to 2 per
  problem)
• referrals (up to 2 per encounter)
• pathology (up to 5) imaging (up to 2) ordered

10
The GP sample frame

• Universal medical insurance scheme, Australian
• Medicare claim files updated every 3 months -
  includes all Gps who claimed a minimum of 375 GP
  items of service in most recent 3 months.
• Australian Government Department of Health and
  Ageing holds the sample frame and draw sample
  each quarter for BEACH.
• Ensures inclusion of practising GPs, and most
  part-timers (excludes very part-time, one locum
  per year etc)

11
GP sample

• 1000 GPs pa needed - but sample quarterly to
  ensure up to date - random from reshuffled sample
  frame
• Reward to GPs is points towards their quality
  assurance requirements, QA cycle is 3 years -
  therefore restart sample frame every 3 years.
• Once selected GP is removed from sample framefor
  3 years
• Provided with name, address, phone number,
  number of claims previous quarter, number of
  claims previous 12 months.

12
Centralised data entry

• GP profile data (questionnaire) entered in
  Profile database
• Coder enters encounter form data into BEACHHEAD
  (Access database)- free to ask questions of
  senior staff
• BEACH-HEAD specifically designed for speedand
  accuracy- data coded and auto-classified(staffing
  , training, speed, discussed later)
• QA programe for data entry

13
Data entry QA

• QA checks of 1/1 to 1/10 depending on coders
  experience highlight issues for training
• Corrections made in data errors by coder
• Standard programs run in Access for errors (e.g.
  lt12 years, pregnancy/family planning)
• Identified issues checked on paper record
• Data imported to SAS each 3 months
• Standard SAS programs applied for more complex
  checks
• If the GP wrote outlier script it is not changed,
  if data entry error, it is changed)

14
(No Transcript)
15
SAND (Supplementary Analysis of Nominated Data )

• For each 100 forms - questions asked of the
  patient
• 40 patient risk factors
• 30 Subject 1
• 30 Subject 2
• Ten x 5 week blocks
• Sample size for Subject about 3,000
• Sometimes repeat subject to increase sample size
  (e.g. estimates of indigenous encounters
  estimates of disease prevalence, management of
  asthma)
• Range prevalence, management, effectiveness can
  be semi-longitudinal

16
Classification ... ICPC-2
The International Classification of Primary Care
- the World Organisation of Family Doctors
(Wonca) - part of WHO Family of
Classifications - Australian standard for
reporting GP and patient self reported
data ICPC-2 PLUS The Family Medicine Research
Centre -
an interface terminology based on ?1 million
encounter records - classified according to
ICPC-2, completing classification to ICD 10.
17
ICPC-2 PLUS

• An interface terminology developed from the
  language of GPs - released in 1995
• from gt1 million encounter records
• including free text descriptions of
• 1.5 million problem labels, 1.5 million RFEs,
• 250,000 other treatments, 150,000 pathology
  orders
• 40,000 imaging test orders, 40,000 referrals
• ? input from 1500 GPs in gt400 practices
  using it in their EHRs on a daily basis

18
BEACH pharmaceutical data

• Drug name
• Strength ?
• Dose ? PDD
• Frequency ?
• Number of repeats
• Prescribed/ advised/ supplied
• New Vs continued

19
Classifying Drugs - CAPS Coding Atlas for
Pharmaceutical Substances

• Classifies medications prescribed, advised,
  provided
• Drug Class (eg antibiotics)
• Drug sub group (eg broad spectrum penicillins)
• Generic (amoxycillin)
• Brand (Amoxil)
• At generic level also classified to
• Anatomic Therapeutic Chemical (ATC)
  classification (WHO)

20
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 lt35 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.

21
Observed vs expected age distribution by state
Expected MBS (A1) claims, observed BEACH
sample
22
Observed vs expected age distribution by state
23
Standard reports(examples see - www.fmrc.org.au)

• Problem based report
• group ( diabetes - all, arthritis - all)
• more specific group (osteoarthritis)
• ICPC-2 level (IDDM)
• ICPC-2 PLUS (osteoarthritis of shoulder)
• Covers
• GPs, patients, RFEs, other morbidity managed,
  medications, other treatments, referrals,
  investigations/tests

24
RFEs Depression 54.2 Prescriptions
(all) 14.1 Weakness/tiredness 4.7 Sleep
disturbance 4.6 Anxiety 4.2
Acute stress reaction 3.8 Back complaint
3.5 Cardiac check-up 3.1
Hypertension/High BP 2.3 General check-up
2.2
Patients Total New Female 67.9
67.8 lt1-14 0.6 0.8 1524
7.7 11.7 2544
39.9 42.0 4564 31.8
31.5 6574 10.1 7.1 75
10.0 6.8
Referrals Psychiatrist 4.4 Psychologist
1.6 Counsellor 0.8 Other health
prof l0.4 Hospital 0.4 Mental health team
0.3 Hospital psychiatrist 0.3 Clinic
psychiatrist 0.2
Pathology Chemistry 5.3 Haematology
2.4 Microbiology 0.3
DEPRESSION N 3,367 (2.4 of all problems
managed) New 666 (19.8 of all depression)
Other problems managed with depression Hypertensio
n 6.7 Back complaint 3.1 Menopause complaints
2.7 Diabetes 2.2 Sleep disturbance
2.1 Osteoarthritis 2.0 Lipid disorder
2.0 URTI 1.9 Immunisation/vacc (all)
1.9 Oesophageal disease 1.8 Asthma 1.5
Prescriptions Sertraline 14.8 Paroxetine
9.0 Fluoxetine hcl 7.1 Moclobemide
6.9 Dothiepin 6.3 Venlafaxine 3.7
Temazepam 3.4 Citalopram 3.3
Amitriptyline 3.3
Clinical treatments Counsel psych NOS
34.2 Counsel NOS 3.7 Advice medication
3.6 Advice NOS 1.5 Advice
treatment 1.2 Counsel relationship 0.9
Counsel relaxation 0.5
25
Standard reports

• Participating organisations up to 20 standard
  reports (or equivalent) per year
• Delivery within a week (or a day if desperate)
• Drug based reports
• Drug group (Serum lipid reducing agents)
• Sub-group (HMG CoA reductase inhibitors)
• Generic (simvastatin)
• Brand
• Covers
• GPs, patients, RFEs, morbidity, PDD

26
(No Transcript)
27
GP characteristicsBEACH (2003-2004) (n983)
28
Patient characteristics - BEACH
29
Rates per 100 encountersBEACH (2004-05)
30
Management rates per 100 problems BEACH (2004-05)
31
Most common problems managed (2004-05)

• Rate/100 encs 95CIs
• Hypertension 8.9 (8.4 - 9.4)
• URTI 5.6 (5.1 - 6.0)
• Immunisation/vacc 4.6 (4.1 - 5.2)
• Depression 3.7 (3.5 - 3.9)
• Lipid disorder 3.3 (3.1 - 3.6)
• Diabetes 3.2 (3.0 - 3.4)
• Back complaint 2.8 (2.6 - 3.0)
• Osteoarthritis 2.8 (2.6 - 3.0)
• Acute bronchitis 2.4 (2.1 2.7)
• Asthma 2.3 (2.2 - 2.5)

32
Most common prescribed medications(2004-05)

• Rate/100 encs 95CIs
• Amoxycillin 3.5 (3.2 - 3.8)
• Paracetamol 2.7 (2.4 - 3.0)
• Cephalexin 2.4 (2.2 - 2.6)
• Paracetamolcodeine 2.0 (1.8 - 2.2)
• Salbutamol 2.0 (1.8 - 2.2)
• Amoxycillin/potass.clav.1.7 (1.5 - 1.9)
• Salbutamol 1.4 (1.3 - 1.6)
• Atorvastatin 1.4 (1.2 - 1.5)

33
The BEACH drug database
Extrapolated total annual prescriptions by GPs
34
Prescribing rates of NSAIDs for all arthritis
problems
35
Dissemination

• Report of annual results - week 1 December
• Single page abstracts for each SAND on web
• 1 or 2 topic books per year (e.g. Locality
  matter, pathology ordering, 65 yrs )
• ? 10 peer reviewed papers p.a.
• Monthly topic summaries in Aust. Fam Physician
• Bytes from BEACH in GP Review (RACGP quarterly)
• Regular input into GP weekly newspapers
• 15-20 paper presentations at conferences p.a.

36
The interactive web-server

• Data available April 2000-September 2005 (access
  only to periods purchased)
• queries - problem based, medication based,
  pathology imaging based
• limited selected analyses available
• only accessible to participating organisations
• only accessible using SecureID

37
Users AIHW (CIHI)

• burden of disease
• disease specific reports (e.g. Asthma)- use
  multiple data sources (with mortality,
  hospitalisations etc)
• Indigenous health
• workforce assessment
• measuring equity and access
• Veterans health

38
Users- Government

• WAMTCs
• monitor health priority areas
• post-marketing surveillance (Drug Utilisation
  Sub-Committee)
• workforce planning
• costing of services - indiginous health
• studying leakage
• pre-post evaluation of government initiatives
• quality use of medicines program (National
  Prescribing Service)
• costing of quality use of pathology imaging
  (costly!)
• patient safetly adverse events monitoring
• measuring Government performance (Productivity
  Commission)
• etc. etc. etc.

39
Other users

• The profession
• curriculum development (undergrad, vocational,
  CME)
• workforce planning
• comparison of groups of GPs (e.g. rural-met,
  FRACGP -not)
• assessment of quality of care
• pre-post measurement of interventions (e.g.
  Indigenous)
• Health economics consultants
• the pharmaceutical industry
• academic researchers (planning and validation)
• Consultants to Governments

40
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 [phone]
8151 email gpstats_at_fmrc.org.au