Scottish Disinvestment Project

1
Scottish Disinvestment Project
ESRC Seminar Birmingham March 2007

• Dr Sheila N Scott
• Director of Public Health
• NHS Western Isles
• 37 South Beach Street
• Stornoway, Isle of Lewis,
• HS1 2BB, Scotland
• Sheila.Scott_at_wihb.scot.nhs.uk

2
CONTENTS

• Review previous objectives and
• results
• Examine changes in practice
• Describe current approaches
• The next steps

3
ORIGINAL OBJECTIVES

• To undertake a stocktake across Scotland of
  Scottish Health Boards about any work to
  stop/curtail interventions of low or no health
  gain
• If so what and by whom?
• What initiated such work - guidelines, HTA
  guidance, audit, Quality Improvement Scotland
  etc.?
• Quantify nationally what the number of procedures
  might be
• Calculate the opportunity cost/disinvestment
  potential
• This project was done with cooperation of
  Scottish Directors of Public Health

4
BACKGROUND
Orkney
Highland
Shetland
Western Isles
Grampian

14 Health Boards

• Each Health Board has a Director of Public Health
  and denominator population from (25,000 - 1.1
  million)
• Hospital activity for past 30 years

5
METHODOLOGY
Literature review in November 02 while in
Argyll and Clyde Clinical Evidence A
compendium of the best available research
findings on common and important clinical
questions, which is updated and expanded every
six months. Published jointly by the BMJ
Publishing Group and the American College of
Physicians. Sample pages from the website
http//www.bmjpg.com/evid99/index.html Evidence
Based Medicine Bi-monthly, to survey at least 70
international medical journals to identify the
key research papers that are scientifically valid
and relevant to practice. These articles are
selected according to scientific criteria and
only those papers with direct message for
practice are included. Covering internal
medicine, general surgery, paediatrics,
obstetrics, gynaecology, psychiatry, general
practice, anaesthesiology and ophthalmology.
http//www.bmjpg.com/data/ebm.htm
6
ANALYSIS OF LITERATURE REVIEW
Nb. unknown effectiveness does not no
effectiveness
7
METHODOLOGY

• Questionnaire for each Health Board designed and
  piloted in 3 Health Boards - one large two
  small.
• Information Services Division, National Services
  Scotland approached re analysis of some sentinel
  procedures chosen from literature review, as a
  pilot.

8
RESULTS
Structured questionnaire pilot

• FEEDBACK
• too much work required to make stocktake
  comprehensive
• no central repository/responsibility for such
  work
• confusion over disinvestment rationale -
  efficiency (access and
• frequency) vs low health gain
• people too busy completing forms for other
  purposes
• but happy with principle if supported

9
Health Technology Section of Quality Improvement
Scotland re sentinel conditions and progress on
full HTAs

• Tonsillectomy and grommets (ENT)
• Dilatation Curettage (Gynaecological)
• Varicose Veins (Surgical)
• Grommets (ENT)

But not reporting for some time
10
Results from ISD Rate per 100,000 population of
sentinel conditions
Variation x2 - x9 (DCs) between each Health Board
11
Secular trends downwards Scotland
but not uniform.
Total number of sentinel procedures was 17,000
in 2003/04
12
Clinicians will say Some procedures will always
be necessary Number of procedures that could
be avoided if the lowest rate was applied to all
NHS Boards for sentinel conditions. (total in
Scotland in 2003/416,687)
13
Are there significant correlations and
differences between the rates in different Health
Boards for different procedures? 1. Spearmans
product moment correlation - There is a
significant correlation between the rates of
tonsillectomy and rates of Varicose veins
operations carried out by HBs (r.587,p0.022) 2.
ANOVA (1 way) Looked for significant variation
in the rates between different HB types (RR,
Mixed and Urban) - Results show that there is
significant differences in rates for
tonsillectomies (p0.002) and varicose veins
(p0.001) between different HB types. 3.Post hoc
Tukey test shows that there are no significant
differences between Urban and Mixed but both are
significantly different from Remote and Rural
14
Possible explanations?

• HB definitions (a mixture of density and no
  conurbation gt 20,000)
• Capacity/referral issues
• Data population based not hospital based
• Weighting by 1 or 2 clinicians
• Other
• Effect small in terms of numbers nationally eg in
  WOS

15
Conclusions from original analysis

• Great potential to save 6,500 operations from 4
  sentinel conditions, perhaps more
• Health Boards and QIS, Scottish Medicines
  Consortium and others providing resources to
  promulgate/evaluate new technologies
• Publication bias towards newer procedures/technolo
  gies
• Nobody appears to have responsibility for
  stopping things that do not work or which may
  harm patients (reducing demand)
• Dedicated resource to facilitate this required in
  Scotland

16
Next stages were

• Discussion of data to DsPH and Scottish
  Association of Medical Directors
• Senior Civil Servants at Scottish Executive
• Some funding from National Public Health Network
  (11K)
• Disinvestment Group set up Medical Directors,
  Health Economists, Statisicians, Information
  Services Division,QIS, SMC but orphan group
  though CMO backed
• Literature Review commissioned

17
Impact of initial analysis and feedback
18
(No Transcript)
19
Conclusions and current steps

• Need core funding and buy in
• Literature review and evidence notes
• Need to understand how clinical practice changes
  / ? link to performance management/clinical
  audit/repeatability
• Speaking to QIS, SEHD, NSS, WOSRP, ESRC, CMO/SAMD
• Great potential if systematically dealt with