SAMASTIPUR CALL GIRL 7857803690 LOW PRICE ESCORT SERVICE
Investigating the relationship between quality of primary care and premature mortality in England
1. Investigating the relationship between quality of
primary care and premature mortality in England
a spatial whole-population study
Evangelos Kontopantelis David Springate Mark Ashworth
Roger Webb Iain Buchan Tim Doran
Centre for Health Informatics, Institute of Population Health
Faculty of Medicine, University of Manchester
HSCIC public board meeting, 28th January 2015
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 1 / 33
2. Outline
1 Background
2 Methods
3 Findings
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 2 / 33
3. Improving quality of care
or quality of recorded care?
A pay-for-performance (p4p) program kicked off in April 2004 with
the introduction of a new GP contract
General practices are rewarded for achieving a set of quality targets
for patients with chronic conditions
The aim was to increase overall quality of care and to reduce
variation in quality between practices
The incentive scheme for payment of GPs was named the Quality
and Outcomes Framework (QOF)
Initial investment estimated at £1.8 bn for 3 years (increasing GP
income by up to 25%)
QOF is reviewed at least every two years
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 4 / 33
4. Quality and Outcomes Framework
details for years 1 (2004/5) and 7 (2010/11)
Domains and indicators in year 1 (year 7):
Clinical care for 10 (19) chronic diseases, with 76 (80) indicators
Organisation of care, with 56 (36) indicators
Additional services, with 10 (8) indicators
Patient experience, with 4 (5) indicators
Implemented simultaneously in all practices (a control group was
out of the question)
Into the 11th year now (01Mar14/31Apr15); cost for the first 10
years was above the estimate at £10 bn approximately
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 5 / 33
5. Investigated relentlessly
in Manchester and elsewhere
Main driver for complete computerisation in primary care
Although a voluntary scheme, participation is almost complete and
computerisation is a prerequisite
Led to improvement in quality more quickly, but the benefits
diminish over time
Reduced inequalities of care
Led to some deterioration in unincentivised aspects of care
Contradictory evidence on its effect on hospital admissions
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 6 / 33
6. But what about ‘harder’ outcomes
namely, mortality
Aimed to quantify the relationship between performance on the
Quality and Outcomes Framework, and:
all cause premature mortality
cause-specific premature mortality linked closely with conditions
included in the scheme
No academic access to the practice mortality database
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 7 / 33
7. Design and setting
Design: Longitudinal spatial study, at the Lower Super Output
Area (LSOA) level
Setting: 32482 LSOAs (neighbourhoods of 1500 people on
average), covering the whole population of England (≈ 53.5
million), from 2007 to 2012
Participants: 8647 English general practices participating in the
QOF for at least one year of the study period, including over 99%
of registered patients
Intervention: National pay-for-performance programme
incentivising performance on over 100 quality-of-care indicators
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 9 / 33
8. Main outcome measures
All-cause mortality
Cause-specific mortality rates for six chronic conditions:
diabetes
heart failure
hypertension
ischaemic heart disease
stroke
chronic kidney disease
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 10 / 33
9. Generating the outcome variables
using ONS data
Revised annual LSOA population estimates, 2005-2012:
based on 2001 and 2011 census information
broken down by age and sex
Got annual death counts at the LSOA level, 2005-2012:
broken down by age and sex
Calculated annual and 2-year age and sex standardised mortality
rates at the LSOA level:
all-cause
cause-specific
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 11 / 33
10. Other data
and sources
LSOA level
Index of Multiple Deprivation, 2007 and 2010 (ONS neighbourhood
statistics)
Rural vs urban (ONS neighbourhood statistics)
Lots of collinear 2011 census variables (ONS census)
At the practice level (to be attributed to the LSOA level)
QOF performance (HSCIC)
QOF disease burden (HSCIC)
practice list size (HSCIC)
Spatial shapefile data maps (ONS Geoportal)
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 12 / 33
11. Spatial estimation
first approach: complete local attendance
32482 English LSOAs with complete census, rurality and
deprivation data
≈ 6500 practice-hub LSOAs (at least one practice)
QOF achievement and morbidity burden calculated as sum of all
numerators over sum of all practice denominators
Get longitude-latitude centroid coordinates for all LSOAs
QOF achievement and morbidity scores estimated for the LSOAs
with no practices as weighted means from the 5 ‘closest’ hubs (on
inverse distance ∗ listsize)
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 13 / 33
12. Spatial estimation
first approach: complete local attendance
Bolton Bury
Manchester
Oldham
Rochdale
Salford
Stockport
Tameside
Trafford
Wigan
(87.0,91.2]
(84.6,87.0]
(82.1,84.6]
[66.2,82.1]
No data
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 14 / 33
14. Spatial estimation
first approach: complete local attendance
007B
009C
014B014E
020B
020G
(87.0,91.2]
(84.6,87.0]
(82.1,84.6]
[66.2,82.1]
No data
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 16 / 33
15. Spatial estimation
second approach: attribution dataset
Complete local attendance assumption difficult to justify for all
patients in all areas, especially urban
HSCIC released information on the attribution of general practice
populations to LSOAs and vice versa
Only covered 2014 but used it as a blueprint to generate annual
attribution datasets from 2011/12 to 2006/7
Poisson and negative binomial regression models
attributed population over time was adjusted for practice list size in
the respective year
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 17 / 33
16. Analyses
Three sets of multiple linear regressions used to investigate the
relationship between QOF quality of care and all-cause and
condition specific mortality:
relationship between QOF scores and 2011-12 SMRs
relationship between changes in QOF scores over a 3 or 5-year
period and 2011-12 SMRs
sensitivity analysis, relationship between QOF quality of care and
mortality over time
Following spatial weighted estimation data were complete for all
32482 English 2001 LSOAs
Each analysis set was applied to both spatial weighted estimation
approaches
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 18 / 33
17. Mortality
by region
North East North West Yorkshire & East MidlandWest MidlanEast EnglandLondon South East South
All-cause death% (2011-12) 1.09 1.05 1.01 0.96 0.98 0.91 0.74 0.93
Condition-specific death% (2011-12) 0.39 0.4 0.4 0.39 0.38 0.38 0.24 0.37
0 0.2 0.4 0.6 0.8 1 1.2
North East
North West
Yorkshire & Humber
East Midlands
West Midlands
East England
London
South East
South Central
South West Coast
England
Condition-specific death% (2011-12) All-cause death% (2011-12)
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 20 / 33
18. Standardised mortality rates
by region
North East North West Yorkshire & East MidlandWest MidlanEast EnglandLondon South East South
All-cause SMR (2011-12) 574 580 541 508 528 466 563 456
Condition-specific SMR (2011-12) 184 198 194 184 184 167 166 154
0 100 200 300 400 500 600 700
North East
North West
Yorkshire & Humber
East Midlands
West Midlands
East England
London
South East
South Central
South West Coast
England
Condition-specific SMR (2011-12) All-cause SMR (2011-12)
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 21 / 33
19. Overall health burden
Greater London
(1.8,3.9]
(1.6,1.8]
(1.4,1.6]
[0.4,1.4]
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 22 / 33
20. Overall quality of care (PA)
Greater London
(83.8,90.9]
(82.5,83.8]
(81.1,82.5]
[68.7,81.1]
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 23 / 33
21. Overall health burden
Greater Manchester
(2.2,2.5]
(2.0,2.2]
(1.9,2.0]
[1.0,1.9]
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 24 / 33
22. Overall quality of care (PA)
Greater Manchester
(84.9,89.8]
(83.6,84.9]
(82.1,83.6]
[73.6,82.1]
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 25 / 33
23. Overall health burden
West Midlands
(2.2,2.7]
(2.1,2.2]
(1.9,2.1]
[0.7,1.9]
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 26 / 33
24. Overall quality of care (PA)
West Midlands
(84.4,88.3]
(83.4,84.4]
(82.4,83.4]
[77.5,82.4]
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 27 / 33
25. Spatial analyses
on all-cause SMRs
QOF Year 8 (2011/12)* QOF Year 7 (2010/11)* QOF Year 6 (2009/10)* QOF Year 5 (2008/9)*
Outcome: all cause SMR; QOF predictors: overall population achievement, overall morbidity load
Index of Multiple
Deprivation 2010
7.44(7.24,7.65)
<0.001(93.67)
7.45(7.24,7.65)
<0.001(93.60)
7.41(7.21,7.62)
<0.001(93.25)
7.46(7.26,7.67)
<0.001(93.40)
Rural (vs urban)
-44.52(-52.17,-36.86)
<0.001(-14.98)
-44.20(-51.86,-36.53)
<0.001(-14.86)
-44.62(-52.28,-36.96)
<0.001(-15.01)
-43.76(-51.42,-36.10)
<0.001(-14.72)
% White population
-0.45(-0.63,-0.28)
<0.001(-6.60)
-0.42(-0.60,-0.25)
<0.001(-6.13)
-0.47(-0.65,-0.29)
<0.001(-6.82)
-0.44(-0.62,-0.27)
<0.001(-6.48)
% Population
achievement (PAoval)†
0.73(-0.60,2.07)
0.158(1.41)
0.47(-0.93,1.87)
0.39(0.86)
-0.06(-1.36,1.24)
0.903(-0.12)
-0.26(-1.67,1.15
)0.636(-0.47)
Morbidity load
(MLtot)†
-77.38(-86.91,-67.84)
<0.001(-20.90)
-72.06(-80.97,-63.15)
<0.001(-20.84)
-69.55(-78.42,-60.69)
<0.001(-20.20)
-82.27(-92.11,-72.43)
<0.001(-21.54)
Model intercept
530.88(421.25,640.50)
<0.001(12.47)
548.14(431.23,665.04)
<0.001(12.08)
588.22(480.14,696.31)
<0.001(14.02)
609.38(490.74,728.03)
<0.001(13.23)
Adjusted R2
29.1% 29.1% 29.0% 29.1%
* Year of reference for the two QOF variables in the model.
†
Time-varying QOF variables across different models. All other variables do not vary over time.
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 28 / 33
26. Spatial analyses
on cause-specific SMRs
QOF Year 8 (2011/12)* QOF Year 7 (2010/11)* QOF Year 6 (2009/10)* QOF Year 5 (2008/9)*
Outcome: condition specific SMR; QOF predictors: nine indicator outcome population achievement, five domains morbidity load
Index of Multiple
Deprivation 2010
2.41(2.27,2.55)
<0.001(43.48)
2.40(2.26,2.54)
<0.001(43.30)
2.41(2.27,2.55)
<0.001(43.49)
2.41(2.27,2.55)
<0.001(43.43)
Rural (vs urban)
-3.67(-9.11,1.77)
0.082(-1.74)
-3.81(-9.25,1.64)
0.072(-1.80)
-3.70(-9.14,1.75)
0.08(-1.75)
-3.61(-9.06,1.83)
0.087(-1.71)
% White population
-0.32(-0.45,-0.20)
<0.001(-6.68)
-0.34(-0.46,-0.21)
<0.001(-6.82)
-0.32(-0.44,-0.19)
<0.001(-6.43)
-0.32(-0.45,-0.20)
<0.001(-6.59)
% Population
achievement (PAoutx)†
0.26(-0.47,0.98)
0.359(0.92)
0.11(-0.57,0.78)
0.688(0.40)
-0.08(-0.74,0.59)
0.768(-0.30)
0.21(-0.46,0.87)
0.419(0.81)
Morbidity load (ML9)†
32.37(4.87,59.87)
0.002(3.03)
40.39(12.02,68.76)
<0.001(3.67)
34.76(6.64,62.89)
0.001(3.18)
33.09(3.20,62.98)
0.004(2.85)
Model intercept
140.23(87.17,193.29)
<0.001(6.81)
150.12(102.79,197.45)
<0.001(8.17)
163.96(117.51,210.40)
<0.001(9.09)
144.46(97.14,191.77)
<0.001(7.86)
Adjusted R2
7.8% 7.8% 7.8% 7.8%
* Year of reference for the two QOF variables in the model.
†
Time-varying QOF variables across different models. All other variables do not vary over time.
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 29 / 33
27. Spatial analyses
summary of results
All-cause and cause-specific mortality rates declined over the
study period
Higher mortality associated with:
greater area deprivation
urban location
proportion of a non-white population
No relationship between practice performance on QOF quality
indicators and all-cause or cause-specific mortality rates in the
practice locality
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 30 / 33
28. Conclusions
Higher reported achievement of activities, incentivised under a
major, nationwide pay-for-performance programme for primary
care, did not appear to result in reduced incidence of premature
death in the population
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 31 / 33
29. Future work
complex methods that need to be re-used to answer more questions
Spatial analysis linking pollution, smoking, BMI, IMD and other
census variables to:
all deaths
cancer related deaths
Spatial analysis linking QOF, distance to practice, patient
satisfaction, IMD (except health sub-domain) to:
standardised all hospital admissions
standardised emergency hospital admissions
Structural equation modelling (SEM) to investigate IMD subscales
on all-cause mortality at the population level
SEM to investigate obesity at the population level
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 32 / 33
30. yCareResearchGroup
[Poster tit
ABSTRACT
TITLE:
[Add text here.]
BACKGROUND:
[Add text here.]
OBJECTIVE:
[Add text here.]
METHODS:
[Add text here.]
RESULTS:
[Add text here.]
CONCLUSIONS:
[Add text here.]
BACKGROUND
[Add title, if necessary.]
Label One
[Replace the following names and titles with those of the actual contributors: Helge Hoeing, PhD1; Carol Phi
1[Add affiliation for first contributor], 2[Add affiliation for second contributor], 3[Add af
METHODS
[Add title, if necessary.]
[Add key point.]
[Add description of key point.]
[Add key point.]
[Add description of key point.]
[Add key point.]
[Add description of key point.]
RESULTS
[Add title, if necessary.]
[Add key point.]
[Add key point.]
[Add key point.]
[Add key point.]
[Add key point.]
0% 20% 40% 60%
1st Qtr
2nd Qtr
3rd Qtr
4th Qtr
4th Qtr
[Replace, move, resize, or delete graphic, as necessary.]
[Replace, move, resize, or delete graphic, as necessary.]
[Replace, move, resize, or delete graphic, as neces
Excepteur Sint
Lkl
(n=212)
Controls
(n=27)
Lorum Wt (kg) 18 (SD 10) 29
(SD 07)
Ipsum (wk) 31 (SD 5) 37 (SD 2)
Irure: B
W
H
HB
O
Unknown
79 (373%)
121 (571%)
2 (09%)
0
1 (05%)
9 (42%)
7 (259%)
18 (667%)
0
1 (37%)
1 (37%)
0
Proident
F
106 (50%)
101 (476%)
17 (63%)
10 (37%)
Kontopantelis E, Springate DA, Ashworth M, Webb RT, Buchan IE and Doran T.
Investigating the relationship between quality of primary care and premature
mortality in England: a spatial whole-population study. BMJ, in print
Comments, suggestions: e.kontopantelis@manchester.ac.uk
Kontopantelis (University of Manchester) quality of primary care & mortality 28 Jan 2015 33 / 33