tim tenbensel and esther willing, university of auckland
TRANSCRIPT
The Implementation of New Zealand’s Health Targets: Generating Traction or Driving to Distraction?
Tim Tenbensel and Esther Willing, University of Auckland
New Zealand’s health targets2009-2012
95 percent of patients will be admitted, discharged, or transferred from an Emergency Department (ED) within six hours.
95 percent of two year olds will be fully immunised by July 2012
Overview
The debate about health targets
Our questions and our ways of researching them
How were the targets implemented?
Assessing the ‘balance of consequences’ of each target
How to explain different consequences?
Implications and Discussion
The Use of Performance Measures
give account (to external
stakeholders)
steer and control learning
Targets (command and control)
Intelligence (collegial)
Rankings (choice,
competition)
Sources: van Dooren et al 2011;Hood 2007
‘High stakes’
NZ’s health targets
Health Targets: The Case For
Generate sustained improvements (when other policy instruments do not)• cf the UK ‘natural experiment’ of the 2000s
Enable clearer focus – prioritising what is deemed important
Promote distributed ownership and collective problem-solving
Encourage creativity and innovation across services
Improve intelligence of organisation(s) and stimulate learning
Enable learning across organisations
Health Targets: The Case Against
Fosters behaviour that ‘hits the target but misses the point’• Gaming (and possibly cheating)
Generate resistance from clinicians• An extrinsic motivation crowding out an intrinsic one• Undermine clinical autonomy
Focus and attention lead to ‘effort substitution’
Promotes ‘myopic’ behaviour• Focus on the ‘parts’ of the service to the detriment of the whole• Jeopardises inter-organisational relationships
Measured target performance 2009-2012
July 2009 performance
July 2012 performance
Number of DHBs that reached the target (95%)
Immunisation
77% 93% 8 /20
Emergency Department
80% 94% 12 /20
We are not evaluating the targets based on achievement against the targetWe are seeking to understand and learn from implementation processes
Our questions
What conditions are conducive to the application of health targets?
Under what conditions should they be avoided, or used with caution?
Our Research What we can’t answer (at this stage)
o Effects (positive or negative) on health outcomeso The extent, and scope of ‘effort substitution’
What we can answero What implementers did, and how they responded (eg…
buy-in, relationships, systems and processes, gaming)
Immunisation Target Research Methods
Four DHBs as case study sites Quantitative data from the National Immunisation
Register Qualitative data on the implementation process Two phases of interviews
o Phase one – 33 interviews between August and December 2011
o Phase two – 16 interviews between September to October 2012
Key informants across each level of the local health systemo DHB, PHO, National Immunisation Register, Public Health
(Medical Officers)
Identified immunisation champions
Created immunisation health target networks
Improved local systems and processes for immunisation
Data matching between practice records and the National Immunisation Register
Outreach Immunisation Services
What did DHBs and PHOs do in response to the target?
Implementer Experience of the Immunisation Target
Focused attention on
child immunisatio
n
Target Pressure
Target Fatigue
Improved relationships between
local organisation
s
Collegial Competition
Researching the Shorter Stays in ED Target
Stream 4: Integratio
n
Stream 1a: Development of
patient-flow model in conjunction with
sector clinicians and managers
Stream 1b: Survey of target-related
resource use in all NZ hospitals;
Stream 2: utilisation and data for all NZ
hospitals 2006-2012clinical quality
markers in 4 case study hospitals
2006-2012
Stream 3: Semi-structured
interviews with key informants at 4 case
study hospitals 03/04 2011, 06/07
2012, 68 interviews in total
http://www.akhdem.co.nz/ssed-research/
ED Target Research Methods
Four DHBs as case study sites Qualitative data on the implementation process Two phases of interviews
o Phase one – 47 interviews between March to May 2011o Phase two – 21 interviews between July to September
2012 Key informants within case study site hospitals
o ED nurses, doctors, o inpatient specialists, nurseso Allied professionals (hospital)o DHB / hospital managers
Jones, P., E. Sopina, et al. (2014). "Resource implications of a national health target: The New Zealand experience of a Shorter Stays in Emergency Departments target." Emerg Med Australas 26(6): 579-
584.
ED Target: What did DHBs do?
Put more resources into Emergency Departments
Created or beefed up ‘short stay’ wards
More attention to discharge planning from wards
Some changes to referral procedures between EDs and wards
Developed internal information systems about ED LOS
What did hospitals do in response to the target?
Implementer Experience of the Emergency Department Target
Target Focus
Target Pressure
Focus on the Front
of the Hospital
Demands on
Emergency Nursing
and Medicine
Dealing with
Resistance
Push Through
the Hospital
July – September 2011
July – September 2008
Balance of Consequences
Table derived fromo Literature reviews of targets and performance managemento Inductive coding of case study interviewso Comparison across the two targets
Positive (Gaining Traction) Negative (Driven to Distraction)
1) Focus / pressure
Target enables organisation to focus, clarifies organisational and individual priorities
Implementers experience (undue) pressure related to focus on target
2) Buy-in / Resistance
Widespread buy-in and ownership from key staff involved in implementation
Active and/or passive resistance (e.g target is not my problem”)
3) Relationships Target stimulates formal and informal relationships and networks between implementers, facilitates trust, collective problem-solving
Target exacerbates pre-existing tensions, and/or creates new tensions between individuals/ units/organisations
Balance of Consequences
Positive (Gaining Traction) Negative (Driven to Distraction)
4) Systems and Processes
Target generates new and improved systems and processes of monitoring, and follow-up,
Failure to generate new systems and processes and/or innovations experienced as negative
5) Learning Across Sites
Case study sites learning from others’ experiences
Absence of learning across sites, or failed attempts at translation
6) Effects on other services
Target implementation has beneficial consequences for other services
Target implementation has adverse consequences for other services
7) Gaming No evidence of gaming (Substantial ) evidence of gaming
Balance of Consequences: Immunisation Target
1. Focus / pressure ++/- (significant focus and commitment to target, but pressure leads to target fatigue at end)
2. Buy-in / Resistance
+++ (buy-in from all key staff across all organisations)
3. Relationships +++ (target fosters the development of immunisation networks in each sites leading to collective problem-solving)
4. Systems and Processes
+++ (extensive development of new systems and processes across all sites)
5. Learning Across Sites
+++ (sites actively learning from early achievers – adapting solutions to local contexts)
6. Effects on other services
++ (perceived enhancement of access primary care through engagement with practice nurses)
7. Gaming + (no evidence of gaming)
Balance of Consequences: ED Target
Emergency Department target1. Focus / pressure ++/-- (strong focus and clarity in EDs, pressure
experienced negatively in EDs and wards)
2. Buy-in / Resistance
++/-- (buy-in from management and ED, resistance from inpatient specialists)
3. Relationships +/-- (mixed effect on relationships – some new relationships strengthened, many tensions exacerbated)
4. Systems and Processes
++/- (development of new systems and processes in all sites, some experienced as micro-management)
5. Learning Across Sites
+ (limited attempts to transfer learning across sites)
6. Effects on other services
- (perception of chaotic environment and churn in inpatient wards)
7. Gaming -- (clear evidence of gaming: clock-stopping, and decanting patients to short stay units)
Similarities and Differences
Similarities Staff experienced
targets as ‘high stakes’
Significant levels of staff ‘buy-in’ (esp when staff agreed with problem definition)
Stimulation of novel systems and processes
Differences Bottom-up ‘networking’,
stronger relationships across implementation ‘units’ (immunisation only)
Resistance (ED target only)
Flow-on effect to other services (+ve for imms, -ve for ED target)
Gaming (ED target only)
Why are the ‘Balances of Consequences’ Different?
Is it something about the
performance measure?
Is the target measuring the
right thing?
What else is the measure
doing?
Is it something about the
implementation context?
Are inter / intra-
organisational relationships
complex?
Does the target encroach on
medical autonomy?
Is it the Performance Measure?
Synecdoche (Bevan and Hood 2006) – how good a proxy is the target for the desired outcome?
Is it the Performance Measure?
Immunisation Output measure –
highly effective proxy for desired outcome (herd immunity)
No scope for gaming o Numerator is clear (and
audited)o Denominator is clear
(and controlled/audited independently)
Emergency Department
Process/efficiency measure – captures one dimension of ED quality
There is scope for gamingo Particularly regarding
the numerator
Is it the Implementation Environment?
Immunisation Service (and service
context) is not complex
Inter-organisational links:o Facilitated by buy-in,
shared professional identity
At DHB & PHO level, requires only a few key implementers
Emergency Department
Complexity of ED demand & supply
Intra-organisational dynamics are complex
Power and autonomyo Solutions required
changes in inpatient specialists’ conditions
Our questions
What conditions are conducive to the application of health targets?o When the measure is an effective proxy
for the desired outcomeo When the implementation context is fairly
simple Under what conditions should they be
avoided, or approached with caution?o Where the measure is only an
‘approximate’ proxyo When implementation context is complex
The Good News
Targets can work, and have an even wider range of benefits• Reducing inequalities• Harnessing ‘synergy’ between ‘hierarchical’
and ‘network’ governance• Can be used for accountability and learning
(across the ‘target use’ spectrum)How common are the circumstances that enable the effective use of targets?
But……
Targets can have adverse consequences
A ‘screening’ process is necessary
• Synecdoche test• How closely does the measure represent the policy objective?• To what extent is variation (over time, between sites) attributable to ‘external’ factors?
• Gaming opportunities• Implementation environment• Anticipating resistance• Mapping the complexity of inter and intra-organisational relationships
Current health sector approach: if something is a priority, then establish a target performance measure
Consider use of targets alongside other instruments