The role of actuaries in the
healthcare system
November 2014
Emile Stipp
Agenda
The role of actuaries in different healthcare systems around the world
Providing understanding: The drivers of healthcare inflation
Developing solutions: Wellness programmes
The role of actuaries
Our skills
Who we advise
Actuaries
The questions we ask
The IAA
Define actuary
a c · t u · a r · y, n o u n
A person who compiles and analyzes statistics and uses them to calculate insurance risks and premiums
Defining characteristics
• Pragmatic
• Numerate, including good understanding of statistics
• But we are data skeptical, and often accept that we work with imperfect data
• We have a very good understanding of how healthcare systems, or administration of healthcare
arrangements, impact on costs and risks, and we incorporate this in our analysis and projections
• We have a deep understanding of demographic trends and how they impact on costs and risks
• Strong emphasis on professional standards and ethics
•
•
•
•
Avoid conflicts of interests
Balanced, objective advice
With full disclosure
Doing our work for the benefit of Society
The actuarial toolkit
Projecting mortality & morbidity &
financial outcomes
Matching of assets and liabilities
Exposure
Risk immunisation & mitigation
Frequency / severity analyses
Optimisation
Good statistical understanding
Anti-selection and its antidotes
Reserving for liabilities incurred
Advisory services
Public healthcare
Private healthcare
Supply side organisations
Insurers
Governments
Government
bodies
Health
indemnity
Critical
illness
Disability
Long term
care
NGOs
Managed care organisations
Policy makers
Third party
administrators
The role of health actuaries
Public healthcare
•
•
•
•
•
•
Budgeting and risk
adjustment
Risk equalisation
Demographic and financial
projections
Funding sustainability
Public Private Partnerships
Analysis of cost drivers
Private healthcare
•
•
•
•
•
•
•
•
•
Analysis
Capital
Disclosure
Valuation
Product design
Pricing
Risk management & managed
care
Optimisation
Projections
The questions actuaries ask
•
•
•
•
•
•
What is the current and prospective burden of healthcare in the context of GDP, household income, and other economic
indicators
What drives disability claims experience?
What drives healthcare inflation?
What is the impact of anti-selection on health insurance risks?
How can costs be managed?
Can wellness programmes make a real difference to medical inflation?
•
•
•
•
•
•
How should products be designed to introduce the right incentives?
What premium should be charged? How to optimise it?
How do we design and select networks of providers to improve efficiencies and quality?
Can alternative reimbursement models be designed to control costs without compromising on quality?
What is the best way to detect and prevent health fraud and abuses in healthcare?
What are risk-adjusted cost differentials between different service providers?
•
•
•
How can private / public partnerships be structured?
How do we insure low income individuals?
Are out-of-pocket expenses equitably distributed between different levels of income?
•
•
•
What are the risk consequences of catastrophic events, such as a pandemic?
What capital is required to protect against adverse events?
How will the HIV epidemic affect insurance costs?
•
How do we ensure that more people have access to health services and do not suffer financial hardship paying for them?
The 3 dimensions of Universal Coverage
Source: WHO WHR 2010.
Our contribution
– Applying the Mathematical / statistical skills of actuaries to the quantification
of cashflow and capital and their associated risks
– Our role is to support policy makers and managers by quantifying expected
outcomes and the risks of deviations both in terms of costs and demand on
resources
– Expected outcomes are estimated by applying actuarial methodology to
factual data and assumptions including the presumed impact of policy
decisions
– Enabling decision makers and managers to compare ex-ante the expected
impact of policy decisions or strategic interventions facilitate optimisation
– As outcomes are explicitly linked to the various drivers there is value added in
the possibility of monitoring the actual outcomes against the expected to
identify the causes of the deviations and apply the feed back to improve the
decision making
– Our methodology helps understand how incentives of role players affect risks
and outcomes
– Our modelling approach tends to be bottom-up & stochastic, rather than topdown & deterministic. We typically don’t assume equilibrium
..
The role of the IAA
Association of worldwide
actuarial professional
associations, with special
interest sections for
individual members
Mission:
 To promote the profession
to the benefit of Society
 Promote professionalism,
develop education,
encourage research
Six strategic objectives:
 Build relationships with key
supranational organisations
 Expand scientific knowledge and skills
of actuaries
 Promote common standards of
actuarial education and professional
conduct
 Develop actuarial profession
worldwide
 Provide a forum for discussion for
actuaries
 Improve recognition of actuarial
profession
The role of the IAA
IAA Health Committee:
• Representatives of member associations
• Purpose to :
o Represent the IAA in international debates on health actuarial matters
o Raise profile of health actuaries
o Support actuaries working in private and public health systems
IAA Health Section:
• Individual membership
• Main objectives: library of actuarial papers, research presented at conferences and
webcasts
• See example papers on risk equalisation
(http://www.actuaries.org/IAAHS/Webcast/RiskAdjustment/RiskAdjustment_Slides.pdf)
• And on stochastic modelling
(http://www.actuaries.org/IAAHS/Webcast/Stochastic/IAAHS_11-15-2010-wocartoons.pdf)
Agenda
The role of actuaries in different healthcare systems around the world
Providing understanding: The drivers of healthcare inflation
Developing solutions: Wellness programmes
Using inflation as an example....
• Of how actuaries analyse problems
• Insights to be gained from actuarial analysis,
and techniques used
• Using South African private health for illustration,
with some references to international experience
• And also offering a solution with roots in South Africa, but which has
spread across the world in different forms
14
Some preliminaries
•
•
Adjusting for exposure is crucial
Consider Simpson’s Paradox:
– In the context of a health insurer with two benefit plans / levels
– Average inflation is 0%, and yet each plan’s inflation rate is 10%!
Number of
members in
Year 1
Premium per
member in
Year 1
Number of
members in
Year 2
Premium per
member in
Year 2
Increase in per
member
contribution
from
Year 1 to Year 2
Plan 1
100
1000
200
1100
10%
Plan 2
50
2000
54
2200
10%
Insurer
150
1333.33
254
1333.86
0%
15
Some preliminaries
• Simpson’s paradox is relevant to:
– health insurers with more than one plan,
– policymakers, considering health inflation across a health
insurance markets
– Governments, considering health inflation in a country
(e.g. public and private spending)
• It implies:
– All inflation studies should adjust for demographic
movements between insurance markets / insurers /
benefit packages
– And not look only at overall average
– Otherwise it will understate inflation where there are
downgrades and overstate where there are upgrades
Some preliminaries
• Consider frequency and severity separately
– As this could provide insight into the reasons for cost increases
• Consider price and utilisation separately
– Price measures tariff increases
• And how that is set by legislation / competition
– And utilisation should be broken down into
• demand side factors and
• supply side factors
SA healthcare inflation exceeds CPI but relatively
low compared to other countries
Net healthcare costs inflation in 2011
9.0
8.0
Net inflation (%)
7.0
6.0
5.0
4.0
3.0
2.0
1.0
India
SA
Russia
Chile
China
Italy
Singapore
UK
France
Brazil
Mexico
Malaysia
UAE
Canada
US
0.0
Out of 52 countries surveyed, SA had the 8th lowest net healthcare cost inflation.
Only India, Philippines, Bulgaria, Cyprus, Romania, Ukraine and Egypt had lower levels
18
Source: Towers Watson 2012 Global Medical Trends
S o u t h A f r i c a : u t i l i s a t i o n i s ke y d r i v e r o f
healthcare inflation (after plan mix adjustment)
Inflation in South Africa: 2008 to 2013 – in Rand amounts
NHE = non-health expenses (cost of administration and managed care
S o u t h A f r i c a : u t i l i s a t i o n i s ke y d r i v e r o f
healthcare inflation (after plan mix adjustment)
Inflation in South Africa: 2008 to 2013
Drivers of the medical inflation
differential
B
Supply-side:
• Fee for service system
• Undersupply of doctors
• New technology and procedures
• New hospitals
• Changes in coding / billing
A
Demand-side :
• Adverse selection
• Increased disease burden
• Ageing
NHE = non-health expenses (cost of administration and managed care
A
Demand side: 2002 to 2012
A
Demand-side: Adverse selection
conundrum
Adverse selection in open medical schemes
1• Young people opt out
of medical schemes
2• Medical schemes have
higher proportions of
older people
1
2
“Impact of adverse
selection estimated at
R13.5bn – 23% of total
contributions for open
medical schemes”
Barry Childs, Lighthouse
Actuarial Consulting
A
Demand side: Increasing burden of
disease
Epidemic of lifestyle diseases
3
4
50
Source: DHMS data, indexed to 2008
Three controllable
behaviours
Four chronic diseases
of lifestyle
Fifty percent of
deaths worldwide
Increasing disease burden in medical
schemes
A
Demand side inflation in South Africa
• Mostly attributable to:
– Lack of a mandate
– Open enrolment, guaranteed acceptance and
community rating
– Very limited underwriting allowed
– Resulting adverse selection – age and chronic
• Roughly 3% to 4% per year attributable to
demand side inflation
B
Supply side: Shortage of doctors
Doctors per 10,000 lives
Developed
BRICS
43
• SA needs to train 2,400
doctors p.a. just to remain
on par with current low
figures
37
35
27
• Average age of specialists
in SA = 55 years
21
17
14
10
Australia
UK
US
Germany
France
Developed
economies
China
Brazil
Russia
BRIC
SA
Source: World Health Stats 2012
India
6
5.5
• SA’s graduates have
remained at 1,200 p.a. for
the last 2 decades
25
B
Supply side: High cost of new medicines
Growth in claimants for high-cost drugs
26
B
Supply side: relationship between market concentration and
bed supply per 1000 lives
In some regions: high bed supply even where low concentration
27
B
Supply side inflation in South Africa
• Attributable to:
– Radiology / pathology
– Increases in hospital beds
– Price of new technologies
• About 1% to 2% per year
• Overall utilisation therefore 4% to 6% per year
above inflation
Another view of healthcare inflation
• CPI is an average of different inflation indices
• Some components of inflation are always higher
than others, e.g. healthcare vs electronic
consumer goods
– Especially those aspects linked to skilled services
• Wages generally keep up with inflation
• Hence all that happens is that people devote a
larger proportion of their salaries to healthcare
over time
• “The Baumol Effect”, after William Baumol’s “The
cost disease”, 2011
But....
• It may be true that healthcare inflation is and
always will be higher than average inflation
• But it is not true that people will continue to
spend a larger proportion of their salaries on
healthcare
• In South Africa, we see that people effectively
buy down their cover to maintain a roughly
similar percentage of their salaries devoted to
healthcare
Affordability – projecting current
trends
35%
Contributions as % of household income
32.3%
30%
Plan mix
impact
25%
22.6%
Family
size
impact
20%
18.9%
15%
2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033
Base Scenario
Base Scenario (constant family size)
Base Scenario (constant plan mix and family size)
Observed plan mix changes and family size changes compensate for above CPI
contribution increases – Baumol effect not observed!
Agenda
The role of actuaries in different healthcare systems around the world
Providing understanding: The drivers of healthcare inflation
Developing solutions: Wellness programmes
Solutions?
One solution to mitigate all of these effects:
Keep people healthy!
Do we have evidence that incentive based wellness
programmes work?
A case for wellness:
Low-levels of wellness engagement; impact of behavioural factors
Benefits are immediate,
price is hidden
Under consumption
of preventive care
Sickness
Benefits are hidden, price
is immediate
Wellness
Lack of information
True efficacy of different health and wellness
approaches is not well understood
Over-optimism
People tend to overestimate their abilities and
health status
Hyperbolic discounting
Future rewards of a healthy lifestyle are
significantly undervalued relative to cost today
A behavioural solution to the under-consumption of wellness:
Member experience
1
2
Know your health
3
Improve your health
Enjoy the rewards
Obtain a Personal
Pathway
Complete an
HRA
Determine Vitality Age
and set health goals
Age
Earn points and
achieve a Vitality status
Earn Vitality rewards
Results of the model:
Vitality benefit utilisation
Gym visits
(million, calendar year)
Cumulative HealthyFood cashback
(R Million since inception)
kulula.com flights
(calendar year)
360
320
800,000
25
280
700,000
240
20
600,000
200
500,000
15
160
400,000
10
120
300,000
200,000
80
100,000
40
5
Jun-12
Feb-12
Oct-11
Jun-11
Feb-11
2011
Oct-10
2010
Jun-10
2009
Feb-10
2008
Oct-09
2007
Jun-09
2006
2006 2007 2008 2009 2010 2011
Feb-09
-
0
-
Vitality clinical foundation: Published research
Date
Date
published
published
Hypotheses/ research
research
Hypotheses/
objective
objective
Preventing
Chronic Disease
October 2009
Fitness engagement and
health and cost outcomes
The Association Between Medical Costs
and Participation in the Vitality Health
Promotion Program Among 948,974
Members of a South African Health
Insurance Company
American
Journal of
Health
Promotion
January/
February 2010
Vitality engagement and
cost outcomes
Participation in Fitness-Related Activities
of an Incentive-Based Health Promotion
Program and Hospital Costs: A
Retrospective Longitudinal Study
American
Journal of
Health
Promotion
May/June 2011
Longitudinal assessment
of fitness engagement
and health and cost
outcomes
Eating Better for Less: A National Discount
Program for Healthy Food Purchases in
South Africa
American
Journal of
Health
Behaviour
In Press
To assess impact of the
discount on healthy food
on fruit and vegetable
intake
Nameofofstudy
study
Name
Journal
Journal
Fitness-Related Activities and Medical
Claims Related to Hospital Admissions —
South Africa, 2006
Engaged members experienced lower costs per patient compared to
other groups
Risk-adjusted hospital admission costs: engaged vs. not engaged Vitality members
Not engaged benchmark
100%
30-40%
90%
15-20%
10-15%
80%
70%
60%
50%
40%
30%
20%
10%
Overall
Cardiovascular
Digestive
Nervous and
musculoskeletal
Respiratory
Endocrine, nutritional
and metabolic
Kidney and UTI
Mental
Cancer
0%
P < 0.001 for all categories (including overall result) except cancer where P < 0.01
Note: Categorisation based on diagnosis-related groupers using ICD-10, CPT-4 and local procedural codes
Source: “Participation in an Incentive-based Wellness Program and health care costs: Results of the Discovery Vitality Insured Persons Study”. Please do not quote without written permission from
Discovery or PruHealth.
Engaged members experienced lower costs per patient, shorter stays in
hospital, and fewer admissions compared to all other groups
Length of stay in hospital (days)
Number of admissions*
100%
100%
98%
Cost per patient
105%
95%
96%
100%
90%
94%
85%
92%
90%
95%
80%
88%
90%
75%
86%
70%
85%
84%
65%
82%
80%
80%
60%
Inactive
Low
Medium
High
engaged engaged engaged
Not registered
Inactive
Low
Medium
High
engaged engaged engaged
Not registered
Fit people make better patients on a risk-adjusted basis
*Patients with at least one admission event
Inactive
Low
Medium
High
engaged engaged engaged
Not registered
Engaged chronic members experienced lower costs per patient compared to
other groups
Not engaged benchmark
100%
90%
20-30%
8-10%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Multiple metabolic
conditions
Mental illness
Cancer
Hypertension
Dyslipidaemia
Beneficiaries with single conditions
P = 0.001 for multiple metabolic conditions, all single conditions are not statistically significant
First model to test impact of increasing engagement
This analysis looks at all Health & Vitality members who subscribed uninterruptedly for a full three year
period to both products. The VEM1 is tracked over this period and the life is categorised into a longitudinal
engagement category. An outcome or target is then explored during the course of the fourth year.
The analysis explores the association between changes in engagement and morbidity measures
1
Vitality engagement metric is the before limit point summation tracking true engagement in the form of activity related events
excluding carry over points, bonus points and outcome points
A longitudinal view of engagement
Compile a three year monthly VEM per beneficiary
Regress a straight line through the data points
Use the results parameters – i.e. the intercept parameter β0 and the slope parameter β1
The parameters are then grouped to create a progression matrix of engagement where the top right hand
corner represent the most engaged group of people
– This was done for various iterations but a 6 month cumulative VEM view was chosen to be the
champion of a three year engagement view
Consider
20 000
VEM
Starting point
category
10 000
y = 1652.9x + 2580
0
Engaged
vs
Most &
increasingly
engaged
y = 1059.1x + 8408
6
12
18
Year 1
Agent A's points
3 year trend for Agent B
24
Year 2
30
36
Year 3
Constantly
unengaged
Agent B's points
20 000
3 year trend for Agent A
VEM
•
•
•
•
10 000
y = 71.429x + 166.67
y = 30x + 570
0
6
Unengaged
12
18
Year 1
24
Year 2
30
36
Year 3
Agent C's points
Agent D's points
3 year trend for Agent C
3 year trend for Agent D
Rate of change
category
Heat map of morbidity rates with risk adjustment
Started high
0.87
0.87
0.79
0.79
0.79
0.91
0.91
0.87
0.79
0.97
0.87
0.87
1
0.97
0.91
Started low
Decreased
engagement
Notes:
1 - estimated parameter for each cell
Starting level refers to the intercept of the straight line
Rate of change refers to the slope of the line, where the line refers to the regression fitted on each individual’s VEM
Increased
engagement
Heat map of relative effect on mortality
Started high
11
0.82
0.67
0.46
0.46
1.17
1.17
0.69
0.46
1.17
0.74
1.48
1.48
0.46
0.46
Started low
Decreased
engagement
Notes:
1 - estimated parameter for each cell
Starting level refers to the intercept of the straight line
Rate of change refers to the slope of the line, where the line refers to the regression fitted on each individual’s VEM
Increased
engagement
Concluding remarks – inflation & wellness programmes
1.
Both demand and supply side causes of medical inflation in South Africa
2.
Incentivised wellness programme is one of the potential solutions
3.
We observe the impact of both selection and behaviour change
4.
Increased engagement in wellness activities is associated with significantly lower mortality and morbidity risks
5.
The combined financial impact of positive risk selection / retention and health engagement generate significant
financial benefits for members of medical schemes
6.
But also leads to significant mortality improvements over time....
BOTTOM LINE: EVERYONE SHOULD EXERCISE!
Conclusion
•
We believe actuaries have deep insight into
healthcare systems that could be of value to
policy makers
• Whether in the Public or Private sector
•
Our insights are based on detailed but
pragmatic analyses, and we are “data
sceptical”
•
We emphasise context: role players’
incentives, impact of administration
arrangements
Conclusion
•
We place strong emphasis on
professionalism and ethics, and we are
objective and balanced in our advice
•
We focus on understanding long and short
term risks and how to mitigate them
•
We aim to do our work to the benefit of
Society