A regulators perspective:

evidence of anti-selection and

experience in addressing risk

pooling failures and benefit design

Council for Medical Schemes

1

Contents

• Introduction

• Anti-selection evidence

• Experience in risk pool regulation

• Conclusion

Impact analysis 2

In SA there is limited empirical evidence on the factors

influencing member movement between different

options and medical schemes.

• International evidence - Health Insurance (Netherlands,

China, Belgium)

• 5 case studies (2004- 2014)

These studies describes the impact of anti-selection

within particular conditions and circumstances

3

Introduction

Introduction

• Anti-selection behaviour can be analysed according to the following categories:– Age: young people defer scheme membership

– Gender: females during child-bearing ages

– Disease burden: people with expensive illnesses to treat or multiple illnesses

– Benefit options: members selecting options with comprehensive benefits only when their likelihood of needing those benefits is high

– Voluntary membership where mandatory participation is not enforced through legislation

– Employer group preferences and splitting of the risk pools

4

• Case study 1 (Chronic Renal Failure)

There is anecdotal evidence that older people with chronic renal failure needing

dialysis are encouraged to join medical schemes in order to get dialysis in the

private sector, as there are limited resources in the public sector. The impact on a

medical scheme is substantial. In 2017, the average cost of chronic renal

failure was R27 590 per patient per month according the data from the Scheme

Risk Measurement (SRM, formerly REF). Such costs have an impact on

contribution increases for all members within a benefit option and a scheme as a

whole.

5

Anti-selection evidence

Anti-selection evidence

• Case study 2: (Pregnancy)

“There are instances where medical scheme cover will be taken by couples or

single women who are planning to have children. The medical and

accommodation costs associated with the pregnancy will be covered by the

medical scheme. Because of information asymmetry, plans to start a family are

considered privileged information and deliberately not disclosed to the medical

scheme.”

6

Case Study 3: Age specific anti-selection: Polmed

7

Anti-selection evidence

The analysis of Polmed claims data for the years 2006 and 2007 showed that

various categories of anti-selection have affected the cost of Polmed options, namely:

“Buy-downs” from the higher to the lower plans by older members. This category

of anti-selection shows that the number of members in age categories from 35

upwards remained constant or declined in the higher plan, while growing in the

lower plan. The data also showed that the level of hospital costs per beneficiary

increased markedly with age. This trend was apparent in all the other benefit

categories.

Anti-selection evidence

• Case study 4: Mcleod and Ramjee (2007)

Undertook a study which compared the relationship between income, age and

gender to expose the anti-selection behaviour that occurs in the voluntary medical

schemes environment. This analysis was undertaken in 2007 where CMS’ Risk

Equalisation Fund (REF) data was analysed.

This study showed higher numbers of maternities than expected in the REF pricing

each year. Furthermore, unpublished scheme investigations also showed

substantially higher maternities than expected as well as some evidence of

increasing numbers of women who join schemes before giving birth and

leaving schemes thereafter (Mcleod & Ramjee, 2007).

8

Case study 5 : Large scheme experience (2014)

9

Anti-selection evidence

0%

2%

4%

6%

8%

10%

12%

14%

16%

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 > 15

Years on Scheme

Duration on Scheme – Musculoskeletal conditions 14.31% who claimed biologics for musculoskeletal conditions had been on the Scheme for <1 year

Medical Scheme presentation to CMS , 2014

10

Anti-selection evidence

Medical Scheme presentation to CMS , 2014

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

Years on Scheme

Duration on Scheme - Multiple Sclerosis (MS)

16.52% who claimed interferon had been on the Scheme for <1 year

Anti-selection evidence

11

0-3 months 4-6 months 7-9 months 10-12 months 13-18 months 19-24 months 2-3 years 3-4 years 4-5 years >5 years

Other underwriting categories 7 155 8 933 9 976 8 084 13 424 9 315 11 172 4 742 1 496 153

7 155

8 933 9 976

8 084

13 424

9 315

11 172

4 742

1 496 153 -

2 000

4 000

6 000

8 000

10 000

12 000

14 000

16 000

Nu

mb

er

of

live

s

Duration on the scheme before first maternity admission (2008- 2013)

34% join within 9 months of the maternity event

12

Anti-selection evidence

0-3 months 4-6 months 7-9 months 10-12 months

Underwriting category A 29,8% 32,1% 29,9% 25,1%

29,8%

32,1%

29,9%

25,1%

0,0%

5,0%

10,0%

15,0%

20,0%

25,0%

30,0%

35,0%

Wit

hd

raw

al r

ate

Duration on scheme before first maternity admission

1-year withdrawal rate of lives who joined <12 months before maternity admission

25% of these leave the scheme within 12 months of the maternity event

Experience in addressing risk

pooling failures

13

Contents

• Introduction

• Risk pool size

• Case studies

• Benefit option registration

• Regulatory gaps

• Conclusion

Impact analysis 14

Introduction

⁻ Pooling involves the accumulation of health revenues on behalf

of a population for eventual transfer to providers

⁻ Core objective for pooling arrangements: maximize redistributive

capacity, financial protection and equity in service use relative to

need for care

⁻ Size (bigger) and diversity (more) are key characteristics to

consider in reform design

WHO Advanced Course on Health Financing for Universal Health Coverage, Barcelona,

Spain 12-16 June 2017 15

Introduction: SA healthcare system

Revenue

sourcesGeneral taxation

Medical schemes

(voluntary contribution)

Pooling

Purchasing

Provision

Population

MOH

uninsured VHI

poor (84%) middle class & rich (16%)

OOPS82 medical schemes

&

287 benefit options

9 provinces

private & some public

Introduction: SA healthcare system

Revenue

sourcesGeneral taxation

Medical schemes

(voluntary contribution)

Pooling

Purchasing

Provision

Population

MOH

uninsured VHI

poor (84%) middle class & rich (16%)

OOP

S82 medical schemes

&

287 benefit options

9 provinces

private & some public

Fragmented risk pools

Duplicated administrative systems

Fee-for-service

Profit motive

Poorly designed

PMBsUp-coding

Affordability challenges!Fiscal constraints!

Passive purchasing

Line-item budgeting

Poor central bureaucracy

RWOPS

High unemployment

rate

Some “signposts” that tell us we have

a problem

• Pooling

– Is pooling highly fragmented and complex?

• Benefit design

– Population not aware of entitlements and obligations?

18

Risk pool size

19

Risk pool size

20

0

50

100

150

200

250

300

350

400

450

2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

Tota

l nu

mb

er

of

be

ne

fit

op

tio

ns

Number of benefit options (2006-2015)

Open schemes Restricted schemes All Schemes

21

Risk pool size

231

247

218 223

192

156145

132

145 150

0

50

100

150

200

250

300

2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

Num

ber

of b

enef

it op

tions

Number of loss making options (2006 -2015)

Open schemes Restricted schemes All Schemes

(231 – 150)

Risk pool size

22

27

83

115

48

0

20

40

60

80

100

120

140

<=1 000 >1 000 & <= 6 000 >6 000 & < = 30 000 >30 000

Num

ber

of o

ptio

ns

Option size (number of beneficiaries) categories, 2015

Community rate analysis - schemes

Impact analysis 23

Community rate analysis - options

Impact analysis 24

R 745,6

R 0

R 500

R1 000

R1 500

R2 000

R2 500

R3 000

R3 500

1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 103

109

115

121

127

133

139

145

151

157

163

169

175

181

187

193

199

205

211

217

223

229

235

241

247

253

259

265

271

277

283

289

295

301

Com

mun

ity r

ate

pbpm

, Dec

ebm

er 2

016

Benefit Options

Open Restricted Industry Community Rate

Case studies

COMMED (2016):

• Total membership – 7981

• Number of option – 4

• Membership on a per option level: Deluxe – 869

Standard – 2195

Roots – 4894

Shina – 23

25

Case studies

• Naspers Medical Fund (2016):– Total membership – 16100

– Number of option – 3

• Membership on a per option level– Option A: 2900

– Option B :3500

– Option C : 9700

• 1 case Gaucher's disease crippling the schemes26

Benefit option registration

• Deregistration of an option is normally considered

as a last resort, such regulatory intervention has a

potential of creating pricing uncertainty and can

lead to member dissatisfaction.

• Close monitoring of loss making options.

Impact analysis 27

Benefit option registration

• If the benefit option continues to be non-

compliant as per provisions outlined above, the

Registrar reserves the right to deregister those

benefit options.

28

Section 63, Section 24 (2) (d) ,

Reg 2 (3)

29

“…Minimum number of members required for the registration of a medical scheme

established after these regulations have come into operation is 6000..”

SUREMED case :

• The judge set aside the decisions of the Registrar and the Appeal Board.

• The high court held that it was not competent for the Registrar to confirm the exposition because

the parties’ merger agreement was rendered void when Suremed’s members voted against the

merger.

• Section 63(11) does not authorise a medical scheme to enter into a transaction that is in conflict

with its rules.

• The high court further held that the Registrar did not have the power to confirm an exposition

which was not underpinned by a valid and binding agreement.

Regulatory gaps

• Amendment of the MSA :

– Regulation 2(3) to explicitly state that all registered

medical schemes should always have the minimum

of 6000 members whether or not these medical

schemes were registered before the amendment of

the 1967 Act.

– Explicitly state corrective measures to be followed by

the Regulator in addressing non-compliance .

30

• Amendment of the MSA:

⁻ Outline of the required membership base at a benefit option l

evel.

⁻ Section 33 (2) (c) does not permit withdrawal of the

benefit option if that option is financially sound, even

if that benefit option has low membership.

31

Regulatory gaps

• Amendment of the MSA:

⁻ A clear interpretation of “public interest” as outlined

within Section 24 (2) (f) is also required.

⁻ Such interpretation needs to take into consideration

membership growth requirements, consumer

preference and the impact of option selection by

employer groups

32

Regulatory gaps

Conclusion

• World Health Organisation recommends that health

financing reforms should not only focus on increasing

the level of prepayment funding for the risk pools, but

should also consider policy options to encourage risk

pool consolidation. Implementing such measures

without paying proper attention to changes in risk pooling

can result in increased fragmentation and

compromised equity and efficiency goals

(WHO 2010).

33

Reference

• Buchmueller T C, Feldstein P J, 1997. The effect of price on switching among health

plans. Journal of health Economics 16 231-247

• Cutler D, Zeckhauzer R, 1997. Adverse Selection in Health Insurance. Working Paper

6107

• FinMark Trust 2009 “Making health insurance work for the low-income market in

South Africa :Cost drivers and strategies”

• Mcleod H, Ramjee S. Medical Schemes. In. Harrison S, Bhana R, Ntuli A, eds. South

African Health Review 2007 . Durban Health Systems Trust 2007

• Mcleod H, Globler P. The role of risk equalisation in moving from voluntary private

health insurance to mandatory coverage: the experience in South Africa. Advances in

Health Economics & Health services Research Vol 19 2009.

• WHO. Essential benefit packages: What are they for? What do they change? Draft

Technical brief No.2. July 2008

34