<Nice cover picture

HMI Seminar

WHO / OECD Working Paper 85 review

30 August 2016

Healthcare focused actuarial consulting firm

Insight Actuaries & Consultants

2

Manage Risk | Develop Opportunity

Do more than ‘keep score’

Data centric

ActuarialConsulting

ProviderBenchmarking

Managed CareAdvisory

PRMAValuations

BusinessIntelligence

DiagnosisRelated Groups

Technical Policy Advisory

StrategicAnalytics

ProductDevelopment

Insight Actuaries & Consultants

Wide range of clients across the spectrum

Independence and trust built up over many years of conduct across

market segments

Sample of current and recent Insight clients

March 2016

Engaged the authors, WHO,

DOH and HMI to establish the

process of formal engagement

Timeline and Insight involvement in OECD Health Working paper 85

4

May 2014

Engaged by WHO to

assist in studying SA

private hospital prices

Sept 2014

Assisted with data from GEMS,

clinical cross walks,

regular query support and refinement

June 2015

Shown initial findings of the

report and verbal feedback

given

Oct 2015

Report released but

subsequently withdrawn

Dec 2015

OECD health Working

Paper 85 referenced in

the NHI white paper

Feb 2016

WHO & OECD presents results

of the Working Paper at the

HMI and the paper is released

Submitted written

response to authors

and the WHO 10

March 2016

Written feedback

received from the WHO

31 March 2016 and

republished working

paper

July 2016

Critiques and responses

released on the CompCom

website

9 May 2016 wrote

back to the WHO

to clarify and

respond and ask

to meet

Clarifying our role in the Working Paper

5

We offered ongoing assistance to the WHO on request, independently and at our own cost.

We facilitated the handover of GEMS data to the WHO at the request of GEMS as part of our actuarial function

for the scheme.

We offered a critique of the Working Paper directly to the authors independently and at our own cost.

Clarifying our role in the Health Market Inquiry

6

Informal briefing session with HMI panel and advisors before public hearings and data collection commenced

Prepared an expert report for Netcare; and presented with Netcare, Bestmed and the IPA Foundation at the

public hearings.

Points of disagreement with the Working Paper

Context

7

Framing

Affordability

Comparability

Test of reasonability

Case mix

Risk Profile

Inferences and suggestions

Two amendments were made to the Working Paper arising from our feedback and correspondence.

1. The data on pathology was corrected.

2. The graph with South African income deciles was included as an appendix (figure 20).

The following areas raised in our letter remain points of disagreement with the Working Paper:

The main points

The paper is mainly concerned with 2 points:

8

1. How do “private hospital prices” in South Africa compare to those in OECD countries?

2. Are these prices affordable for South Africans?

Context

Insufficient context is given as to the reasons why South Africa’s health system looks the way it does.

9

Context does not legitimise the current system but aids in understanding how we got here and why

things look the way they do.

The role of the private sector is not discussed in any detail.

The public sector receives no attention in the paper.

South Africa’s (very) low out-of-pocket expenditure gets a mention, as well as a footnote that it is not

unusual for developing countries to have well developed private health insurance; but private health

insurance as a share of total current health expenditure for South Africa against OECD countries is

the graphic plotted (figure 1) making South Africa look like the outlier.

Framing

10

“Private hospital prices” are the theme of the paper whereas all services

performed in hospital are considered.

While the issue is mentioned in the paper, it is not emphasised, and the

reader is left with the label of “Private hospital prices” throughout the

paper.

Hospital entities are confused with hospitals as a place of service, which

brings significant interpretation risks for readers unfamiliar with the

South African private provider structures.

62.6%

17.7%

7.0%5.2%5.1%2.4%

Pathology

Doctors

Allieds

Radiology

Hospital

Other

Split of spend for

hospital admissions

The Working Paper makes the argument that “prices set in the private

sector set labour benchmarks that doctors face in choosing between

working in a public or private facility” – but this has nothing to do with

private hospitals. If that were the question, doctor tariffs and earnings

should be compared to doctor salaries in the public sector.

Affordability

11

Inequality in South Africa is not discussed in the paper.

Including the income decile figure goes some way, but not far enough in highlighting the issue and its

bearing on the analysis.

Cross subsidies between private and public sectors are ignored (medicines being the main example).

Income cross subsidies are ignored.

Prices are deemed to relate to income when comparing between countries, but not within countries.

Affordability and Correspondence

12Source: Stats SA, Income and Expenditure survey 2010/2011

0.0%

2.0%

4.0%

6.0%

8.0%

10.0%

12.0%

14.0%

16.0%

18.0%

R 0

R 8

33

R 1

,667

R 2

,500

R 3

,333

R 4

,167

R 5

,000

R 5

,833

R 6

,667

R 7

,500

R 8

,333

R 1

0,0

00

R 1

1,6

67

R 1

3,3

33

R 1

5,0

00

R 1

6,6

67

R 1

8,3

33

R 2

0,0

00

R 2

1,6

67

R 2

3,3

33

R 2

5,0

00

R 2

9,1

67

R 3

3,3

33

R 3

7,5

00

R 4

1,6

67

R 4

5,8

33

R 5

0,0

00

R 5

4,1

67

R 5

8,3

33

R 6

2,5

00

R 6

6,6

67

Proportion of population Proportion of population (grant income removed)

27 million people live below

the upper bound poverty line

(R780 per capita in 2015)

Household income per month

Affordability and Correspondence

OECD countries have more equitable societies and health systems.

Hospital price data from the OECD countries represents hospital usage from each country’s overall population.

Hospital price data for South Africa represents hospital usage from the wealthiest portion of South Africa’s population.

13

1.4% 0.9% 1.3% 1.0% 1.6%2.9%

5.9%

12.8%

26.1%

46.2%

0.0%

5.0%

10.0%

15.0%

20.0%

25.0%

30.0%

35.0%

40.0%

45.0%

50.0%

1 2 3 4 5 6 7 8 9 10

% of medical scheme membership from each household income decile

Source: Stats SA, Income and Expenditure survey 2010/2011

85% of medical scheme members are in

the top 3 income deciles

On average, medical schemes members

income decile is 8.8 (9.4 if weighted by

income)

Inequality in South Africa cannot be

ignored in such analyses, especially when

data from only a wealthy subset of the

population is considered.

Comparability

14

Its not immediately obvious why South Africa should be compared to 20 of the most developed countries

in the world, particularly with respect to the maturity of their health systems.

The data being available for only these countries does not validate the comparison.

There are undeniable structural and economic differences between South Africa and the 20 Eurostat countries.

Even low-income OECD countries are very different from South Africa.

If one was asking the question of whether South Africa’s “private hospital prices” were appropriate one

should start by looking for suitable comparators.

Either the subsection of the South African economy that approximates the OECD should be compared to

the OECD, or South Africa as a whole should be compared to more comparable countries.

Comparability

15

0

20000

40000

60000

80000

100000

120000

GDP per capita, US$, PPP

South Africa’s GDP per capita is 70% below the

comparator countries. And 55% below the

“low-income” OECD countries compared to.

Source: OECD, Eurostat countries & South Africa

Comparability

16

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Gini Coefficient, South Africa versus comparator set

South Africa’s Gini coefficient stands in stark

contrast to the comparator countries,

including “low-income” countries.

Source: OECD, Eurostat countries & South Africa

South Africa has a small subsection of its

economy and population that might proxy

European living standards, but this is not

representative of the population as a whole.

Affordability and Comparability

The working paper insists that “private hospital prices” are unaffordable for

South Africans, including upper income South Africans.

17

SA

SA(Decile7) SA(Decile8) SA(Decile9) SA(Decile10)

PolandHungary

Portugal

Slovenia

Czech Republic Estonia

Spain United Kingdom

Italy

France

IcelandNetherlands

Germany

Austria

SwedenFinland

Ireland

SwitzerlandNorway

Luxembourg

0

50

100

150

200

250

0 5000 10000 15000 20000 25000 30000 35000

(13%)(6%)

Given private hospitals are accessed

predominantly through medical

schemes, and the previously shown

income distribution of medical

scheme membership, the statement

is incorrect.

Furthermore, the analysis ignores

income cross subsidy in the system

in terms of differential hospital

prices on some so called ‘low

income options’ and income banded

scheme contributions.

(46%)(26%)Ho

spit

al co

mp

ara

tive p

rice

level

Household final consumption expenditure per capita (US$ PPP)

Figure 20 - Working Paper 85, recreated

Tests of reasonability

We remain of the view that the results of the paper as seen in figure 5

fail the test of reasonability.

18

Consider what South African

“private hospital prices”

would have to be to not be

deemed an outlier. The

consequences of such an

assertion are unreasonable.

South Africa

Poland

Hungary

Portugal

Slovenia

Czech Republic Estonia

Spain United Kingdom

Italy France

Iceland Netherlands

Germany

AustriaSweden

FinlandIreland

SwitzerlandNorway

Luxembourg

-10%

-25%

-50%

-75%

0

50

100

150

200

250

0 10,000 20,000 30,000 40,000 50,000 60,000 70,000 80,000 90,000 100,000

Ho

spit

al co

mp

ara

tive p

rice

level

GDP per capita (US$ PPP)

Figure 5 - Working Paper 85, recreated

Case mix

Some prices are higher and some are lower which makes case mix

important, as well as the weights used.

19

-80

-60

-40

-20

0

20

40

60

Acutemyocardialinfarction

Anginapectoris

Normaldelivery

Malignantneoplasm of

bronchusand lungs

Heart failure Cholelitiasis Pneumonia

% difference in cost per case for Medical cases SA vs OECD

-60

-40

-20

0

20

40

60

% difference in cost per case for Surgial cases SA vs OECD

-40

-30

-20

-10

0

10

20

30

40

50

Cataract surgery Arthroscopic excisionof meniscus of knee

Ligation and strippingof varicose veins -

lower limb

Tonsillectomy and/oradenoidectomy

% difference in cost per case for Outpatient cases SA vs OECD

The study uses “typical” cases making up around 30% of total

admissions types.

The study data has 60% of cases as surgical rather than 50% in

medical scheme data which would affect the weighting and result.

Many “typical” case types missing, including mental health

admissions.

Change in profile

GEMS data over the period included a large

take-on of pensioners which had a noticeable

effect on the age profile and burden of disease.

However the working paper cites no significant

change in age profile when considering

increases in certain procedures like hip

replacements.

20

Jan

-11

Ap

r-11

Jul-

11

Oct

-11

Jan

-12

Ap

r-12

Jul-

12

Oct

-12

Jan

-13

Ap

r-13

Jul-

13

Oct

-13

Jan

-14

Ap

r-14

Jul-

14

Oct

-14

Jan

-15

Ap

r-15

Jul-

15

Oct

-15

Jan

-16

Ap

r-16

Dis

ease

bu

rden

in

dex

“pre-92” pensioner group joined

GEMS

Source: Dr Guni Goolab, BHF 2016

Inferences and conclusions

21

South Africa is not alone in having

“hospital prices” above “general

price” levels.

Norway, Switzerland, Luxembourg

and Ireland show similar (or higher)

differentials.

Source: OECD Health Working Paper 85

Inferences and conclusions

22

Working Paper 85 and its precursor working paper 75 mention notable shortcomings in the data used

for the analysis. For observation and study this is tolerable, but caution should be exercised when

extending the conclusions of such studies to high impact policy recommendations.

South Africa’s short length of stay compared to other countries is highlighted as a potential quality

concern without considering quality data.

Considering only price differences without considering output or quality differences presents an

incomplete picture.

Clarifications on the Ramjee paper

23

The purpose of the research was to highlight the considerations for setting equivalent prices across the

public and private sectors in a pluralistic purchasing arrangement.

The research illustrated that structural differences between the two sectors accounts for a large

component of the cost differential.

Very clear from the report that the analysis was constrained by data availability. Part of the intention of the

research was to raise awareness about the data requirements to enable price setting across the two

sectors – an essential part of the preparatory work for NHI.

Unfortunately three years later we face similar data constraints – the absence of reliable case-mix data in

the public sector, as well as meaningful cost data for various types of admissions in the public sector.

In our view it is not appropriate to use the Ramjee paper to defend prices in the private sector – and we

did not make reference to it in our communications with the WHO.

Overall conclusions

We note the WHO responses to the criticisms. In our view these responses do not adequately

address the concerns and we remain of the view that the paper has interpretive flaws.

24

Data collected from only upper income South Africans is weighted against the countries overall GDP,

overall comparative price index, and compared to data from OECD countries where data is

representative of the whole population.

The paper’s findings reflect on how poor South Africans are relative to OECD country residents,

rather than whether private hospitals are expensive.

Extension of the findings to upper income South Africa’s seems a particularly big stretch.