The relevance of the Purple Book for current debates

Pia Mäkelä, THL. Nordic Alcohol and Drug Researchers’ Assembly (NADRA), 2.9.2016, Helsinki

1. The relevance of the Purple Book for
current debates
Pia Mäkelä
Alcohol and Drugs Unit
Nordic Alcohol and Drug Researchers’ Assembly (NADRA), 2.9.2016, Helsinki
24.8.2017 Pia Mäkelä 1

2. Policycontextin Finland
•In Finland, availabilityof alcoholand alcoholpriceshavebeen
usedas a toolin alcoholpolicyto minimizeharmfrom
drinkingeversincethe Prohibition (1932)
•Thispolicywasstrengthenedby:
–Experiencefrom”Mieto linja” policy(”The mildpolicy”) in 1950’s &
1960’s favouringmildbeverages(wine; 1969 beer): the aimto
replacespiritsand reduceintoxicationdidnotsucceed, on the
contrary
–Researchevidence& theoryfirstgatheredin the purplebook
24.8.2017 Pia Mäkelä 2

3. Policycontextin Finland
•A new AlcoholAct to replacethe obsoleteonefrom1995
(and partly1932) in the makingsince2011 –firstgovernment
failed
•Recentcompromiseof the 3 partiesin government(Centre
party, National CoalitionParty/Moderates, TrueFinns):
–Maximum limitfor alcoholicbeveragessoldin grocerystores:
4.7% -> 5.5% (46% of recordedsalesin thiscategory)
–Alsospirits-baseddrinks(<5.5%) allowedto besoldin grocery
stores
–”de-bureaucratization” of variousaspectsof on-premise
sales
–”farmsales” frommicro-breweries(for theirownbrewed
products, withoutupperalcohollimit) allowed; goesto EU
notification.
24.8.2017 Pia Mäkelä 3

4. Policycontext/ heateddebateabouteffects
•Alcoholresearchers& publichealthpeople: increasein
consumptionand harmsto beexpected
•Aggressivelobbyingbybeerindustryin particular, with
suggestionsthatthe onlyimpactwouldbeto movesalesfrom
Estonia and Alko to grocerystores
•Alcoholindustrynotalone–a retiredalcoholresearcher(KP)
as an activecontributorin popularizingmedia (and many
activistsin social media)
–KP: ”8 mythsof alcoholpolicy”
–3 of thesequestionthe messagesin the Purple Book
24.8.2017 Pia Mäkelä 4

5. Main messageof the Purple Book
•“Changes in the overall consumption of alcoholic
beverages have a bearing on the health of the peoplein any
society. Alcohol control measures can be used to limit
consumption: thus, control of alcohol availability
becomes a public health issue”
= a population / universalistic approach should complement an
individualistic / treatment approach
24.8.2017 Pia Mäkelä 5

6. Purple book& successors(esp. AlcoholPolicy
and the Public Good)
•PB --main postulatedmechanism:
–Increase/decrease in mean consumption -> increase/decrease
in heavy drinking -> increase/decrease in mortality/harm
•APPG –morethoroughunderstandingof and supportfor
mechanisms:
+ prevention paradox(i.e. harmarisesalsofromothersthanheavy
drinkers–and theyaremorenumerous)
& collectivityof drinking(whenmeanconsumptionchanges,
drinkingpopulationmoves’in concert’ upand downthe
consumptionscale)
24.8.2017 Pia Mäkelä 6

7. Questionsto answer(bysession organizerTomi Lintonen)
•Has the main theses of PB endured the test of time?
•Is contemporary research still behind the claim that the
average amount of alcohol consumption in a society directly
affects the amount of problems experienced by the society?
•Is there still reason to believe that the most effective way to
reduce alcohol problems is through policies aimed at reducing
total consumption?
I use the counter-arguments in the Finnish debate (3 “myths”) to
inspect this
24.8.2017 Pia Mäkelä 7

8. ”KP / Myth1: Total consumptionshouldbe
reducedin orderto reducealcohol-related
harm”
Counter-claimsbyKP: [iftotalconsumptionis reduced…]
–”The reductionin consumptionis mainlylimitedto moderate
drinkers”
–”harmsarenotaffectedmuch, buttaxrevenueis reduced”
24.8.2017 Pia Mäkelä 8

9. 24.8.2017 Pia Mäkelä 9
”[iftotalconsumptionis reduced] the reductionin consumptionis
mainlylimitedto moderatedrinkers”?
–quitethe contrary!
The higherthe
population’s
mean
consumption, the
greateris the
proportion of
heavy drinkers!
Data: Finland, drinking
habitsurveys1968-
2008

10. 24.8.2017 10Lähde: Rossow, Mäkelä & Kerr, Addiction2014
Populationmeanconsumption, litres100% alcohol(log-scale)
Volume of alcoholconsumedin givengroupsof drinkers(Finland, surveyyears1968-2008):
1) Lightdrinkers/ P25: 25% consumeless, 75% consumemore
The amountsof
alcoholthat
prototypicallight
drinkersdrink
closelyfollows
changesin per
capita
consumption

11. 24.8.2017 11
Boththe population’s
heaviestdrinkers(gray,
P95) and lighest
drinkers(black, P25)
and allgroupsin
betweenchange
similarly
The wholedrinking
distributionmoves
collectivelywhen
consumptionchanges
Volume of alcoholconsumedin the givengroupsof drinkers(Finland, surveyyears1968-2008)
Populationmeanconsumption, litres100% alcohol(log-scale)
Source: Rossow, Mäkelä & Kerr, Addiction109:1447–55, 2014
1968 drinkingdistribution
2008 drinkingdistribution

12. ”[iftotalconsumptionis reduced]…harmsare
notaffected”? –quitethe contrary!
Per capitaconsumption, assaultsper 1000 and alcohol-inducedmortality, 1969-2012
24.8.2017 Pia Mäkelä 12
0
2
4
6
8
10
12
0
5
10
15
20
25
30
35
40
45
50
196919721975197819811984198719901993199619992002200520082011
Mortality
Per capita
consumption
Assaults
Litres100% alcohol; Assaultsper 1000Mortalityper 100 000

13. Per capitaconsumption, assaultsper 1000 and
liverdiseasemortality, 1969-2012
24.8.2017 ALHU 13
0
2
4
6
8
10
12
14
16
18
20
22
19691973197719811985198919931997200120052009
Litres
; Assaults
per 1000;
Deaths
per 100 000
Assaults
Liver diseases
Per capita
consumption

14. Alcoholtaxrevenuein Finland, 2002-2013
million€
0
200
400
600
800
1 000
1 200
1 400
1 600
milj. euroa
Toteutuma
Ennuste
14
”harmsarenotaffectedmuch, buttaxrevenueis reduced”? –Quitethe contrary!

15. ”KP/ Myth2: A tightalcoholcontrolpolicy
protectspublichealth”
•KP: ”a tightalcoholpolicydoesnot, accordingto a
comparisonof 30 OECD countries, reducedisabilityadjusted
life yearsdueto alcohol”. [Poikolainen, Alcohol& Alcoholism
2015]
24.8.2017 Pia Mäkelä 15

16. Poikolainen2015: Alcohol-related DALYs per
100 of population by per capita alcohol
consumption (no evidence of positive correlation)
24.8.2017 Pia Mäkelä 16Source: Poikolainen, Alcohol& Alcoholism2015

17. Problemin inference
No inferenceon causaleffectorlackthereofcanbemade on the
basisof this:
-confoundingfactorsarenotcontrolledfor –the societiesdiffer
in innumberableways, the effectof onefactorcannotbe
separatedbycross-sectionalcorrelationdiagram
-Imaginesimilarcomparisonin anotherpolicyarena:
Canthe UN policyof usinginsecticidesagainstmalaria-
spreadingmosquitosbeassessedbythisdesign?
Inferencewouldbe: In the countriesthatinvestin insecticide-
favouringpolicies, thereis no less(butevenmore!) malaria
=> ineffectivepolicy?!
24.8.2017 Pia Mäkelä 17

18. A goodexampleof culturalconfoundingin a
similardesign to Poikolainen’s
(Mats Ramstedt, EuropeanJournal of Population2002):
24.8.2017 Pia Mäkelä 18
Per capitaconsumptionand ”AAA” mortalityin 14 Europeancountries:
no (ora negative) association!
(alcoholdependence, alcoholpsychosis, alcoholpoisoning)
Per capitaalcoholconsumption

19. A goodexampleof culturalconfoundingin a
similardesign to Poikolainen’s
(Mats Ramstedt, EuropeanJournal of Population2002):
24.8.2017 Pia Mäkelä 19

Butonceculturaldifferences(south, north, central) aretakeninto account
a clearpositiveassociation emergesin each!
óCountry comparisonsareNOT a gooddesign for assessingpolicyimpacts
Samedots!

20. ”KP/ Myth3: The largestproportion of alcohol-
relatedharmarisesfromthe largemajorityof
drinkers, whodrinklessthanthe smallnumber
of heavy drinkers” (prevention paradox)
24.8.2017 Pia Mäkelä 20

21. Tworealexamples:
Fromwhichdrinkinggroupsdoesthe burdencomefrom?
24.8.2017 Pia Mäkelä 21
0
10
20
30
40
50
60
70
80
0
1
2
3
4
5
6
0102030405060708090100
RR
0
5
10
15
20
25
30
35
40
45
0
1
2
3
4
5
6
0102030405060708090100
RR
Alcohol intake, gramsper day Alcohol intake, gramsper day
Women/ breastcancerMen/ livercirrhosis
RR RR
Source: Meta-analysis / Rehmet al., Drug and Alcohol Review 2010
Source: Meta-analysis / Corraoet al., Preventive Medicine 2004
Risk

22. Fromwhichdrinkinggroupsdoesthe burdencomefrom?
24.8.2017 Pia Mäkelä 22
Alcohol intake, gramsper day Alcohol intake, gramsper day
Women/ breastcancerMen/ livercirrhosis
RR RR
0
10
20
30
40
50
60
70
80
0
1
2
3
4
5
6
0102030405060708090100
RR
Preval.
0
5
10
15
20
25
30
35
40
45
0
1
2
3
4
5
6
0102030405060708090100
RR
Preval.
% %
Drinking
distributions

23. Burden: notonlyfromheaviestdrinkers!
24.8.2017 Pia Mäkelä 23
Alcohol intake, gramsper day Alcohol intake, gramsper day
Women/ breastcancerMen/ livercirrhosis
RR & alcohol-attributableburden(%) RR & alcohol-attributableburden(%)%
%
0
10
20
30
40
50
60
70
80
0
1
2
3
4
5
6
0102030405060708090100
RR
PAF
Preval.
0
5
10
15
20
25
30
35
40
45
0
1
2
3
4
5
6
0102030405060708090100
RR
PAF
Preval.
Burden

24. 24.8.2017 Pia Mäkelä 24
Whydowepunishthe
moderatelydrinkingmajority
becauseof the fewwho
haveproblems?
Werecognizethe problem
drinkers–weshouldjust
concentrateoureffortson
helpingthem

25. Employmenthistoryof…
24.8.2017 25Source: Paljärvi, Martikainen, Leinonen, Pensola, Mäkelä; PlosOne 2014
Just beforedeath:
¼ wereemployed
0
10
20
30
40
50
60
70
80
90
100
-1-2-3-4-5-6-7-8-9-10-11-12-13-14-15-16-17
%
Years before death/end of study period
The general population
Those
whodied
from
alcohol
causes
in year0
(aged45-64)

26. 0
10
20
30
40
50
60
70
80
90
100
-1-2-3-4-5-6-7-8-9-10-11-12-13-14-15-16-17
%
Years before death/end of study period
24.8.2017 26Source: Paljärvi, Martikainen, Leinonen, Pensola, Mäkelä; PLoSONE 2014
Nearlysame
employment
with general
population!
The general population
Those
whodied
from
alcohol
causesin
year0

27. 0
10
20
30
40
50
60
70
80
90
100
-1-2-3-4-5-6-7-8-9-10-11-12-13-14-15-16-17
%
Years before death/end of study period
24.8.2017 27Source: Paljärvi, Martikainen, Leinonen, Pensola, Mäkelä; PLoSONE 2014
Thesepeople
werecognize
and cantryto
treat
For prevention,
weshould
makean
intervention
here!

28. Summary
–whydouniversalalcoholpolicieswork?
•Collectivity: whatothersdrink, i.e. the drinkingculture affects
heavy drinkers
•Prevention paradox: alsomoderatelyelevatedrisksmatter,
becausetheyapplyto a largenumberof people
•Process/ slowdevelopmentof harms: prevention ratherthan
treatment.
•Drinkingpatternsareimportant! Butit’sverydifficultto make
themchange& changeis slow
24.8.2017 Pia Mäkelä 28

29. Questionsposed:
•Has the main theses of PB endured the test of time?
–Yes: “Changes in the overall consumption of alcoholic
beverages have a bearing on the health of the peoplein any
society. Alcohol control measures can be used to limit
consumption: thus, control of alcohol availability becomes a
public health issue”
•Is contemporary research still behind the claim that the
average amount of alcohol consumption in a society directly
affects the amount of problems experienced by the society?
–Not necessarily in 100% countries 100% of time, but in Finland in
the 2000’s: yes.
•Is there still reason to believe that the most effective way to
reduce alcohol problems is through policies aimed at reducing
total consumption?
–Thisseemsto bethe case
24.8.2017 Pia Mäkelä 29