Kypros Kypri

1
Effects of lowering the alcohol minimum
purchasing age on weekend hospitalised assault in
New Zealand Applied Research in Crime and
Justice Conference Sydney 18-19 February 2015

• Kypros Kypri
• School of Medicine and Public Health
• University of Newcastle, Australia
• Injury Prevention Research Unit
• Department of Preventive Social Medicine,
• University of Otago, New Zealand

2

• Co-investigators
• Patrick McElduff
• University of Newcastle, Australia
• Gabrielle Davie, Jennie Connor, John Langley
• University of Otago, New Zealand
• Funding Health Research Council Project Grant
  2012-15

3
Background

• Minimum purchasing age (MPA) reduced from 20 to
  18 years in December 1999
• Previous studies show deleterious effects on
  traffic injury outcomes consistent with USA,
  Canada, Australia 1970s and 1980s
• Few studies on intentional injury
• Data quality and volume are barriers
• No studies of effects on Indigenous people

4
Evidence on the minimum legal drinking age (MLDA)
/ minimum purchase age (MPA)

• During and after the Vietnam war, 29 states of
  the USA, 3 Canadian provinces and 3 Australian
  states reduced their MLDA/MPAs
• By 1988 all 50 states of the USA increased their
  MLDAs to 21 (note the variation in laws by state)
• Over 100 studies have been published on the
  effects of lowering and increasing the MLDA / MPA
• Evidence shows an inverse relationship between
  the change in MLDA / MPA and levels of alcohol
  consumption and traffic among 18-20 year-olds

5
Logic framework for reviews of interventions to
reduce alcohol-impaired driving

• Shults et al. Reviews of evidence regarding
  interventions to reduce alcohol-impaired driving.
  American Journal of Preventive Medicine
  20012166-88.

6

• Shults et al. Reviews of evidence regarding
  interventions to reduce alcohol-impaired driving.
  American Journal of Preventive Medicine
  20012166-88.

7
Aims

• Estimate effects on the target age group (18-19
  years) and a younger age group (15-17 years) from
  trickle down
• Estimate effects separately for males and females
• Estimate effects separately for Maori (Mana
  Whakamarama equal explanatory power)

8
Causal model
9
Methods

• Pre-post design with age control (20-21 years)
  for economic and other factors affecting drinking
  among young people
• Pre-change period 1996-1999
• (1992 0.03 g/dL law for drivers under 20)
• Three four-year post-change periods
• 2000-2003
• 2004-2007
• 2008-2011 (0.00 g/dL law from August 2011)

10
Patients

• Admitted to public hospitals (97 of acute injury
  cases) in NZ from 0001 Friday to 2400 Sunday
  (weekends)
• Note no alcohol involvement nor any time of
  injury indicator is routinely recorded, thus
  assaults between e.g., 10pm-6am cannot be
  identified
• Cases patients aged 15-17 or 18-19 years
• Controls patients aged 20-21 years

11
Maori ethnicity

• Self-identified ethnic group mandatory in the
  National Minimum Data Set
• Can change over time thus ethnicity data are
  recorded for each hospital admission
• Prioritisation determined using Statistics NZ
  algorithm (NZ Maori highest priority code)
• Ethnicity data in health sector collected in same
  way as Census allowing for valid population
  hospitalisation rate estimates

12
Analysis

• Poisson regression to model change in each age
  group relative to the 20-21 year-olds
• Exponents of fitted coefficients are equivalent
  to Incidence Rate Ratios (IRR) with the
  pre-postage group interaction terms providing
  pre-post IRRs relative to the comparison age
  group

13
Results Males (all)
Age group (years) Period December to November Mean assaults per year Population (per year) Rate (per 10,000 persons per year) Within age group Post/Pre IRR (95 CI)   Effect estimate Ratio of IRRs (95 CI) Target and trickle down groups relative to 20-21 year-olds
  15-17 (Trickle down) Pre 1995-1999 133 83453 15.9 1 1
  15-17 (Trickle down) Post 1 1999-2003 199 87531 22.8 1.43 (1.28 to 1.60) 1.28 (1.10 to 1.49)
  15-17 (Trickle down) Post 2 2003-2007 234 97036 24.1 1.52 (1.36 to 1.69) 1.25 (1.08 to 1.45)
  15-17 (Trickle down) Post 3 2007-2011 214 96858 22.0 1.39 (1.24 to 1.54) 1.04 (0.90 to 1.21)
  18-19 (Target) Pre 1995-1999 166 54726 30.3 1 1
  18-19 (Target) Post 1 1999-2003 211 57422 36.5 1.20 (1.09 to 1.33) 1.08 (0.93 to 1.24)
  18-19 (Target) Post 2 2003-2007 274 61698 44.4 1.46 (1.33 to 1.61) 1.21 (1.05 to 1.39)
  18-19 (Target) Post 3 2007-2011 324 67319 48.2 1.59 (1.45 to 1.74) 1.20 (1.05 to 1.37)
  20-21 (Control) Pre 1995-1999 170 53735 31.5 1 1
  20-21 (Control) Post 1 1999-2003 200 56734 35.2 1.12 (1.01 to 1.24) 1
  20-21 (Control) Post 2 2003-2007 229 60008 38.1 1.21 (1.09 to 1.33) 1
  20-21 (Control) Post 3 2007-2011 281 67196 41.9 1.33 (1.21 to 1.46) 1
14
Results Females (all)
Age group (years) Period December to November Mean assaults per year Population (per year) Rate (per 10,000 persons per year) Within age group Post/Pre IRR (95 CI)   Effect estimate Ratio of IRRs (95 CI) Target and trickle down groups relative to 20-21 year-olds
  15-17 (Trickle down) Pre 1995-1999 29.8 79658 3.7 1 1
  15-17 (Trickle down) Post 1 1999-2003 38.0 84211 4.5 1.21 (0.95 to 1.54) 0.82 (0.58 to 1.15)
  15-17 (Trickle down) Post 2 2003-2007 51.3 93529 5.5 1.47 (1.17 to 1.84) 0.96 (0.69 to 1.33)
  15-17 (Trickle down) Post 3 2007-2011 56.0 92071 6.1 1.63 (1.30 to 2.03) 0.79 (0.58 to 1.09)
  18-19 (Target) Pre 1995-1999 26.0 53142 4.9 1 1
  18-19 (Target) Post 1 1999-2003 37.3 55951 6.7 1.36 (1.06 to 1.75) 0.92 (0.65 to 1.30)
  18-19 (Target) Post 2 2003-2007 43.0 59847 7.2 1.47 (1.15 to 1.87) 0.96 (0.68 to 1.35)
  18-19 (Target) Post 3 2007-2011 69.3 63970 10.8 2.21 (1.78 to 2.77) 1.08 (0.78 to 1.48)
  20-21 (Control) Pre 1995-1999 27.0 53055 5.1 1 1
  20-21 (Control) Post 1 1999-2003 41.8 55355 7.5 1.48 (1.16 to 1.89) 1
  20-21 (Control) Post 2 2003-2007 46.0 59032 7.8 1.53 (1.21 to 1.94) 1
  20-21 (Control) Post 3 2007-2011 66.5 63684 10.4 2.05 (1.64 to 2.57) 1
15
Results Maori Males
Age group   Period   Mean assaults per year Population (per year) Rate (per 10,000 persons per year) Within age group Post/Pre IRR (95 CI)   Effect estimate Ratio of IRRs (95 CI) Target and trickle down groups relative to 20-21 year-olds
  15-17 years Pre 1995-1999 32 16640 18.9 1 1
  15-17 years Post 1 1999-2003 57 17955 31.7 1.68 (1.35, 2.01) 1.13 (0.82, 1.55)
  15-17 years Post 2 2003-2007 64 20563 30.9 1.63 (1.32, 2.02) 1.03 (0.76, 1.41)
  15-17 years Post 3 2007-2011 64 21115 30.4 1.61 (1.30, 1.99) 0.85 (0.63, 1.15)
  18-19 years Pre 1995-1999 38 10893 34.9 1 1
  18-19 years Post 1 1999-2003 47 10850 43.3 1.24 (1.00, 1.54) 0.83 (0.61, 1.14)
  18-19 years Post 2 2003-2007 67 12123 55.5 1.59 (1.30, 1.94) 1.01 (0.74, 1.36)
  18-19 years Post 3 2007-2011 84 13588 61.8 1.77 (1.46, 2.15) 0.93 (0.70, 1.25)
  20-21 years (Control) Pre 1995-1999 30 10075 29.5 1 1
  20-21 years (Control) Post 1 1999-2003 44 9900 43.9 1.49 (1.18, 1.88) 1
  20-21 years (Control) Post 2 2003-2007 49 10453 46.6 1.58 (1.26, 1.98) 1
  20-21 years (Control) Post 3 2007-2011 69 12225 56.0 1.90 (1.53, 2.35) 1
16
Results Maori Females
Age group   Period   Mean assaults per year Population (per year) Rate (per 10,000 persons per year) Within age group Post/Pre IRR (95 CI)   Effect estimate Ratio of IRRs (95 CI) Target and trickle down groups relative to 20-21 year-olds
  15-17 years Pre 1995-1999 12 16390 7.0 1 1
  15-17 years Post 1 1999-2003 13 17880 7.4 1.06 (0.71, 1.57) 0.60 (0.35, 1.03)
  15-17 years Post 2 2003-2007 24 20378 11.5 1.64 (1.16, 2.34) 1.09 (0.65, 1.82)
  15-17 years Post 3 2007-2011 28 19888 13.8 1.97 (1.40, 2.78) 0.78 (0.48, 1.27)
  18-19 years Pre 1995-1999 11 11035 9.5 1 1
  18-19 years Post 1 1999-2003 16 10873 14.9 1.57 (1.07, 2.32) 0.89 (0.52, 1.53)
  18-19 years Post 2 2003-2007 18 12393 14.7 1.55 (1.06, 2.26) 1.02 (0.60, 1.75)
  18-19 years Post 3 2007-2011 29 13210 22.0 2.31 (1.62, 3.28) 0.92 (0.56, 1.50)
  20-21 years (control) Pre 1995-1999 11 10458 10.3 1 1
  20-21 years (control) Post 1 1999-2003 19 10203 18.1 1.8 (1.21, 2.57) 1
  20-21 years (control) Post 2 2003-2007 17 11085 15.6 1.51 (1.03, 2.22) 1
  20-21 years (control) Post 3 2007-2011 32 12358 25.9 2.52 (1.78, 3.56) 1
17
Summary

• Compared with 20-21 year-old males
• assaults increased significantly among 18-19
  year-old males (IRRs 1.04 to 1.21) relative to
  the pre-change period.
• assaults increased significantly among 15-17
  year-old males (IRRs 1.08 to 1.28) relative to
  the pre-change period
• No significant effects for females (note lower
  incidence rates for females 14 ratio)
• No effects detected among Maori

18
Limitations

• Statistical power restricted by sensitivity of
  outcome indicator (some cases will not have been
  alcohol involved) bias toward the null
• Inferences should not be made about trends
  because of change in ED coding over time not
  expected to differ by age and therefore would not
  bias effect estimate
• Lack of effect for females may reflect different
  victim / perpetrator dynamics by gender (age gap
  greater for females)

19
In relation to Maori

• No large effects but small effects in either
  direction cannot be ruled out because of small
  numbers
• There may be differences in informal access to
  alcohol between Maori and non-Maori that made the
  MPA less important for the former
• Findings underline the importance of government
  evaluation planning BEFORE major policy changes,
  especially for Maori (Mana Whakamarama)

20
Implications

• The rate of serious assault is increasing in New
  Zealand, particularly among young people, Maori
  and people living in deprived areas
• Contrast with trend in traffic injury
• For intentional injury (assault and deliberate
  self-harm) we lack the countermeasures we have
  for traffic injury (e.g., RBT)
• Increasing the MPA / MLDA should be considered
  for reducing assault

21
(No Transcript)
22
Minister of Justice
Otago Daily Times May 2004
23

• Papers available on request (kypros.kypri_at_newcastl
  e.edu.au)
• Effects of lowering the alcohol minimum
  purchasing age on weekend hospitalised assault.
  American Journal of Public Health, 2014, 104(8)
  1396-1401
• Effects of lowering the alcohol minimum
  purchasing age on weekend hospitalised assaults
  of young Maori in New Zealand. Drug Alcohol
  Review (in press 2015).
• Long-term effects of lowering the alcohol minimum
  purchasing age on traffic crash injury rates in
  New Zealand (under review).