The Early Home Learning Study:

Collaboration with government and providers to develop and evaluate a parenting approach that improves learning outcomes for young children experiencing disadvantage.

Presentation to Brotherhood of St Laurence 28th August, 2014

Presenter: Jan Matthews

Aims of the Early Home Learning Study

Funded by Victorian Government Department of Education and Early Childhood Development

(DEECD)

1. Improve early learning and development foundations for children in over 2,000 families living in vulnerable circumstances

2. Find out how best, in the Victorian early childhood context, to support parents to provide positive home learning environments to their children (focus on 0 – 3 years)

Our Brief

Build capacity in existing service sector

Embed new evidence based strategies in existing service platforms

Provide practitioner training and site support

Build positive home learning environments in families

Parents increase quantity & quality of parent-child interactions &

Provide stimulating learning environments

Children’s learning and development

AddressVictorian

Early Years Learning &

Development Framework

Acknowledgements

Principal Investigators

Jan Nicholson

Naomi Hackworth

Jan Matthews

Warren Cann

Donna Berthelsen

Advisors

Obe Ukoumunne(statistics, randomisation)

Lisa Gold (economics)

Jay Buzhardt & Kathleen Baggett (observational methods)

EHLS Team

Victoria Hamilton

Elizabeth Westrupp

Amanda Sicluna

Misel Trajanovska

Shannon Bennetts

Jane Bennetto

Kim Campbell

Emily Gutierrez

Nicola Johnson

Tracey Phan

Durga Shrestha

Maggie Yu, Olivia Clayton

Eloise Cameron

Hilary Davis, Simon Fox

Tessa Hillgrove, Laurel Johnson

Monique Seymour

Site Coordinators/Managers

Lauren BeecroftNatacha BuchananAnn BurkeIris CrookDonna FeoreVanessa FrogleyJacquie GilmourBarbara HayesAnnita KeeversBeth KershawDaniel Leach-McGillJulie McKenzieTanya McMaster-ThieleKaren MitchellAnne Monichon

Joanne NovakLuisa PerezMargaret PhillipsCarmel PillingerJenny RukuwaiDanielle SchofieldSally ThomsonKristy WaltonKaren WheelerKay Widdicombe

Framework for enhancing early learning

Quality everyday interactions• tuning in • following the child’s lead • listening and talking more• using teachable moments• being warm and gentle

Stimulating home learning environment

• shared reading at home• learning through everyday routines• supporting children’s play at home • using community resources• monitoring use of media at home.

accessing needed services supported by and contributing to their community

Developing self-efficacy and self-sufficiency, self-monitoring & personal agency

Parent wellbeing stress

management

Quality everyday

interactions

Stimulating environment

Parent self-careParenting

confidence

Community and service connectedness

The smalltalk intervention

Two levels of intensity • smalltalk – group only• smalltalk plus – group plus home coaching (6 home visits)

Two service platforms: MCH SPG

Parents of:Infants

6 – 12 months

Toddlers

12- 36 months

No. weekly group

sessions:6 10

Compared to ‘usual care’• Standard parenting or playgroups

Study questions

The Service

• Was the intervention provided to the families it was designed for?

• Was it delivered the way it was intended?

• Did the families and the service providers like it?

• Did families remain engaged over time?

The Outcomes

• Did smalltalk families show greater gains over time compared to families attending standard groups?

• How effective was smalltalk compared to smalltalk plus?

Cluster RCT Design

PLATFORM

SITE

LOCATION

smalltalkplus

Families

smalltalk

Families

‘usual care’

Families

2 Platforms (infant & toddler)

20 sites: Local Government Areas(10 infant ; 10 toddler)

101 locations: physical place where the groups were run = UNIT OF RANDOMISATION

Families received the variant offered in their local area

Each location ran 1 type of research condition

Metropolitan sites (n=13)

Melton

(44 km)

Mornington Peninsula

(90 km)

Yarra Ranges

(30-110 km)

Regional sites (n=7)

Bendigo

(153 km)

Shepparton

(191 km)

Engaging parents & practitioners in intervention development

through consultation and collaboration:

Impact on engagement and process.

Contextually mediated practice

“An approach to early childhood intervention that

uses everyday family and community activities as the

contexts for children’s learning, and child interests as the

foundation for involving children in everyday learning

activities “ (Dunst, 2006)

Collaborative-consultative approach to intervention development

• Dissemination failure - attributed to inadequate early attention to critical aspects of program design and planning

(Dunst et al., 2008; Durlak & DuPre, 2008)

• Importance of consumer engagement (e.g. Boyd, Diamond & Bourjolly, 2006; Goodare & Smith, 1995; Greenhalgh et al, 2004; Nilsen et al 2013; Sanders & Kirby 2012)

Factors influencing intervention design

Ease of Implementation• Fit with existing, widely accessed, services

• Modelled around Maternal and Child Health parenting groups & supported playgroups

• Fit with the funding models: Low intensity/light touch

Narrow Focus• Focus on fairly narrow range of parenting skills

Fidelity• Build on existing staff competencies

• Keep core elements simple and within staff capacity

• Provision of training, detailed manuals & resources

• Trained in ONE variant only

How did we develop smalltalk?

Development Process

smalltalk materials

Practitioner manuals

Advertising

Practitioner resources to use with parents

Parent workbooks

DVD examples of the smalltalk strategies

Conversation cards

Materials to support change

Parenting intervention - triadic

model

19

PRACTITIONER

Facilitate groups

Conduct home coaching

PARENT

Implement smalltalkstrategies

CHILD

Enhance learning & communication

Principles

•Partnership with parents

•Implementable – ‘light touch’

•Ensuring ‘fit’ with existing service delivery

•Building on approaches and strategies already in use

•Providing tools/resources appropriate for services and families

•Focus on skill development using methods known to bring about change

•Practice model emphasis on building parent’s self-efficacy

Primary parent change mechanism

Coaching*Joint planning

Observation – live and filmed modeling (Parent dvd)

Action/practice

Reflection

Feedback

*Practice characteristics of coaching (Rush & Shelden, 2005)

Plus

Written/pictorial information

Group discussion

Practitioner role – introduce & coach

MCH groups – smalltalk strategies introduced via cooperative learning

activities & practice in group.

SP groups – strategies introduced informally during brief ‘one on one’

or ‘one on two’ conversations + practice in group

Both platforms - Between session practice at home

Home coaching

DVD based – DVD = ‘3rd party’

Coach works with parent on DVD content

Facilitates parent practice with video feedback

Collaborative relationship between parent/practitioner –parent expert on their own child

Results: Process data

Reach, retention, participation

Program fidelity

Exposure (quantity): Were the components of the intervention covered sufficiently?

Adherence: Did parents use the strategies in session and at home?

Exposure (quality): Were the parents and facilitators satisfied with the content and delivery

Process measures

Site & facilitator/coach reportsParent recruitment/Attendance

Post – session ratings of parent participation, content covered & session conduct

Post–intervention ratings – satisfaction, benefit

Parent reportssmalltalk strategy use in home/other setting –weekly verbal report

Post intervention telephone interview –satisfaction, benefit

Reach and retention

Recruitment 2,661 expressed interest

2,228 (84%) were eligible, contactable and

consented to participate

Reach Successful in recruiting targeted families:

84% at least 1 risk factor

20% had 4+ risk factors

Retention 83% retained to post-intervention

78% retained to 5-month follow-up

26

Baseline risk factors for less than optimal child development

Clin

≤ 36,400

<Year 12

<<20

Sympt

< 52,000

Year 12 or Equivalent

<=25

0 20 40 60 80 100

Par. psychological distress

Low parent self efficacy

No parent employed

Govt pension/benefits

Low family income

Low par. education

Language other than English

Single parent

Young parent (<20; <=25)

Proportion

1111

N=2187

Participation: Average proportion of sessions attended

0

20

40

60

80

100

MCH SPG

% s

essi

ons

atte

nded

standard

smalltalk

smalltalk plus

home coaching

59-64% of group sessions (first 3 bars)78-81% of home coaching (4th bar) 67% attended all

0

20

40

60

80

100

Warm and gentle

Tuning in Follow child’s lead

Listening & talking more

Teachable moments

Reading

Par

tici

pan

ts (

%)

smalltalk strategy

Discussed at group

Practiced at group

Used at home

Data for the MCH (infant) platform

Same pattern for Supported Playgroups (toddler)

platform

Exposure to and use of strategies

Parent satisfaction with intervention, facilitator & home coach

Parent satisfaction with intervention, facilitator & home coach

Parent satisfaction with intervention, facilitator & home coach

Impact on knowledge – parent report

0 20 40 60 80 100

smalltalk plus

smalltalk

standard

smalltalk plus

smalltalk

standard

smalltalk plus

smalltalk

standard

smalltalk plus

smalltalk

standard

Met

oth

er

par

ents

Fou

nd

ou

t ab

ou

t o

ther

se

rvic

es

Un

der

stan

din

g o

f ch

ild

dev

elo

pm

ent

New

way

s to

te

ach

ch

ild

Proportion of parents (%)

A little

A lot

Data for the MCH (infant) platform

Similar, but stronger differences for SP (toddler)

platform

Facilitator satisfaction

Rated as going “well” or “very well”

Standard smalltalk smalltalk plus

MCH (infant)70% 92% 81%

Supported

playgroups (toddler) 77% 80% 90%

Home coaching

sessionsn/a n/a 85%

Outcomes of CRCT

Research Questions

Did families show greater improvements:

1. For smalltalk vs. usual care?

2. For smalltalk plus home coaching vs smalltalk

group only?

Outcomes: parenting behaviours and the quality of

the home learning environment

Design

Two parallel cRCTs

• Infants – 6 week MCH parent groups

• Toddlers – 10 week supported playgroups

Three conditions

• Standard

• smalltalk (group only)

• smalltalk plus (group + 6 home coaching sessions)

Three time-points

• Pre, post, 5-month follow-up

Outcome measures – self-report

Parenting

• Warmth*

• Irritability*

• Verbal responsivity

Home environment

• Activities with child*

• Literacy environment

• Household confusion

Computer-assisted telephone interview (CATI)

Collected and analysed for full sample

*Reliable, validated measures from the

Longitudinal Study of Australian Children

Outcomes measures – observational

Individual Growth & Development Indicators

Indicator of Parent-Child Interaction: Parent ‘facilitator’ behaviours

• Acceptance and warmth

• Descriptive language

• Follow child’s lead

• Maintains child’s interest(Carta, Greenwood, Walker & Buzhardt, 2010)

Four videotaped, structured parent-child play & tasks (8-10mins)

Collected for full sample; pre, post & 5-mthsCoded and analysed for random subsampleTrained coders, blind to intervention and time-point20% checked for coder agreement (87% agreement)

Analyses

Within platform (i.e. separately for infant and toddler)

1. Participant characteristics

• Was randomisation successful?

• Was there selective attrition?

2. Adjusted regression analyses

• Pre to post; pre to follow-up

• Presented estimates are Effect Sizes (ES) for continuous data & Odds Ratios (OR) for categorical data

Adjusted for: child age, single parent family, language other than English, young parent (<25yrs), low parent education, parent unemployment, low income, government benefits

Interpretation

Effect Sizes (ES) – summary of the magnitude of difference between the groups. For a community samples and a preventive intervention, we interpret these as:

Large: >0.50

Medium: 0.25 - 0.50

Small: <0.25

Odds Ratio (OR) – likelihood of the outcome cf. reference group

= 1.0 equal probability of the outcome

< 1.0 reduced probability of the outcome

> 1.0 increased probability of the outcome

Analyses

3. Adjusted regression: Reliable change*

• Summary statistic, applied to the self-report data only

• For each individual, calculates whether change on a given measures over time is greater than could be expected by chance alone

• We have categorised each participant as showing reliable change on at least 1 measure (Y/N)

• Comparisons by condition expressed as OR

* Jacobson & Truax (1991) Journal of Consulting & Clinical Psychology

Study Participants

2228 parent/child dyads

Infant = 1002 Toddler = 1226

Randomisation was successful

There was selective attrition – families with higher risk were more likely to be lost at post and follow-up

Final sample for analysis:

* Varies due to tapes excluded as ‘not able to be coded’

Infant (n) Toddler (n)

Self-report (CATI) 750 – 800 940 – 1010

Observations (videotaped)* 90 – 100 120 – 130

Overall pattern:

• smalltalk plus was better than (or the same as) smalktalk

• smalltalk was better than (or the same as) standard

For simplicity, we present:

• smalltalk plus vs. standard comparisons

Infant Platform – self report data

smalltalk plus vs. standard

pre to post

N=798

pre to follow-up

N=757

ES p ES p

Parenting warmth 0.11 0.223 0.01 0.862

Parenting irritability -0.03 0.703 0.13 0.075

Parent verbal responsivity 0.18 0.012 0.08 0.285

Home activities 0.20 0.005 0.00 0.971

Literacy environment 0.19 0.007 0.03 0.671

OR (CI) p OR (CI) p

High household confusion0.80

(0.43, 1.48)0.478

1.09

(0.56, 2.12)0.807

Reliable change (1 or more

areas)

1.77

(1.26, 2.48)0.001

1.14

(0.83, 1.58)0.419

Infant Platform – observational data

smalltalk plus vs. standard

pre to post

N=100

pre to follow-up

N=88

ES p ES p

Acceptance & warmth -0.05 0.856 0.20 0.379

Descriptive language 0.63 0.014 0.16 0.635

Follow child’s lead 0.56 0.022 0.25 0.482

Maintains child’s interest 0.03 0.907 0.32 0.362

Toddler Platform – self report datasmalltalk plus vs. standard

pre to post

N=1014

pre to follow-up

N=940

ES p ES p

Parenting warmth -0.04 0.679 0.05 0.532

Parenting irritability -0.10 0.080 -0.09 0.145

Parent verbal responsivity 0.08 0.233 0.04 0.630

Home activities 0.16 0.011 0.11 0.152

Literacy environment -0.05 0.444 0.05 0.379

OR (CI) p OR (CI) p

High household confusion1.04

(0.68, 1.60)0.858

1.38 (0.88, 2.18) 0.163

Reliable change (1 or more

areas)

1.26 (0.80, 1.98)

0.3231.94

(1.36, 2.78) 0.000

Toddler Platform – observational data

smalltalk plus vs. standard

pre to post

N=119

pre to follow-up

N=129

ES p ES p

Acceptance & warmth 0.20 0.298 0.19 0.291

Descriptive language 0.77 0.000 0.46 0.038

Follow child’s lead 0.18 0.440 0.35 0.071

Maintains child’s interest 0.52 0.029 0.55 0.008

Summary of findings:

Infant platform

• Early improvements in parent reported and observed behaviours

• Differences not sustained to follow up

Toddler platform

• Fewer early improvements parent reported and observed behaviours

• Differences were sustained and strengthened to follow up

Conclusions

1. Coproduction is the ‘way to go’

• High level of ‘buy in’ from families , practitioners & services

• Referral, reach & attendance/ retention indicate high acceptability/relevance for families

• Good fit to existing service model and staff training

Conclusions

2. It is possible to change parent behaviour to improve the quality of the home learning environment through a relatively light touch intervention

• although more intensive intervention is required for the most vulnerable families

Recommendation: Use data to inform effective ‘triage’ processes to ensure that the families who receive this limited resource are the ones who need it most

Conclusions

3. It is possible to embed rigorous and complex research design into real world service delivery

• Engage services as meaningful research partners.

• Provide intensive support around research design, study implementation & data collection

Services and practitioners took on ‘ownership’ of the study.

What next?

Impact on long term child outcomes

• Awarded research grant funding to assess child outcomes at school entry

PRC is supporting the transition to full implementation

• State government have secured ongoing funding for smalltalk

Aims are to achieve

• evidence-informed scale-up – match services to community profiles

• make smalltalk supported playgroups more broadly available

• effectively triage the additional home coaching component to the families who would benefit the most