Experiencing  Engaging  Educa.on  Feb  4-­‐  6,  2015  

Mia  A.  Papas,  PhD  Behavioral  Health  and  Nutri.on  Department    

College  of  Health  Sciences  mpapas@udel.edu  

Project-­‐Based  Learning  

•  Team  based,  problem-­‐solving  ac.vity  

•  Delivering  a  product  that  is:  – Significant  – Clearly  communicated  – Useful  

Big  data:  Small  projects  •  Objec.ve:  – Explore  large  datasets  of  secondary  data  

– Pose  a  testable  research  ques.on  

– Visualize  the  analysis  with  mul.ple  tools  •  Sta.s.cal  soXware:  JMP  

– Communicate/present  the  results  

3P’s  Approach  to  Science  Educa.on  

•  Problem  Posing  

•  Problem  Solving  

•  Peer  Persuasion  

h]p://bioquest.org/the-­‐3-­‐peas-­‐concept/  

HLPR632:  Applied  Data  Analysis  Objec.ves:  1)  Provides  an  overview  of  descrip.ve  and  

inferen.al  sta.s.cs  needed  to  interpret  health  related  data    

2)  Apply  sta.s.cs  needed  to  analyze  and  evaluate  health-­‐related  literature  and  conduct  public  health  research  

Epidemiology  

Distribu.on  and  Determinants  of  Health  and  Disease  in  Human  Popula.ons    

Study  of….   To  enable….  

Preven.on  and  Control  of  Health  Problems    

Last  JM:  A  Dic-onary  of  Epidemiology,  4th  ed.  New  York,  Oxford  University  Press,  2000.  

Epidemiology  and  the  3Ps  •  Problem  Posing  – Addressing  an  e.ologic  ques.on  of  public  health  relevance  

 •  Problem  Solving  – Using  epidemiologic  methods  to  design  studies  and  analyze  data  

•  Peer  Persuasion  –  Interpreta.on  of  evidence  and  communica.on  of  results  

Posing  the  right  ques.on?  

Developing  a  research  ques.on:  FINER  (Hulley  et  al.,  2007)  

F:  Feasibility    –  Sufficient  resources  in  terms  of  .me,  staff,  and  funding;  Use  of  

appropriate  study  design;  Manageable  in  scope;  Adequate  sample  size;  Trained  research  staff    

I:  Interes.ng  –   Interes.ng  as  a  researcher  or  collaborator;  Inves.gator’s  mo.va.on  to  

make  it  interes.ng    N:  Novel  

–  Thorough  literature  search;  New  findings  or  extension  of  previous  findings;  Guidance  from  mentors  and  experts    

E:  Ethical    –  Following  ethical  guidelines;  Regulatory  approval  from  Ins.tu.onal  

Review  Board    R:  Relevant    

–  Influence  on  clinical  prac.ce;  furthering  research  and  health  policy    

Problem  Solving  •  Tes.ng  the  hypothesis  using  exis.ng  secondary  data  sources  

•  “Drowning  in  data,  but  thirs.ng  for  informa.on”  

•  Sta.s.cs  provides  the  key  to  unlock  and  make  sense  of  data    

JMP  Sta.s.cal  soXware  

•  JMP  is  a  freely  available  sta.s.cal  soXware  program  for  students  and  faculty  at  UD  

•  JMP  has  a  graph  builder  app  that  creates,  edits  and  views  graphs  right  on  your  iPad  

•  JMP  works  in  both  MAC  and  Windows  formats  

Secondary  Data  Sources  •  US  Census  Bureau  •  Na.onal  Health  Interview  Survey  •  Na.onal  Center  for  Health  Sta.s.cs  •  Behavioral  Risk  Factor  Surveillance  Survey  •  Youth  Risk  Behavior  Survey  •  Na.onal  Immuniza.on  Survey  •  World  Health  Organiza.on  

Raw  Data  for  Analysis  

Data  Analysis  •  Descrip.ve  Sta.s.cs  – Summarize  and  report  on  individual  variables  – Numbers  and  graphs  – What  is  the  mean  body  mass  index  of  study  par.cipants?  

– Examples:    •  Mean  BMI  of  study  sample  •  Graphics  and  summary  sta.s.cs  

 

Data  Analysis,  cont.  

•  Inferen.al  sta.s.cs  – Asking  a  ques.on  about  the  associa.on  between  two  variables  

– Does  maternal  body  mass  index  differ  by  child’s  gender?    

Peer  Persuasion  

•  Appropriately  communica.ng  results  

•  Produce  and  present  “the  virtual  poster”  

•  Engage  in  peer  discussion    

The  specific  aims  of  the  inves5ga5on  were  to:    1.  Examine   the   prevalence   of   poor   diet,   lack   of   physical   ac.vity,  

unhealthy   weight   loss   behaviors,   and   obesity   in   a   popula.on  sample  of  adolescents  in  the  United  States.  

2.  Explore   differences   in   these   factors   between   adolescents   with  and  without  disabili.es.  

Purpose  

Introduc5on  

Results  

Conclusions  and  Recommenda5ons  1.  Adolescents with a broad range of disabilities are more likely to be

obese and engage in unhealthy weight loss behaviors that those without disabilities.

2.  Provision of successful weight loss strategies that address barriers for healthy dieting and physical activity for this population is critical to develop effective prevention programs.

Study  Sample  and  Measures  

•  Childhood obesity remains a major public health concern.

•  Children with disabilities have a higher prevalence of obesity compared to children without disabilities.

•  Due to physical and environmental barriers, children with disabilities may have a difficult time controlling their weight leading to a higher prevalence of obesity.

•  22% (1986/9775) of all survey participants reported an emotional or physical disability.

•  Compared to adolescents without disabilities, adolescents with disabilities were more likely to:

1.  Be obese (OR=1.7; 95% CI: 1.3, 2.1) 2.  Actively try and lose weight (OR=1.3; 95% CI: 1.1, 1.6) 3.  State that they were overweight (OR=1.4; 95% CI: 1.2, 1.6)

Figure  1.  Prevalence  of  unhealthy  weight  loss  strategies  by  disability  status  for  9775  adolescent  study  par5cipants  in  2011  YRBS  

* p-value < 0.05, ** p-value < 0.01

Table  1  .  Odds  ra5os  (OR)  and  95%  confidence  intervals  (CI)  for  associa5ons  among  diet,  physical  ac5vity  and    disability  for  9775  adolescent  study  par5cipants  in  2011  YRBS  

  Disability (n=1986)  

No Disability (n=7789)  

OR    

95% CI  

  % (95% CI)   % (95% CI)      Physical Activity          Days per week active at least 60 minutes **  

     

 

0 days   20 (16, 24)   13 (11, 15)   1.0   ref  >= 1 day & < 5 days   42 (38, 45)   35 (34, 37)   0.8   0.6, 1.0  

>= 5 days   38 (34, 43)   52 (49, 54)   0.5   0.4, 0.6  Hours per day watch television **          

< 2 hours   49 (45, 54)   43 (40, 46)   1.0   ref  >= 2 hours   51 (46, 55)   57 (54, 60)   0.8   0.7, 0.9  

Hours per day play video games          < 2 hours   46 (40, 51)   45 (42, 48)   1.0   ref  

>= 2 hours   54 (49, 60)   55 (52, 58)   1.1   0.9, 1.3  Dietary Consumption          Fruit Juice (100% fruit juice) **          

Never   29 (25, 33)   24 (21, 26)   1.0   ref   >=1 time in the past week   71 (67, 75)   76 (74, 79)   0.8   0.6, 0.9  

Soda (not including diet soda)          Never   24 (21, 27)   21 (18, 23)   1.0   ref  

>=1 time in the past week   76 (73. 79)   79 (77, 82)   0.8   0.7, 1.1  Fruit **          

Never   19 (16, 23)   12 (11, 14)   1.0   ref   >=1 time in the past week   81 (77, 84)   88 (86. 89)   0.6   0.5, 0.7  

Green Salad **          Never   43 (38, 48)   39 (35, 42)   1.0   ref  

>=1 time in the past week   57 (52, 62)   61 (58, 65)   0.8   0.7, 1.0  Other Vegetables **          

Never   22 (17, 27)   15 (13, 18)   1.0   ref   >=1 time in the past week   78 (73, 83)   85 (82, 87)   0.7   0.5, 0.9  

!"#

$$#$!#

%#

&#'#

"#

(#

$"#

$(#

!"#

!(#

)*+,-*.*/0#12#)*+,-*.*/0#

Fast in the Past 24 Hrs

Take Diet Pills, Powders

or Liquids

Vomit or Take Laxatives

Percen

t  (%)  

* p-value < 0.05, ** p-value < 0.01

 **    

 **      **    

Figure  2.    Associa5on  (OR  and  95%  CI)  between  number  of  unhealthy  weight  loss  behaviors  and  obesity  by  disability  status  for  9775  adolescent  study  par5cipants  in  2011  YRBS  

1 2 3

Number of Unhealthy Weight Loss Behaviors

Disability No Disability

4.5

1.0

0.8

15.0

1.4

2.5

8.0 Data from 2011 US National Youth Risk Behavior Surveillance Survey among US high school students in North Carolina, North Dakota, Rhode Island, and Delaware. Unhealthy weight loss behaviors included fasting within the past 24 hours, taking diet pills, powders, or liquids, or vomiting/taking laxatives. Obesity was defined as an age and gender adjusted body mass index (BMI) at or above the 95th percentile Disability was defined as at least one positive response to: 1)  “Do you have any physical disabilities or long-term health

problems (long-term means 6 months or more)?” 2)  “Do you have any long-term emotional problems or learning

disabilities (long-term means 6 months or more)?”

Session  Objec.ves  1.  Work  in  teams  to  develop  a  testable  hypothesis  

using  the  available  data  

2.  Analyze  data  and  produce  data  tables  and  graphs  

3.  Communicate  results    

Case  study  

Problem  Posing    1.  Background  informa.on  

–  Measles  Case  Defini.on  •  h]p://www.cdc.gov/measles/hcp/index.html  

–  Current  Measles  Outbreak  •  h]p://www.cdc.gov/measles/cases-­‐outbreaks.html    

–  History  of  Vaccines  •  h]p://www.historyofvaccines.org/content/.melines/measles  

   

Problem  Solving  •  What  data  are  available  that  might  shed  light  on  the  current  measles  outbreak?  

•  What  ques.ons  can  we  ask  of  that  data?  

•  How  can  we  best  analyze  that  data  to  produce  meaningful  sta.s.cs?  

Datasets  

•  Two  datasets:  – Dataset  1  has  3  variables    

•  Year  (2000  to  2014)  •  Number  of  measles  cases    •  Percent  of  children  immunized  

 – Dataset  2  is  a  subset  of  data  from  the  2012  Na.onal  Immuniza.on  Survey  

Na.onal  Immuniza.on  Survey  2011  

•  Telephone  survey  (27,305  households)  

•  Asks  about  vaccina.on  coverage  of  children  aged  19  to  35  months  in  the  household  

•  Demographic  characteris.cs,  household  characteris.cs  and  vaccina.on  coverage  data  are  included  in  the  dataset  

Data  Explora.on  

•  Examine  data  •  Pose  ques.ons  •  Develop  hypothesis  •  Test  hypothesis  •  Interpret  data  

Variables  •  HH_MCV  –  Number  of  measles  containing  shots  (None=0,  at  least  one=1,  

all=2)  •  AGE_GRP  –  Age  Group  of  child  (19-­‐23  months=1,  24  –  29  months=2,  30  –  

35  months=3)  •  BF_ENDR06  –  Dura.on  in  days  of  breasteeding  •  CBF_01  –  Was  child  ever  breasted  (yes=1/no=2)  •  EDUC_1  -­‐  Maternal  educa.on  (<  12  years=1,  12  years=2,  >  12  years=3,  

college=4)  •  FRSTBRN  -­‐  First  born  child  (yes=2/no=1)  •  INCPORAR  -­‐  (income  to  poverty  ra.o)  •  M_AGEGRP  –  age  of  mother  (<=19=1,  20-­‐29=2,  >=30  years=3)  •  RACE_K  -­‐  (White  only=1,  black  only=2,  Other  plus  mul.ple  race=3)  •  SEX  -­‐  (Male=1/female=2)  •  STATE  -­‐  (FIPS  CODE)  

Variable  Defini.on  •  Income-­‐to-­‐poverty  ra.os  represent  the  ra.o  of  family  or  unrelated  individual  income  to  their  appropriate  poverty  threshold.  Ra.os  below  1.00  indicate  that  the  income  for  the  respec.ve  family  or  unrelated  individual  is  below  the  official  defini.on  of  poverty,  while  a  ra.o  of  1.00  or  greater  indicates  income  above  the  poverty  level.  A  ra.o  of  1.25,  for  example,  indicates  that  income  was  125  percent  above  the  appropriate  poverty  threshold”  (U.S.  Census  Bureau,  2004).  

Peer  Persuasion  •  Final  Product  

– Develop  a  virtual  poster  using  powerpoint  that  communicates  the  results    

– Use  graphs  and  numbers  to  present  analy.c  informa.on  

Have  Fun!