21

PHARMAC / TE - SAS€¦ · PHARMAC / TE PĀTAKA WHAIORANGA Sankey 1 Nodes contain the events in each path. The node displays the width of each link that enters and exits the node

  • Upload
    others

  • View
    6

  • Download
    0

Embed Size (px)

Citation preview

Page 1: PHARMAC / TE - SAS€¦ · PHARMAC / TE PĀTAKA WHAIORANGA Sankey 1 Nodes contain the events in each path. The node displays the width of each link that enters and exits the node
Page 2: PHARMAC / TE - SAS€¦ · PHARMAC / TE PĀTAKA WHAIORANGA Sankey 1 Nodes contain the events in each path. The node displays the width of each link that enters and exits the node

PHARMAC / TE PĀTAKA WHAIORANGA

Page 3: PHARMAC / TE - SAS€¦ · PHARMAC / TE PĀTAKA WHAIORANGA Sankey 1 Nodes contain the events in each path. The node displays the width of each link that enters and exits the node

PHARMAC / TE PĀTAKA WHAIORANGA

PHARMAC / TE PĀTAKA WHAIORANGA

Our objective

“…to secure for eligible people in need of pharmaceuticals, the best health outcomes that are reasonably achievable from pharmaceutical treatment and from within the amount of funding provided.”

New Zealand Public Health and Disability Act 2000

Page 4: PHARMAC / TE - SAS€¦ · PHARMAC / TE PĀTAKA WHAIORANGA Sankey 1 Nodes contain the events in each path. The node displays the width of each link that enters and exits the node

PHARMAC / TE PĀTAKA WHAIORANGA

PHARMAC Environment

• SAS 9.4

• SAS Enterprise Guide 7.15

• SAS Visual Analytics 7.3

• SAS Visual Statistics 7.3

Page 5: PHARMAC / TE - SAS€¦ · PHARMAC / TE PĀTAKA WHAIORANGA Sankey 1 Nodes contain the events in each path. The node displays the width of each link that enters and exits the node

PHARMAC / TE PĀTAKA WHAIORANGA

Pharmaceutical Data

https://www.health.govt.nz/nz-health-statistics/national-collections-and-surveys/collections/pharmaceutical-collection

Page 6: PHARMAC / TE - SAS€¦ · PHARMAC / TE PĀTAKA WHAIORANGA Sankey 1 Nodes contain the events in each path. The node displays the width of each link that enters and exits the node

PHARMAC / TE PĀTAKA WHAIORANGA

Type 2 Diabetes is the

single biggest medical

‘equity’ issue we are

facing in New Zealand.Dr Bryan Betty - PHARMAC

A rising tide of type 2 diabetes in younger people

Page 7: PHARMAC / TE - SAS€¦ · PHARMAC / TE PĀTAKA WHAIORANGA Sankey 1 Nodes contain the events in each path. The node displays the width of each link that enters and exits the node

PHARMAC / TE PĀTAKA WHAIORANGA

Diabetes Treatment

Progression

M= Metformin

S = Sulphonylurea

I = Insulin

Page 8: PHARMAC / TE - SAS€¦ · PHARMAC / TE PĀTAKA WHAIORANGA Sankey 1 Nodes contain the events in each path. The node displays the width of each link that enters and exits the node

PHARMAC / TE PĀTAKA WHAIORANGA

Sankey

1 Nodes contain the events in each path. The node displays the width of each link that enters and exits the node. The same event can appear at multiple nodes in the diagram.

2 Links between nodes make up the paths in the diagram. The width of each

link can represent either the frequency of the path or the value of a weight measure.

3 Drop-off links are links that end at the current node. Drop-off links are displayed only if some links continue onward from the current node.

Parts of a Sankey Diagram

Page 9: PHARMAC / TE - SAS€¦ · PHARMAC / TE PĀTAKA WHAIORANGA Sankey 1 Nodes contain the events in each path. The node displays the width of each link that enters and exits the node

PHARMAC / TE PĀTAKA WHAIORANGA

SAS EG Sankey code

/*--Render the Diagram--*/

ods graphics / reset attrpriority=color width=5in height=3in

imagename='Sankey_2_940';

footnote j=l 'SAS 9.40';

proc sgplot data=sankey2 noborder noautolegend nocycleattrs;

styleattrs datacontrastcolors=(darkred blue cx5f9f1f)

datacolors=(darkred blue cx5f9f1f);

series x=x y=y / group=id lineattrs=(pattern=solid thickness=12)

nomissinggroup transparency=0.8 smoothconnect;

series x=x y=y / group=id2 lineattrs=(pattern=solid thickness=62 color=cx5f9f1f)

nomissinggroup transparency=0.8 smoothconnect;

highlow y=y2 low=xl high=xh / highlabel=hlabel lowlabel=llabel type=bar

intervalbarwidth=10 group=id transparency=0.3 nooutline

labelattrs=(color=black size=10 weight=bold);

highlow y=y3 low=xl high=xh / highlabel=hlabel lowlabel=llabel type=bar

intervalbarwidth=62 group=id2 transparency=0.3 nooutline

fillattrs=(color=cx5f9f1f) labelattrs=(color=black size=10 weight=bold);

scatter x=xlbl y=ylbl / datalabel=label datalabelpos=topright markerattrs=(size=0)

datalabelattrs=(size=12 weight=bold);

series x=xln y=yln / group=lid nomissinggroup lineattrs=(color=black);

scatter x=xa y=ya / datalabel=anno datalabelpos=center

markerattrs=(symbol=circlefilled color=lightblue size=18)

datalabelattrs=(size=12 weight=bold);

xaxis min=0 max=1 offsetmin=0 offsetmax=0 display=none;

yaxis min=0 max=1 offsetmin=0 offsetmax=0 display=none;

run;

footnote;

Page 10: PHARMAC / TE - SAS€¦ · PHARMAC / TE PĀTAKA WHAIORANGA Sankey 1 Nodes contain the events in each path. The node displays the width of each link that enters and exits the node

PHARMAC / TE PĀTAKA WHAIORANGA

SAS Visual Analytics® Sankey

• Event

• Sequence Order

• Transaction Identifier

• Class of treatment

• Years since starting treatment

• NHI person identifier

• BGTOnly = Blood Glucose Testing only

• M= Metformin

• S = Sulphonylurea

• I = Insulin

Page 11: PHARMAC / TE - SAS€¦ · PHARMAC / TE PĀTAKA WHAIORANGA Sankey 1 Nodes contain the events in each path. The node displays the width of each link that enters and exits the node

PHARMAC / TE PĀTAKA WHAIORANGA

First AttemptIs this

year 1?

Page 12: PHARMAC / TE - SAS€¦ · PHARMAC / TE PĀTAKA WHAIORANGA Sankey 1 Nodes contain the events in each path. The node displays the width of each link that enters and exits the node

PHARMAC / TE PĀTAKA WHAIORANGA

The issue

MASTER_ENCRYPTED_HCU_ID FYR First_FYR Years_from_Start AlignedTypeCombAdj

/QiRj9KBRpA 2012 2012 0 M -> 0 YR

/QiRj9KBRpA 2013 2012 1 M -> 1 YR

/QiRj9KBRpA 2017 2012 5 I+OralAH -> 5 YR

Page 13: PHARMAC / TE - SAS€¦ · PHARMAC / TE PĀTAKA WHAIORANGA Sankey 1 Nodes contain the events in each path. The node displays the width of each link that enters and exits the node

PHARMAC / TE PĀTAKA WHAIORANGA

Some Data Manipulation

Give everyone a -1 year bucket.

MASTER_ENCRYPTED_HCU_ID First_FYR FYR Years_from_Start AlignedTypeCombAdj

/QiRj9KBRpA 2012 2011 -1 None -1 YR

/QiRj9KBRpA 2012 2012 0 M -> 0 YR

/QiRj9KBRpA 2012 2013 1 M -> 1 YR

/QiRj9KBRpA 2012 2014 2 None 2 YR

/QiRj9KBRpA 2012 2015 3 None 3 YR

/QiRj9KBRpA 2012 2016 4 None 4 YR

/QiRj9KBRpA 2012 2017 5 I+OralAH -> 5 YR

Page 14: PHARMAC / TE - SAS€¦ · PHARMAC / TE PĀTAKA WHAIORANGA Sankey 1 Nodes contain the events in each path. The node displays the width of each link that enters and exits the node

PHARMAC / TE PĀTAKA WHAIORANGA

A better product – with a sense of timeStart Year Year 1 Year 2 Year 3

What about year

-1?

Page 15: PHARMAC / TE - SAS€¦ · PHARMAC / TE PĀTAKA WHAIORANGA Sankey 1 Nodes contain the events in each path. The node displays the width of each link that enters and exits the node

PHARMAC / TE PĀTAKA WHAIORANGA

Splitting the start by gender

Page 16: PHARMAC / TE - SAS€¦ · PHARMAC / TE PĀTAKA WHAIORANGA Sankey 1 Nodes contain the events in each path. The node displays the width of each link that enters and exits the node

PHARMAC / TE PĀTAKA WHAIORANGA

With paths

Page 17: PHARMAC / TE - SAS€¦ · PHARMAC / TE PĀTAKA WHAIORANGA Sankey 1 Nodes contain the events in each path. The node displays the width of each link that enters and exits the node

PHARMAC / TE PĀTAKA WHAIORANGA

Video

Page 18: PHARMAC / TE - SAS€¦ · PHARMAC / TE PĀTAKA WHAIORANGA Sankey 1 Nodes contain the events in each path. The node displays the width of each link that enters and exits the node

PHARMAC / TE PĀTAKA WHAIORANGA

When we don’t have

facts, all we can hope for

is combining data with

assumptions to make

reasonable decisions. Cassie Kozyrkov

Page 19: PHARMAC / TE - SAS€¦ · PHARMAC / TE PĀTAKA WHAIORANGA Sankey 1 Nodes contain the events in each path. The node displays the width of each link that enters and exits the node

PHARMAC / TE PĀTAKA WHAIORANGA

Final thoughts

Page 20: PHARMAC / TE - SAS€¦ · PHARMAC / TE PĀTAKA WHAIORANGA Sankey 1 Nodes contain the events in each path. The node displays the width of each link that enters and exits the node

PHARMAC / TE PĀTAKA WHAIORANGA

Further Reading

Sankey Diagrams in SAS Visual Analytics® - 2015

Taking the Path More Travelled – SAS Visual

Analytics® and Path Analysis - 2015

Sankey Diagram- A Compelling, Convenient, and

Informational Path Analysis with SAS Visual

Analytics® - 2017

Page 21: PHARMAC / TE - SAS€¦ · PHARMAC / TE PĀTAKA WHAIORANGA Sankey 1 Nodes contain the events in each path. The node displays the width of each link that enters and exits the node

Tēnā koutou katoa

Contact Details: [email protected]

Jason @ LinkedIn

• Website – www.pharmac.govt.nz

• Twitter - PHARMACnz

• Facebook - PHARMACnz

• YouTube - PHARMACgovtnz