►
From YouTube: Integrated Data & Coordinated Access: Boston's Open Source Solution Towards Ending Homelessness
Description
Laila Bernstein, Jennifer Flynn, and Ian Kozak discuss the advantages of using open source systems and coordinated access HMIS data to better serve the homeless citizens of Boston.
A
Hi,
my
name
is
Lila
Bernstein
I
am
advisor
to
the
mayor
for
the
initiative
to
end
chronic
homelessness
here
in
the
city
of
Boston
and
I'm.
Here
with
my
colleagues,
recording
a
presentation
called
integrated
data
and
coordinated
access,
Boston's
open
source
solution.
This
is
a
presentation
we
gave
at
the
National
Human
Services
data
consortium
back
in
April.
A
I
am
here
with
my
colleague,
Jennifer
Flynn,
with
HMIS
administrator
for
the
City
of
Boston
and
Ian
Kozak,
who
is
director
of
strategic
development
for
Green
River
and
we're
talking
about
our
integrated
data
and
coordinated
access,
our
open
source
solution
for
ending
homelessness
and
so
I'm
going
to
get
a
little
bit
of
background
and
then
I'm
going
to
hand
it
over
to
Jen
to
talk
about
our
data
system
and
then
we'll
top
it
off.
Ian
talk
about
the
technology
and
the
open
source,
build
that
that
we've
built
together
so
next
slide.
A
So
a
little
bit
of
history.
So
in
Boston,
the
City
of
Boston
is
the
lead
agency
for
the
time
of
care.
Hud
dollars
we
get
about
twenty
four
million
dollars
annually
for
for
housing,
homeless
and
ending
homelessness
were
also
the
lead
agency
for
AIA,
homeless
management.
Information
system
and
the
city
runs
about
40%
of
the
single
adult
shelter
beds
in
a
city
run
shelter,
so
we're
the
lead
in
multiple
ways
and
back
in
October
of
2014.
A
There's
a
there's,
a
crisis
in
our
homeless
system
and
the
city
run
shelter
that
was
located
on
an
island
in
our
harbor.
The
bridge
going
to
that
island
was
condemned,
it
was
unsafe,
and
so
in
a
moment's
notice,
we
had
to
relocate
those
shelter
beds
somewhere
else
in
the
city
and
it
posed
a
real
crisis
for
the
system.
But
you
know
even
more
so,
for
the
people
experiencing
homelessness
who
had
been
staying
in
those
beds,
and
it
was
an
opportunity
for
the
city
to
think
about
how
we
were
responding
to
homelessness.
Not
just.
A
How
are
we
going
to
respond
to
this
crisis?
How
we're
going
to
relocate
the
bed?
But
how
are
we
ending
homelessness
in
the
city
of
Boston
and
what
could
we
be
doing
differently
and
more
effectively?
So
we
we
took
time
to
convene
partners,
bring
in
experts
and
think
about
what
what
we
have
and
what
we're
doing
and
where
we
need
to
add.
A
So
we
created
this
based
on
a
visual
actually
that
Corporation
for
supportive
housing
invented,
and
it's
really
just
the
tip
of
the
iceberg,
but
it
shows
the
chaos
that
that
people
go
through
and
the
kind
of
illogic
of
our
system
when
people
are
trying
to
find
their
way
out
and
really
the
solution.
The
bottom
resources
are
all
housing
resources
or
permanent,
supportive
housing
resources,
and
that
is
how
someone
and
their
own
homelessness
or
how
a
system
end
homelessness,
is
by
providing
housing
but
to
get
to
those
resources.
A
So,
instead,
what
what
we
kind
of
came
out
of
that
planning
process
with
is
that
we
came
out
with
an
action
plan
to
end
chronic
and
veteran
homelessness,
and
it's
really
about
creating
the
single
homeless
response
system
out
of
our
many
different
programs,
some
of
which
are
best
in
the
nation,
but
but
really
need
to
be
working
arm
in
arm
with
every
other
resource
that
we
have.
So
that's.
A
That's
our
that's
been
our
goal
for
the
last
few
years,
so
the
action
plan
put
forward
these
needs
very
ambitious
goals:
they're
they're,
very
familiar
to
communities
across
the
country
and
veteran
homelessness
and
chronic
homelessness,
and
we
knew
to
accomplish
either
of
those
goals.
We
would
have
to
transform
our
entire
system
and
transform
how
we
were
responding
to
home.
A
We
did,
we
did
have
an
early
win
at
the
end
of
2015.
The
mayor
was
able
to
announce
that
we
had
ended
chronic
homelessness
among
veterans,
and
that
gave
us
a
lot
of
momentum
and
learning
that
we've
used
to
propel
the
rest
of
the
initiative.
So
where
are
we
right
now?
So
we've
actually
updated
the
sense
that
we
gave
the
presentation
in
April
we've
held
331,
chronically
homeless
individuals
since
January
of
2016,
and
when
we
tally
up
the
number
of
years
homeless
that
those
331
individuals
had
experienced.
A
We
know
that
we've
ended
over
2,000
years
of
homelessness,
so
we're
housing,
really
long-term
homeless.
Folks,
with
our
system
change
and
we're
we're,
you
know
we
remapping
that
maze
and
making
sure
that
long-term
homeless
individuals
are
getting
housed.
So
we
put
forward
in
the
plan
for
just
system
transformation
elements.
We
are
going
to
be
focused
on
on
kind
of
the
two
or
actually
coordinated
access
and
the
data
behind
this
really
to
do
any
kind
of
system
transformation.
A
We've
had
to
also
transform
our
data
system
make
sure
our
data
system
is
also
driving
the
change
and
is
as
unified
in
and
efficient
as
it
can
be
so
I
put
red
to
the
fore.
First,
let
me
say
the
four
buff
boxes
are
front
door
triage,
which
is
really
about
triaging
folks,
as
they
first
become
homeless,
making
sure
we're
connecting
them
with
the
resources
they
need.
Just
like
an
emergency
room.
A
Rapid
rehousing
is
about
giving
short-term
resources
to
folks
who
need
a
little
bit
to
get
back
into
housing
like
first
month
last
month's
rent
and
triaging.
The
right
folks
for
that
intervention.
Coordinated
access
is
our
housing
matching
engine
that
we'll
talk
more
about
in
a
minute
and,
of
course,
permanent.
A
Secondly,
on
the
HMIS
data
on
the
other,
we
have
our
housing
providers
telling
us
when
there
are
vacancies
and
and
when
they
alert
us
to
a
vacancy.
We
know
specifically
what
the
rules
are.
The
eligibility
criteria
for
getting
into
that
unit
if
it's
for
an
elderly
person
and
if
it's
for
someone
who's
chronically
homeless,
we're
making
sure
that
we're
using
the
data
to
match
the
most
vulnerable
person
to
that
vacancy
and
and
make
sure
that
they
get
the
first
opportunity
at
of
being
housed,
and
so
the
the
system
makes
the
match.
A
It
creates
a
ranked
list
based
on
vulnerability
and
we're
using
first
state
homeless
as
the
way
that
we're
ranking
that
list,
and
then
it
generates
emails
along
the
way
we
didn't
want
our
HMIS
users
to
have
to
log
into
another
system,
and
so
we're
still.
You
know.
The
message
is
a
straight
of
my
own
fate
to
my
us.
A
So
what
are
we
accomplishing
by
having
this
kind
of
housing
matching
engine?
First,
where
we're
targeting
resources
to
the
most
vulnerable?
First,
so
by
creating
this
list
no
longer
is
it
yeah?
You
serve
chronically
homeless
individuals
go
ahead
and
find
one
we're
saying:
how
is
this
most
vulnerable
person
first
and
and
house
them?
You
know
as
quickly
as
you
can,
and
so
you
know
that
dot
piece
really
is
working.
This
coordinated
access
system
really
has
flipped
the
system
on
its
head
in
that
way,
and
it
also
is
bringing
transparency
to
our
work.
A
So
it's
bringing
transparency
in
that
we
know.
We
know
when
a
match
is
made.
We
know
you
know
first
of
all
are
where
rules
are
very
transparent,
so
our
contractors
know
what
the
expectations
are
for
making
that
match,
but
also
the
the
person
experiencing
homelessness
knows
how
to
transplant
see
around
okay.
I
have
been
offered
this
resource.
These
are
my
opportunity.
These
are
my
options.
This
is
how
how
the
process
will
work
so,
there's
more
transparency
for
the
clients
and
then
last
as
a
system
there's
more
transparency.
A
We
now,
as
a
COC,
see
the
practices
and
the
speed
and
the
steps
that
our
housing
providers
take
to
help.
Someone
and
we've
in
the
process
negotiated
that
those
barriers
be
lowered
and
that
we
we
really
get
to
the
most.
You
know
the
most
housing
first
model
that
we
can
across
the
board.
So
the
transparency
has
brought
that
level
of
accountability
for
our
system
and
it's
also
more
efficient
to
have
a
system
that
is
automatically
looping
in
all
the
staff
that
need
to
be
involved
in
a
housing
match.
A
We're
certainly
using
our
resources
to
house
the
most
vulnerable.
There's,
an
efficiency
there
and
we're
continuously
working
on
improving
this
match
so
that
the
process
is
faster
and
and
streamlined
and
more
efficient.
So
we've
gotten
rid
of
some
of
our
paper
applications.
We've
gotten
rid
of
some
of
the
steps
that
people
need
to
go
through
and
we're
really
streamlining
the
process
of
of
getting
folks
into
housing.
So
I'm
going
to
turn
over
to
Jen
Flynn
a
27
minutes
trigger
for
the
city
of
Boston
hi.
B
I'm
Jennifer
Flynn
eh-
administrator
for
the
City
of
Austin
and
I'm,
going
to
give
you
a
little
bit
of
info
about
what
our
system
actually
looks
like
in
the
matter
of
number.
So
annually
we
see
about
11,000
single
adults
in
our
emergency
shelter
system,
2,600
families
1,700
singles
on
any
given
night.
We
have
27
service
agencies
that
use
rh-
system
for
HMS
installations,
which
I'll
talk
about
a
little
bit
in
a
minute
and
then
about
700
users.
That,
for
each
installations,
is
really
the
backbone
of
why
we
needed
an
HMIS
data
warehouse.
B
B
Therefore,
we
took
advantage
of
HUD
CSV
specifications
and
built
a
data
warehouse
for
data
transfer,
another
wrinkle
and
not
just
on
the
legacy
systems,
but
Massachusetts
is
a
right
to
shelter
state
and
our
State
Housing
and
Community
Development
Partners
run
our
emergency
shelter
system
for
families
statewide.
So
all
family
providers
are
required
to
use
the
state's
HMIS
system
again.
Therefore,
we
needed
a
data
warehouse
in
order
to
do
COC
wide
HUD
required
reporting.
We
do
provide
front-end
software
license
for
providers.
They
can
use
whichever
system
they
want,
but
to
be
a
nh-
participating
provider.
B
They
must
upload
their
CSV
data
into
our
H
at
least
monthly.
So
the
warehouse
consolidates
all
of
our
H
reporting.
It
gives
us
a
360
degree
view
of
our
clients
and
our
performance
across
the
COC
and
also
is
the
backbone
now
of
our
housing,
coordinated
access
process
and,
as
Lyla
said
during
you
know,
a
few
years
ago,
we
had
a
number
of
things
happen
that
led
us
to
making
the
decision
that
we
really
needed
to
make
an
investment
to
update
our
technical
backbone
of
the
COC
Boston.
B
All
of
that
led
to
a
crisis
in
our
data,
and
our
data
warehouse
had
not
been
updated
in
ten
years.
It
was
about
2005
when
it
was
built
and,
as
we
all
know,
technology
has
greatly
advanced
in
these
years.
So
through
the
mayor's
action
plan,
a
significant
financial
investment
was
made.
In
order
to
do
this,
it
was
the
first
time
we
at
the
City
of
Austin
and
as
a
CFC
lead
in
each
lead
had
had
such
a
significant
financial
investment
in
our
technical
backbone.
B
What
that
meant
was
we
wanted
to
preserve
our
H
infrastructure,
minimize,
the
duplication
of
data
entry
for
our
clients
and
then
going
forward
be
able
to
leverage
some
new
analytics
and
move
our
system
forward
again.
Technology
is
advancing
so
quickly
nowadays
that
something
that
we
create
today
is
obsolete.
Tomorrow,.
B
So,
with
our
large
financial
investment,
as
a
public
agency,
we
had
to
put
out
a
publicly
advertised
RFP
with
very
explicit
instructions
for
what
we
needed
in
the
warehouse
and
spent
a
lot
of
time.
Writing
our
RFP.
It
reads
more
like
a
story
than
an
RFP,
because
again
we
needed
to
be
able
to
tell
a
vendor
how
our
system
worked
and
be
able
to
design
a
system
that
worked
for
us
and
not
an
off-the-shelf
product
that
the
vendor
thought
should
be
working.
This
way.
B
We
worked
with
the
mayor's
office
of
new
urban
mechanics
in
the
Department
of
innovation
and
technology
on
ownership
rights,
which
is
a
very
exciting
idea
for
us.
As
we
put
out
the
RFP,
we
had
a
number
of
options
as
how
you
could
respond,
the
you
know
an
off-the-shelf
vendor
and
off-the-shelf
product
that
we
could
customize
or
work-for-hire,
which
is
the
solution
that
we
ended
up
going
with
the
really
cool
part
of
that
is
that
Boston
owns
the
code
for
the
system
and
we
have
put
it
out
on
the
open
source.
B
Something
also
it
was
very
different
for
us
was
an
agile
development.
We'd,
never
gone
through
that
process
before
we
always
had
fairly
off-the-shelf
products
and
then
customized
as
needed.
With
this
agile
development
process,
we
started
from
scratch,
put
something
together
very
quickly,
put
it
out
there
in
the
works,
tested
it
and
then
started
making
refinements.
As
we
saw
how
the
system
was
working.
B
I'm
going
to
share
a
few
screenshots
now
of
our
warehouse
that
have
become
so
useful
to
us
it
to
be
able
to
see
a
360-degree
view
of
a
client
amongst
all
shelters
and
providers
that
they're
using
has
been
an
incredible
wealth
of
information
both
for
our
service
providers,
but
also
has
helped
the
clients
take
out
some
steps
right
and
really
make
sure
that
we're
serving
them
at
the
curve.
The
most
is
the
most
appropriate
service
level
that
they
have
so
here's
a
screenshot
of
my
administrative
view.
B
As
the
warehouse
administrator
we've
been
able
to
bring
in
client
records
and
do
some
automated
joining
systems
from
our
four
sources.
To
probably
mention
that
previously,
our
warehouse
was
a
monthly
manually,
updated
system.
When
we
kind
of
flip
the
switch
on
this
new
warehouse,
we
are
now
getting
nightly.
Data
feeds
from
three
different
HMIS
source
providers,
both
off-the-shelf
vendors
that
are
out
there
in
your
world,
as
well
as
some
homegrown
systems
from
some
other
agencies,
they're
being
fed
through
nightly
to
a
secure,
SFTP
server.
And
at
this
point
in
time
we
are
ingesting.
B
B
Some
things
that
are
very
useful
for
us
is
that
not
only
can
we
just
be
program
enrollments,
but
we
can
drill
down
on
actual
assessment
data
and
have
a
client
history
calendar,
which
basically
shows
every
point
of
service
every
day
that
a
client
has
come
in.
We've
been
able
to
start
using
this
for
housing
history
verifications
as
well
as
chronic
homeless
verifications
with
our
housing
providers.
B
So,
as
I
said
earlier,
our
warehouse
is
our
designated
HMIS
for
the
COC.
It
means
a
couple
of
different
things,
and
that
means
that
the
warehouse
is
our
reporting
structure
for
HUD
and
also
our
HMIS
warehouse
has
to
act
as
a
HUD
vendor,
because
anytime
HUD
makes
data,
dictionary,
changes
or
programming
changes
in
the
reporting.
B
So,
as
the
designated
HMIS,
we
have
the
requirements
to
produce
many
HUD
required
reports,
so
system
performance
measures
are
in
here.
Housing
inventory
chart
point
in
time,
count
annual
homeless
assessment
report,
they're
all
generated
now
for
Boston
through
the
HMIS
data
warehouse.
All
the
reports
that
we
can
view
here.
Some
data
quality
reports.
You
know
what
clients
entry
date
is
the
same
of
their
birthday.
We
have
a
potentially
chronic
homeless
report,
which
has
been
invaluable
in
determining
our
most
vulnerable
clients
and
also
first-time
homeless
and
I.
B
Just
want
to
touch
on
that
for
a
moment,
because
this
is
something
that
we
had
never
known
before.
We,
you
know
we're
working
hard
housing,
our
folks
housing,
our
chronic
and
our
numbers
just
weren't
going
down,
and
we
were
able
to
determine
from
our
first-time
homeless
report
at
this
point
in
time,
Boston
is
receiving
one
brand
new
veteran
to
our
system
every
day
that
has
never
taken
a
bed
and
they're
coming
here
now,
as
well
as
roughly
just
about
12
brand
new
people
to
our
system
as
a
whole
a
day.
B
That
is
a
number
that
none
of
us
were
expecting
and
not
quite
sure
how
to
kind
of
handle
that
and
deal
with
that
right.
That's
what
front
door
triage
is
about,
but
because
Boston
is
the
large
city
in
not
only
Massachusetts
but
in
whole
New
England
region.
We
have
clients
coming
from
all
over
both
for
support
services,
or
you
know,
hospital
stays
any
kind
of
health
care
that
they
need.
B
We
also
at
this
point
have
somebody
in
our
shelter
system
from
every
state
in
the
nation.
Our
providers
really
needed
to
know
who
they
were
serving
and
where
their
clients
were
being
served
at.
We
at
Boston,
even
though
we're
a
large
city,
we're
really
a
small
city
where
you
know
we're
a
very
walkable
city,
and
that
means
that
folks
can
bounce
from
agency
to
agency.
B
So
after
I
would
say
maybe
18
months
of
dealing
with
lawyers,
HIPAA
lawyers
providers,
city
lawyers,
we
finally
managed
to
get
to
the
point
where
we
can
share
data
in
our
system,
both
with
client
release,
end
without
client
releases
and
that's
a
very
important
distinction
to
make.
So
we
formed
a
Boston
homeless
assistance
network
and
basically,
it's
kind
of
like
a
business
service
agreement,
business
associate
agreement
and
it
allows
our
providers
to
share
primary
personal
information
about
our
clients
in
order
to
coordinate
and
provide
services
to
them.
B
This
window
has
been
recently
rolled
out
both
to
our
street
outreach
teams,
our
front
door
triage
and
our
rapid
rehousing
programs,
as
we
move
forward
and
clients
begin
to
sign
releases
that
will
open
up
some
more
information
and
providers
will
be
able
to
view
front
door,
triage
assessments,
HUD
intake
assessments
and
any
other
kind
of
contact
information
that
needs
to
happen
in
our
window.
We
can
see
their
first
date
that
they
were
homeless
and
their
last
place
of
contact.
B
That
has
been
extremely
important
as
we
move
through
this
coordinated
access
system
in
our
housing
placement
system.
A
housing
opportunity
arrives,
for
you,
know,
Joe
Smith,
over
at
shelter
a
and
shelter
a
has
not
seen
them
in
a
while
they're
able
to
go
into
our
window
into
the
warehouse
and
find
out
that
the
client
may
have
had
lunch
at
a
participating,
shelter
yesterday.
So
we
can
put
that
flag
in
that
shelter
is
front-end
HMIS
system
and
alert
and
say
you
know.
Joe
Smith
has
a
housing
opportunity
next
time.
You
shows
up
that
your
agency.
B
C
You
Jen
I'm
in
from
Green
River
we've
been
the
software
development
partner
with
Boston
on
this
project
and
when
we
started
we
had
done
no
work
for
any
municipality
or
in
the
homeless
data
we
have
done
Public,
Health
sort
of
at
the
state
and
regional
level,
but
this
was
completely
new
to
us.
So
we've
learned
a
lot
and
when
we
first
made
the
RFP,
we
first
had
a
meetings
about
this
project.
C
It
seems
fairly
straightforward,
HMIS
was
well
defined,
I
mean
there's
all
of
these
federal
guidelines
about
what
the
data
needs
to
look
like
how
its
collected.
So
we
said
sure
we
can
put
together
a
data
warehouse.
The
federal
government
actually
publishes
even
a
data
warehouse
specification
for
a
schema
and
that's
where
we
started.
So
we
built
out
this
warehouse
a
database
that
began
to
consume
the
legal
specification
HMIS
data
from
all
of
the
different
systems
that
used
and
as
genocide.
C
There
were
quite
a
few
of
them
and
that
had
its
own
bumps
that
we
sort
of
didn't
anticipate.
Actually,
that
seemed
to
be
a
simple
process,
but
that
ended
up
being
a
bit
of
a
bureaucratic
challenge,
but
we
got
over
the
bureaucratic
challenge.
We
got
the
data
we
began
consuming
it
and
what
we
quickly
realized
was
the
warehouse,
then,
could
be
really
the
vehicle
for
integration
beyond
just
a
core
HMIS
specification.
C
What's
in
HMIS
sure
it's
the
agency,
the
service
providing
history,
it's
some
income
history,
but
it's
not
the
full
assessment.
It's
not
all
of
the
forms
or
things
of
data
that
was
collected,
documents
that
were
collected
and
there
aren't
little
things
like
client
photos
and
release
forms
that
actually
are
needed
if
you're
using
it
as
an
operational
system.
So
we
started
building
out
different
ways
of
consuming
data,
and
the
system
itself
is
set
up
to
basically
accept
any
kind
of
data
in
any
kind
of
form.
It's
this
modular
data
construction.
C
So
there's
core
HMIS
data
that
comes
in
in
the
exports
that
Jen
was
talking
about.
We
also
call
back
through
the
API,
is
available
for
the
HMIS
systems
to
pull
down
assessment
documents
and
photos
and
have
a
system
where
we
can
start
to
consume
pretty
much
any
data
shape.
What
that
meant
was
the
underlying
database
itself
quickly
evolved.
It
went
from
that
hub
specification
that
was
very
clean
and
very
simple
and
sort
of
could
follow
the
programming
recommendations
to
something
that
was
kind
of
sprawling,
that
that
accepted
all
the
different
data
types.
C
As
Jen
mentioned
every
night,
the
system
compares
over
a
million
records.
That's
not
a
million
client
records,
but
it's
record
of
service
of
enrollment
of
people
and
a
huge
challenge.
Very
clearly,
we
stumbled
upon
was
deduplication
both
within
a
single
HMIS,
and
there
were
plenty
of
instances
where
a
given
client
was
in
there
multiple
times
and
obviously
I
cross
multiply
HL
assets.
The
whole
point
was
to
merge
records
from
different
systems
and
again
that
was
one
of
those
things
that
seemed
easier
said
than
done.
So
what
we
built
were
three
tiers
of
deduplication.
C
The
first
was
completely
automatic.
There's
three
data
fields.
There's
your
birthday
here,
there's
your
naming
your
social
security
number.
If
two
of
those
three
things
match
boom
in
the
system
automatically
merges
you.
The
next
here
is
where
the
system
can
suggest
based
on
sort
of
the
magic
of
data
analytics,
a
similarity
between
clients,
so
it
goes
through
and
it
calculates
any
new
record.
It's
easy,
but
this
wait
a
minute.
C
This
record,
based
on
all
the
different
attributes
available
in
the
system,
appear
to
be
like
these
other
records
and
the
system
isn't
sure
because
the
name
might
not
be
the
same.
The
social
security
number
might
not
be
the
same,
but
it's
close
enough
that
it
recommends
that
a
human
then
goes
it
matches
it,
and
it's
actually
smart
enough
to
then
to
learn
from
those
patterns
of
matching
so
long
term.
It
should
get
more
and
more
accurate
to
be
a
little
data
geeky
and
it's
basically
looking
at
similarity
coefficient.
C
We
quickly
realized
again
that
the
warehouse
provided
this
vehicle
for
adding
all
kinds
of
different
information
and,
as
a
side
note
there's
this
project
that
was
started
by
Boston
healthcare
for
the
homeless,
to
bring
in
medical
records
that
they
have
all
of
their
providers
in
with
HMIS
data.
So
the
idea
is
all
of
the
service
agencies.
Providing
services
and
housing
to
homeless
can
now
join
together
in
making
an
integrated
care
plan
for
individuals.
C
The
neat
part
about
that
is
it's
targeting
people
who
are
both
chronically
homeless
and
have
chronic
medical
needs
and
for
the
first
time,
pretty
much
in
this
area.
I
believe
all
of
that
information
is
pulled
together.
So
in
a
grand
scheme
that
means
that
each
individual
can
have
a
very
targeted
care
plan
that
multiple
agencies
are
cooperating
with,
there's
also
small
benefits,
which
means
that
someone
walks
into
a
day
shelter
their
case
manager.
There
can
say:
hey
Joe,
do
you
know?
C
This
is
actually
just
a
wireframe
of
what
the
integrated
care
plan
from
the
healthcare
for
the
homeless
was
going
to
look
like
this
is
from
health
care,
for
the
homeless
and
I
say
was
because
it's
now
been
implemented.
It's
very
similar
to
this,
but
the
system
is
up
and
running
and
deployed,
and
actually
now
is
starting
to
serve
clients.
C
So
that
was
a
little
side
jaunt
into
some
of
the
benefits
of
the
warehouse.
But
the
purpose
of
it,
as
Lila
said
starting
off,
was
really
to
do
coordinated
access
and
coordination
often
means
communication,
and
we
spent
a
long
time
figuring
out
how
to
loop
in
the
different
players
and
actors
that
needed
to
be
involved
in
the
process
didn't
need
another
system
that
was
very
clear.
C
So
how
can
all
of
these
people
in
all
of
the
different
agencies
and
all
of
the
different
capacities
be
involved
in
a
way
that
was
meaningful
and
we
came
up
with
a
bunch
of
different
solutions,
and
this
is
pretty
illegible,
but
I
think
what
it
does
demonstrate
is
the
number
of
pathways
and
actors.
So
each
one
of
these
channels
in
the
swingline
document
represents
a
type
of
user.
So
there's
a
variety
of
actors
involved
in
the
process
and
each
one
of
these
swimlanes
represents
a
different
user
group
involved
in
the
process.
C
So
they
can
be
involved
in
different
ways.
They
can
be
simply
notified
that
that
some
housing
opportunity
has
become
available
and
offered
to
a
client.
Maybe
that's
just
an
FYI
that
they
need
to
know
they
can
be
involved
in
a
way
that
they
can
accept
or
make
a
comment
on
or
actually
reject
the
match
and
say
we
know
about
that,
and
it's
not
going
to
work
out
or
they
can
actually
log
in
and
sort
of,
be
involved
in
the
multiple
client
process
and
get
a
fuller
view
of.
C
What's
going
on
for
the
notifications
that
go
out,
we've
worked
in
several
ways.
One
is
just
a
passive
email
that
that
viola
described
that
sort
of
gives
a
notification
its
redacted,
so
that
it's
private,
the
other
way
is
to
send
out
an
email.
A
time
expiring,
link
that
if
someone
clicked
on
it
gets
additional
information.
But
again
that
is
sort
of
limited
in
privacy
and
permission
and
then
the
final
solution
will
someone
can
actually
log
into
the
system
and
see
the
full
details
of
the
client
and
the
match
opportunity?
C
C
The
internal
workings
of
coordinated
access
rely
a
lot
on
trying
to
figure
out
the
best
match
between
a
given
individual
and
a
housing
opportunity
that
comes
available
and
that
again
kind
of
easier
said
than
done.
So
what
it
allows
is
eligibility
rules
sort
of?
What's
what
are
the
requirements
and
restrictions
on
a
given
opportunity
to
be
set
at
and
read
this
at
the
program
level?
The
sub
program
for
funding
sources
for
sub
grantees
for
sites
and
for
individual
vacancies?
C
Well,
how
does
that
play
in
when
you're
trying
to
come
up
with
a
priority?
So
do
you
try
to
target
someone
who
may
not
be
the
most
vulnerable
individual
technically,
but
fits
a
very
restrictive
criteria,
base
and
say
wait
a
minute?
This
is
someone
who's,
probably
going
to
succeed
in
this
particular
opportunity,
or
do
you
sort
of
do
it
straight
down?
As
they
will
know,
the
most
full
level
comes
to
the
first
opportunity
that
that
person
is
eligible
for
so
that's
sort
of
an
ongoing
dialogue
and
again
it
begins.
C
And
this
is
just
a
little
insight
into
sort
of
our
agile
process,
different
designs,
that
we've
come
up
with
to
sort
of
show
the
steps
visually
within
the
process
of
matching
and
haven't
implemented,
actually
any
of
those,
but
use
them
as
an
illustration
of
sort
of
saying
wait
a
minute.
One
of
these
would
probably
be
useful.
A
So
successes
and
challenges:
where
have
we?
Where
have
we
come
so
since
we
started
matching
chronically
homeless
individuals
through
this
coordinated
access
system,
51
people
have
successfully
matched
to
a
housing
resource.
Most
of
those
housing
resources
have
been
10,
apiece
vouchers,
so
17
of
those
folks
are
actually
housed
and
the
rest
are
in
the
housing
search.
A
We
now
have
this
nightly
data
coming
from
our
HMIS
warehouse
into
the
housing
matching
engine
and
also
loading,
so
people
can
see
exactly
where
someone
slept
last
night
and
perhaps
where
they
had
lunch
yesterday,
which
is
helping
quite
a
bit
with
coordination
in
our
system.
Our
coordinated
access
system
is
live
linked
to
that
HMIS
data.
So
it's
also
in
real-time.
A
We
now
have
the
basis
for
a
data-driven
system
and
we're
starting
to
use
data
to
hold
providers
accountable
to
drive
towards
housing,
first
to
drive
towards
outcomes,
to
reach
our
goals,
and
we
know
there's
more.
We
can
do
that
do
in
that
realm.
Now
that
we
have
the
data
in
this
format,
we
have
reporting
to
HUD
all
coded
into
our
warehouse.
Our
systems
performance
measures
are
coded,
we're
starting
to
use
that
back
out
to
our
providers
again
so
that
they
see
where
they're
stacking
up
in
the
system.
A
Performance
measures
we're
starting
to
create
performance
dashboards
around
our
goals,
so
we're
working
on
a
dashboard
for
homeless
veterans.
How
are
we
doing?
What
are
we
learning
from
the
data?
How
can
we
improve
based
on
on
what
we
see
and
we
have
as
Jen
talks
about
this
network
sharing
agreement
amongst
our
partners
and
we're
working
on
other
sharing
agreements
beyond
that
as
well,
including
the
county
houses
of
Corrections
and
Medicaid
data
and
other
data
sets
that
will
help
us
improve
our
performance
and
improve
our
outcomes?
A
Where
have
we
run
into
some
challenges?
Certainly,
there
are
a
lot
of
moving
parts.
There
are
a
lot
of
stakeholders
and
there
are
a
lot
of
tech
stakeholders.
There
are
multiple
hy
offenders
and
so
coordinating
those
has
has
been
a
challenge,
we're
in
a
much
better
place,
but
that
has
taken
some
time.
Certainly,
a
lot
of
this
data
sharing
the
complex
part
gets
bogged
down
in
HIPAA
rules
and
you
know
who's
a
HIPAA
covered
entity
who
it's
not
a
HIPAA
covered
entity.
A
So
beyond
the
COC
providers
to
see
housing
providers,
you
know
look,
we
can
match
on
average
in
this
number
of
days,
so
come
put
your
units
into
our
system,
we'll
be
able
to
get
you
tenants
fairly
quickly
and
we're
just
now
starting
to
look
at
performance
reports
out
of
our
housing
mapping
engine.
So
we
can
start
to
see
where
we
can
improve
our
time
for
housing
and
so
we're
we
had
it.
A
So
certainly,
we
want
to
expand
beyond
matching
chronically
homeless
individuals
to
permanent
supportive
housing
to
other
populations.
We
want
to
be
able
to
match
veterans
to
housing,
but
also
we
want
to
bring
in
our
rapid
rehousing
resources
into
the
system
and
and
then
we
want
to
bring
in
housing
resources
that
we
don't
fund
as
a
continuum
of
care,
but
that
are
dedicated
to
homeless
households.
So
we
have
affordable
housing
in
our
community
that
are
reserved
for
homeless
households.
A
Our
Housing
Authority
has
a
homeless
preference,
so
there
is
a
lot
of
opportunity
to
expand
this
and
bring
in
a
lot
more
a
lot
more
housing
to
be
able
to
match
I've
even
talked
about
where
we're,
starting
with
this
pilot,
with
healthcare
for
the
homeless,
to
integrate
healthcare
data
with
homeless
data.
But
we
we
see
that
that
could
could
be
something
that
that
we
expand
quite
a
bit.
More
certainly
is
a
lot
of
value
in
understanding
how
housing
improves
folks,
health
outcomes
and
how
to
target
housing
very
vulnerable
people
based
on
their
medical
status.
A
We
certainly
want
to
bring
in
more
qualitative
information
as
needed,
so
we
may
want
to
bring
in
case
manager
notes
into
our
warehouse
window
and
there's
opportunity
here
now
that
we
have
this
this
data
system
to
bring
in
new
users.
So
we
can,
you
know,
kind
of
conceptually
open-source
our
homeless
response.
Maybe
there
are
community
groups
that
have
not
typically
been
able
to
help
people
get
into
housing.
That
could
be
part
of
it.
A
So
if
someone
shows
up
at
a
library
and
they
have
a
housing
match,
but
they
need
a
an
ID,
because
that
librarian
help
this
person
get
their
photo
ID
and
help
move
them
to
housing.
So
there's
a
lot
of
possibilities
now
that
we
have
these
different
tools
and
different
windows
for
our
information
and
I'm
going
to
pass
it
back
to
n
to
talk
first
last
slide
so.
C
We
have
all
this
data
now
and
it's
growing
and
growing
and
growing
and
we
keep
adding
more.
We
add
different
types,
we
add
more
records
and
we
sort
of
live
in
the
world
now
of
machine
learning
and
artificial
intelligence
and
predictive
analytics
and
software
developers.
Our
heads
immediately
go
to
what
can
we
do
with
all
of
this
data
or
what
are
the
sort
of
secrets
within
that?
We
can
start
to
leverage.
C
All
of
the
data
in
there
and
Boston
has
a
tremendous
data
set
that
goes
back
years
and
years
and
years
theoretically
perhaps
could
yield
interesting
questions.
Interesting
answers
to
things
like
where
is
chronicity
coming
from.
So
all
of
the
the
clients
that
now
are
being
targeted
for
coordinated
access
are
on
the
chronic
list,
and
the
calculation
of
that
is
very
straightforward.
But
is
there
something
in
those
clients
histories
that
would
have
been
a
predictor?
C
That
could
be
say
if
this
pattern
of
services
was
identified
early
on
that
that
person
could
be
interacted,
there
could
be
an
intervention
before
the
state
of
credit
city
was
reached
that
there
was
a
consistent
pattern
that
could
be
identified
within
the
data
that
would
actually
allow
for
early
identification
and
as
then,
the
system
is
being
used
for
different
purposes
like
coordinated
access
to
housing.
Can
we
learn
from
the
successes
and
failures,
because
these
folks
have
been
in
housing
before
those
opportunities
have
come
up
and
they
failed,
and
why
was
that?
C
And
is
there
something
in
the
larger
data
set
of
their
actions,
their
interactions
with
agency,
the
opportunity
itself?
The
attributes
around
that
that
we
can
sort
of
mine
out
to
start
to
predict
opportunity,
match
success
and,
to
sort
of
say
well,
wait
a
minute.
What
is
the
best
match?
What
is
the
best
opportunity
for
this
individual
for
success
and
then
obviously
being
able
to
look
at
program
performance
of
agencies
and
services
that
are
actually
being
provided,
because
now,
with
this
whole
integrated
history,
you
can
start
to
see
a
wait.
A
minute
is
this
agency's
actions.
C
Is
this
shelter
program
leading
to
success?
Is
it
related
to
other
other
activities
and
that
gets
again
very
exciting
when
you
start
to
go
into
the
healthcare
side
of
things,
because
you
have
so
many
different
data
from
these
from
so
many
different
types
of
organizations,
you
can
sort
of
look
for
patterns,
so
that's
all
the
hope
we're
not
there.
Yet
we
are
finally
getting
the
larger
data
set
that
that's
certainly
where
we
hope
to
go
very
soon
and
just
destroyed
on
the
follow-up.
C
Both
while
injen
mentioned,
the
open
source
Boston
was
pretty
progressive
in
their
licensing
choice
because
it's
open
source
under
the
GPL
license,
which
means
it's
available
for
everyone
for
free.
Any
community
any
organization
can
adopt
and
use
it
as
is,
and
any
modification
that
is
made
any
improvement
or
enhancement
either
for
commercial
or
nonprofit
needs
has
to
also
go
into
the
open
source.
So
the
hope
really
is
that
another
community
or
organization
will
adopt
it
and
start
to
contribute
back
to
it
and
again
it's
there's.
C
No,
what
royalties
there's
no
license
fee
so
technically,
it
is
freely
available
to
everyone
and
actually,
since
this
presentation
there
has
been
another
community
in
Massachusetts-
that's
adopted
it
and
the
hope
is
that
more
will
a
little
mechanical
bit.
It's
a
Ruby
on
Rails
application
built
on
top
of
originally
Microsoft
sequel
server,
and
that
was
a
technical
requirement
of
the
city
of
Boston.
But
it's
since
been
migrated
to
Postgres
for
Boston.
We
never
actually
handles
it
as
a
hosted
solution.
C
A
Thank
you.
We
are
really
excited
for
the
possibility
of
partnering
with
other
folks.
We
built
this
using
open
source
code
so
that
we
could
collaborate
with
other
communities
and
what
we
ask
is
any
community
that
does
decide
to
take
up
the
code
that
anything
you
build
or
modify.
You
contribute
back
to
the
open
source
code
and,
in
that
way,
we're
collectively
learning
from
each
other
and
building
improving
together
on
our
collective
mission
to
end
homelessness
in
our
communities.
So
please
do
reach
out
to
us.