Professional Documents
Culture Documents
TO DECISIONS III
Lessons from Early Analytics Programs
N OVEMBER 2013
The IBM Center for The Business of Government connects public management research with practice. Since 1998, the
Center has helped public sector executives improve the effectiveness of government with practical ideas and original
thinking. The Center sponsors independent research by top minds in academe and the non-profit sector, and creates
opportunities for dialogue on a broad range of public management topics.
b PARTNERSHIP FOR PUBLIC SERVICE | IBM CENTER FOR THE BUSINESS OF GOVERNMENT
INTRODUCTION
Kenneth Dickie still shudders at the memory. One day in 1979, someone snuck into
his secret office in the basement of the [Veterans Affairs Departments] Washington
Medical Center. The intruder stacked piles of patients records around Dickies DEC
minicomputer, doused them with a flammable material and set them on fire. Smoke
filled the room, but fortunately for Dr. Dickie and for the future of American health
care, an alarm went off in time, and the computer he was using to build the countrys
first practical electronic medical record system was spared.1
Its hard to imagine in todays data-drenched govern- Analytics is the study of data to discover patterns,
ment, but early analytics users sometimes were scorned opportunities and linkages that enable prediction and
by agency staff as rogues and renegades. Dickie and a inform decisions. Data trailblazers like VA have much to
band of other amateur programmers overcame attacks teach agencies being pressed to use analytics now. Thats
from VAs central computer office by winning over doc- why we examined how early programs got started, what
tors with their fledgling health record software and sustained them and how data use altered mission-critical
thereby getting senior leaders support. programs. We purposely focused on true mission analyt-
Had Dickies machine burned or had higher-ups ics programs that apply data-based analysis directly to
failed to protect the secret programmers, one of todays improve mission delivery or performance.
most widely used electronic medical records, the Veter- Government, like industry and commerce, is inun-
ans Health Information Systems and Technology Archi- dated with data demands, especially with Barack Obama,
tecture (VISTA), might never have been created. With- who has been dubbed the big data president, at the
out it, the VA might never have adopted the wide array helm.2
of data-based performance measures and tools that have On his first day in office in 2009, Obama issued a
helped transform it into a hospital system whose quality memo on transparency ordering agencies to put op-
and safety scores are among the best in the United States. erational information online.3 In March 2012, he is-
Dickies story is a fitting beginning for the third re- sued a $200 million big data research and development
port in our From Data to Decisions series, a collabo- initiative.4
ration between the Partnership for Public Service and And in May 2013, he set a new standard requiring
IBMs public sector business analytics and optimization
practice. Our focus this time is mature analytics pro-
grams, such as the VAs, that started before the terms big 2 Nancy Scola, Obama, the Big Data President, Washington Post, 14
June 2013, http://bit.ly/18Ghl3F.
data and analytics were in use, let alone in vogue.
3 White House Office of the Press Secretary, Transparency and Open
Government Memorandum, January 21, 2009, http://bit.ly/lfWWOgD.
1 Phillip Longman. Best Care Anywhere: Why VA Health Care Would 4 White House Office of Science and Technology Policy, Obama Ad-
Work Better For Everyone, (Bk Currents, Berrett-Koehler Publishers, ministration Unveils Big Data Initiative, March 29, 2012, http://l.usa.
San Francisco 2012), Kindle Edition, 23. gov.1blTsoW.
2 PARTNERSHIP FOR PUBLIC SERVICE | IBM CENTER FOR THE BUSINESS OF GOVERNMENT
Earlier programs, like todays, Collaborate with other agencies analytics programs, this has been a
needed senior support. One lesson to collect data and share struggle.
they offer is that turning leaders analytics expertise Lacking cost-benefit data, some
into allies entails delivering findings Save money and effort, and increase programs turned for evidence to
quickly and tailoring them to execu- the speed of analytics adoption, by academic studies and reports from
tives needs, especially when the goal acquiring data and services, such professional organizations. Other
is persuading them to act on analysis as collection, analysis and modeling sources can include federal statisti-
that challenges long-held percep- tools, from other agencies. cal agencies, such as the Census and
tions or suggests that practices or Search for existing authorities Labor Statistics bureaus, and indus-
whole programs must change. that allow you to pay for help. They try groups.
Now, as in the past, data can can range from interagency acquisi- Programs that can point to im-
produce unwelcome results. Prac- tion of data and services under the proved outcomes from the use of
titioners confront resistance from Economy Act, to provisions specific mission analytics still are challenged
those who gather data and stand to to your agency, such as USAIDs in putting a dollar value on those
be affected by it. Thats why its im- participating agency service and results, particularly when they are
portant to ensure that collectors see program agreements. Grants are measured in terms of costs avoided.
the results of analytics and to com- another source of funds for engag- Some program managers are
municate how using data often can ing partners. And when analytics experimenting with methods for es-
improve the ways jobs are carried programs can help other agencies timating the value of cost-avoidance,
out and missions are achieved. achieve their missions, consider for example, by using billing records
Analytics has always been a striking memorandums of under- of trusted vendors to estimate costs
collaborative effort. The best pro- standing so each partner can per- avoided by cutting off defraud-
grams look to others for additional form the work that suits it best. ers. Analysts must be careful when
data sets, funding, ideas and labor. Marketing data-driven products calculating ROI for analytics to re-
Methods employed by early users to other agencies also can bring in port not only projected savings, but
for building and sustaining those funds and assistance. Research and also projected costs for their data
relationshipsfrom marketing their development agencies are accus- programs.
analytics projects to potential col- tomed to sharing their discoveries In some cases, improving ROI
laborators to using grants to equip to get them into operation, and the estimates requires increased and
and staff partnersoffer tested mod- approach is worth expanding, both enhanced data collection. Programs
els for today. to prevent duplication and to share also can employ surveys and audits,
And in todays austere budget increasingly tight resources. and conduct secondary screening,
environment, analytics programs which captures what analytics-
must justify their costs. They can Develop data to determine based efforts might have missed, to
take lessons from older programs return on investment for assess the success of analytics pro-
techniques, such as relying on the analytics programs grams and compare their results
great power of output and out- Reporting improved outcomes, such with alternate strategies for achiev-
come metrics to make the case for as increased numbers of foodborne ing the same goals.
mission-critical data projects. The illness outbreaks detected or enemy
danger is in relying on these metrics combatants identified, is a bottom- Give agency leaders clear, concise
exclusively, without also measuring line requirement for mission ana- analysis and proof of adoption,
monetary benefits. lytics programs. But just reporting and results they can use to
When benefits data was hard better outcomes is not sufficient, support data-driven programs
to collect, some programs turned to especially now that sequestration Much as most analytics users wish
academic studies or the work of pro- is compelling programs to compete everyone would immediately under-
fessional organizations for support. fiercely for scarce dollars. stand and appreciate their findings,
Perhaps the most important lesson Agency leaders need cost-bene- that doesnt always happen. Pre-
is to expend analytical effort not just fit metrics and measures of ROI to sentation is especially important for
on mission improvement but also on prove that data-based efforts com- top officials whose time and atten-
demonstrating return on investment. pare favorably with other programs tion are limited, but whose support
These cases offer lessons in during budget reviews. So dont fo- is vital. Data visualizationcharts,
making analytics a default approach cus on core analysis so single-mind- graphs, maps and modelsmake
for accomplishing mission goals. edly that you fail to develop data to analytical findings easier and faster
demonstrate ROI. Even for mature to comprehend.
4 PARTNERSHIP FOR PUBLIC SERVICE | IBM CENTER FOR THE BUSINESS OF GOVERNMENT
Findings from Previous Reports
PAGE 8
6 PARTNERSHIP FOR PUBLIC SERVICE | IBM CENTER FOR THE BUSINESS OF GOVERNMENT
Defense Department Animal Plant Health Veterans Health
Biometrics Inspection Service Administration
The Automated Biometric Identification Agricultural Quarantine Care Assessment Needs (CAN) Score
System (ABIS), fed by field collection using Activity System (AQAS) Patient Care Assessment System (PCAS)
the Biometrics Automated Toolset (BAT),
Handheld Interagency Identity Detec- The USDA Animal Plant Health Inspec- The Veterans Health Administration (VHA)
tion Equipment (HIIDE), Secure Electronic tion Service (APHIS), in cooperation uses the care assessment needs (CAN) score,
Enrollment Kit (SEEK), and other data with the Homeland Security Department a weekly analytic predicting the likelihood
Customs and Border Patrol, uses AQAS of hospitalization or death within 90 days
The U.S. armed forces have collected to make risk-based determinations about and a year, to identify high-risk patients. The
biometric informationideally 10 finger- which shipping containers at U.S. ports patient care assessment system (PCAS) uses
prints, iris scans, facial photo and biographic to examine for plant-borne pests. the CAN score and a host of other data to
informationfrom non-U.S. citizens in Iraq enable patient-centered care teams to coor-
and Afghanistan since 2003. Biometrics Began: 2007 dinate services to prevent hospitalizations.
are used for access to the United States,
U.S. facilities and coalition-controlled areas Cost: Unavailable Began: CAN2011, PCAS in pilot testing
in-country, identification of enemy fight- since Dec. 2012, rollout slated for 2014
ers, forensics (e.g., to identify makers and ROI: Invasive species cause estimated losses
implanters of improvised explosive devices) of $136 billion annually. Automating the Cost: Unavailable
and intelligence. Biometrics collected by emergency action notification reporting has
members of the Army, U.S. Marine Corps, enabled APHIS to redeploy one full-time ana- ROI: Patients with CAN scores in the
Special Operations Command and inter- lyst into a more valuable role. Replacing du- top 10 percent who saw their assigned
national governments are shared with the plicative reporting tools with a single analyt- primary care providers for more than 60
State and Homeland Security departments. ics solution has reduced costs by 30 percent. percent of scheduled visits were 10 percent
less likely to die or be hospitalized than
Began: 2003 Data Size: Unavailable similar risk patients who did not see their
providers during the preceding year.
Cost: $3 billion from 2007 to 2012
PAGE 14
Data Size: The CAN process collects more
ROI: From 2004 to 2012, approximately than 14 GB of patient-level data (120 unique
3,000 enemy combatants identified, elements for each score) on 5.25 million
950 high-value individuals captured primary care patients. CAN and PCAS are
or killed, 2,300 detainees denied early fed by the VHAs 80-terabyte corporate
release; added 190,000 identities to DODs data warehouse, which aggregates the
biometric-enabled watch list; through electronic health records in the Veterans
February 2011, 538 people seeking asylum Health Information Systems and Technol-
in the U.S. turned away due to biometric ogy Architectureapproximately 30 million
matches with negative information. records, 3.2 billion clinical orders, 1.8 billion
prescriptions, 2.3 billion vital-sign measure-
Data Size: 7 million records in ments and 2 billion clinical text notes.
ABIS, 4.4 million unique identities
PAGE 16
PAGE 12
The world awoke slowly to the 19831985 While the network does warehouse data Program, humanitarian assistance organiza-
famine in Ethiopia and Sudan. Only after in computerized environments, it does not tions and other governments.
shocking images of masses of starving and analyze all of them using automated analyti- Combining physical and social science
dying people were televised did humanitar- cal algorithms, preferring instead to merge takes close interaction, wrote NASA research
ian aid flow. Even then, Ethiopias embattled human and software analysis. Theoretically scientist Molly Brown. For example, an analy-
Marxist government was slow to distribute it. it could be computerized, Eilerts says. We sis of the impact of drought as measured both
Between 400,000 and 1 million people died. have had people offer to organize the data, by vegetation anomalies and rainfall deficits
The U.S. Agency for International De- but were afraid that the overhead and the needs to be integrated with information on
velopment (USAID) created the Famine Early outcomes arent worth the risk of disrupting elevated food prices, migration patterns and
Warning Systems Network (FEWS NET) in our activities. water scarcity.9
1985 to speed response to future famines and Whats more, he isnt sure software Research and development agencies,
prevent a repeat of Ethiopias massive loss of could capture and make sense of the subtle- such as NASA, look for opportunities to get
life. FEWS NET was created to help USAID ties and nuances that long-experienced ob- their new techniques for using satellite data
deliver its $1 billion in annual food aid where servers can apply. For example, he says, all and modeling put into use, and the famine
it was most needed and would do the most the science and much of the social data may network provides one. We help them see
good, said Gary Eilerts, the networks pro- show that the food crisis is approaching fam- ways in which their science and technology
gram manager. ine level in Nigeria. But I know the area, and can be effectively applied for an activity like
Humanitarian aid is enmeshed in poli- a lot of people are sending remittances home FEWS NET, which is quite operational in na-
tics. In Africa, where the network focused first, from outside Nigeria, so its not, Eilerts says. ture, said Verdin.
governments can be reluctant to announce a Each situation is very local; its about human Same thing with NOAA, he added. The
food emergency, which can be read as an ad- behavior. weather service is very operational. We help
mission of policy failure. Aid-giving govern- The USGS is the networks most active them by contributing land surface monitoring
ments face withering criticism if they drum up science partner, but its hardly alone. Even products to the atmospheric, which is their
assistance for a crisis that fails to materialize Eilerts, who is housed in USAID, is a USDA bread and butter.
or if it is seen to be bolstering enemies. And employee. USAID built the famine network The agencies now use data they devel-
famines are slow-onset disasters in places through interagency agreements with USGS, oped for famine warning to analyze condi-
where drought, poverty, illness, malnourish- USDA, the National Aeronautics and Space tions elsewhere, according to Verdin. Our ex-
ment and lack of sanitation are endemic. Administration (NASA) and the National pertise with the vegetation index has helped
When a crisis call is unpopular, difficult and Oceanic and Atmospheric Administration us to monitor indications of drought and fire
fraught with political implications, it has to (NOAA). The science agencies provide satel- hazard in the U.S, he said. Another example
be correct. Thats where the famine warning lite remote-sensing data, modeling, forecast- youve got later is the land surface tempera-
network comes in. ing, geographic information systems, training ture data. We first used it to help USAID in
From the beginning, the network relied and analysis. Afghanistan and now use it in regional offices.
on a mix of social and physical science data We focused very early on evidence and The opportunity for scientific communi-
to determine, more precisely than ever be- data and science, Eilerts said. We didnt flail cation, rather than authority over one another,
fore, which parts of the population, in which around. It was, How do we get data and infor- sustains the networkthat and just a com-
regions of which countries, would suffer most mation to say if there is a [food] access, avail- mon desire to do the best job we can do by
from environmental shocks, usually drought. ability or utilization problem. FEWS, he said.
The FEWS decision-support system can Network food security analysts in Africa, NASA is tailoring a version of its land
be seen as an interactive filtering process by Central America and Afghanistan track mar- information system (LIS) just for FEWS NET.
which enormous amounts of data are trans- ket, vulnerability, livelihood and agricultural The software framework contains several dif-
formed into fair, objective, reproducible and conditions. A contractor compiles the analy- ferent land surface models that each help
defensible analyses, wrote two of its scien- sis into tightly written food security outlooks, characterize a continent or country in terms
tific supporters, Chris Funk and James Verdin alerts and briefs for decision-makers in US- of soil, vegetation coverage and other land
from the U.S. Geological Survey (USGS). Ef- AIDs Food for Peace and disaster assistance attributes. Its a way to transfer your mea-
fective early warning combines a successful offices, the State Department, Congress, the surements into land surface conditionsthe
blend of Earth observations, hydrologic mod- White House, the United Nations World Food depth of rivers, snowpack, etc., Verdin said.
eling, food economics, weather and climate The advantage of having an LIS, [is that] if
modeling, and much more.8
Network, Satellite Rainfall Applications for Surface
Hydrology, eds. Mekonnen Gebremichael and Faisal 9 Molly Brown, Famine Early Warning Systems
8 Chris Funk and Jim Verdin, Real-time Decision Hossain, (Springer Science+Business Media B.V., and Remote Sensing Data (Springer-Verlag Berlin
Support Systems: The Famine Early Warning Systems 2009). Heidelberg, 2008), Kindle location 595.
8 PARTNERSHIP FOR PUBLIC SERVICE | IBM CENTER FOR THE BUSINESS OF GOVERNMENT
SATELLITES
Between November 1992 and February 1993, of foodborne illness ranges from $51 billion to Ohio State Universitys Robert Scharff
four children died and 732 people became ill more than $77 billion.11 estimated that PulseNet costs about $10 mil-
after eating E. colicontaminated hamburgers In its first year, PulseNet interacted with lion a year and saves $291 million.14
served at Jack in the Box restaurants in Wash- four public health laboratories and tracked a The CDC pointed out that during the
ington, Idaho, California and Nevada. The single pathogen. By 2012, PulseNet included 1997 E. coli outbreak in Colorado, If 15 cases
outbreak inspired a national effort to speed 87 labs and was tracking eight pathogens. were averted by the recall of potentially con-
detection of foodborne illnesses. PulseNet-certified Food and Drug Adminis- taminated ground beef, the PulseNet system
In 1994, the Centers for Disease Control tration (FDA) and Agriculture Department [in that state] would have recovered all costs
and Prevention (CDC) and the Association (USDA) laboratories also use PulseNet to of start-up and five years of operation.15
of Public Health Laboratories (APHL) began track pathogens collected from food or ani- In an era of continuing fiscal uncertainty,
work on a database containing the DNA fin- mals in an attempt to catch illness-causing PulseNet must continue to demonstrate its
gerprints of the E. coli bacterium. By 1996, bacteria earlier, before they infect people. value. A risk to PulseNets effectiveness is the
the database, known as PulseNetthe same These partnerships flourish because rise of very fast and inexpensive clinical labo-
name as the network of state laboratories CDC meets regularly with FDA and USDA to ratory tests that do not produce a pure bac-
working with CDCwas up and running, pro- discuss data standards. CDC branch chief Ian teria culture. Without bacteria to fingerprint,
cessing 154 bacterial DNA samples and iden- Williams reflected: Like any good relation- PulseNet cant track outbreaks.
tifying several multistate outbreaks in its first ship or marriage, it requires working together So CDC plans to ask labs to continue
year. Since then, PulseNet has accumulated to identify and resolve problems as they submitting cultures and preserve finger-
more than 500,000 bacteria isolates. come up, and were good at that. prints of samples to help halt future out-
Eighty-seven public health laboratories, In 2011, officials traced Listeria to can- breaks. Meanwhile, CDC is working on de-
including at least one in every state, partici- taloupes from a farm in Colorado. It was the veloping sequencing technology for genetic
pate in PulseNet, creating partial images of deadliest foodborne disease outbreak in material that doesnt require a pure culture of
bacterial DNA using equipment purchased the United States in almost 90 yearscaus- bacteria.
with CDC grants. Fecal, blood or urine sam- ing 29 deaths and a miscarriage. In just 10 Whole genome sequencingmapping
ples from sick patients are sent from hospi- days, officials spotted an unusual increase an entire strand of DNAoffers PulseNet a
tals and doctors offices to local labs, which in Listeria cases in local hospitals, identified tantalizing opportunity. Genome sequencing
extract cultures of bacteria and send them to contaminated cantaloupes as the source and produces more DNA data than the current
the public health labs for DNA fingerprinting. issued a national consumer warning. Officials testing, which would increase the speed and
The prints then go to PulseNet for analysis. said it was the fastest Listeria investigation accuracy of PulseNets bacteria identification.
The CDC requires all the PulseNet labo- [theyd] ever seen.12 Up to twice as many As sequencing becomes further automated
ratories to use the same data standards, en- would have been infected had officials not and problems with transmitting and storing
suring that bacteria strains can be compared had the tools, people and systems in place, its large images are solved, it will allow CDC
in a single, shared database. PulseNet scien- estimated CDC deputy director Robert to include in PulseNets database all known
tists run these images against the database Tauxe.13 pathogens instead of the eight currently
for matches with samples from other patients. Food is not the only bacteria carrier tracked.
More than two matches form a cluster, rep- PulseNet can trace. In 2012, PulseNet discov-
resenting a possible outbreak. ered a Salmonella outbreak; 26 people in 12
When CDC analysts discover a cluster states were infected. Eight were hospitalized
that is larger, faster-forming or more dan- and one died. Epidemiological investigations
gerous than is typically expected during that led to pet hedgehogs.
time of year, they alert state public health
epidemiologists to investigate. Interviewing
patients about their recent food handling, United States, http://1.usa.gov/17iULmd.
exposure and ingestion helps state health of-
11 Robert L. Scharff, Economic Burden from
ficials identify the source of the bacteria and Health Losses Due to Foodborne Illness in the Unit-
stop the outbreak. 14 Robert L. Scharff, A Model of Economic Ben-
ed States, Journal of Food Protection, Vol. 75 No. 1,
efits and Costs from PulseNet, The Ohio State
The CDC estimates that 47.8 million 2012, 123131.
University Department of Consumer Sciences, 2010,
people a year get foodborne illnesses, result- 12 Quoting Robert Tauxe, MD, MPH, deputy direc- presentation slides, http://bit.ly/18GMGmU.
ing in 127,839 hospitalizations and more than tor of CDCs Division of Foodborne, Waterborne
15 EH Elbasha, TD Fitzsimmons, and MI Meltzer,
3,000 deaths.10 The annual economic burden and Environmental Diseases, from Deadly Listeria
Costs and Benefits of a Subtype-specific Surveil-
Halted in Record Time, http://1.usa.gov/17OlNfl.
lance system for Identifying Escherichia coli O157:H7
13 Tauxe, Deadly Listeria Halted in Record Time, Outbreaks, Emerging Infectious Diseases, 2000,
10 CDC, 2011 Estimates of Foodborne Illness in the available http://1.usa.gov/17OlNfl. Vol. 6 No. 3, 293297.
10 PARTNERSHIP FOR PUBLIC SERVICE | IBM CENTER FOR THE BUSINESS OF GOVERNMENT
The CDC estimates that 47.8 million people in the United
States (15%) fall ill due to foodborne illness annually,
resulting in 127,839 hospitalizations and 3,037 deaths.
200
CLUSTERS
IDENTIFIED
If a cluster is atypical, state and local
epidemiologists work to identify the source of the
167 outbreak. CDC may coordinate efforts involving
multiple states.
IN 2012 investigations
PERFORMED
60
SOURCES
IDENTIFIED
9
RECALLS
PERFORMED
Once the source is identified,
the CDC works closely with FDA
and USDA regulators to stop the
outbreak, determine when the
outbreak is over and prevent
future outbreaks.
Afghanistan government forcesprimarily als, denied access to over 64,000 potential Not being a program of record kept bio-
police and militaryhave been killing a larger threats and resulted in over 200 interdictions metrics from being taught in Army school-
percentage of soldiers from the U.S.-led co- through collaboration with interagency part- houses, so pre-deployment training is ad hoc
alition each year. The percentage of so-called ners, said John Boyd, Pentagon biometrics and troops often arent proficient at captur-
green-on-blue killings grew from less than 1 director. Leveraging biometrics enables on- ing full fingerprints or identifying which hand
percent in 2008 to 15 percent of all coalition going operations and translates to real suc- or person the prints are from.
deaths in 2012. The 80 attacks between Janu- cesses for the security of our nation. If service members fail to collect bio-
ary1, 2008 and July 15, 2013, left 134 coalition Yet defense biometrics analysis has graphical information from subjects or it is
troops dead and 153 wounded.16 been criticized for lacking a single strong lost, the data become useless, such as the
So when biometrics helped prevent a sponsor within the Pentagon, insufficiently approximately 4,000 biometrics collected
man on a U.S. terrorist watch list from joining training collectors and their commanders, from 2004 to 2008 that were separated
the Afghan police this June, Don Salo, direc- buying mismatched collection devices, and from their associated identities, according to
tor of the Defense Department Forensics and not fixing slow transmissions to and from the GAO.22 When money was no object, the ser-
Biometrics Agency, counted it as a victory. central databasethe Automated Biometrics vices could pay contractors to collect the data,
You can imagine the implications for avoiding Identification System (ABIS). but that didnt build a DOD knowledge base,
potential green-on-blue attacks [of] catching Begun when war funding for Iraq and RAND noted.
the bad guys before they get approved by Afghanistan was flowing freely, the biomet- When theres a match between biomet-
the Afghan army or local police from gaining rics program is on shaky ground now that U.S. rics collected in the field and a person on a
access to our bases, he said during a June troops have left Iraq and will leave Afghani- watch list, the information goes to analysts
biometrics symposium.17 stan after next year. The special pot of money and intelligence units, but often not to the
By 2012, there had been at least 3,000 for overseas contingency operations is dwin- troops who collected the data.
matches on biometrics collected from 1.2 mil- dling as deployed U.S. forces come home. That reduces the enthusiasm of the men
lion non-U.S. persons in Afghanistan, accord- The Army holds the responsibility for and women on the ground doing the collect-
ing to the Government Accountability Office biometrics for all the military services, but au- ing but not hearing the success stories, the
(GAO).18 thority has been split among eight Army or- guys that go into a bad-actor house in the
Including biometrics collected in Iraq ganizations. A lack of collaboration between foothills of Afghanistan and theres 10 bad
beginning in 2002 through December 2011, [Army] components has led to direct failures, guys there that theyve got to enroll before
the U.S. military has created more than 7.1 according to a 2012 report by the RAND Na- mortars start coming in, said Boyd.
million recordssets of facial photographs, tional Defense Research Institute.20 If we prevented a bad actor from enter-
iris scans and 10 fingerprintsfrom 4.5 mil- The program was whipped up quickly ing the country on the information and col-
lion people. Between fiscal years 2007 and in reaction to soldiers urgent needs on the laboration from Afghanistan, we need to feed
2012, DOD invested $3 billion in the program.19 frontlines in Iraq. The Army never made it a that back, Boyd said. If a warfighter sees no
Over the past decade, the U.S. Army program of record, instead purchasing bio- value in what hes doing, eventually hell stop
grew a program that led to the successful metrics collection devices as quick-response doing it.23
targeting of over 850 high-value individu- capabilities. The training that troops do receive pre-
This led to a proliferation of collection deployment and in the field doesnt extend to
equipmentmore than 7,000 of three differ- their leaders, GAO found. Unit commanders
16 Green-on-Blue Attacks in Afghanistan: The ent types of devices have been fielded, much arent taught how to use biometrics effectively,
Data, The Long War Journal, August 23, 2013, of it not interoperable because the Army get their soldiers trained to collect the data or
http://bit.ly/19m3HAV. rapidly purchased whatever companies had find troops who already have those skills. But
17 Biometrics, Analytics and Big Data Symposium, available. even when commanders effectively deploy
Technical Training Corporation, June 18, 2013, Ross- The Army has been forced to respond biometrics in the field, their support cant
lyn, Va. to urgent operational needs from Iraq and overcome lack of knowledge in the top ranks.
18 Government Accountability Office, Defense Bio- Afghanistan, which resulted in tools and tech- We have a lot of people at the major and
metrics: Additional Training for Leaders and More nology being rapidly developed and fielded lieutenant colonel level who get it, who saw
Timely Transmission of Data Could Enhance the Use without adhering to DOD standards, formal biometrics achieve great success in theater.
of Biometrics in Afghanistan (Washington, DC: GAO-
12-442, April 2012), http://1.usa.gov/15Gteoi.
performance measures and operational test- But some of the senior civilians and [senior
ing and evaluation requirements, RAND people] in uniform may not appreciate the
19 Douglas Shontz, Martin C. Libicki, Rena Ru-
davzky and Melissa A. Bradley, An Assessment of
found.21
the Assignments and Arrangements of the Executive
Agent for DoD Biometrics and Status Report on the
20 Ibid., 24. 22 GAO, 2012.
Biometrics Enterprise (RAND National Defense Re-
search Institute, 2012), 70. 21 Ibid., 25. 23 Biometrics, Analytics and Big Data Symposium.
12 PARTNERSHIP FOR PUBLIC SERVICE | IBM CENTER FOR THE BUSINESS OF GOVERNMENT
7.1 MILLION BIOMETRICS FROM 4.5 MILLION PEOPLE ENROLLED
(20022011)
Iraq
Afghanistan
Biometrics Biometrics are also collected
from IEDs.
FINGERPRINTS
In 1909, the long-awaited shipment of 2,000 ally. The Asian longhorned beetle, for exam- CBP officials send their findings to the
Japanese cherry trees arrived at United States ple, arrived from China on a wood shipping quarantine system, where it is incorporated
ports, a gift symbolizing the friendship be- pallet and since 1985 has destroyed more in the risk analysis that guides future inspec-
tween the two nations. The U.S. Department than 80,000 ash, maple and other trees it has tions. They send pests to the APHIS Plant
of Agriculture (USDA) found non-native pests, bored holes into. It has caused $269 million in Protection Quarantine National Inspection
including white peach scale, black thrips, damages in five states, an amount that could Service for final confirmation, information
clearwing moths and San Jose scale in the exceed $41 billion if the beetle spreads to the that also goes into the quarantine system.
trees and had to destroy them.25 The USDA entire country.28 Todd Schroeder, director of business
worked with Japan to treat a new batch of Another pest suspected of hitching a systems at the inspection agency said, The
trees, which arrived pest free in 1912 and con- ride from Asia was discovered in Michigan in whole idea here [is] to protect American
tinue to blossom, drawing tourists from all 2002. The emerald ash borer has killed some agriculture, youve got to be able to make
over the world each spring. 58 million ash trees in more than 10 states, risk-based decisions about imports of [agri-
A hundred years later, more than a bil- causing tens of millions of dollars in damage.29 cultural] products.
lion plant imports come into the country each APHIS helps prevent pest damage using Improved decision-making has in-
year, and USDAs Animal Plant Health Inspec- two databases: the Agricultural Quarantine creased the inspection systems find rates
tion Service (APHIS) Plant Protection and Activity System (AQAS) and Integrated Plant and productivity and changed its approach to
Quarantine Division uses data analytics to Health Information System. The quarantine core mission activities. For instance, by know-
inspect shipments, detect pests and prevent system helps manage shipment data and tar- ing which shipments may have what prob-
them from spreading to U.S. agriculture.26 get inspections; the plant health information lems, APHIS can turn its attention to working
As shipments arrive at U.S. ports, APHIS system manages information from infesta- with offshore partners to prevent pests from
officials oversee inspections by Department tions already in the country. boarding shipments in the first place.
of Homeland Security (DHS) Customs and CBP inspectors feed their findings into APHIS provided 1 million classical swine
Border Protection (CBP) officials. By analyz- the quarantine system and also are guided fever vaccines to Guatemala in March, for ex-
ing data about the type and date of shipment by it. The risk-based system, adopted in 2007, ample, to prevent the disease from spreading.
and country of origin, the inspection service provides data that can be used in probability With trade opportunities increasing each
can target inspections to containers most distributions to aid in determining how many year, the promotion of animal health across
likely to contain pests. of which crates to target, based on known borders is important now more than ever,
If inspectors find pests, APHIS can de- pest or disease outbreaks in other countries, said John Clifford, USDAs chief veterinarian.
stroy the shipment, send it back to the export- weather patterns and other information.30 APHIS recognizes that the prevalence of ani-
ing country or treat it and send it to market. With the quarantine system, the animal mal disease in one country could easily trans-
APHIS officials also set traps around ports to and plant agency can better allocate inspec- pose to another.32
catch pests that escape detection. APHIS tion resources, improving its find rates at U.S. Analytics could improve the inspection
intercepted more than 565,000 insects in ports and its trapping rates. systems pest detection to the point that it
baggage and cargo between 1984 and 2002, The 2,360 CBP officials who use the could invest fewer resources in inspection
though it always misses some.27 Accordingly, quarantine system for inspections at 167 U.S. and turn toward prevention, working with
the agency works with state agricultural de- ports (out of 329) do so under a 2003 memo- food companies and other countries, said
partments and universities to conduct delim- randum of understanding between DHS and Schroeder. You may be making decisions
iting surveys, which determine the scope of USDA. It gives the inspection agency the job that impact different industries in different
existing infestations in the United States. of maintaining the quarantine system and ways, he said. You have less resources to
The Government Accountability Office training CBP inspectors, who check contain- spread across different programs, so lets fo-
estimates that non-native pests cause about ers and collect data.31 cus our resources on those areas [where] we
$136 billion in lost agricultural revenue annu- can have the largest overall impact.
14 PARTNERSHIP FOR PUBLIC SERVICE | IBM CENTER FOR THE BUSINESS OF GOVERNMENT
APHIS also lays traps
Commodities are flagged around the ports of entry
for inspection at port. to serve as a secondary
defense against flying
insects that may have
evaded inspection.
AQAS
Agricultural Quarantine
Activity System
RISK ANALYSIS
Source country
Commodity
Date of shipment
Like most primary care physicians, Stephan only way we would see we were not doing it only if they see it as a help, not a hindrance,
Fihn is always busy, regularly behind and of- well was [high] hospitalization rates and wed says VHA health information technology lead
ten frazzled. have to figure out why, he said. Now were Tami Box. The best and most optimized use
Left to his own devices, when seeing a looking at our patients at high risk of being in remains be seen. We are trying to put tools in
patient whose age and chronic illness might the hospital and saying, What services can the hands of the people to use them and then
qualify him for home health care, Fihn was we provide to keep them out? find out how they use them in the best ways.
lucky if he even remembered that the Vet- Both the scores and the assessment Box and her team take pains to ask little
erans Affairs Departments Puget Sound system draw from the 80 terabytes of patient of first-time users so as not to interrupt their
Healthcare System offered it. statistics and other information collected in delivery of care. The first release of PCAS,
A patient would come in every three VHAs corporate data warehouse, which sucks now in use across VHA hospitals, doesnt
to four months to the clinic, and if I wasnt up data from electronic health records (EHRs) have a whole lot of places where users have
too distracted, if I wasnt too overwhelmingly stored in the Veterans Health Information Sys- to enter data, she adds.
busy, and not too behind on my schedule, I tems and Technology Architecture (VISTA) at
would say maybe this patient should go to VHA medical facilities. VISTA was created 20
home care, Fihn recalled. So whether that years ago by rogue software tinkerers, many
patient got into home care depended whether of them doctors, who saw in early personal
I thought about it, or whether they asked for computers an opportunity to improve how
it, or whether I was familiar with the program. they managed and cared for patients.
It largely bordered on being a random event. The software they designed evolved
But home care keeps patients out of the into VISTA, which VHA offers free online. The
hospital, and can keep them happier, healthier EHR system is one of the most widely used
and alive longer, while saving the VHA money. in the world. The presence of all that data
So the VHA needs its primary care providers has inspired VHA analysts to continue inno-
to get eligible patients enrolled. vating and building tools to use it. The data
Fortunately, Fihn directs VHAs Office warehouse made the assessment scores and
of Analytics and Business Intelligence, posi- system possible, and helped VHA do greater
tioning him to do something to make doctors comparisons of our patients as they move
referrals routine. across our system and as we see differences
First, Fihn helped come up with the care in treatment, said Gail Graham, the VHAs
assessments needs (CAN) score, a predictive deputy undersecretary for health for infor-
analytic tool intended to identify which of matics and analytics, and Fihns boss.
VHAs 6.5 million primary care patients are at She credits him with building the culture
highest risk of hospitalization or death. that makes data users out of VHA staffers 1970s
Now his office is creating an online care
coordination tool, the patient care assess-
at all levels. Fihn is really growing analytics
from a basic level of making sure that there is 1980s
ment system (PCAS), which uses these as- a cadre of people in medical centers and clin-
sessment scores and hundreds of other data ics that know how to use Excel, all the way to
points to help 7,000 VHA medical teams co- offering university-level courses in using the
ordinate 900 to 1,200 patients each. data and advanced analysis. Tinkering doctors start using minicomputers
We have a [registered nurse] care coor- As a result, VHA employees refine and to store patient files. After initial
dinator. Shes got this list in front of her with invent new uses for informatics products. The resistance, the homemade electronic
patients who are at highest risk and she can PCAS, for example, was built for registered health records are adopted VHA-wide.
go down it systematically now and say, Oh I nurses in concert with registered nurses, who
see Mr. So-and-So was in the hospital twice in coordinate care for most teams, says clinical
the last month and hes got a really high risk program manager Joanne Shear. Dr. Fihns
score and were not giving him any of these office heard the hue and cry across the coun-
services. Maybe I will go talk with his primary try: How do we know exactly who we have to
care provider and figure out in our daily hud- manage, how do we identify them? Wheres
dle which of his patients should be referred to the tool that we can interact with the medical
one of these programs, Fihn said. record?
Its really changed the paradigm from System developers fully expect the as-
a reactive one, where maybe in the end the sessment tool will morph as nurses use itbut
16 PARTNERSHIP FOR PUBLIC SERVICE | IBM CENTER FOR THE BUSINESS OF GOVERNMENT
RANKING
1990s
Electronic health records
enable VHA to measure its
quality of health care against
SCORECARD
U.S. standards and improve.
20
1
23 12 2
11
21 19
4
3
10
22 15
5
1994
10 9
22 6 VHA is decentralized to
increase autonomy and
18 16 improve regional care,
7
forming 22 (now 21) Veterans
Integrated Service Networks
(VISNs), comprising VA
20 17 hospitals, health care
centers, ambulatory care
21 8
centers and community-
based outpatient clinics.
2010
2011
A Urgent VHA introduces care
assessment needs (CAN) scores
B Monitor to help care teams prioritize
VHA begins a patient-centered medical home model that organizes primary services to high-risk patients. In
care around interdisciplinary Patient Aligned Care Teams (PACTs), each with a 2012 the CAN score becomes a
C Stable
primary care provider, registered nurse, licensed practical nurse and medical part of the new care scheduling
clerk who share responsibility for managing a group of nearly 1,200 patients. system for the teams.
Most analytics path- the socioeconomic conditions that turn them into food
breakers are strikingly emergencies. To find out, the network draws data from
collaborative. They seek a variety of sources, including meteorology and clima-
out like-minded souls tology, agricultural monitoring and harvest assessment,
who might already have food market and trade analysis, health and nutrition out-
collected data they can comes, livelihoods analysis and food-needs tracking.
adapt to their purposes or From its beginning in 1986, FEWS NET used data
who have developed new from social and physical scientists to improve how US-
methods for prying out or AID targets the most vulnerable populations for food aid.
combining it. For more than 15 years, it has used participating agency
The U.S. Agency for program agreements (PAPA) and participating agency
International Develop- service agreements (PASA) to pay other agencies for peo-
ment (USAID), for ex- ple and information.
ample, specifically created the Famine Early Warning Somebody would say, I understand these guys have
Systems Network (FEWS NET) to be collaborative. From a new way of measuring rainfall, lets go talk with them,
the start, USAID paid other agencies for expertise, data and they would say, Theres this guy over at NASA that
and analysis. The network comprises staff from the Agri- actually is in control of the imagery, so why dont you
culture Department (USDA), scientists and data from the make an arrangement with him, said FEWS NET pro-
U.S. Geological Survey (USGS), analysts and data from the gram manager Gary Eilerts. So we made an arrange-
National Aeronautics and Space Administration (NASA) ment with [NASAs] Goddard Space Flight Center to get
and the National Oceanic and Atmospheric Administra- imagery then we had [NOAA] giving us better rainfall
tion (NOAA), contractor decision support and in-country estimates and USGS trying to help us put it together in
staff, as well as alliances with a host of humanitarian aid spatial perspective.
and government agencies. Eilerts, a USDA employee who works at USAID man-
Most famines are caused by environmental shocks, aging the famine network, draws up and oversees agree-
such as droughts. USAID knew plenty about food assis- ments with NASA, NOAA, USGS and USDA. The agree-
tance, but far less about the causes of those shocks and ments permit USAID to pay other agencies for inherently
18 PARTNERSHIP FOR PUBLIC SERVICE | IBM CENTER FOR THE BUSINESS OF GOVERNMENT
governmental services and facilities and data products, but also by get- containers to inspect based on data
that are particularly or uniquely ting feedback on those products. about known pest or disease out-
suitable for technical assistance, are The networks in-country staffers breaks in other countries, weather
not competitive with private en- compare satellite data with what patterns and other information that
terprise, and can be made available they see on the ground, for example, increases the likelihood that a ship-
without interfering unduly with do- a quality check whose importance ment is infested.
mestic programs.33 cannot be overstated in the devel- In 2003, DHS and USDA signed
The famine network has a pro- opment and maturation of remote a memorandum of agreement as-
gram agreement with the geologi- sensing, wrote NASA biospheric re- signing about 3,000 APHIS inspec-
cal survey for six scientists and two search scientist Molly Brown.34 tors to CBP. Under the agreement,
managers, for example, while the Every year, the network holds APHIS trains the inspectors.35 The
NASA agreement covers satellite- a science day to keep its social and service also relies on state and local
collected data products on vegeta- physical scientists aware of each governments, academics and alert
tion, rainfall and other underlying others activities, discoveries and citizens to report unusual instances
environmental causes of food inse- capabilities, says USGSs Jim Ver- of damage to plants or other indica-
curity and famine, particularly cli- din. We bring each other up to date tions that bugs or diseases escaped
mate change. through a series of presentations, the ports.
he said. It keeps us abreast of what CDC also turns to others for
our colleagues are doing, but it also data, relying on 87 public-health
Dont Reinvent
informs the social science food ana- laboratories across the country, at
the Wheel
lysts as well. Learning about ad- least one per state, to provide sam-
NASA is building a version of its land vances in physical science data col- ples to compare against the database
information system (LIS) for the lection and analysis lets FEWS NET of more than 500,000 bacterial DNA
famine network. Its Goddard Space social scientists understand what fingerprints in PulseNet. CDC pro-
Flight Center developed the soft- new or more sophisticated ques- vides grants to the labs to buy the
ware framework to improve assess- tions they can get answered. equipment and to pay some of the
ment of ground conditions. It lets The meetings also uncover new employees who use it to prepare
scientists run more than one model data that leads to better analysis. For samples.
of a portion of the Earths surface at example, Verdin discovered at a re- CDC scientists run the DNA im-
a time to come up with precise pre- cent meeting that NASA had funded ages captured from samples taken
dictions of soil moisture and other one of the geological surveys uni- from suspected victims of foodborne
interactions between the atmo- versity partners to review past satel- illness against the database to de-
sphere and the land. The land sys- lite images to determine the amount tect multistate outbreaks. Scientists
tem is itself a collaboration: The Air of water accumulated seasonally in analyze matches of more than two
Force Weather Agency (AFWA) and the United States and Central Asia. similar patterns, which are consid-
Army Corps of Engineers helped pay Its going to give us the opportunity ered possible outbreaks. Confirmed
for NASA to develop it. to go back and compare the model in results trigger investigations by local
When I provide an investment real-time and retrospective analysis. public health department epidemi-
into the LIS system, thats because Some analytics-driven agencies ologists to find the source.
I need a certain capability added to rely on others not just to share, but In addition, Food and Drug Ad-
it, said John Eylander, a co-founder also to collect their data. For exam- ministration and USDA laboratories
of the collaboration between NASA ple, the Agriculture Departments have begun using PulseNet to track
and the weather agency. Everybody Animal Plant Health Inspection data on pathogens to catch illness-
basically said, Lets not reinvent Service (APHIS) populates its Agri- causing bacteria before people get
the wheel. Lets go and collaborate cultural Quarantine Activity System sick. USDA and FDA also have the
with NASA and benefit from that (AQAS) with data collected by agri- authority to recall products or regu-
partnership. cultural import inspectors working late meatpacking plants and farms
The science agencies in the for the DHSs Customs and Border as a result of PulseNets and public
famine network benefit not only by Protection (CBP) bureau. health officials findings. The wealth
receiving funds for remote sensing The quarantine system, adopted of information and collaboration
in 2007, helps CBP determine which
33 Famine Early Warning Systems Network 35 2003 Memorandum of Agreement be-
U.S. Geological Survey Participating Agency 34 Brown, Famine Early Warning Systems tween DHS and USDA, http://1.usa.gov/
Program Agreement. and Remote Sensing Data. GzzeZN.
Analytics pioneers shared and added to one anothers data and expertise in a variety of ways:
Most often, they used legal authorities to buy data and the experts it, helping both agencies meet their mission goals.
and software to analyze it.
NASA has created research and development programs whose
Some used the government-wide provisions for interagency funding is contingent on recipients promoting their products to
acquisitions under the 1932 Economy Act. other agencies that can apply them and might invest in developing
them further.
Others relied on agency-specific authority, such as the partici-
pating agency program and service agreements provided for CDC used grant money to help public health labs acquire the equip-
under the 1961 Foreign Assistance Act, which created USAID ment they use to process DNA samples for matching against the
and permits it to use other agencies resources when they are PulseNet database.
uniquely suitable for technical assistance in education, health,
housing or agriculture. Annual science days give FEWS NET collaborators insights into
each others work on famine, preventing duplication and augment-
Another form of interagency agreement, a memorandum of under- ing other projects across participating agencies.
standing, enabled CBP to collect data and USDAs APHIS to analyze
20 PARTNERSHIP FOR PUBLIC SERVICE | IBM CENTER FOR THE BUSINESS OF GOVERNMENT
Develop data
to demonstrate
return on investment
22 PARTNERSHIP FOR PUBLIC SERVICE | IBM CENTER FOR THE BUSINESS OF GOVERNMENT
overstated the savings in some cases and projected program costs.
while understating them in others, Preventive analytics programs
and failed to report some costs, the also might consider adopting ROI
IG found. estimation methods similar to those
CMS concurred with the IG and proposed by John Whitley in his
is taking corrective action. Since 2012 study of federal law enforce-
CMSs fraud protection system is the ment agencies. Whitley offers sev-
largest predictive analytics-based eral approaches for gathering and
program of its kind in government, analyzing data to help determine the
the IGs recommendations offer effect of federal programs on fraud,
practices other predictive analytics tax evasion, drug smuggling or other
users should consider.43 unobserved activities.
For example, he suggested that He recommends inventorying
CMS require its contractors to track applicable data available within the
the amounts of money recovered as agency or from others, and looking
a result of leads generated by the for ways to estimate program effects
analytics system. CMS also should by combining it or enhancing its col-
coordinate with law enforcement lection. Surveys and audits can help
agencies to improve the reporting of evaluate how much to invest in dif-
outcomes of investigations and pros- ferent programs, including analytics.
ecutions stemming from leads gen- Data-driven inspection programs
erated by the system. He noted that might experiment with secondary
because CMS used savingsactual screenings to infer their success or
and projectedto calculate ROI, it failure.44
should also have reported all actual
Long-term data users often had no ROI measures when they began, but developed and adapted them as their projects evolved.
Most initially reported improved mission outputs and outcomes, To demonstrate ROI, mission analytics programs learned to devote
for example, the CDCs increasing numbers of foodborne illness resources to develop data to track financial and other results re-
outbreaks identified and DODs numbers of biometrics matched lated in whole or in part to analytics.
with the subjects on the high-value-target watch list.
Predictive analytics programs still are refining their cost-bene-
Increasingly, however, they are called upon to deliver cost-bene- fit metrics and findings and must take care in estimating costs
fit and return-on-investment metrics in monetary terms so agen- avoided, for example, making certain they report all actual and
cy leaders can compare program costs to determine whether projected costs.
data-based efforts are more or less cost effective than alternate
strategies. To improve their estimates of return on investment, analytics pro-
grams can employ surveys and audits, use experimental methods
such as secondary screening, and increase and enhance the data
they collect.
45 U.S. Institute for Peace, American Interests and U.N. Reform: A Re-
port of the Task Force on the United Nations (Washington, DC, June
2005), 125, http://bit.ly/17kxZYa.
24 PARTNERSHIP FOR PUBLIC SERVICE | IBM CENTER FOR THE BUSINESS OF GOVERNMENT
their too-slow response to famine
in Somalia in 2011. By the time they
ics keeps facts front and center and
difficult to ignore even when they Providing
took noticeafter FEWS NET en-
abled the first-ever real-time dec-
point to unwelcome conclusions. In
Somalia, the famine network was objective
laration of famine in July of that
yeartens of thousands of people
able to perform market analyses
that led to delivery of monetary aid data in cases
already had died. The final tally was
258,000 deaths4.6 percent of the
for food purchases even when it was
almost impossible to send in food of political
populationdue to the food crisis,
10 percent of them children under
itself.
contention
age five.46
FEWS NET had issued 17 in-
Use Data to Drive
Home Hard Truths
can help
creasingly urgent warnings begin-
ning in August 2010, but to many aid A few months later, another fam- focus debate
ine appeared to be brewing in the
officials, dire conditions looked little
different than normal for Somalia.47 Sahelthe continent-wide region on choices
between the Sahara Desert to the
Whats more, the hardest-hit areas
were controlled by al Shabaab, a north and the Sudanian savanna between
to the southparticularly in Niger.
group the United States had labeled
terrorists, and its fighters were kill- This time FEWS NET, almost alone, fact-based
said the problem was not acute but
ing aid workers. FEWS NET hadnt
been able to overcome aid agencies chronic, and advised against mas- alternatives.
crisis fatigue and caution. sive humanitarian assistance.
When we made the declaration Everybody was embarrassed
of famine we did have discussions by having missed the Somalia issue,
about whether this was going to sink Eilerts said. Then, all of a sudden,
us, Eilerts recalled. When we tried within a month or two, the Sahel
to convince people, we thought it starts heating up and people say,
was a fairly difficult case and people Aha, were not going to get caught
were really not on board. Then, on again. Were going to start moving
July 11, when we made the declara- and acting and doing everything we
tion, by that time everybody was can. And then here we are saying,
starting to see people walking out Wait a minute guys, no, you dont
of Somalia; they were dying on the need to.
road. Everybody kind of fell in line. It was just hard for them to ra-
As John Whitley has pointed out, tionalize. Wait a minute, we just had
providing objective data in cases of a famine over here and we missed
political contention can help focus it, and heres one now and it looks
debate on choices between fact- like its been here forever and its the
based alternatives.48 Persistence in same characteristics, and now you
presenting data derived from analyt- dont want us to respond?
In the end, it was analytics that
46 London School of Hygiene and Tropical saved FEWS NET. The data proved
and the Johns Hopkins University Bloomberg out on the ground.
School of Public Health, Mortality Among We had a lot of data. I went out
Populations of Southern and Central Somalia on the road with a slide show that I
During 2010-2012 (Rome, Washington, DC,
May 2, 2013) 8, http://uni.cf/16DwQNK. showed 50 times. I said, Heres how
47 Mija-Tesse Ververs, East Africa Food Se- this year stacks up against the con-
curity Crisis An Overview of What We Knew text, said Eilerts. What happened
and When Before June 2011 (Assessment Ca- was that a lot of the governments out
pacities Project, Geneva, 12 July 2011), Annex there agreed with us a lot of the
1, http://bit.ly/1eUPmDJ.
people in the government of Niger
48 Whitley, Five Methods for Measuring
Unobserved Events: A Case Study of Federal even. They said, You know, this is
Law Enforcement, 10. not a famine.
Data programs with long track records found they had to deliver analysis leaders could use and support.
The absence of a powerful sponsor can hobble an analytical ef- Program managers learned to use leaner, punchier and more visual
fort even when it shows mission achievements, as DOD biometrics methods for presenting their findings so senior officials could ab-
backers have discoveredespecially now, when programs vie for sorb them and get the main points quickly.
funding as budgets are cut.
Programs that grew from the grassroots, such as VHAs VISTA, sur-
Mature programs struggled when delivering analytics-based mes- vived resistance by demonstrating their effectiveness in terms of
sages leaders didnt want to hear, but made headway when they broad user adoption.
persisted in presenting the supporting data.
26 PARTNERSHIP FOR PUBLIC SERVICE | IBM CENTER FOR THE BUSINESS OF GOVERNMENT
To encourage data use
and spark insights, enable
employees to easily see,
combine and analyze it
In mature data programs, example, [Show me] the imports in a yearly cycle so we
non-managers often ap- can be prepared on what we should be inspecting domes-
plied analytics spontane- tically as a result of possible pests coming into the coun-
ously, even before leaders try, because were importing more guavas from wherever
received training on driv- during the months February and March.
ing data use, so user inno- From such demands grew an analytics program
vation continues to be wel- that directs inspectors to target the import containers
comed. most likely to be carrying pests and lets state and local
If there are not lead- governments, academics and citizens help catch those
ers at the top that want [data-based] information, youre that evade inspection by matching them with images on
not going to be able to let go of the reins enough to allow APHISs online directory.
the brilliant thinkers at the grassroots to take these tools We roll a lot of these tools out to the people who are
and provide the types of outputs or analytic products that responding to customers on a daily basis. Its the insights
would change your organization, said Todd Schroeder of and the products they put together to support local needs
the Department of Agricultures Animal Plant Health In- that provided the insights needed at higher levels to raise
spection Service (APHIS). new questions, said Schroeder.
At APHIS, the idea of analytics took off long be- User insights also include direct human interven-
fore we knew it, in limited locations where larger tion when no algorithm can do the job. As yet, nothing
amounts of data were being collected, he said. It really beats the eye when it comes to verifying fingerprints,
started with [business intelligence software]. To save whether of bacterial DNA or human beings.
money, the agency introduced the same software across Some very bright person said the best way to detect
organizations. images is the eyeball-a-metric method, said the CDCs
So domestic pest surveyors were using it to manage Williams. Weve also developed some algorithms to help
insect traps, while import inspectors were using it to supplement the eyeball method. Nobodys been able to
track how many of which types of commodities they had design one thats actually better than the human eye.
checked and with what results. People started seeing Similarly, the Department of Defense (DOD) uses
the linkages, Schroeder said. people rather than formulas to inspect a portion of the
They also began asking more of the tool, he said. For fingerprints that must be analyzed. About 9.5 percent
28 PARTNERSHIP FOR PUBLIC SERVICE | IBM CENTER FOR THE BUSINESS OF GOVERNMENT
is improving agency effectiveness, said U.S. Geological Surveys Verdin
not punishment. [Use of analyt- of the Famine Early Warning Sys-
ics changes] daily work because tems Network. Some scientists are
now theyre getting more direction content to do their research, publish
on where to go, what to do, said their papers, move on to the next
APHISs Schroeder. The risk, he questions. There also are people
added, is that this gives insights who like to see their findings make
into who is performing well and their way into practice and have an
who isnt, and leaves people feeling impact on decision-making pro-
vulnerable. cesses. And Im one of them.
If you can tell these [insect Others were engaged by the op-
trap] surveyors where to go and portunity to tackle the worlds knot-
what to do, what do you need all of tiest problems.
these other layers of managers for? As a weather guy, Im very con-
Some people Im sure are thinking cerned about making sure I provide
that, he added. He also pointed out the best information possible, said
that analytics can cause changes in Eylander. I dont have all the an-
how agencies do business, thereby swers myself, but theyre very, very
altering jobs. difficult problems. Weather is such
Analytics is disruptive innova- a difficult thing that if you try to do
tion at its finest, Schroeder said. it all yourself youd just never get
Maybe it changes the dynamic of there. So the only option you have is
how you work . Maybe you need to share and collaborate.
less inspection because you have
more information, he said.
Analytics is disruptive innovation
Capitalize on the
Power of the Mission at its finest.
For early analytics adopters, the
power of improved delivery of their
inspiring missions overcame fear of
change. Even usually dispassionate
scientists were energized when us-
ing analytics meant saving lives.
Among us are people who ap-
preciate the idea that the work they
do makes a difference in programs
that touch a large number of people,
Moved beyond using data exclusively to measure or compare em- Made sure those who collected data also benefited directly from it
ployee and organizational performance by providing tools that en- or clearly understood how it improved mission delivery.
able staff to combine, analyze and use data when, where and how
they needed it to speed and ease the work process. Capitalized on employees zeal for the agencys mission to help
them overcome reluctance about adopting analytics.
Were guided by users insights, implementing good ideas from the
grassroots and recognizing those who suggested them. Were honest about the potential for analytics to change agency
operations and the jobs of those performing them.
Refined analytics tools by watching how employees used them to
greatest effect, but without disrupting work flow.
30 PARTNERSHIP FOR PUBLIC SERVICE | IBM CENTER FOR THE BUSINESS OF GOVERNMENT
To that end, Fihns office of Because the biometrics initiative is Medicaid Services (CMS) fraud pre-
analytics and business intelligence not yet an official program of record, vention system built in an analytics
at VHA is teaching analytics skills, biometrics classes are not offered laboratory from the beginning, said
from the basic, such as making sure in Army schoolhouses. As a result, Gent. We brought in 10 statisticians,
each medical center and clinic has pre-deployment training is ad hoc economists, programmers who re-
a cadre of adept Excel users, to the and troops dont always collect data ally understand the art of predic-
expert, such as offering university- properly. For example, they have tive analytics. In program integrity,
level courses on advanced analysis. misidentified fingerprints from the things evolve and youre forever
Fihn and his business intelligence left hand as the right and vice versa.54 learning new information and new
team members also offer to speak Analytics training enables em- schemes that need to be identified
and present webinars about clinical ployees to ask better questions of and targeted.
applications of analytics53 as part of data, scrutinize it more effectively The Animal Plant Health In-
the VA Information Research Cen- for patterns and linkages and offers spection Service (APHIS) now is
ters extensive online, printed and opportunities to improve operations, building a business analytics compe-
live educational resources. collect data more efficiently, become tency center in the hope of helping
At the Centers for Disease Con- comfortable using it and incorpo- units across the agency adopt data
trol and Prevention (CDC), analysts rate it in more aspects of their work. programs. The idea is that this in-
have to pass an external course in Data wont be pervasively used until formation and the use of those ana-
order to learn how to identify and analytics is standard operating pro- lytical products will span all of those
analyze bacterial DNA data. It gen- cedure; that cant happen until em- functional areas, Schroeder said.
erally takes at least six months of ployees adapt to using data and ana- Schroeder also has led efforts to
training to develop the expertise to lytics as part of everyday operations. standardize data across APHIS so it
independently classify, said CDC can be linked more easily by com-
branch chief Peter Gerner-Smidt. modity, geography, type of pest and
Create Centers of
They also must pass an external business entity. We have to have
Excellence to Spread
quality check. It takes a long time commonalities in our taxonomy of in-
Adoption
to get to this proficiency level, he formation to really analyze big sets of
added. Long-standing data programs, such data to show us the patterns, themes
Failure to properly train em- as VHAs, have corporate data ware- we might need to be more risk-based,
ployees can undermine both data houses and specialized analytics to make better decisions sooner, to
collection and adoption of analyt- organizations. Others are creating have larger impacts, he said.
ics, as biometrics leaders at the such centers, some after years of
Defense Department have learned. data-driven success.
The Centers for Medicare and
53 Tamra Box and Stephen Fihn, Care As-
sessment Need (CAN) Score and the Patient 54 Shontz et al., An Assessment of the As-
Care Assessment System (PCAS): Tools for signments and Arrangements of the Executive
Care Management, June 27, 2013, http://1.usa. Agent for DoD Biometrics and Status Report on
gov/18H54fu. the Biometrics Enterprise, 34.
Instilling analytics in all agency activities became a goal once early programs demonstrated gains. Its an ongoing process involving:
Standardization of data to enable users to look across collections funds, staff, space and other resources.
by time, entity, geography, source and other attributes to find link-
ages and patterns and to share information. Centers of excellence with expertise in data analytics, the organi-
zations operations and policies. CMS, for example, houses policy
Formal and on-the-job education. experts along with statisticians in its analytics laboratory. Policy
people provide expertise on what is appropriate to bill to Medicare
Training thats appropriate to the organization and the employees so the fraud prevention system can be trained to identify what isnt.
position; for example, VHA dieticians learn to analyze dietetic data.
Data evangelists who encourage use of data-driven techniques and
Teaching leaders to base their decisions on data, so they, in turn, tools beyond their own units across organizations.
require employees to muster analytics to support their cases for
32 PARTNERSHIP FOR PUBLIC SERVICE | IBM CENTER FOR THE BUSINESS OF GOVERNMENT
Conclusion
From Data to Decisions III: Lessons from Analytics Pioneers is Association of Public Health Laboratories
the third in a series of reports examining federal agencies that Kristy Kubota
have been successful in applying data analytics solutions to fur- Senior Specialist, PulseNet Program
ther their missions. In the first report, we highlighted promising
practices in using analytics to save money, improve services and Kirsten Larson
more effectively achieve agency goals. In the second report, we Senior Specialist, Food Safety Program
focused on outlining the steps necessary to begin to use data
analytics and how an agency can begin to foster an analytics Cablespeed
culture. In this report, we examine how data analytics pioneers
Wayne Wheeles
began their efforts, how these efforts matured over time, how
Analytic, Infrastructure and Enrichment Developer, Red
their ROI is defined and how it helps drive program success.
Alpha
To accomplish this, we spoke with more than 30 experts
from more than 15 agencies, agency subcomponents, offices
and private organizations. We conducted more than 50 in- Centers for Disease Control and Prevention
depth phone and in-person interviews between April and Au- John Besser
gust 2013. We attended four local conferences and several on- Deputy Chief, Enteric Diseases Laboratory Branch,
line webinars on big data or data analytics. We also conducted Division of Foodborne, Waterborne, and Environmental
an extensive literature review examining recent laws, OMB cir- Diseases
culars and memorandums, private sector best practices, news
Peter Gerner-Smidt
articles and other publicly available agency documentation.
Chief, Enteric Diseases Laboratory Branch
Alvin Shultz
Program Coordinator, Epidemiology and Laboratory
Capacity for Infectious Diseases, Division of Preparedness
and Emerging Infections, National Center for Emerging
and Zoonotic Infectious Diseases
Ian Williams, PhD
Chief, Outbreak Response and Prevention Branch
CTOVision
Robert Gourley
Chief Technology Officer, Crucial Point, LLP
Department of Defense
Department of the Army
John Boyd
Director, Defense Biometrics and Forensics Office of the
Assistant Secretary of Defense, Research and Engineering
John Eylander
Cold Regions Research and Engineering Lab, U.S. Army
Engineer Research and Development Center
34 PARTNERSHIP FOR PUBLIC SERVICE | IBM CENTER FOR THE BUSINESS OF GOVERNMENT
Department of Defense RAND Corporation
Department of the Air Force Gregory F. Treverton
Jeff Cetola Director, Center for Global Risk and Security
Chief, Environmental Characterization Flight
16th Weather Squadron (16WS/WXE), 2nd Weather United States Agency for International Development
Group (Air Force Weather Agency)
Cara Christie
Chris Franks Team Lead, East and Central Africa, Office of U.S. Foreign
Land Information System Team Lead, Air Force Weather Disaster Assistance
Agency
Gary Eilerts
Program Manager, FEWS NET
Department of Health and Human Services
Centers for Medicare and Medicaid Services
United States Geological Survey
Kelly Gent
Director, Data Analytics and Control Group, Center for James P. Verdin
Program Integrity Physical Scientist, Manager of USGS FEWS NET Activities,
Earth Resources Observation and Science
Executive Office of the President
Office of Science and Technology Policy Department of Veterans Affairs
George Strawn Veterans Health Administration
Director, National Coordination Office for Networking and Jack Bates
Information Technology Research and Development Director, Business Intelligence Product Line, Office of
Wendy Wigen Information and Technology
Technical Coordinator, National coordination Office for Tamra L. Box
Networking and Technology Research and Development Lead, Health IT, Clinical Assessment, Reporting and
Tracking Program
National Aeronautics and Space Administration National Project Manager, Patient Care Assessment
Goddard Space Flight Center System (PCAS)
Office of Informatics and Analytics, Office of Analytics
Dr. Christa D. Peters-Lidard and Business Intelligence
Physical Scientist, Hydrological Sciences Laboratory
Stephan Fihn
Director, Analytics and Business Intelligence
Office of Management and Budget
Gail Graham
Shelley Metzenbaum Assistant Deputy Undersecretary for Health for
Former Associate Director of Performance and Personnel Informatics and Analytics
Management
President, The Volcker Alliance Joanne M. Shear
Primary Care Clinical Program Manager, Offices of PC
Operations (10NC3) and PC Services/Policy (10P4F)
United States Department of Agriculture
Animal and Plant Health Inspection Service
Todd Schroeder
Director, Business Systems Management
Lee Spaulding
Cognos Administrator, Plant Protection and Quarantine
Judy England-Joseph
Cynthia Heckman
Bevin Johnston
Stephanie Mabrey
Seth Melling
Ellen Perlman
Audrey Pfund
Lara Shane
Max Stier
Melissa Wiak
IBM
Daniel J. Chenok
Executive Director, IBM Center for the Business of Government
Gregory Greben
Vice President, Public Sector Business Analytics and Optimization
John Kamensky
Senior Fellow, IBM Center for the Business of Government
Brian D. Murrow
Associate Partner, Business Analytics and Optimization
36 PARTNERSHIP FOR PUBLIC SERVICE | IBM CENTER FOR THE BUSINESS OF GOVERNMENT
1100 New York Avenue nw 600 14th Street NW
Suite 200 East Second Floor
Washington DC 20005 Washington DC 20005