The Urban Institute Data Catalog


Data@Urban: “We believe that data make the biggest impact when they are accessible to everyone.

Today, we are excited to announce the public launch of the Urban Institute Data Catalog, a place to discover, learn about, and download open data provided by Urban Institute researchers and data scientists. You can find data that reflect the breadth of Urban’s expertise — health, education, the workforce, nonprofits, local government finances, and so much more.

Built using open source technology, the catalog holds valuable data and metadata that Urban Institute staff have created, enhanced, cleaned, or otherwise added value to as part of our work. And it will provide, for the first time, a central, searchable resource to find many of Urban’s published open data assets.

We hope that researchers, data analysts, civic tech actors, application developers, and many others will use this tool to enhance their work, save time, and generate insights that elevate the policy debate. As Urban produces data for research, analysis, and data visualization, and as new data are released, we will continue to update the catalog.

We’re thrilled to put the power of data in your hands to better understand and respond to many critical issues facing us locally and nationally. If you have comments about the tool or the data it contains, or if you would like to share examples of how you are using these data, please feel free to contact us at [email protected].

Here are some current highlights of the Urban Data Catalog — both the data and research products we’ve built using the data — as of this writing:

– LODES data: The Longitudinal Employer-Household Dynamics Origin-Destination Employment Statistics (LODES) from the US Census Bureau provide detailed information on workers and jobs by census block. We have summarized these large, dispersed data into a set of census tract and census place datasets to make them easier to use. For more information, read our earlier Data@Urban blog post.

– Medicaid opioid data: Our Medicaid Spending and Prescriptions for the Treatment of Opioid Use Disorder and Opioid Overdose dataset is sourced from state drug utilization data and provides breakdowns by state, year, quarter, drug type, and brand name or generic drug status. For more information and to view our data visualization using the data, see the complete project page.

– Nonprofit and foundation data: Members of Urban’s National Center for Charitable Statistics (NCCS) compile, clean, and standardize data from the Internal Revenue Service (IRS) on organizations filing IRS forms 990 or 990-EZ, including private charities, foundations, and other tax-exempt organizations. To read more about these data, see our previous blog posts on redesigning our Nonprofit Sector in Brief Report in R and repurposing our open code and data to create your own custom summary tables….(More)”.

Nudging the Nudger: Toward a Choice Architecture for Regulators


Working Paper by Susan E. Dudley and Zhoudan Xie: “Behavioral research has shown that individuals do not always behave in ways that match textbook definitions of rationality. Recognizing that “bounded rationality” also occurs in the regulatory process and building on public choice insights that focus on how institutional incentives affect behavior, this article explores the interaction between the institutions in which regulators operate and their cognitive biases. It attempts to understand the extent to which the “choice architecture” regulators face reinforces or counteracts predictable cognitive biases. Just as behavioral insights are increasingly used to design choice architecture to frame individual decisions in ways that encourage welfare-enhancing choices, consciously designing the institutions that influence regulators’ policy decisions with behavioral insights in mind could lead to more public-welfare-enhancing policies. The article concludes with some modest ideas for improving regulators’ choice architecture and suggestions for further research….(More)”.

Index: Secondary Uses of Personal Data


By Alexandra Shaw, Andrew Zahuranec, Andrew Young, Stefaan Verhulst

The Living Library Index–inspired by the Harper’s Index–provides important statistics and highlights global trends in governance innovation. This installment focuses on public perceptions regarding secondary uses of personal data (or the re-use of data initially collected for a different purpose). It provides a summary of societal perspectives toward personal data usage, sharing, and control. It is not meant to be comprehensive–rather, it intends to illustrate conflicting, and often confusing, attitudes toward the re-use of personal data. 

Please share any additional, illustrative statistics on data, or other issues at the nexus of technology and governance, with us at [email protected]

Data ownership and control 

  • Percentage of Americans who say it is “very important” they control information collected about them: 74% – 2016
  • Americans who think that today’s privacy laws are not good enough at protecting people’s privacy online: 68% – 2016
  • Americans who say they have “a lot” of control over how companies collect and use their information: 9% – 2015
  • In a survey of 507 online shoppers, the number of respondents who indicated they don’t want brands tracking their location: 62% – 2015
  • In a survey of 507 online shoppers, the amount who “prefer offers that are targeted to where they are and what they are doing:” 60% – 2015 
  • Number of surveyed American consumers willing to provide data to corporations under the following conditions: 
    • “Data about my social concerns to better connect me with non-profit organizations that advance those causes:” 19% – 2018
    • “Data about my DNA to help me uncover any hereditary illnesses:” 21% – 2018
    • “Data about my interests and hobbies to receive relevant information and offers from online sellers:” 32% – 2018
    • “Data about my location to help me find the fastest route to my destination:” 40% – 2018
    • “My email address to receive exclusive offers from my favorite brands:”  56% – 2018  

Consumer Attitudes 

  • Academic study participants willing to donate personal data to research if it could lead to public good: 60% – 2014
  • Academic study participants willing to share personal data for research purposes in the interest of public good: 25% – 2014
  • Percentage who expect companies to “treat [them] like an individual, not as a member of some segment like ‘millennials’ or ‘suburban mothers:’” 74% – 2018 
    • Percentage who believe that brands should understand a “consumer’s individual situation (e.g. marital status, age, location, etc.)” when they’re being marketed to: 70% – 2018 Number who are “more annoyed” by companies now compared to 5 years ago: 40% – 2018Percentage worried their data is shared across companies without their permission: 88% – 2018Amount worried about a brand’s ability to track their behavior while on the brand’s website, app, or neither: 75% – 2018 
  • Consumers globally who expect brands to anticipate needs before they arise: 33%  – 2018 
  • Surveyed residents of the United Kingdom who identify as:
    • “Data pragmatists” willing to share personal data “under the right circumstances:” 58% – 2017
    • “Fundamentalists,” who would not share personal data for better services: 24% – 2017
    • Respondents who think data sharing is part of participating in the modern economy: 62% – 2018
    • Respondents who believe that data sharing benefits enterprises more than consumers: 75% – 2018
    • People who want more control over their data that enterprises collect: 84% – 2018
    • Percentage “unconcerned” about personal data protection: 18% – 2018
  • Percentage of Americans who think that government should do more to regulate large technology companies: 55% – 2018
  • Registered American voters who trust broadband companies with personal data “a great deal” or “a fair amount”: 43% – 2017
  • Americans who report experiencing a major data breach: 64% – 2017
  • Number of Americans who believe that their personal data is less secure than it was 5 years ago: 49% – 2019
  • Amount of surveyed American citizens who consider trust in a company an important factor for sharing data: 54% – 2018

Convenience

Microsoft’s 2015 Consumer Data Value Exchange Report attempts to understand consumer attitudes on the exchange of personal data across the global markets of Australia, Brazil, Canada, Colombia, Egypt, Germany, Kenya, Mexico, Nigeria, Spain, South Africa, United Kingdom and the United States. From their survey of 16,500 users, they find:

  • The most popular incentives for sharing data are: 
    • Cash rewards: 64% – 2015
    • Significant discounts: 49% – 2015
    • Streamlined processes: 29% – 2015
    • New ideas: 28% – 2015
  • Respondents who would prefer to see more ads to get new services: 34% – 2015
  • Respondents willing to share search terms for a service that enabled fewer steps to get things done: 70% – 2015 
  • Respondents willing to share activity data for such an improvement: 82% – 2015
  • Respondents willing to share their gender for “a service that inspires something new based on others like them:” 79% – 2015

A 2015 Pew Research Center survey presented Americans with several data-sharing scenarios related to convenience. Participants could respond: “acceptable,” “it depends,” or “not acceptable” to the following scenarios: 

  • Share health information to get access to personal health records and arrange appointments more easily:
    • Acceptable: 52% – 2015
    • It depends: 20% – 2015
    • Not acceptable: 26% – 2015
  • Share data for discounted auto insurance rates: 
    • Acceptable: 37% – 2015
    • It depends: 16% – 2015
    • Not acceptable: 45% – 2015
  • Share data for free social media services: 
    • Acceptable: 33% – 2015
    • It depends: 15% – 2015
    • Not acceptable: 51% – 2015
  • Share data on smart thermostats for cheaper energy bills: 
    • Acceptable: 33% – 2015
    • It depends: 15% – 2015
    • Not acceptable: 51% – 2015

Other Studies

  • Surveyed banking and insurance customers who would exchange personal data for:
    • Targeted auto insurance premiums: 64% – 2019
    • Better life insurance premiums for healthy lifestyle choices: 52% – 2019 
  • Surveyed banking and insurance customers willing to share data specifically related to income, location and lifestyle habits to: 
    • Secure faster loan approvals: 81.3% – 2019
    • Lower the chances of injury or loss: 79.7% – 2019 
    • Receive discounts on non-insurance products or services: 74.6% – 2019
    • Receive text alerts related to banking account activity: 59.8% – 2019 
    • Get saving advice based on spending patterns: 56.6% – 2019
  • In a survey of over 7,000 members of the public around the globe, respondents indicated:
    • They thought “smartphone and tablet apps used for navigation, chat, and news that can access your contacts, photos, and browsing history” is “creepy;” 16% – 2016
    • Emailing a friend about a trip to Paris and receiving advertisements for hotels, restaurants and excursions in Paris is “creepy:” 32% – 2016
    • A free fitness-tracking device that monitors your well-being and sends a monthly report to you and your employer is “creepy:” 45% – 2016
    • A telematics device that allows emergency services to track your vehicle is “creepy:” 78% – 2016
  • The number of British residents who do not want to work with virtual agents of any kind: 48% – 2017
  • Americans who disagree that “if companies give me a discount, it is a fair exchange for them to collect information about me without my knowing”: 91% – 2015

Data Brokers, Intermediaries, and Third Parties 

  • Americans who consider it acceptable for a grocery store to offer a free loyalty card in exchange for selling their shopping data to third parties: 47% – 2016
  • Number of people who know that “searches, site visits and purchases” are reviewed without consent:  55% – 2015
  • The number of people in 1991 who wanted companies to ask them for permission first before collecting their personal information and selling that data to intermediaries: 93% – 1991
    • Number of Americans who “would be very concerned if the company at which their data were stored sold it to another party:” 90% – 2008
    • Percentage of Americans who think it’s unacceptable for their grocery store to share their shopping data with third parties in exchange for a free loyalty card: 32% – 2016
  • Percentage of Americans who think that government needs to do more to regulate advertisers: 64% – 2016
    • Number of Americans who “want to have control over what marketers can learn about” them online: 84% – 2015
    • Percentage of Americans who think they have no power over marketers to figure out what they’re learning about them: 58% – 2015
  • Registered American voters who are “somewhat uncomfortable” or “very uncomfortable” with companies like Internet service providers or websites using personal data to recommend stories, articles, or videos:  56% – 2017
  • Registered American voters who are “somewhat uncomfortable” or “very uncomfortable” with companies like Internet service providers or websites selling their personal information to third parties for advertising purposes: 64% – 2017

Personal Health Data

The Robert Wood Johnson Foundation’s 2014 Health Data Exploration Project Report analyzes attitudes about personal health data (PHD). PHD is self-tracking data related to health that is traceable through wearable devices and sensors. The three major stakeholder groups involved in using PHD for public good are users, companies that track the users’ data, and researchers. 

  • Overall Respondents:
    • Percentage who believe anonymity is “very” or “extremely” important: 67% – 2014
    • Percentage who “probably would” or “definitely would” share their personal data with researchers: 78% – 2014
    • Percentage who believe that they own—or should own—all the data about them, even when it is indirectly collected: 54% – 2014
    • Percentage who think they share or ought to share ownership with the company: 30% – 2014
    • Percentage who think companies alone own or should own all the data about them: 4% – 2014
    • Percentage for whom data ownership “is not something I care about”: 13% – 2014
    • Percentage who indicated they wanted to own their data: 75% – 2014 
    • Percentage who would share data only if “privacy were assured:” 68% – 2014
    • People who would supply data regardless of privacy or compensation: 27% – 2014
      • Percentage of participants who mentioned privacy, anonymity, or confidentiality when asked under what conditions they would share their data:  63% – 2014
      • Percentage who would be “more” or “much more” likely to share data for compensation: 56% – 2014
      • Percentage who indicated compensation would make no difference: 38% – 2014
      • Amount opposed to commercial  or profit-making use of their data: 13% – 2014
    • Percentage of people who would only share personal health data with a guarantee of:
      • Privacy: 57% – 2014
      • Anonymization: 90% – 2014
  • Surveyed Researchers: 
    • Percentage who agree or strongly agree that self-tracking data would help provide more insights in their research: 89% – 2014
    • Percentage who say PHD could answer questions that other data sources could not: 95% – 2014
    • Percentage who have used public datasets: 57% – 2014
    • Percentage who have paid for data for research: 19% – 2014
    • Percentage who have used self-tracking data before for research purposes: 46% – 2014
    • Percentage who have worked with application, device, or social media companies: 23% – 2014
    • Percentage who “somewhat disagree” or “strongly disagree” there are barriers that cannot be overcome to using self-tracking data in their research: 82% – 2014 

SOURCES: 

“2019 Accenture Global Financial Services Consumer Study: Discover the Patterns in Personality”, Accenture, 2019. 

“Americans’ Views About Data Collection and Security”, Pew Research Center, 2015. 

“Data Donation: Sharing Personal Data for Public Good?”, ResearchGate, 2014.

Data privacy: What the consumer really thinks,” Acxiom, 2018.

“Exclusive: Public wants Big Tech regulated”, Axios, 2018.

Consumer data value exchange,” Microsoft, 2015.

Crossing the Line: Staying on the right side of consumer privacy,” KPMG International Cooperative, 2016.

“How do you feel about the government sharing our personal data? – livechat”, The Guardian, 2017. 

“Personal data for public good: using health information in medical research”, The Academy of Medical Sciences, 2006. 

“Personal Data for the Public Good: New Opportunities to Enrich Understanding of Individual and Population Health”, Robert Wood Johnson Foundation, Health Data Exploration Project, Calit2, UC Irvine and UC San Diego, 2014. 

“Pew Internet and American Life Project: Cloud Computing Raises Privacy Concerns”, Pew Research Center, 2008. 

“Poll: Little Trust That Tech Giants Will Keep Personal Data Private”, Morning Consult & Politico, 2017. 

“Privacy and Information Sharing”, Pew Research Center, 2016. 

“Privacy, Data and the Consumer: What US Thinks About Sharing Data”, MarTech Advisor, 2018. 

“Public Opinion on Privacy”, Electronic Privacy Information Center, 2019. 

“Selligent Marketing Cloud Study Finds Consumer Expectations and Marketer Challenges are Rising in Tandem”, Selligent Marketing Cloud, 2018. 

The Data-Sharing Disconnect: The Impact of Context, Consumer Trust, and Relevance in Retail Marketing,” Boxever, 2015. 

Microsoft Research reveals understanding gap in the brand-consumer data exchange,” Microsoft Research, 2015.

“Survey: 58% will share personal data under the right circumstances”, Marketing Land: Third Door Media, 2019. 

“The state of privacy in post-Snowden America”, Pew Research Center, 2016. 

The Tradeoff Fallacy: How Marketers Are Misrepresenting American Consumers And Opening Them Up to Exploitation”, University of Pennsylvania, 2015.

Andrew Yang proposes that your digital data be considered personal property


Michael Grothaus at Fast Company: “2020 Democratic presidential candidate Andrew Yang may not be at the top of the race when it comes to polling (Politico currently has him ranked as the 7th most-popular Democratic contender), but his policies, including support for universal basic income, have made him popular among a subset of young, liberal-leaning, tech-savvy voters. Yang’s latest proposal, too, is sure to strike a chord with them.

The presidential candidate published his latest policy proposal today: to treat data as a property right. Announcing the proposal on his website, Yang lamented how our data is collected, used, and abused by companies, often with little awareness or consent from us. “This needs to stop,” Yang says. “Data generated by each individual needs to be owned by them, with certain rights conveyed that will allow them to know how it’s used and protect it.”

The rights Yang is proposing:

  • The right to be informed as to what data will be collected, and how it will be used
  • The right to opt out of data collection or sharing
  • The right to be told if a website has data on you, and what that data is
  • The right to be forgotten; to have all data related to you deleted upon request
  • The right to be informed if ownership of your data changes hands
  • The right to be informed of any data breaches including your information in a timely manner
  • The right to download all data in a standardized format to port to another platform…(More)”.

Tracking the Labor Market with “Big Data”


Tomaz Cajner, Leland Crane, Ryan Decker, Adrian Hamins-Puertolas, and Christopher Kurz at FEDSNotes: “Payroll employment growth is one of the most reliable business cycle indicators. Each postwar recession in the United States has been characterized by a year-on-year drop in payroll employment as measured by the BLS Current Employment Statistics (CES) survey, and, outside of these recessionary declines, the year-on-year payroll employment growth has always been positive. Thus, it is not surprising that policymakers, financial markets, and the general public pay a great deal of attention to the CES payroll employment gains reported at the beginning of each month.

However, while the CES survey is one of the most carefully conducted measures of labor market activity and uses an extremely large sample, it is still subject to significant sampling error and nonsampling errors. For example, when the BLS first reported that private nonfarm payroll gains were 148,000 in July 2019, the associated 90 percent confidence interval was +/- 100,000 due to sampling error alone….

One such source of alternative labor market data is the payroll-processing company ADP, which covers 20 percent of the private workforce. These are the data that underlie ADP’s monthly National Employment Report (NER), which forecasts BLS payroll employment changes by using a combination of ADP-derived data and other publicly available data. In our research, we explore the information content of the ADP microdata alone by producing an estimate of employment changes independent from the BLS payroll series as well as from other data sources.

A potential concern when using the ADP data is that only the firms which hire ADP to manage their payrolls will appear in the data, and this may introduce sample selection issues….(More)”

Citizens’ voices for better health and social policies


Olivia Biermann et al at PLOS Blog Speaking of Medicine: “Citizen engagement is important to make health and social policies more inclusive and equitable, and to contribute to learning and responsive health and social systems. It is also valuable in understanding the complexities of the social structure and how to adequately respond to them with policies. By engaging citizens, we ensure that their tacit knowledge feeds into the policy-making process. What citizens know can be valuable in identifying feasible policy options, understanding contextual factors, and putting policies into practice. In addition, the benefit of citizen engagement extends much beyond improving health policy-making processes by making them more participatory and inclusive; being engaged in policy-making processes can build patients’ capacity and empower them to speak up for their own and their families’ health and social needs, and to hold policy-makers accountable. Moreover, apart from being involved in their own care, citizen-patients can contribute to quality improvement, research and education.

Most studies on citizen engagement to date originate from high-income countries. The engagement methods used are not necessarily applicable in low- and middle-income countries, and even the political support, the culture of engagement and established citizen engagement processes might be different. Still, published processes of engaging citizens can be helpful in identifying key components across different settings, e.g. in terms of levels of engagement, communication channels and methods of recruitment. Contextualizing the modes of engagement between and within countries is a must.

Examples of citizen engagement

There are many examples of ad hoc citizen engagement initiatives at local, national and international levels. Participedia, a repository of public participation initiatives around the globe, showcases that the field of citizen engagement is extremely vibrant.  In the United Kingdom, the Citizens’ Council of the National Institute for Health and Clinical Excellence (NICE) provides NICE with a public perspective on overarching moral and ethical issues that NICE has to take into account when producing guidance. In the United States of America, the National Issues Forum supports the implementation of deliberative forums on pressing national policy issues. Yet, there are few examples that have long-standing programs of engagement and that engage citizens in evidence-informed policymaking.

A pioneer in engaging citizens in health policy-making processes is the McMaster Health Forum in Hamilton, Canada. The citizens who are invited to engage in a “citizen panel” first receive a pre-circulated, plain-language briefing document to spark deliberation about a pressing health and social-system issue. During the panels, citizens then discuss the problem and its causes, options to address it and implementation considerations. The values that they believe should underpin action to address the issue are captured in a panel summary which is used to inform a policy dialogue on the same topic, also organized by the McMaster Health Forum….(More)”.

We Need a PBS for Social Media


Mark Coatney at the New York Times: “Social media is an opportunity wrapped in a problem. YouTube spreads propaganda and is toxic to children. Twitter spreads propaganda and is toxic to racial relationsFacebook spreads propaganda and is toxic to democracy itself.

Such problems aren’t surprising when you consider that all these companies operate on the same basic model: Create a product that maximizes the attention you can command from a person, collect as much data as you can about that person, and sell it.

Proposed solutions like breaking up companies and imposing regulation have been met with resistance: The platforms, understandably, worry that their profits might be reduced from staggering to merely amazing. And this may not be the best course of action anyway.

What if the problem is something that can’t be solved by existing for-profit media platforms? Maybe the answer to fixing social media isn’t trying to change companies with business models built around products that hijack our attention, and instead work to create a less toxic alternative.

Nonprofit public media is part of the answer. More than 50 years ago, President Lyndon Johnson signed the Public Broadcasting Act, committing federal funds to create public television and radio that would “be responsive to the interests of people.”

It isn’t a big leap to expand “public media” to include not just television and radio but also social media. In 2019, the definition of “media” is considerably larger than it was in 1967. Commentary on Twitter, memes on Instagram and performances on TikTok are all as much a part of the media landscape today as newspapers and television news.

Public media came out of a recognition that the broadcasting spectrum is a finite resource. TV broadcasters given licenses to use the spectrum were expected to provide programming like news and educational shows in return. But that was not enough. To make sure that some of that finite resource would always be used in the public interest, Congress established public media.

Today, the limited resource isn’t the spectrum — it’s our attention….(More)”.

Mobility Data Sharing: Challenges and Policy Recommendations


Paper by Mollie D’Agostino, Paige Pellaton, and Austin Brown: “Dynamic and responsive transportation systems are a core pillar of equitable and sustainable communities. Achieving such systems requires comprehensive mobility data, or data that reports the movement of individuals and vehicles. Such data enable planners and policymakers to make informed decisions and enable researchers to model the effects of various transportation solutions. However, collecting mobility data also raises concerns about privacy and proprietary interests.

This issue paper provides an overview of the top needs and challenges surrounding mobility data sharing and presents four relevant policy strategies: (1) Foster voluntary agreement among mobility providers for a set of standardized data specifications; (2) Develop clear data-sharing requirements designed for transportation network companies and other mobility providers; (3) Establish publicly held big-data repositories, managed by third parties, to securely hold mobility data and provide structured access by states, cities, and researchers; (4) Leverage innovative land-use and transportation-planning tools….(More)”.

New York Report Studies Risks, Rewards of the Smart City


GovTech: “The New York state comptroller tasked his staff with analyzing the deployment of new technologies at the municipal level while cautioning local leaders and the public about cyberthreats.

New York Comptroller Thomas DiNapoli announced the reportSmart Solutions Across the State: Advanced Technology in Local Governments, during a press conference last week in Schenectady, which was featured in the 25-page document for its deployment of an advanced streetlight network.

“New technologies are reshaping how local government services are delivered,” DiNapoli said during the announcement. “Local officials are stepping up to meet the evolving expectations of residents who want their interactions with government to be easy and convenient.”

The report showcases online bill payment for people to resolve parking tickets, utilities and property taxes; bike-share programs using mobile apps to access bicycles in downtown areas; public Wi-Fi through partnerships with telecommunication companies; and more….The modernization of communities across New York could create possibilities for partnerships between municipalities, counties and the state, she said. The report details how a city might attempt to emulate some of the projects included. Martinez said local government leaders should collaborate and share best practices if they decide to innovate their jurisdictions in similar ways….(More)”.

A fairer way forward for AI in health care


Linda Nordling at Nature: “When data scientists in Chicago, Illinois, set out to test whether a machine-learning algorithm could predict how long people would stay in hospital, they thought that they were doing everyone a favour. Keeping people in hospital is expensive, and if managers knew which patients were most likely to be eligible for discharge, they could move them to the top of doctors’ priority lists to avoid unnecessary delays. It would be a win–win situation: the hospital would save money and people could leave as soon as possible.

Starting their work at the end of 2017, the scientists trained their algorithm on patient data from the University of Chicago academic hospital system. Taking data from the previous three years, they crunched the numbers to see what combination of factors best predicted length of stay. At first they only looked at clinical data. But when they expanded their analysis to other patient information, they discovered that one of the best predictors for length of stay was the person’s postal code. This was puzzling. What did the duration of a person’s stay in hospital have to do with where they lived?

As the researchers dug deeper, they became increasingly concerned. The postal codes that correlated to longer hospital stays were in poor and predominantly African American neighbourhoods. People from these areas stayed in hospitals longer than did those from more affluent, predominantly white areas. The reason for this disparity evaded the team. Perhaps people from the poorer areas were admitted with more severe conditions. Or perhaps they were less likely to be prescribed the drugs they needed.

The finding threw up an ethical conundrum. If optimizing hospital resources was the sole aim of their programme, people’s postal codes would clearly be a powerful predictor for length of hospital stay. But using them would, in practice, divert hospital resources away from poor, black people towards wealthy white people, exacerbating existing biases in the system.

“The initial goal was efficiency, which in isolation is a worthy goal,” says Marshall Chin, who studies health-care ethics at University of Chicago Medicine and was one of the scientists who worked on the project. But fairness is also important, he says, and this was not explicitly considered in the algorithm’s design….(More)”.