Difference between revisions of "Project Nightingale"

Revision as of 23:52, 9 February 2022

Project Nightingale

"Partnerships Logo" text Type Healthcare Data Transformation Launch Date 2018 Status Active Product Line Artificial Intelligence Platform Online Website Google Website

Project Nightingale is a Google initiative that is meant to collect and analyze millions of patients’ personal-health information in the United States.^[1] Google is one of the largest technology companies in the world that is most known for its internet search engine, however, it continues to look to expand upon its capabilities. The project sits under the Google Cloud division, which is run by Thomas Kurian; Project Nightingale started in 2018, and has only become bigger as time has progressed.

With Project Nightingale, Google works with Ascension, the second largest healthcare system in the country that has 2,600 hospitals, to obtain access to over 50 million patients’ medical records.^[2] As Ascension is handing over data, in return they receive any new services that Google creates from the data. Ascension’s goal of working with Google is to improve upon its patient care system to centralize its electronic record-keeping processes by moving its data to the cloud; there is a rapid growth in the volume of healthcare information as many healthcare professionals are overwhelmed within the United States healthcare system.^[3] Google’s end goal is to be able to create tools (using the data and information) to identify and predict health problems before a patient visits a doctor, which helps with decision making for patient treatment.

Project Nightingale has become an increasingly debated topic as lawmakers are seeking to gain more insight into the operations and ethics of this Google project to see whether it is driving medical innovation or prioritizing Google’s profits; Google has a large advertisement division, which makes its money from knowing as much information on consumers as possible. Large numbers of people are worried about the privacy of their health data as it gets into the hands of one of the largest technology companies in the world. In addition, many people of color are concerned about the algorithms being generated from this project that could systematically discriminate against them.

The Rise and Fall of Google Health

How Google Health Started

Google was started on the foundation of providing information to consumers in convenient ways, in which the technology firm has gained a major interest in using this within the healthcare industry. Project Nightingale is not the first healthcare venture that has started in Google.

Google Health started its development in 2006 as an online personal health records service that allowed people to enter, update and edit their health and wellness information.^[1] One of Google’s main goals was to reduce medical errors and cut the costs of healthcare in the United States.^[4] As there is a ton of patient data, data management contributes to rising healthcare administrative costs.^[5] Google continued to develop this product for two years before live testing with 1,600 patients at the Cleveland Clinic. From there, the Google Health Beta Edition was released to the public in 2008.^[5] The product allowed a user to create an account and enter their personal medical records (manually or automatically from hospitals, laboratories, and pharmacies that were Google Health partners).^[6] During this time, there were fears over the safety and security of users’ stored health information; Google released a statement saying “no personal or medical information in your Google Health profile is used to customize your Google.com search results or used for advertising."^[6]

Due to its lack of adoption and scalability, the electronic personal health records (PHR) platform closed its operations in 2012. Patients did not always receive dependable information as many users did not take the time needed to update the details of their healthcare records.^[7] Google Health did not have strong relationships with big labs, so users were not able to access their test results; the key was to build strong partnerships with providers to be able to acquire patients.^[7] This is something Google learned as it began to scale Project Nightingale.

Other Google Health Projects

In 2019, Google acquired wearable fitness device company Fitbit for $2.1 Billion giving it access to user data such as daily steps, heart rate, calories burned, and location. Many people in the healthcare industry are worried about the amount of health data that sits in the hands of Google as it acquires new companies and information.

Google also uses “emergent medical data."^[8] Emergent medical data is health data analyzed by artifical intelligence from normal consumer behavior. Anytime a user interacts with technology, they leave behind a digital footprint of their behavior that can serve as raw materials for companies like Google to mine for emergent medical data. Other companies mine for emergent medical data as well; a study found that Facebook analyzed health records and social media posts of almost 1,000 Facebook users to discover that religious language such as “God” or “Lord” were strong indicators of diabetes.^[8]

In 2019, Google patented a smart home device that mined emergent medical data from occupants’ behavior to conclude whether they are developing Alzheimer’s disease or substance use disorders.^[8] While Project Nightingale takes on a massive amount of data, outsiders believe it is a piece to a larger puzzle that will help Google increase their personalization tactics.

Launching Project Nightingale

In 2018, the company attempted to re-launch its separate health division to consolidate the company’s health projects which spanned across web services, software, and AI-backed products. These projects consisted of topics such as Google Brain (focused on deep learning), Google Fit (focused on wearable fitness devices), and Project Nightingale.^[4] However, it closed down its operations in 2021 as the company felt it was best to split up the combined healthcare strategy into other segments. These projects have continued to be worked on, just within the research division of Google as Project Nightingale is within Google Cloud.^[4]

What Project Nightingale Uses and its Potential Benefits

Before partnering with Google, Ascension used 40 data centers in more than a dozen states. Project Nightingale uses lab results, doctor diagnoses, hospitalization records, and lots of other medical information to analyze patient data; the amount of data compares to a complete health history, such as patient names, family member addresses, and dates of birth.^[5] This data is used with artificial intelligence and machine learning to create strategies that will improve patients’ treatment. These algorithms can help predict diseases a patient might have before a doctor has even diagnosed them; clinicians are able to make data-driven decisions that can save their patients’ lives.

A pathology report is a document that contains the diagnosis determined by examining cells and tissues under a microscope; pathologists provide one of the most significant sources of diagnostic data for providers within the world of care delivery.^[9] “Seventy percent of all decisions in healthcare are based on a pathology result and between 70-75% of all data in an electronic health record are from a pathology result” states Jeffrey Golden, MD, the Chair of the Department of Pathology at Brigham & Women’s Hospital.^[9] Analytics on the data from artificial intelligence can drill down “to the pixel level” on extremely large digital images, which allow healthcare professionals to identify variations that the human eye cannot catch.^[9] Golden states that these analytics can help providers get more accurate diagnosis, and the sooner the correct diagnosis, the higher likelihood of a better outcome for the patient.^[9] It is tools such as this that Project Nightingale is capable of producing that will become extremely beneficial to hospitals.

Project Nightingale is keeping up with the healthcare industry trends as artificial intelligence and automation is one of the biggest trends in the sector.^[10] Real-time analytics are being used to expedite care as other organizations are also using up-to-date information to drive the treatment process. Healthcare workers have become frustrated with the manual aspect of paperwork and find it extremely helpful to have processes automated.

Ethics of Project Nightingale

Misuse of Data Concerns

When the Wall Street Journal broke the news about the secret operation of Project Nightingale, doctors nor patients were not formally notified of the partnership between Google and Ascension.^[11] It was this secret aspect that worried most groups of people. In 2019, The Office for Civil Rights in the Department of Health and Human Services sought to learn more information about Project Nightingale’s data collection to ensure that HIPAA protections were fully implemented. Under the HIPAA Privacy Rule, hospitals are allowed to share data with business partners without telling patients as long as the information is used ‘only to help the covered entity carry out its health-care functions.' ^[11] Experts state that Project Nightingale appears to be permissible under federal law as a spokeswoman stated the Ascension data wouldn’t be used to sell ads. ^[11] However Kirsten Ostherr, a co-founder and the director of the Medical Futures Lab at Rice University, states that “it is widely agreed that HIPAA is out of date."^[12] It also noted that searching for a dermatologist or stores that sell Plan B is not protected under HIPAA; these results are data that Google already tracks alongside users’ IP address and location.^[12] Google already has access to millions of patient records through relationships with plenty of healthcare partners, such as the Mayo Clinic, the University of Chicago, and the Cleveland Clinic. What makes this project with Ascension different is the magnitude and scale of it.^[8] United States Senators Amy Klobuchar and Lisa Murkowski proposed a law, the Protecting Personal Health Data Act, which addresses health privacy concerns by requiring the Secretary of Health and Human Services to publicize regulations for new health technologies such as health apps and wearable devices that are not regulated by existing laws.^[13]

Ellen Wright Clayton, a professor of biomedical ethics at Vanderbilt University is worried that Project Nightingale will use the health data to perform independent research outside the direct scope of treatment for the patients.^[11] An anonymous whistleblower from the Google side came out and stated that those on the project team did not necessarily know which artificial intelligence algorithms were being used as the data was being transferred from the hospital groups to Google. ^[14] When the report of the Project first came out, it had been claimed that over 100 Google employees had been able to access the data of patients tied to the Project ^[2]. While Google states it will not use the Ascension data outside the scope of the project, there is a major difference between the data itself and the knowledge Google can get access to from analyzing the data; this gives Google the opportunity to apply this knowledge to other aspects while technically still playing within the rules.^[8] Those that oppose Project Nightingale speculate that Google will take the health data and identify new indicators of health that can be used to infer consumers’ medical conditions outside of the health care system.^[8] From here, Google could use its Google Ads business to target different users; in 2019, the company expected $120 billion of its revenue to come from advertising (making up ~83% of its parent company’s total revenue).^[15]

Security of the Data Concerns

With all data privacy issues, patients and health care providers are worried about the protection of their information. In 2018, Google was the victim of a major data privacy offense in which five million Google+API users had their private data exposed. As the company stores massive amounts of personalized information, it leaves itself extremely vulnerable for more attacks. In 2021, one of the biggest cloud computing challenges for IT service providers was security of data.^[16] There are several situations that can stem from stolen health data: finding details about specific diseases & illnesses of a patient and long-term identity theft.^[17] For example, a hacker can use information about one’s private sexually transmitted disease to coerce them into paying a lump sum of cash or doing what they tell them to do.^[17] With long-term identity theft, unlike bank information, the health information of one does not change, so once the hacker has it, they have it forever. Brandon Reagin, a victim of stolen health data, states that his hacker had multiple medical procedures completed, which amounted to nearly $20,000 in hospital bills; this is something he has dealt with for over 15 years.^[17] As it recognizes these concerns, Google has learned from the past and continues to strengthen its security protocols through its $10 billion investment in advancing cybersecurity over the next five years; it plans on expanding zero-trust programs, helping secure the software supply chain, and enhancing open-source security.^[18]

Discrimination Concerns

As Google builds out algorithms in Project Nightingale, a concern is that they may discriminate against various communities of people. A term for this is “algorithmic bias,” which is defined as the application of an algorithm that compounds existing inequities in socioeconomic status, race, ethinic background, religion, gender, disability, or sexual orientation and amplifies inequities in health systems.^[15] In the field of genetics, it is estimated that Caucasians make up almost 80% of collected data, and thus studies may be more applicable for that group than for other underrepresented groups.^[15] While this number does not translate directly over to the Ascension data, some argue that the distributions are not drastically different.

In 2019, a study published in Science, concluded that an algorithm widely used in US hospitals was less likely to refer black people than white people who were equally sick to programs that aim to improve care for patients with complex medical records.^[19] For researchers a study like this is important because it is extremely hard to gain access to a large set of health data to run full tests; there have only been smaller studies conducted.^[19] Scientists theorize that the reasoning for the lack of access to care is due to the effects of systemic racism, which ranges from doubt of the health-care system to direct racial discrimination by health-care providers.^[19] This study shines a light on a major problem that worries patients who have their medical information tied with the Ascension data. While Google has not consciously produced discriminatory algorithms, this 2019 study shows there is a likelihood of inequality in the data resulting in unjust practices from the algorithms. As Project Nightingale is working with the largest healthcare provider in the country, these algorithms could have major unintended effects if not addressed.

Competition

Google has a stronghold on the healthcare data segment. On May 26th, 2021, Google and HCA Healthcare, a national hospital chain, announced they are teaming up in a data sharing partnership.^[20] This partnership will allow Google to gain more access to a host of patient records with real-time medical information as it already has a lot of knowledge with emergent medical data and Fitbit data ^[20]. Accessing medical records has given Google a major leg-up against its competitors who are trying to replicate the technology company’s work. Facebook does not use health data to make predictions, it uses user-generated content; with this lack of access, the accuracy of its health predictions are extremely variable ^[20].

Microsoft

Microsoft had its own personal health record system called, “HealthVault” that helped users store health data and fitness information in a centralized location. What happened with the original Google Health platform, held true for HealthVault as well; it did not have enough adoption since it did not help users understand and address their health concerns and problems. What differentiates Google is the fact that now it has patient-acquired data, which can help it create innovative services and products. ^[21]

IBM

Other companies have or are starting to follow suit and use healthcare to create artificial intelligence tools and services. In 2016, IBM bought Truven Health Analytics for $2.6 billion for their healthcare data of tens of millions of medical records and years of health insurance data. ^[22] This acquisition of Truven will help IBM find health cost savings, improve healthcare outcomes, fight fraud and make operations more efficient ^[22].

Small Company Partnerships

As there are many people with concerns over the ethicalness of tech giants (like Google, Facebook, Microsoft, and IBM), small companies have arisen to work with patient-acquired data. The Indiana University School of Medicine has partnered with a small Indianapolis-based company, LifeOmic, to create a cloud-based service that uses artificial intelligence to deliver insights to patients, providers, and clinical researchers ^[21].

CloudMedx

There are also companies whose sole business model is to create predictive analyses for healthcare providers using the medical data. CloudMedx is a startup in California that uses disparate healthcare data and converts into actionable treatment for patients. ^[23] The founder of the company advocates for uses of data like this as his father almost died as a result of a misdiagnosis; he believes it is technology like CloudMedx (and Project Nightingale) that can shape the future of the healthcare industry and save more and more patients.

Looking Ahead

As the world continues to increase its technological capabilities, humans’ lives will only continue to digitize more. When this is occurring, there are different perspectives on the extent to which this digitalization influences and controls lives. As with Project Nightingale and all of the Google Health initiatives, it comes down to the protection and use of data and whether it will benefit society, a concern that many researchers, scientists, and citizens have for these types of projects.

In a statement last year, a Google spokesperson stated that “Google deeply believes in the power of technology to improve health and wellness and we have increased our health investments across the company. Today, health is growing, company-wide effort and the Google Health name will continue and encompass our projects that share the common purpose to improve global health outcomes." ^[24]

References

1. ↑ ^{1.0 1.1} https://www.wsj.com/articles/google-s-secret-project-nightingale-gathers-personal-health-data-on-millions-of-americans-11573496790/
2. ↑ ^{2.0 2.1} https://www.proquest.com/docview/2370268359?parentSessionId=RsaEYDrrnNfdICB%2BRmNouaDI7%2Fj0qBHzaZnxVJsvXQg%3D&pq-origsite=primo&accountid=14667
3. ↑ https://www-sciencedirect-com.proxy.lib.umich.edu/science/article/pii/S221188371400032X
4. ↑ ^{4.0 4.1 4.2} https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2405844/
5. ↑ ^{5.0 5.1 5.2} https://www-sciencedirect-com.proxy.lib.umich.edu/science/article/pii/S221188371400032X
6. ↑ ^{6.0 6.1} https://www.advisory.com/daily-briefing/2021/08/24/google-health
7. ↑ ^{7.0 7.1} https://www.recordnations.com/2015/11/3-reasons-why-google-health-failed/#:~:text=Lack%20of%20Provider%20Relationships,often%20get%20patients%20using%20PHRs.
8. ↑ ^{8.0 8.1 8.2 8.3 8.4 8.5} https://slate.com/technology/2019/11/google-ascension-project-nightingale-emergent-medical-data.html
9. ↑ ^{9.0 9.1 9.2 9.3} https://healthitanalytics.com/news/top-12-ways-artificial-intelligence-will-impact-healthcare
10. ↑ https://trustees.aha.org/top-10-emerging-trends-health-care-2021-new-normal
11. ↑ ^{11.0 11.1 11.2 11.3} https://bi-gale-com.proxy.lib.umich.edu/essentials/article/GALE%7CA616992563/70a0b51115d05f6514b1584a626adac6?u=umuser
12. ↑ ^{12.0 12.1} https://www.theatlantic.com/technology/archive/2019/11/google-project-nightingale-all-your-health-data/601999/
13. ↑ Cite error: Invalid <ref> tag; no text was provided for refs named name
14. ↑ https://www.theguardian.com/commentisfree/2019/nov/14/im-the-google-whistleblower-the-medical-data-of-millions-of-americans-is-at-risk
15. ↑ ^{15.0 15.1 15.2} https://www.hsph.harvard.edu/ecpe/how-to-prevent-algorithmic-bias-in-health-care/
16. ↑ https://www.mindinventory.com/blog/cloud-computing-challenges/
17. ↑ ^{17.0 17.1 17.2} https://healthtechmagazine.net/article/2019/10/what-happens-stolen-healthcare-data-perfcon
18. ↑ https://blog.google/technology/safety-security/why-were-committing-10-billion-to-advance-cybersecurity/
19. ↑ ^{19.0 19.1 19.2} https://www.nature.com/articles/d41586-019-03228-6
20. ↑ ^{20.0 20.1 20.2} https://venturebeat.com/2021/06/20/googles-health-care-data-sharing-partnership-is-a-problem/
21. ↑ ^{21.0 21.1} https://hitconsultant.net/2019/04/19/what-the-failure-of-microsofts-healthvault-means-for-the-future-of-ehrs/
22. ↑ ^{22.0 22.1} https://www.reuters.com/article/us-truven-m-a-ibm/ibm-to-acquire-truven-health-analytics-for-2-6-billion-idUSKCN0VR1SS
23. ↑ https://cloudmedxhealth.com/about/
24. ↑ https://www.fiercehealthcare.com/tech/google-dissolved-its-unified-health-division-what-s-next-for-its-health-tech-strategy