Information & Ethics Committee on Feb. 17th, 2022

[Dr. Kamran Khan]

Alex, did you want to take this one?

Mr. Chair, neither of us is the technical leader. We have our head of technology. Everything that we have is managed in highly secure cloud environments. We have full levels of encryption. We work with independent third parties to enhance our data security practices. We have administrative and operational procedures with how all our data are managed.

Alex, I don't know if there's anything else you wanted to add.

[Alex Demarsh]

We would have to check the geographic locations of the physical servers that our cloud providers use in our case.

It would be Canada or the U.S., and we'd confer with our head of technology to be sure. I'd be happy to report back with a definitive answer.

[Jeremy Patzer, Conservative, Cypress Hills—Grasslands, SK]

The reason I'm wondering is that the context of this study is de-identified data. We've seen other reports that between 90% and 95% of individuals who had their data de-identified can be re-identified.

The reason I'm asking is that quite often data is only as secure as the person who's trying to find it. When there are multiple avenues, that's what I'm trying to get at here.

I guess this would be a broad, industry-based question, and we see breaches of security within the government from time to time. We see it in the private sector from time to time, and quite often, actually. Industry-wide, what risks are there to this data being taken by a nefarious character?

[Alex Demarsh]

Mr. Chair, we've spoken to our general cybersecurity and data security practices. As the honourable member will be aware, innovative Canadian start-ups need to be quite conscious of data privacy, as our innovations are globally sought after for a number of reasons. It's top of mind. We use industry best practices, select secure cloud environments, internal auditing, access control processes and multifactor authentication. These are the normal suite of cybersecurity practices required by our type of business.

[Jeremy Patzer, Conservative, Cypress Hills—Grasslands, SK]

I've one more question here for you guys.

The federal response plan in April 2021, under the heading of surveillance, states that:

COVID-19 surveillance is a pan-Canadian initiative...numerous data [systems] including existing surveillance systems with novel, non-traditional data sources.

It sounds like a lot of things going on. It's a very vague, very broad definition for surveillance.

Is there any issue that a definition like that might be too all-encompassing, too broad, and not narrowly focused enough for the framework of what you guys are doing, which is providing specifically for infectious disease?

[Alex Demarsh]

If it's helpful to clarify, in public health, surveillance is used as a catch-all term for infectious disease case data or other disease case data.

I'm not certain which document he's referring to, but that is a general term well understood within public health collection of data about individual cases in the context of an issue of public health importance.

[The Chair, Conservative, Pat Kelly]

I'm afraid that's time.

For the final round, we have Ms. Khalid for five minutes.

[Iqra Khalid, Liberal, Mississauga—Erin Mills, ON]

Thank you very much, Mr. Chair, and thank you to the witnesses.

Perhaps I'll start with Dr. Khan.

Dr. Khan, how helpful do you think this collection of mobility data was in shaping a good response to the pandemic in Canada?

[Dr. Kamran Khan]

I think it was actually quite helpful. As I was describing earlier, the counterfactual, or how this would have played out in the absence of these types of data, would have been trying to catch up after we saw that cases were surging in a particular area. This is a leading indicator by being a proxy of contacts that are forthcoming.

I will say that in our anecdotal assessments of the past two years, we've seen many instances of the analytics that we have generated and provided to the Public Health Agency being precursors of subsequent surges or providing really important actionable insights.

What I think, and I say this as a scientist, is that to answer that question appropriately and fully would require a full retrospective after the pandemic is over to understand what worked and what didn't. Just as in any other instance, there's an opportunity to learn from this.

The simple answer is that I think a lot of lives were positively impacted and benefited from this type of work. This is notwithstanding the importance of privacy, which we're discussing here today, but I do firmly believe it had a very strong impact on protecting lives and livelihoods across the country.

[Iqra Khalid, Liberal, Mississauga—Erin Mills, ON]

Thank you.

Dr. Khan, what would be the impact if PHAC stopped using this type of de-aggregated, anonymized data to inform health policy in Canada?

[Dr. Kamran Khan]

I will start. Alex, feel free if you would like to add your thoughts.

The metaphor I might use is that it provides some timely insights in a rapidly evolving outbreak. The metaphor of not having that information in those types of insights is a bit like fighting an outbreak but with a bit of a blindfold on. For example, if there's a public health intervention on social distancing, you may not even know if it's working, and if it is, whether it's having the intended effect and where, when and how to adapt.

I think in many ways, if I'm just looking at it purely from a public health perspective, lack of these insights could really compromise the public health community's ability to respond to this type of threat.

Alex, is there anything you would like to add?

[Alex Demarsh]

No, I think that's right. It would be removing a tool that's useful in the context of us, as a public health community, not having many tools. Beyond that, I'm not sure. I think Dr. Khan gave a comprehensive answer.

[Iqra Khalid, Liberal, Mississauga—Erin Mills, ON]

Thanks very much for that.

Are you satisfied with your company's policies and protocols around the protection of privacy? Are you satisfied with the government's protocols, including in the RFP in the contract and the role and importance that privacy played within that?

[Dr. Kamran Khan]

Let me start out with our own privacy practices. We are always interested in striving to continuously improve and enhance in areas where we can.

I will say I am quite comfortable and proud of all of the work we have done to ensure privacy at BlueDot, starting with only working with data that has entirely been de-identified from providers that are following the privacy laws and regulations. We're being very thoughtful on how we generate it out to share with the Public Health Agency.

I do want to note that we had been preparing for this years before the pandemic emerged. That gave us the opportunity not only to develop the technical capability but also to have rigorous discussions and develop policies and standard operating procedures on how we would work with these types of data in a crisis. I think we have been very thoughtful throughout the entire process.

I'm probably not well suited to speak on behalf of the Public Health Agency and the RFP.

[The Chair, Conservative, Pat Kelly]

With that, I'm going to bring this panel to a close. We're just a minute over time right now, but I do want to thank our witnesses.

I will suspend, and we will, hopefully, turn this around as quickly as we can for panel two.

[The Chair, Conservative, Pat Kelly]

I call the meeting back to order.

I would now like to welcome everybody to the second panel.

Welcome to our witness. I understand we have one witness from Telus. From Telus Communications, we have Pamela Snively, vice-president, chief data and trust officer.

You have five minutes for your opening statement. Go ahead, please.

[Pamela Snively, Vice-President, Chief Data and Trust Officer, Telus Communications Inc.]

Thank you for the opportunity to provide this committee and Canadians with the facts about our Data for Good program.

Telus launched Data for Good in April 2020 because we believed, as we still do, that our company's responsible use of data can play an impactful role in making more evidence-based and informed decisions. We created Data for Good to provide de-identified—that is, essentially anonymous—data to assist governments and health authorities in their efforts to stem the spread of COVID-19 and better understand the impact of interventions like restrictions and stay-at-home orders.

Data for Good was a natural extension of Telus's broader commitment to using our technology to enable social good in support of Canadians and the communities in which we live and work.

As you heard from Dr. Tam, the Data for Good program provided critical insights that supported more informed policies. In short, it worked.

I want to make one thing perfectly clear, Telus did not share any personal information with government—not one iota. Telus always puts its customers and their privacy first. At no time have we ever relaxed any of our rigorous policies about our treatment of personal information, including when we launched Data for Good during the pandemic.

The Data for Good program operates on a data analytics platform called Telus Insights, which is the only privacy-by-design certified platform of its kind in Canada. This platform uses de-identified datasets to reveal movement trends and patterns while protecting individual privacy. Under the Data for Good program, we allow data scientists from our partners, including the government, to have supervised and guided access to our secure Insights platform, which contains only de-identified datasets from our mobility network. Those datasets never left our systems. The data of our customers, even de-identified data, was not sent to the government.

I want to pause on privacy by design. That Telus Insights is privacy-by-design certified is important. Privacy by design is the international gold standard for privacy protection. It was developed here in Canada by Dr. Ann Cavoukian. Privacy by design goes beyond the requirements of the law to entrench privacy protections into the design and operation of the IT systems, networks and business practices of an organization.

With this certification, our Data for Good program is independently validated as being rigorous in its privacy protections. You also heard Dr. Cavoukian's endorsement of our approach last week.

We have taken a leading role nationally on the development and promotion of de-identification as a critical process to enhance privacy protections. Telus is a founding member of CANON, the Canadian Anonymization Network, whose mission is to promote effective de-identification practices and includes the leading Canadian de-identification experts.

Our commitment to de-identification is at the core of Telus Insights. Thanks to our privacy-first approach, Telus was able to leverage our Insights platform to provide pandemic assistance through Data for Good while fully protecting the privacy of customers. While some may compromise on privacy during a public health emergency, we did not. We are very proud of Data for Good and we were intentional and explicit in our public communication about the program.

We developed and published on our website five core data use commitments on how we would share de-identified data and protect privacy. These accompanied a full description of our program along with an FAQ. We had a banner on the main Telus website that linked to this information. Before and after launching the program, we did op-eds and interviews with The Globe and Mail and other Canadian media outlets and published news releases announcing new collaborations. We later publicized that Data for Good was awarded the International Association of Privacy Professionals Privacy Innovation Award in November 2020.

Apart from our public-facing communications about Data for Good and Insights, we consulted with the Office of the Privacy Commissioner of Canada on our transparency plan. We provided that office with an overview of our program, including the five core commitments. The OPC provided valuable feedback, which we gratefully incorporated.

The final point I'd like to make is that as part of our effort to go beyond simple compliance with the law, we offer our customers the ability to opt out of our data analytics program. We see this as a reflection of our customer-first commitment.

In closing, I want to reiterate that Telus provided access to this de-identified data for the public good. The data contained no personal information, so the privacy of our customers was respected, and we made great efforts to be transparent about the program. All of this is consistent with Telus's long-standing track record of protecting our customers' privacy.

Thank you again for the opportunity to speak with you today. I'd be pleased to answer your questions.

[The Chair, Conservative, Pat Kelly]

Right, and with that we'll begin with Mr. Kurek for six minutes.

[Damien Kurek, Conservative, Battle River—Crowfoot, AB]

Thank you very much to our witness for joining us here today.

I found it very interesting and I'm hoping you can unpack a little about what you meant when you said that data was never sent to government.

I'll give you context for the type of response I'm hoping to get. This committee was provided with a letter from the Parliamentary Secretary to the Minister of Health with a slide deck from BlueDot as an example of some of the information that had been provided to PHAC from that organization. We have not been provided with a similar dataset or information as to what the data looked like that Telus had provided to the government.

I wonder if you can elaborate on what it means when you say that it was never sent to government.

[Pamela Snively]

Absolutely. Thank you for the question.

What I meant by that is that we don't actually send data to the government. The way our program works is that we allow data scientists from our partners to come onto our platform, our de-identified data platform, for supervised and guided access. There, they are able to do the queries that are consistent with the use and purpose that we've discussed with them and that needs to fit with our program, and then they're able to create derived data, or what we call “insights”.

When I talk about an insight, the best thing to do would be to picture a heat map or a graph, a bar chart or line graph that would show movement patterns or trends. After they've done that and pulled out these insights, they would be able to download them. Before they could take them, we would review them to make sure they were consistent, that they met all our reidentification risk metrics and that they were consistent with the purpose for the contract, and then the government would be able to take that derived data or insights with them.

[Damien Kurek, Conservative, Battle River—Crowfoot, AB]

In terms of officials from the Public Health Agency of Canada who went to Telus Data for Good, was it a specific location or virtual location? I'm curious as to exactly what that means.

[Pamela Snively]

It is a virtual location, so they're not physically on the premises. It was data scientists coming onto our platform. We provide guided, supervised access to our platform, but it is virtual.

[Damien Kurek, Conservative, Battle River—Crowfoot, AB]

I'm sure all of us around this table have mobile devices, and there is a tremendous amount of information that is available and exists within the cellular providers that operate within our country, and there has to be a high level of trust there.

I'm curious as to whether you would be willing to share some of what that data looked like when it was sent to the Public Health Agency of Canada.

[Pamela Snively]

Yes, I can certainly take that away and we can share sample reports of what the data would look like.

[Damien Kurek, Conservative, Battle River—Crowfoot, AB]

In terms of those who were able to access the platform, we've heard from some experts that de-identifying and the ability for data to be reidentified is very much on a scale. Anonymized and de-identified data can be names and cell numbers taken out of a dataset, or it can be synthesized in a way that would make it virtually impossible for that to be reidentified.

Can you highlight exactly what the officials from the Public Health Agency of Canada, data scientists or whatever the case was, were able to see when accessing your platform—not just the reports, but what were they able to see?