The Algorithmic Managing of ‘At-risk’ Children
Mad in America; Rob Wipond, 7/9/2014
Part two of a Mad In America investigation into the expansion of psychological screening and electronic surveillance of children and youth. Read part one.
Experts point to mounting evidence that scientifically dubious mental health screening programs are just one part of an international governance shift towards creating all-pervasive surveillance systems for diagnosing ‘pre-crime’ and managing ‘at-risk’ children and youth. And not only is this not helping kids, critics argue, it’s demonstrably harming them.
The coroner’s report on the February 2010 suicide of Victoria, British Columbia teenager Freya Milne seemed to suggest that Freya’s mental health treatment may have contributed to her death. The coroner’s report notes, for example, that Freya and her mother were “not advised of the potential risks” of the benzodiazepine medication Freya was prescribed in mid-January of 2010, such as increased “paradoxical behavior” and “suicidal thoughts and behavior.” And the report describes how “the dosage was increased and the medication refilled” by the treating psychiatrist twice more, even as Freya suddenly became alternately too depressed, hyperactive or panicked to attend her psychiatric appointments.
Nevertheless, the coroner recommended boosting mental health training and interventions in schools. In part from those recommendations emerged the Practice Support Program for Child and Youth Mental Health (PSP-CYMH), a training program for British Columbia family doctors and school staff. Independent experts criticize many of that program’s educational materials, saying that they tend to misrepresent the scientific evidence base and downplay the risks of psychiatric medications (see Part One: The Proactive Search for Mental Illnesses in Children).
The coroner’s report on Freya Milne also advised the provincial Ministry of Education “to ensure that all documentation related to a student be placed on a student file including email correspondence, meeting notes and anything else pertaining to a student,” and to start a “flagging system on all student files where a child or youth has a diagnosed mental health issue.”
The government has begun taking action in line with those recommendations, too. But experts are warning that the well-meaning intentions behind those recommendations are dangerously divorced from what the evidence base shows about the psychological and social impacts of surveillance activities on children and youth. And along the way, they explain why being divorced from any scientific evidence base may not be an accidental feature of these recommendations and programs.
Jo Ann Nolan described what those types of recommendations, and programs like the PSP-CYMH, ultimately lead to on the ground.
“There’s so much pressure on teachers and superintendents and school boards to do this,” said Nolan, a social worker assigned to eight Ontario public schools. Nolan said that all professionals connected to Ontario schools are constantly being deluged with “fact sheets,” training workshops, screening tools, and hand-outs and educational videos that are just like the PSP-CYMH materials (and in some cases they’re the exact same materials), ultimately geared towards identifying students from pre-Kindergarten on up who have mental health problems and getting them into treatment. Typically, said Nolan, these initiatives begin by inviting discussion about the normalcy of anxiety and sadness and the importance of learning coping strategies to maintain mental healthiness, and then transition into stating that when problems endure beyond a short time they may indicate signs of “brain diseases” that require the earliest interventions possible with drug treatments.
His doctor tells Edward he has ‘Social Anxiety Disorder’ which is ‘a real medical condition.’ If therapy doesn’t help, his doctor says he can prescribe some medications for Edward.
“Everything we’re doing is talking about brain illnesses and brain diseases. And everybody, teachers, social workers, psychologists, they’re all onto that language,” said Nolan. “It’s unbelievable how something could be changed that drastically in just a few years.”
Nolan said she watched the number of students taking psychiatric medications rising in the 90s, and that’s when she began researching on her own and discovered prominent critics like the International Society for Ethical Psychology and Psychiatry. But in recent years she’s seen a dramatic increase in school-based mental health programs.
The growing movement is easy to spot even for people who never visit schools, in the corresponding online profusion of child and youth-centric mental health-promotion organizations, resources and initiatives, such as school curriculum guides, help lines, chat rooms, comprehensive educational websites, videos, animations, and apps, many of which are heavily funded by a range of governments, institutions, pharmaceutical companies, and other corporate sponsors and advocacy groups.
Pharmaceutical companies are among the sponsors of When Something’s Wrong, a mental health guide for teachers from the Canadian Psychiatric Research Foundation.
“Most people see it in a positive way, that it’s going to be activities or resources or treatments that are going to help children,” said Nolan. “There doesn’t seem to be any critical analysis. And it’s hard to believe that teachers in schools have no critical thinking skills. If they do, maybe they’re just frightened. It’s a pretty hierarchical system.”
According to Nolan, even trained, experienced social workers like herself are no longer entrusted to take actions that two decades ago were part of her normal day, such as simply talking issues over with children who were feeling suicidal. Today, said Nolan, everyone is instructed to watch for “markers” of risk such as falling grades, disruptive behavior or self-cutting, and then are required to refer students to physicians and mental health professionals at specialized agencies or organizations.
Required? “It’s expected,” she said. And dissidents are kept under control. Nolan recently gave a copy of Robert Whitaker’s book detailing the negative findings from studies of long-term use of psychiatric medications, Anatomy of an Epidemic, to a parent, and then waited with trepidation. “Oh yeah, it’s risky,” said Nolan. One time after she recommended that a parent consider taking her child off an ADHD drug, Nolan said, the parent reported her. “I was called into the principal’s office. I remember [the principal] saying, ‘You’re not a doctor.’”
Nolan said financial interests are a major driving factor, whether people are working in the private, public or non-profit sector, or in psychiatry or psychology. “They’re all little businesses,” Nolan said. “They’re all coming into schools offering their services.”
Other warning signs of schizophrenia in children according to this online tool from HealthCentral include ‘unusual crawling’ in toddlers and low grades at school.
Nolan said she’s also concerned about the increasing amount of intimate psychological information that’s being gathered about students through these surveys, screenings, proactive intervention efforts, and other initiatives. “After a year or two, I used to tear up my notes and shred them and no one would know about Johnnie or Susie being involved with me,” said Nolan. But over the last few years it’s become common for many professionals involved with schools to share the same electronic data systems, said Nolan. The assumption is that more information will lead to better decisions about kids, but Nolan said she more often finds that old information left by other people prejudices others against looking at children with fresh eyes. She said she also wonders who else is looking at that information, and for what purposes. “It’s all getting stored somewhere,” said Nolan.
Nolan isn’t the only one becoming concerned about that. Running these mental health programs in part as a means to collect sensitive personal and psychological information about school children has become for some a crusade, for others big business, and for still others a worrisome threat to privacy. It’s certainly a key part of an important international governance shift. Developments in both BC and the US provide illustrative examples.
Towards better management of “unproductive mindsets”
This year, in line with the recommendations and lobbying after the deaths of Milne, Proctor and Kozeletski to intensify the tracking and psychological monitoring of students, the BC government began pushing all BC school districts to participate in a new, much more comprehensive electronic student information system. The old system already collected test results, grades, report card comments, attendance, discipline incidents, and Individualized Education Plans. This new system was designed to also track assignments, surveys, school-related health and mental health records, ongoing assessment and counselling notes, extracurricular activities, email communications and more.
Then, the government declared that it planned to link all this information to the province’s developing system of federated databases connecting driver’s license, social service and electronic personal health records from the public health care system. What would be the purpose of doing all this?
BC Civil Liberties Association policy director Micheal Vonn, whose organization commissioned two studies of these new data systems, stated that the BC government has “bought into” the Big Data revolution. “This government is convinced that data linkages are going to generate important information in research, quality control and citizen services,” said Vonn. Using schools as an example, Vonn said, while teachers generally are skeptical of this initiative, there is a spreading belief among politicians and managers of education systems that analyzing mass amounts of aggregated student data will eventually lead to improved teaching and learning.
The same goes for the aggregation of health records, said Vonn, and she could only speculate about what sorts of studies or initiatives might emerge in that field, especially if pharmaceutical companies gained access to integrated school, health and mental health records. “There’s no question whenever we’re dealing with health information of any kind that we’re dealing with big moneyed interests. This is incredibly valuable information.”
Governments around North America are chipping away at privacy laws to facilitate these efforts, said Vonn, pointing to increasingly controversial school database systems in the US. “We see these issues highlighted much better, they fall into starker relief with systems that are a little more down the road than we are,” she said.
The goal is to capture more “noncognitive” information about school children like “attributes, dispositions, social skills, attitudes, and intrapersonal resources.”
The Gates and Carnegie Foundations have put $100 million into InBloom, one of the biggest of a growing number of for-profit companies and non-profits gathering and analyzing school records, and in some cases linking those with post-secondary and workforce records. These organizations generate analytical reports and sometimes give or sell direct access to the data to private companies or government agencies that provide goods or services to the education sector or to children and families. It’s a rapidly expanding business in nearly every state, bolstered by federal stimulus funding, while lobbying alliances like the Data Quality Campaign give a mixture of rationales about improving education strategies and school accountability. A 2013 US Department of Education draft report draws a vivid picture of where this movement is heading. Lamenting the analytical limitations of test scores, the report endorses a “growing movement” to implement more comprehensive data gathering systems and use more invasive surveillance methods such as video cameras and sensors in computer mice. The goal is to capture more “noncognitive” information about school children like “attributes, dispositions, social skills, attitudes, and intrapersonal resources.” The hope, the report explains, is to gain better understanding of what traits create “high-achieving individuals,” while improving management of children who are “at-risk” of developing “unproductive mindsets” or mental health problems. In essence, there’s growing interest from many sectors of our society in achieving panoptical surveillance of school children.
Health is one of those sectors. Wendy Armstrong, a health policy analyst and consumer advocate, told Mad In America that many health researchers are among the strongest promoters of testing and screening, accessing school and electronic health records, and even mass surveillance to gather information from a multitude of sources, in hopes of being able to better identify “risk factors” and then intervene early to prevent all manner of ills. “Health policy researchers have actually been the biggest pushers for amalgamating all of these databases, because now we’re big into the ‘determinants of health.’” Your schooling, work history, income, housing, intimate relationships and more affect your physical condition and psychological status, said Armstrong, “It’s all seen as important health information.”
Yet this aspect of population surveillance and health research is nascent and often weak, said Armstrong. “What the public and government are being sold is actually the power of prediction.” And governments, she said, are buying – particularly when it comes to methods of psychologically profiling children and youth.
Surveillance and identification of “at-risk” children
If we want to see where our governments are driving all this, said Valerie Steeves, a criminology professor at the University of Ottawa who has extensively researched the surveillance and “dataveillance” of children and youth, we need only look to Britain.
The UK now tracks all of its eleven million kids through integrating health and mental health data, school records, and information held by social services. In a 2010 special issue of the journal Surveillance and Society co-edited by Steeves, various authors examine how British authorities are increasingly screening and tracking kids, and studying these pools of aggregated data, all with an eye to locating children and youth on a spectrum somewhere between “vulnerable” or “at risk of mental illness” to “potentially delinquent” or “susceptible to extremist radicalization.” In effect, the government surveillance system now operates with an overlapping child protection, health and mental health, anti-bullying, anti-violence, and national security agenda that’s focused on predictive “early intervention” in all these arenas.
When the database was first started, said Steeves, “The idea was to create these detailed dossiers to protect kids from domestic violence and other kinds of harm. They also used artificial intelligence software to identify potential criminals. And the youngest ‘criminal’ they identified was a three-year-old boy. To me, this is the hub of what’s wrong with this kind of thinking.”
A psychological screening tool’s “high risk” scores and a database’s analytical risk algorithms may seem scientific and objective, said Steeves, but in reality they are as discriminatory as the people who design them. “A child who could be just a regular three-year-old kid who has a temper tantrum now and then, if he’s racialized, if he’s living in social housing, if he’s impoverished, suddenly becomes identified as a potential criminal.” And once such a child is identified as being “at risk” of criminality, mental illness or anything else, said Steeves, usually “there’s an obligation to intervene.”
Yet much like the mental health screening programs that contribute to this data pool, Steeves said dataveillance and risk-based algorithmic analysis of children and youth have not been proven in any sorts of large trials to be effective or beneficial, and instead often demonstrably lead to many false positives, unnecessary interventions, and harms.
Even before the introduction of any questionable treatments, said Steeves, the act of surveillance itself isn’t benign. “Everything we know about child development is that children develop well and thrive when they are in strong relationships of mutual trust and mutual respect.” However, kids regard being put under surveillance as a breach of trust, said Steeves, and so surveillance “interferes with the development of those kinds of relationships.”
“The rhetoric in the press is that kids don’t care about privacy,” said Steeves. “If you’ve ever done research in the area, the first thing you learn is kids care a lot about privacy.”
In addition, surveillance discourages the very kind of “help seeking” that mental health professionals behind programs like the PSP-CYMH say they want to encourage. “The rhetoric in the press is that kids don’t care about privacy,” said Steeves. “If you’ve ever done research in the area, the first thing you learn is kids care a lot about privacy.” Confidentiality and trust are essential for creating “safe space” for nuanced conversations about sensitive topics, said Steeves. Conversely, surveillance, dataveillance or other types of systematic monitoring that search for “markers of risk” and automatically trigger interventions by outside experts destroy that safe space. “So put that into the context of mental health issues where it becomes much more difficult to go to the teacher for help because [as a child you know that] as soon as you’re identified, the system – not the individual teacher that you trust and know, but the system – will begin to make all sorts of assumptions about you and slot you in for all sorts of interventions.”
Vonn and Armstrong similarly said that these risk-management approaches haven’t been shown to help kids, and expose too many to unnecessary harms.
“It’s not like we can point to great successes in this model writ large,” said Vonn. “It’s completely predicated on the notion that we actually have effective predictive capacities. And what we find out time and time again is, we don’t. Full on, we don’t. And the mistakes that we make in this system create huge amounts of collateral damage.”
Vonn compared it to the building of anti-terrorist no-fly lists. “Risk logics are by nature grotesquely over inclusive,” said Vonn. “The bureaucratic rationale that you employ when you consider putting somebody on the list, is this: If I miss them and they’re [a terrorist], it’s my neck on the chopping block. If I add them and they’re not [a terrorist], oh well, nothing bad happens to me. Of course, something bad happens to that person.”
“This identity theft is far more profound than the theft of papers, this steals who I am,” writes Wrennall. “I cease to exist and in my place is what is written about me by people who do not know me.”
One of the most common damaging impacts comes from “data shadows,” said Vonn. If a child is once screened and identified as “at risk” for a mental illness – even if he’s simply one of the hundreds of false positives in every thousand – then comprehensive data systems like we’re building mean that changing classes, changing schools, or even quitting school, moving, and entering the workforce soon may not allow young people to escape from repercussions and prescribed interventions. Vonn and Armstrong both described examples of mental health labels recorded in databases that were subsequently shared in other contexts and created unexpected serious consequences for people months or years later, including loss of child custody, denial of health insurance, difficulty having physical complaints taken seriously in doctors’ offices, subjection to involuntary treatment, and loss of the right to make decisions regarding financial affairs. This year, for instance, it was disclosed that Canadian police have been routinely sharing personal mental health information on their files – in some cases dating back to people’s childhoods – with prospective employers and U.S. law enforcement, causing people lost jobs and difficulties crossing the border.
In one essay from Surveillance and Society, criminologist Lynne Wrennall, from Liverpool John Moores University, even likens this kind of dataveillance, labeling and sharing of decontextualized information about children to the damages that can be caused by identity theft. “This identity theft is far more profound than the theft of papers, this steals who I am,” writes Wrennall. “I cease to exist and in my place is what is written about me by people who do not know me.”
Scientific accountability interferes with political power
Even though there’s no evidence any of this ultimately helps children, said Steeves, child surveillance has become a multi-billion dollar global industry involving some of the world’s biggest technology companies, with its own vast machinery of public and political influence preying on people’s fears.
“Over the last five or ten years, there’s this recurrent theme that if you love your kids, you’ll put them under surveillance.” And the spreading of mental health screening and education programs in schools that involve psychological surveying and testing, like the PSP-CYMH training program in British Columbia promotes, said Steeves, is just one part of this broader societal shift towards a surveillance-based, risk-management approach to kids.
But if none of this has been proven to help children, why are governments so supportive of it? Wrennall provides commentary on this question in her essay.
Consequently, writes Wrennall, many public service agencies involved with children now function less like “social welfare” and more like “social policing.”
Governments are increasingly turning towards aggressively monitoring, identifying and intervening ever earlier because such intensive psychological micro-management of “problem” children and youth seems like it may be cheaper – and to some governments it’s ideologically more palatable – than broadly funding affordable housing, poverty reduction, accessible daycare, better schools, or other substantive social support initiatives that demonstrably help children. Consequently, writes Wrennall, many public service agencies involved with children now function less like “social welfare” and more like “social policing.”
Meanwhile, Wrennall argues, the lack of any agreed-upon scientific evidence base and scientific framework in effect gives government, law enforcement, social work and mental health authorities broad leeway and discretionary powers. “Almost every child can be labelled as ‘at risk.’ Almost any narrative event can be construed as a Child Protection ‘concern’. There are no clear constraints on the construction of what constitutes transgression or what justifies investigation,” writes Wrennall. “This surveillance is part of a general strategy of subjugation that is able to arbitrarily summon powers in a flexible but intensely aggressive expression of discipline.”
Essentially, argues Wrennall, surveillance-based risk management is an authoritarian approach to governance. The primary concern of its agents is to maintain the ability to conduct surveillance and, when necessary, exert power over others based on their own beliefs, prejudices, political goals, and financial or other motivations. In that context, being democratically and scientifically accountable for truly helping people tends to get in the way.
Where does the illness lie?
After four years of reflecting on the circumstances surrounding her daughter Freya’s death, Shelley Milne has also come to regard issues of scientific accountability and social responsibility as centrally important.
In conversation with Mad In America, Shelley acknowledged that Freya’s Generalized Anxiety Disorder label at times helped her get valuable extra assistance. However, Shelley added, she has noticed since that time how often the label has provided a means for people to escape accountability for their own actions by blaming Freya’s death on her “personal mental problem.” Shelley pointed to the coroner’s report describing how school staff refused to make a simple accommodation for Freya at a crucial time, and how the treating psychiatrist did not follow best practices nor what health regulators advised. “Freya felt forsaken, and in the long run I have felt that, too,” said Shelley. “Because these were trusted professionals in our lives. And nobody has said, ‘We made terrible mistakes and we’re going to learn from them and make changes.’”
With the rate of mental health problems exploding, said Shelley, the social institutions around children need to learn to better accommodate children’s needs, not “treat” at-risk children ever more aggressively to try to get them to fit in better. “I know it’s a massive challenge but, if we address people’s needs,” said Shelley, “what would we do to the rate of mental health issues?”
She described Freya’s unusually sensitive sensory experience of the world, and her struggles spending hours sitting in uncomfortable chairs and submersed in unnatural lighting, constant bustle, and overcrowded classes, while feeling anxious about the time that school and homework demanded. “I never thought of Freya as mentally ill,” said Shelley. “I understood that Freya was vulnerable. The school system was what was ill.” Shelley pointed to a guide to preventing youth suicide from the BC Ministry of Children and Family Development (notably, not the Health Ministry), which cautions that focusing on “individual-level risks” such as identifying and treating mental illnesses can “mask the role of social, structural and institutional factors” like racism, poverty, lack of social opportunities, and overstretched, unaccommodating school systems.
Steeves agreed. “This discourse around risk is being driven by a neo-liberal approach to governance as a whole: Download the responsibility to the individual, don’t look at the social causes.” And by expanding surveillance-based risk-management approaches, said Steeves, “We are focusing on avoiding bad, instead of creating a nurturing, healthy, socially just society.” However, building a better society requires having more honest public discussions about the type of society that we’ve created, and accepting accountability for it, said Steeves. “And that’s a really hard conversation to have. First of all, it’s about us. It’s about us taking responsibility for the environment.”
http://www.madinamerica.com/2014/07/algorithmic-management/