We Built a System That Makes Children Sick and Then Express Surprise When They Get Sick.

By Ms. Poon · 15 December 2025 · 7 min read

Every year, another set of data documenting the mental health crisis among Hong Kong young people. Every year, the same institutional expressions of concern. Every year, the same response: more counsellors, more hotlines, an expanded protocol here, a new support mechanism there.

This year, the 2025 Policy Address announced the extension of the Three-tier Suicide Emergency Mechanism to upper primary schools. Depression and anxiety indices at CUHK are at all-time record highs. Forty-three percent of young adults showing moderate-to-severe depression. Eighty-three percent of secondary students feeling "more tense than usual." Over four hundred school crisis referrals in a single year.

And the official response, as always, is to hire more people to manage the crisis while leaving intact the system that produces it.

I spent twelve years as Head of Admissions at a competitive kindergarten. I have sat with two-year-olds while their parents watched through one-way glass to see if their child could identify colours in Cantonese and English and Putonghua simultaneously. I know exactly where this pipeline starts. I would like to describe the full length of it, because I think seeing it as a continuous structure — rather than as isolated "pressure points" at K1, P6, and DSE — is the only way to understand why the mental health data looks the way it does.

Stage One: The K1 Interview

It begins before the child's third birthday. Parents who cannot afford to begin earlier — and many begin in the child's first year of life — start preparing their child for the kindergarten assessment process at around eighteen months to two years of age. Preparation classes. Interview coaching. Developmental milestone acceleration.

The child in the assessment room has no understanding of what is happening. The child's parents have communicated, unavoidably and involuntarily, that this room and this stranger and this outcome are extremely important. Children at this age read parental anxiety with perfect fidelity. The child does not know what a kindergarten is. The child knows that something is at stake, and that their parent is frightened.

The assessment measures a fifteen-minute snapshot of a two-and-a-half-year-old. The result determines school placement. School placement, in the collective understanding of Hong Kong families, has implications for primary school pathways, for secondary school access, and ultimately for life outcomes. This belief is not entirely irrational — there is genuine tracking, genuine correlation. It is also substantially exaggerated beyond what the evidence supports. But the belief is real, it is widespread, and it is installed in children before they can walk well.

This is stage one. The child has not yet started school. They are already inside a selection system.

Stage Two: Primary School and the Accumulation of Load

Primary school adds layers. TSA begins formally at P3, but the preparation begins in P1 — because teachers who feel the school will be judged on TSA results cannot simply wait until P3 to start working toward those results.

I want to be specific about what "more tense than usual" means for a primary school child. It means that the educational experience is structured primarily around performance outcomes rather than around learning. The child who asks a question because they are curious is less valued by this system than the child who gives the correct answer on the assessment. Curiosity is inefficient. Correct answers are measurable.

Academic streaming in primary school begins early and hardens early. By P4, most children have received enough information about where they stand to have formed a stable belief about their academic identity. The children at the top of the hierarchy are under pressure to maintain their position. The children in the middle are under pressure to improve it. The children at the bottom are under pressure from a system that has decided they are near the bottom, which is a particularly cruel kind of pressure to place on a nine-year-old.

The P6 Secondary School Place Allocation adds a terminal event to the primary school experience: a formal sorting process that determines secondary school placement and that families understand — again, with some accuracy — as consequential. The preparation for SSPA begins approximately two years before it happens and shapes the entire P5 and P6 experience.

Nothing in this structure is neutral. Every stage adds load. No stage removes any.

Stage Three: Secondary School and the DSE

Secondary school is where the accumulated load becomes clinically significant for enough students that the data becomes unmistakable.

The DSE is a single examination series that determines university placement. In a system with a small number of sought-after university places, a limited number of vocational alternatives that carry comparable social status, and a youth unemployment rate of 12.3% for ages 20–24, the DSE does not merely feel high-stakes. It is high-stakes. The student's perception of the consequences is not disordered thinking. It is an accurate reading of the situation.

The student who sits the DSE in S6 has been inside a selection system for fourteen or fifteen years. They have been assessed, streamed, ranked, and compared continuously since before they started school. They have never experienced an educational environment in which their value was not primarily derived from their performance relative to others.

Forty-three percent moderate-to-severe depression among 18–24 year olds is not a surprise. It is the logical output of the system I have just described.

The Government's Response

Let me be very precise about what the government's mental health response does and does not do.

More school counsellors: useful. Counsellors do valuable, important work. They are also being asked to treat, individually and expensively, symptoms produced by a structural problem. Expanding the counsellor headcount without addressing the structure is the equivalent of hiring more doctors to treat the sick while leaving in place the conditions that make people ill.

The extended suicide prevention mechanism: necessary. Children are in crisis early enough that primary schools need it. The mechanism should exist. It is also an extraordinary admission: the system has produced crisis conditions in upper primary school children, and the response is a formalized management protocol rather than an examination of what is producing the crisis.

More hotlines, more campaigns, more public mental health messaging: fine. Also insufficient. The student who calls a hotline at midnight because they cannot see a reason to continue is not in crisis because they didn't know the number. They are in crisis because they have spent the last decade in a system that organised their entire sense of self worth around a performance outcome that may or may not materialise.

What Would Actually Help

I am under no illusion that what I am about to say will happen. It requires structural change to a system that serves significant institutional interests — property markets, competitive school status hierarchies, university admissions frameworks — and those interests will resist structural change.

But here is what would actually help: removing load at each stage, rather than adding support mechanisms to manage the consequences of load.

Genuine low-stakes TSA — not nominal low-stakes, actually low-stakes, with no school-level publication of results and no institutional consequences for low-scoring schools.

Secondary school allocation processes that do not function as high-stakes ranking events for eleven-year-olds.

A university admissions system that is not organised almost entirely around a single examination taken at seventeen.

Vocational education pathways with genuine social and economic parity — not nominal parity, actual parity, such that the student who goes into skilled trades at sixteen is not understood, by themselves or their family or their community, as having failed.

None of this is coming soon. In the meantime, children will continue to be assessed at two-and-a-half, streamed at nine, ranked through their teenage years, and arrive at adulthood with depression rates that would be recognised as a public health emergency in any other context.

The data will continue to be published. The official expressions of concern will continue. More counsellors will be hired. The mechanism will be extended, perhaps to lower primary, perhaps eventually to kindergarten.

And every year, the same question will not be asked, because asking it leads somewhere that is institutionally inconvenient: if we know the system makes children sick, and we keep running the system, what does that tell us about what we actually value?