Longitudinal Trajectories of Child and Youth Mental Health Symptoms Across Distinct Phases of the COVID-19 Pandemic: A population-based study in Ontario, Canada

This population-based study in Ontario reveals that child and youth mental health symptoms remained elevated during pandemic lockdowns but significantly improved as restrictions eased, with notable variations by symptom type, gender, and pre-existing conditions, highlighting the need for equity-informed policies that balance viral control with the minimization of lockdown impacts.

The Gist

Imagine the COVID-19 pandemic as a massive, sudden storm that hit a small town called "Ontario." For two years, the town had to stay inside its houses, schools were closed like libraries during a fire drill, and the rules changed constantly.

This paper is like a long-term weather report for the mental health of the town's children. Instead of just taking a snapshot of how kids felt on one rainy day, the researchers followed 1,261 families from January 2021 to December 2022. They checked in five times, like a doctor making house calls, to see how the "weather" inside the children's minds changed as the storm raged and then slowly cleared up.

Here is the story of what they found, broken down into simple concepts:

1. The "Stuck in the Basement" Phase (Lockdowns)

When the province went into a full lockdown (like being stuck in a basement with no windows), the children's mental health didn't get worse immediately, but it didn't get better either. It stayed stuck at a high level of stress.

• The Analogy: Think of mental health symptoms like water in a bathtub. During the strict lockdowns, the water level was high (lots of anxiety, anger, or trouble focusing), and it just stayed there, bubbling over. It didn't drain away.

2. The "Drain Plug" Moment (Restrictions Eased)

As soon as the government started loosening the rules—letting kids go back to school and play outside—the water in the bathtub started to drain.

• The Twist: The water didn't drain evenly.
  • The "Behavior" Water: Symptoms like acting out, being defiant, or having trouble sitting still (ADHD-like symptoms) drained away very quickly. It's like opening the drain wide; the water rushed out.
  • The "Worry" Water: Symptoms like sadness, anxiety, and fear (internalizing symptoms) drained much slower. It was like a clogged drain; the water level went down, but it took a long time to get back to normal.

3. Who Got Wettest? (The Different Groups)

Not everyone experienced the storm the same way. The researchers found some interesting patterns:

• The "Newcomers" to Lockdown: Children who were not in lockdown at the very start but got locked down later (when the province went into a total lockdown) saw their stress levels spike the most. It was like being told to suddenly hide under the bed when you were just playing outside.
• The "Veterans": Kids who were already in lockdown at the start seemed to adapt. Their stress levels stayed steady because they had already adjusted to the "basement" life.
• The "Pre-Existing Conditions" Group: Children who already had diagnosed health or mental health issues actually saw less of a spike in stress when lockdowns started. It's as if they were already wearing raincoats, so the sudden storm didn't soak them as badly as it did the others.
• The Gender Gap: As the storm cleared, boys' "worry water" (internalizing symptoms) drained away faster. Girls, however, stayed in the "worry zone" longer. It's like boys ran out of the rain quickly, while girls were still standing in the drizzle.

4. The Parent-Child "Dance"

The study found a strong connection between the parents' stress and the children's stress. They move like a dance partner.

• If the parent was stressed and anxious, the child's stress went up.
• If the child was acting out or worried, the parent's stress went up.
• This happened in both directions, like a feedback loop. When the parents felt better, the kids felt better, and vice versa.

5. The Big Takeaway

The most important lesson from this paper is that mental health isn't a straight line.

Many people thought the pandemic made kids' mental health permanently broken. This study says: No.

• When the rules were tight, stress was high and stuck.
• When the rules loosened, things got better, but not instantly for everyone.

The Metaphor of the Hike:
Imagine the pandemic was a hike up a steep, foggy mountain.

• During the lockdowns, everyone was stuck on a high ledge, looking down at the fog.
• When the fog lifted (restrictions eased), everyone started walking down.
• Some people (kids with behavioral issues) ran down the stairs quickly.
• Others (girls with anxiety) had to walk down a winding, muddy path slowly.
• And some families (parents and kids) were holding hands the whole time; if one stumbled, the other stumbled too.

Why This Matters

This study tells us that we shouldn't panic if a child is still struggling a year after the pandemic. It takes time for the "worry water" to drain. It also tells us that when we make big changes (like locking down a whole province), we need to be careful because the transition into lockdown can be just as stressful as the lockdown itself.

The best way to help? Keep an eye on the parents, because when the parents are calm, the kids are more likely to find their way down the mountain, too.

Technical Summary

1. Problem Statement and Research Gap

While extensive literature exists regarding child and youth mental health during the COVID-19 pandemic, significant methodological limitations persist:

• Temporal Limitations: Most studies focus on the early pandemic phases (2020) and lockdown periods, lacking data on how symptoms evolved as restrictions eased or tightened over time.
• Design Limitations: Existing research relies heavily on cross-sectional designs, convenience sampling, and online data collection, leading to high risks of bias and poor representation of marginalized populations.
• Scope Gaps: Evidence is skewed toward depression and anxiety in adolescents, with limited data on younger children (4–17 years) and other domains such as externalizing (oppositional defiant) and neurodevelopmental (inattention/hyperactivity) symptoms.
• Contextual Gap: There is a scarcity of high-quality, longitudinal, population-based data from Canada, a country that implemented some of the most prolonged and restrictive public health measures globally.

Objective: To estimate longitudinal trajectories of parent-reported internalizing, oppositional defiant, and inattention/hyperactivity symptoms across five distinct phases of the pandemic (Jan 2021–Dec 2022) and identify baseline and time-varying covariates associated with these trajectories.

2. Methodology

Study Design and Sample:

• Design: A prospective, population-based longitudinal study (the ONPATH study).
• Sample: $N = 1,261$ children and youth aged 4–17 years and their parents from Ontario, Canada.
• Sampling: Drawn from the Probit panel (a probabilistic survey panel) to reflect the demographic and socio-economic characteristics of the Canadian population. Stratified by gender, education, and region.
• Data Collection: Five waves of data collection:
  • T1 (Baseline): Jan–Feb 2021 (Regional lockdown variability).
  • T2: May–June 2021 (Provincial complete lockdown).
  • T3: Oct–Nov 2021 (Re-opening).
  • T4: Feb–Mar 2022 (Omicron wave/restrictions).
  • T5: Oct–Dec 2022 (Restrictions largely lifted).
• Retention: 76.0% at T2, 66.1% at T3, 61.8% at T4, and 57.0% at T5.

Measures:

• Primary Outcome: An abbreviated 21-item version of the Ontario Child Health Study Emotional Behavioural Scales (OCHS-EBS).
  • Domains: Internalizing (depression/anxiety), Oppositional Defiant, and Inattention/Hyperactivity.
  • Scoring: Moderated Non-Linear Factor Analysis (MNLFA) was used to derive standardized factor scores (M=0, SD=1) to correct for measurement non-invariance across time and participant characteristics.
• Covariates:
  • Time-Invariant (TICs): Child age/gender, parent gender/race, household income, pre-existing clinician-diagnosed conditions, family structure, and baseline lockdown status.
  • Time-Varying (TVCs): Parental distress (Kessler-6), school learning modality (in-person/hybrid/remote), and daily COVID-19 case rates.

Statistical Analysis:

• Modeling: Latent Growth Curve Modeling (LGCM) with piecewise functional forms to capture non-linear trajectories.
• Approach: Unconditional models were first fitted to identify turning points (knots). Conditional models were then constructed to test the effects of TICs on latent intercepts (baseline levels) and slopes (rates of change), and TVCs on concurrent and lagged symptom levels.
• Estimation: Full-information maximum likelihood (FIML) to handle missing data under a missing-at-random assumption, utilizing sampling weights to ensure population representativeness.

3. Key Results

Trajectory Patterns:

• Stability during Lockdowns: Mental health symptoms remained elevated and stable from T1 to T2 (during peak restrictions).
• Improvement during Re-opening: Significant reductions in symptoms occurred from T2 to T5 as restrictions loosened.
• Domain Differences: The magnitude of improvement was largest for oppositional defiant and inattention/hyperactivity symptoms compared to internalizing symptoms.
  • Effect Sizes (Cohen's d): Internalizing ($d \approx -0.18$), Oppositional ($d \approx -0.39$), Inattention/Hyperactivity ($d \approx -0.39$).

Covariate Associations:

• Lockdown Status (T1): Children not in lockdown at baseline showed a significant increase in symptoms when transitioning to lockdown at T2. Conversely, those already in lockdown at T1 showed relative stability, suggesting an "adaptation" effect to prolonged restrictions.
• Pre-existing Conditions: Children with pre-existing physical, mental, or neurodevelopmental conditions had higher baseline symptoms but showed attenuated increases during the initial lockdown transition (T1–T2) compared to healthy peers.
• Gender Differences:
  • Boys had higher baseline inattention/hyperactivity symptoms.
  • Girls showed smaller reductions in internalizing symptoms from T2 to T5 compared to boys, indicating a slower recovery for girls in this domain.
• Parental Distress: Robust bidirectional associations were found. Parental distress predicted subsequent child symptoms, and child symptoms predicted subsequent parental distress across all time points. Concurrent correlations were strongest at T1 and T5.
• Schooling and Case Rates: Effects were less consistent than parental distress. Remote schooling at T4 predicted higher internalizing symptoms; remote schooling at T1 predicted higher oppositional symptoms at T2. Higher case rates at T1 predicted higher oppositional symptoms.

4. Key Contributions

1. Longitudinal Granularity: This is one of the few studies to track mental health trajectories across the entirety of the pandemic's restrictive and re-opening phases (2021–2022), moving beyond the "snapshot" limitations of cross-sectional studies.
2. Methodological Rigor: The use of a probability-based sample (reducing selection bias) and MNLFA scoring (reducing measurement bias) sets a high standard for epidemiological research in this field.
3. Natural Experiment: The study leveraged the regional variability of lockdowns at baseline to create a quasi-experimental comparison between children transitioning into lockdown versus those already in lockdown, revealing the acute impact of transition versus duration.
4. Domain-Specific Insights: It highlights that externalizing and neurodevelopmental symptoms are more sensitive to environmental structure (e.g., school routines) and recover faster upon re-opening, whereas internalizing symptoms (particularly in girls) show more persistent elevation.

5. Significance and Implications

• Policy Planning: The findings suggest that the transition into lockdown is a critical period of acute mental health risk. Future pandemic planning should prioritize immediate support during the onset of restrictions rather than assuming stability during prolonged lockdowns.
• Equity and Targeted Interventions:
  • Gender: Girls require targeted support for internalizing symptoms as restrictions ease, as their recovery trajectory is slower.
  • Pre-existing Conditions: While these children had high baseline symptoms, they were less reactive to the initial shock of lockdown, suggesting different intervention needs compared to previously healthy children who experienced acute symptom spikes.
• Family Systems: The strong bidirectional link between parent and child distress underscores the necessity of family-centered interventions. Supporting parental mental health is a direct mechanism for improving child outcomes.
• Future Research: The study emphasizes that relying solely on data from lockdown periods may overestimate the persistence of pandemic-related mental health impacts. Long-term monitoring is essential to distinguish between acute stress reactions and chronic deterioration.

Limitations Noted:

• Lack of pre-pandemic baseline data (collection started ~10 months into the pandemic).
• Reliance on parent-reported outcomes (potential informant bias).
• Modest response rate (30%) and attrition (43% by T5), potentially leading to survivorship bias where the most vulnerable families were lost to follow-up.