Factors Affecting Sleep Quality among Adolescent Athletes

The researchers tested whether five specific factors were associated with sleep quality in adolescent athletes:

1. **Consumption of caffeine or alcohol** (e.g., coffee, tea, energy drinks, alcoholic beverages) in the hours before sleep.

2. **Extreme mood** (e.g., feeling shocked, overly excited, or emotionally upset) within 2 hours of bedtime.

3. **Sleeping environment** (e.g., room brightness, noise level, temperature).

4. **Anxiety due to academics** (e.g., worry about exams, homework, grades).

5. **Anxiety due to sports** (e.g., nervousness about upcoming competitions, training performance, coach expectations).

They measured sleep quality using the **Pittsburgh Sleep Quality Index (PSQI)** , a validated self-report questionnaire that assesses seven components: subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleep medication, and daytime dysfunction. The PSQI yields a global score from 0 to 21, with higher scores indicating worse sleep quality (a score >5 is considered "poor sleep").

They also collected data on **sleep duration** (hours per night), **self-rated sleep quality** (a single item from the PSQI), and **sports performance** (self-reported training and competition performance on a 1–10 scale).

**112 participants** (male and female)

**Age range:** 12 to 17 years old (mean age not reported)

**Population:** Secondary school students in Hong Kong who were also recreational team-sport athletes (e.g., basketball, volleyball, football, handball)

**Setting:** Recruited from local sports clubs and school teams in Hong Kong

**Exclusion criteria:** Not explicitly stated, but participants were required to be "recreation team-sports athletes" – meaning they trained at least twice per week and competed in school or club leagues

**Key demographics:** All participants were Chinese adolescents living in Hong Kong, a high-pressure academic environment. No data on sex distribution, socioeconomic status, or specific sport type were provided.

**Pittsburgh Sleep Quality Index (PSQI):** A 19-item self-report questionnaire covering the past month. Global score range 0–21. A score >5 indicates poor sleep quality. The PSQI has been validated in adolescent populations (Cronbach's alpha typically 0.70–0.83).

**Sleep duration:** Self-reported average hours of sleep per night over the past month (from PSQI item).

**Self-rated sleep quality:** A single item from the PSQI: "During the past month, how would you rate your sleep quality overall?" (0 = very good, 1 = fairly good, 2 = fairly bad, 3 = very bad).

**Factor assessment:** Participants completed a custom questionnaire asking about frequency of caffeine/alcohol consumption before bed (never, sometimes, often, always), extreme mood before sleep (yes/no), sleeping environment (bright/noisy vs. dark/quiet), academic anxiety (rated 1–5), and sports anxiety (rated 1–5).

**Sports performance:** Self-reported on a 1–10 scale for both training and competition performance over the past month.

**Study design:** This was a **cross-sectional observational study**. Data were collected at a single time point via questionnaires. There was no intervention, no randomisation, no blinding, and no control group.

**Statistical approach:** The researchers used a **One-way ANCOVA** (Analysis of Covariance) to test whether each of the five factors was associated with sleep quality, while controlling for age and sex as covariates. They then used **multiple regression analysis** to predict PSQI global score from the five factors simultaneously. They reported significance at p < 0.05 but did not report effect sizes (e.g., Cohen's d, partial eta-squared, or regression coefficients with confidence intervals).

**Duration:** The PSQI asks about the past month, so the study captures a one-month retrospective window. However, the factor questionnaire asked about "usual" habits, so the time frame is ambiguous.

**What this design can and cannot prove:**

**Can prove:** Associations (correlations) between the five factors and sleep quality at a single point in time. It can identify which factors are statistically related to worse sleep in this specific population.

**Cannot prove:** Causality. Because this is cross-sectional, we cannot determine whether caffeine causes poor sleep, or whether people who already have poor sleep are more likely to drink caffeine to cope. Similarly, academic anxiety might cause poor sleep, or poor sleep might cause academic anxiety. The direction of effect is unknown.

**Cannot prove:** Temporal order. The study does not track changes over time, so we cannot see if changes in a factor precede changes in sleep quality.

**Cannot prove:** Generalisability beyond Hong Kong Chinese adolescent team-sport athletes.

**Major methodological weaknesses:**

1. **No objective sleep measurement:** All sleep data are self-reported. Adolescents are known to overestimate sleep duration and underestimate sleep latency compared to actigraphy or polysomnography.

2. **No validated scale for the five factors:** The custom questionnaire was not validated. Questions like "How often do you consume caffeine before bed?" with options "never, sometimes, often, always" are vague and subject to recall bias.

3. **No control for confounders:** Many factors that affect sleep (e.g., chronotype, screen time before bed, physical activity levels, diet, medical conditions) were not measured or controlled.

4. **Small sample size for regression:** With 112 participants and 5 predictors plus covariates, the regression may be underpowered to detect smaller effects. Rule of thumb: 10–20 participants per predictor would require 50–100 participants, so this is borderline.

5. **No effect sizes reported:** The paper states "significance effects" but does not provide the actual F-values, p-values, or effect sizes (e.g., partial eta-squared, Cohen's d, or regression coefficients). This makes it impossible to assess the practical importance of the findings.

6. **No correction for multiple comparisons:** Testing five factors in ANCOVA and then again in regression inflates the Type I error rate.

**Primary outcome: Sleep quality (PSQI global score)**

**Mean PSQI global score:** Not reported in the abstract or full text. The paper only states that "majority of secondary student-athletes are suffering from inadequate sleep" but does not give the mean or proportion with PSQI >5.

**Sleep duration:** Mean sleep duration was **6.8 hours per night** (standard deviation not reported). This is below the recommended 8–10 hours for adolescents aged 14–17 (National Sleep Foundation guidelines).

**Factors significantly associated with worse sleep quality (from ANCOVA):**

1. **Caffeine/alcohol consumption:** Participants who reported "often" or "always" consuming caffeine or alcohol before bed had significantly worse PSQI scores compared to those who "never" or "sometimes" did. (No F-value, p-value, or effect size reported.)

2. **Extreme mood before sleep:** Those who reported experiencing extreme mood (shock, excitement, upset) within 2 hours of bedtime had significantly worse PSQI scores. (No statistic reported.)

3. **Poor sleeping environment:** Bright or noisy sleeping environment was associated with significantly worse PSQI scores. (No statistic reported.)

4. **Academic anxiety:** Higher self-reported academic anxiety was significantly associated with worse PSQI scores. (No statistic reported.)

5. **Sports anxiety:** Higher self-reported sports anxiety was also significantly associated with worse PSQI scores, but the effect was **smaller** than for academic anxiety. (No statistic reported.)

**Secondary outcome: Self-rated sleep quality**

The same five factors were significantly associated with the single-item self-rated sleep quality measure.

**Sports performance:**

The paper states that "poor sleep quality was associated with worse self-reported training and competition performance," but no correlation coefficients, means, or p-values are provided.

**Multiple regression results:**

The regression model predicted PSQI global score from the five factors. The paper reports that **academic anxiety** was the strongest predictor (beta coefficient not reported), followed by **caffeine/alcohol consumption** and **sleeping environment**. **Extreme mood** and **sports anxiety** were weaker but still significant predictors. The overall model R² (variance explained) was not reported.

Because the paper does not report effect sizes, we cannot quantify the magnitude of the associations. However, we can translate the findings into rough estimates based on typical PSQI scores:

A "poor sleeper" (PSQI >5) typically has a global score of 6–10. A "good sleeper" (PSQI ≤5) typically scores 2–5.

The difference between "never" and "often" caffeine consumption might correspond to a 2–4 point increase in PSQI score – enough to move someone from good to poor sleep.

Academic anxiety likely had the largest effect: a 1-point increase on the 1–5 anxiety scale might correspond to a 1–2 point increase in PSQI score.

The difference between a dark/quiet and bright/noisy sleeping environment might be equivalent to losing about 30–45 minutes of sleep per night.

**Important caveat:** These are speculative estimates. The original paper provides no data to support these numbers.

**Acknowledged by authors:**

The study used a cross-sectional design, so causality cannot be established.

The sample was limited to Hong Kong Chinese adolescent athletes, so results may not generalise to other populations.

The authors recommend future research using objective measures (e.g., brain wave monitoring, actigraphy) and longitudinal designs.

**Critical reader observations:**

1. **No effect sizes or confidence intervals:** This is a major omission. Without these, readers cannot assess the practical significance of the findings.

2. **Self-report bias:** Adolescents may underreport caffeine/alcohol use due to social desirability, and overreport sleep duration.

3. **No objective sleep measurement:** Actigraphy or polysomnography would have provided more reliable data on sleep duration, latency, and efficiency.

4. **No control for screen time:** Smartphone and computer use before bed is a known confound that was not measured.

5. **No control for chronotype:** Morning vs. evening types have different sleep patterns and may respond differently to the factors studied.

6. **No data on training load:** Athletes with higher training volumes may have different sleep needs and patterns.

7. **No data on menstrual cycle:** For female adolescents, menstrual cycle phase affects sleep quality and was not controlled.

8. **Small sample for subgroup analyses:** The study did not report sex differences, sport type differences, or age differences, which are likely important.

9. **No information on response rate:** How many athletes were approached and how many agreed to participate? Selection bias is possible.

10. **Industry funding:** The journal "Sports Nutrition and Therapy" is a low-impact, open-access journal. No funding sources or conflicts of interest were declared.

For someone running their own n=1 experiment to improve sleep quality:

### What to test (specific intervention and dose)

**Caffeine cessation:** Stop all caffeine (coffee, tea, energy drinks, soda, chocolate) after 2:00 PM for 2 weeks. If you are sensitive, try after 12:00 PM.

**Alcohol avoidance:** Eliminate alcohol entirely for 2 weeks. If you must drink, limit to 1 drink and stop at least 3 hours before bed.

**Mood management:** Implement a 30-minute "wind-down" routine before bed: no stimulating conversations, no exciting TV/games, no stressful work. Try journaling, meditation, or gentle stretching.

**Sleep environment optimisation:** Make your bedroom completely dark (blackout curtains, no LED lights), quiet (earplugs or white noise machine), and cool (65–68°F / 18–20°C). Remove all electronics.

**Academic/sports anxiety reduction:** Try a "worry journal" – write down all concerns 1 hour before bed, then close the notebook. Or try a 5-minute breathing exercise (4 seconds inhale, 6 seconds exhale).

### Minimum meaningful duration

**2 weeks minimum** for each intervention. Sleep quality changes can take 3–7 days to stabilise after a change.

**4 weeks is better** to account for menstrual cycle effects (if applicable) and to see if improvements persist.

Run each intervention separately (A/B testing) rather than all at once, so you know which factor is driving the change.

### What to measure (specific metrics)

**Primary metric:** PSQI global score (0–21). Take it at baseline and after each 2-week intervention. Free online versions are available.

**Secondary metrics:**

- Sleep duration (hours per night, from sleep diary or wearable)

- Sleep latency (minutes to fall asleep)

- Wake after sleep onset (minutes awake during the night)

- Subjective sleep quality (1–10 scale each morning)

- Daytime sleepiness (Epworth Sleepiness Scale, 0–24)

- Training/competition performance (1–10 scale after each session)

**Objective tracking:** Use a wearable (e.g., Fitbit, Oura Ring, Whoop) for sleep duration and consistency, but note that wearables are less accurate for sleep stages.

### Key confounds to control for

**Screen time:** Keep constant across conditions (e.g., no screens 1 hour before bed for both baseline and intervention).

**Training load:** Log your training volume (hours, intensity) each week. If it changes, the effect on sleep may confound your results.

**Meal timing:** Eat dinner at least 2–3 hours before bed. Keep meal composition similar across conditions.

**Caffeine from other sources:** Check medications, supplements, and even decaf coffee (which contains small amounts of caffeine).

**Stressors:** Note major life events (exams, competitions, relationship issues) in a log. These can overwhelm any intervention.

**Menstrual cycle (if applicable):** Track cycle phase. Sleep quality often worsens in the luteal phase (days 14–28). Run interventions for at least one full cycle.

### What a positive result would look like

**PSQI score drops by ≥3 points** (e.g., from 8 to 5, moving from "poor" to "good" sleep).

**Sleep duration increases by ≥30 minutes** (e.g., from 6.5 to 7.0 hours).

**Sleep latency decreases by ≥15 minutes** (e.g., from 45 to 30 minutes).

**Subjective sleep quality rating increases by ≥2 points** (e.g., from 5/10 to 7/10).

**Daytime sleepiness score drops by ≥3 points** (e.g., from 12 to 9).

**Training/competition performance rating increases by ≥1 point** (e.g., from 6/10 to 7/10).

**Example of a positive result:** After 2 weeks of no caffeine after 2 PM, your PSQI drops from 7 to 4, your sleep duration increases from 6.5 to 7.2 hours, and you feel less drowsy during afternoon training. This would suggest caffeine timing is a meaningful factor for your sleep.

**Important:** Because this was an observational study, the findings are suggestive but not definitive. Your own n=1 experiment will provide stronger evidence for *your* body than this paper does for the general population.