Sleep Quality among Medical Students of a Tertiary Care Hospital: A Descriptive Cross-sectional Study

This was not an intervention study. The researchers tested nothing — they simply measured the current state of sleep quality among medical students using a validated questionnaire. Specifically, they assessed:

**Primary outcome:** Prevalence of "poor sleep quality" as defined by the Pittsburgh Sleep Quality Index (PSQI) global score >5.

**Secondary outcomes:** Mean sleep duration (hours per night), and differences in sleep quality between male and female students.

**No comparator group:** There was no control group, no intervention, and no follow-up. This is a single-time-point snapshot.

The study is purely descriptive: it answers "how many students have poor sleep?" but cannot answer "what causes it?" or "what improves it?"

**Sample size:** 217 undergraduate medical students.

**Population:** Students from Kathmandu Medical College, a tertiary care hospital in Nepal.

**Years included:** First, second, third, and final (fourth) year medical students.

**Recruitment:** Simple random sampling from each year.

**Exclusion criteria:** Not explicitly stated in the abstract. The full text may have excluded students with diagnosed sleep disorders, but this is not reported.

**Demographics:** 103 males, 114 females (approximate based on prevalence breakdown). Mean age not reported in abstract, but typical medical student age range is 18–25.

**Key limitation for self-experimenters:** This is a narrow, homogeneous sample — young, Nepalese, medical students in a single institution. Results may not generalise to other populations (e.g., working adults, older adults, non-students).

The researchers used a single instrument:

**Pittsburgh Sleep Quality Index (PSQI):** A self-reported questionnaire that assesses sleep quality over the past month. It yields a global score from 0 to 21, where:

- **0–5:** Good sleep quality

- **6–21:** Poor sleep quality

- The PSQI measures seven components: subjective sleep quality, sleep latency (how long to fall asleep), sleep duration, habitual sleep efficiency (time asleep vs. time in bed), sleep disturbances, use of sleep medication, and daytime dysfunction.

**No objective measures:** No actigraphy (wrist-worn movement sensors), no polysomnography (lab sleep study), no sleep diaries. Everything is based on student recall over the past month, which is prone to memory bias.

**Data collection:** Self-administered paper questionnaires, filled out in a classroom or clinical setting. No mention of whether students were supervised or could ask questions.

**Why this matters for self-experimenters:** The PSQI is a validated, widely used tool that you can use yourself for free. However, it measures *perceived* sleep quality, not objective sleep physiology. A person who feels they sleep poorly may have normal brain wave patterns, and vice versa. For your own experiments, combining PSQI with a sleep diary or a wearable (e.g., Fitbit, Oura Ring) gives a more complete picture.

### Study design

This is a **descriptive cross-sectional study** — a single survey administered at one point in time. It is observational, not experimental.

### Sampling

Simple random sampling from each academic year (first through final).

Total sample of 217. No power calculation is reported, meaning we don't know if the sample size was large enough to detect meaningful differences between subgroups (e.g., male vs. female).

### Duration

Data collection occurred over **two months** (October to November 2019).

Each participant filled out the questionnaire once. There was no follow-up, no repeated measures, no longitudinal tracking.

### Statistical analysis

Descriptive statistics only: frequencies, percentages, means, standard deviations.

Software: SPSS version 24.

No inferential statistics reported in the abstract (no t-tests, chi-square tests, regression, or confidence intervals). The abstract states "prevalence among male and female students as 41 (39.8%) and 55 (48.2%) respectively" but does not test whether this difference is statistically significant.

### What this design can and cannot prove

**Can prove:**

The *prevalence* of poor sleep quality in this specific population at this specific time.

The *average* sleep duration.

Differences in prevalence between subgroups (if statistically tested, which they were not here).

**Cannot prove:**

**Causation:** Cannot say "medical school causes poor sleep" because there is no comparison group (e.g., non-medical students, working adults).

**Temporal direction:** Cannot say whether poor sleep causes academic problems or academic stress causes poor sleep — both are possible.

**Change over time:** A single snapshot cannot show whether sleep quality worsens or improves across the years of medical training.

**Effectiveness of interventions:** No intervention was tested, so nothing can be said about what improves sleep.

### Major methodological weaknesses

1. **No inferential statistics:** The authors report percentages but do not test whether the male-female difference (39.8% vs. 48.2%) is real or due to chance. This is a critical omission.

2. **Self-report bias:** Students may overestimate or underestimate their sleep quality due to social desirability, poor memory, or misunderstanding of questions.

3. **No objective validation:** Without actigraphy or sleep diaries, we cannot verify actual sleep duration or quality.

4. **Single institution:** Results may reflect local culture, curriculum, or living conditions rather than medical students generally.

5. **No response rate reported:** We don't know how many students were approached vs. how many participated. If many refused, the sample may be biased (e.g., students with worse sleep may have been less likely to participate).

6. **No exclusion criteria stated:** Students with diagnosed sleep disorders, shift workers, or those on medications affecting sleep were not excluded, which could inflate the prevalence of poor sleep.

**Overall prevalence:** 96 out of 217 students (44.23%) had poor sleep quality (PSQI global score >5).

**By gender:**

- Males: 41 out of 103 (39.8%) had poor sleep quality.

- Females: 55 out of 114 (48.2%) had poor sleep quality.

- *Note: The authors did not report a p-value or confidence interval for this difference, so we cannot determine if it is statistically significant.*

**Mean sleep duration:** 6.7 hours per night (standard deviation ±1.6 hours).

- This is below the National Sleep Foundation's recommended 7–9 hours for young adults (ages 18–25).

- Approximately 16% of students (based on the standard deviation) likely slept 5.1 hours or less, and another 16% slept 8.3 hours or more.

**No other outcomes reported:** The abstract does not break down PSQI subcomponents (e.g., sleep latency, efficiency, daytime dysfunction). It also does not report differences by academic year, age, or other variables.

**What is missing:** The authors do not report how many students had "good" sleep quality (PSQI ≤5), the full distribution of PSQI scores, or any correlations with academic performance, stress levels, or lifestyle factors.

**44% prevalence:** This means that nearly 1 in 2 medical students in this sample reported clinically significant sleep problems. For context, general population studies using the PSQI typically find poor sleep quality in 20–30% of adults. So medical students here have roughly **1.5 to 2 times** the rate of poor sleep compared to the general population.

**6.7 hours average sleep:** This is about **30–60 minutes less** than the recommended minimum of 7 hours. Over a week, this accumulates to a sleep debt of 3.5–7 hours — roughly equivalent to missing one full night of sleep every 2–4 weeks.

**Gender gap:** The 8.4 percentage point difference between males (39.8%) and females (48.2%) is modest. If this difference were real, it would mean that for every 100 male students with poor sleep, about 121 female students would have poor sleep. However, without a statistical test, this could easily be due to chance.

**Plain English translation:** If you are a medical student (or a similarly stressed young adult), there is about a 50/50 chance that your sleep quality is poor enough to affect your daytime functioning. You are likely sleeping about 6–7 hours per night, which is less than optimal for cognitive performance, memory consolidation, and emotional regulation.

### Acknowledged by authors (likely, based on abstract)

The abstract does not explicitly list limitations, but the authors note that "there are not enough recent studies" and call for "efforts must be directed towards educating about sleep hygiene." This implies they recognise the need for intervention studies.

### Critical reader observations

1. **No causal inference possible:** The cross-sectional design cannot determine whether poor sleep is caused by medical school stress, pre-existing sleep problems, or other factors (e.g., caffeine, screen time, living conditions).

2. **Self-report bias is severe:** Students may underreport sleep problems due to stigma ("I should be able to handle this") or overreport due to a desire to justify poor academic performance.

3. **No objective sleep measurement:** Without actigraphy or polysomnography, we cannot distinguish between perceived poor sleep and actual physiological poor sleep. Some people who feel they sleep poorly actually have normal sleep architecture.

4. **Single institution, single country:** Results may not generalise to medical students in other countries, other universities, or other cultures with different academic schedules, living arrangements, or sleep norms.

5. **No control for confounders:** The study did not measure or control for caffeine intake, alcohol use, screen time before bed, room-sharing, shift work (e.g., clinical rotations), mental health status (depression, anxiety), or medication use — all of which strongly affect sleep quality.

6. **No response rate:** If only 50% of approached students participated, the sample could be biased toward students who are more (or less) concerned about their sleep.

7. **No longitudinal follow-up:** We don't know if sleep quality changes across the academic year (e.g., worse during exam periods) or across years of training (e.g., worse in clinical years vs. preclinical years).

8. **No statistical testing:** The male-female difference is reported without a p-value, confidence interval, or effect size. This is a fundamental statistical flaw.

For someone running their own n=1 experiment to improve sleep quality, here is how to use this study as a starting point:

### What to test

**Specific intervention:** Based on the finding that medical students average 6.7 hours of sleep, test whether **extending time in bed to 8 hours** (i.e., aiming for 7.5–8 hours of actual sleep) improves your daytime functioning.

**Alternative intervention:** Test a **sleep hygiene protocol** (e.g., no screens 1 hour before bed, consistent wake-up time, no caffeine after 2 PM, bedroom temperature 18–20°C). The study implies that poor sleep hygiene is common, so improving it is a logical first step.

**Dose:** For the time-in-bed extension, add 30–60 minutes to your current bedtime (e.g., if you currently sleep 11 PM–6 AM, try 10:30 PM–6 AM or 11 PM–6:30 AM).

### Minimum meaningful duration

**At least 2 weeks per condition.** The PSQI asks about the past month, but for an n=1 experiment, 2 weeks is enough to see a stable pattern. A full menstrual cycle (4 weeks) is better for women, as sleep quality varies with menstrual phase.

**Include a 1-week baseline** before starting the intervention to establish your current sleep quality and duration.

### What to measure

**Primary metric:** Pittsburgh Sleep Quality Index (PSQI) — take it at the end of each 2-week block. You can find the questionnaire and scoring instructions for free online. A change of 3 points or more on the global score is considered clinically meaningful.

**Secondary metrics (daily):**

- **Sleep duration** (hours per night) — from a sleep diary or wearable.

- **Sleep onset latency** (minutes to fall asleep) — from a sleep diary.

- **Wake after sleep onset** (minutes awake during the night) — from a wearable or diary.

- **Daytime sleepiness** — use the Epworth Sleepiness Scale (free online) once per week.

- **Subjective energy/mood** — rate 1–10 each morning and evening.

**Objective metric (optional):** If you have a wearable (Fitbit, Oura, Apple Watch), track **sleep efficiency** (time asleep ÷ time in bed × 100). Aim for >85%.

### Key confounds to control for

1. **Caffeine:** Keep intake constant (same amount, same time of day) across both conditions. Better yet, eliminate caffeine after 2 PM entirely.

2. **Alcohol:** Avoid alcohol entirely during the experiment, or keep it constant. Alcohol fragments sleep even in small amounts.

3. **Exercise:** Keep exercise timing and intensity constant. Evening exercise within 2 hours of bed can delay sleep onset for some people.

4. **Meal timing:** Avoid large meals within 3 hours of bedtime. Keep dinner timing consistent.

5. **Screen use:** If testing sleep hygiene, eliminate screens 1 hour before bed. If testing time extension, keep screen use constant.

6. **Stress:** Track daily stress on a 1–10 scale. Major exams, relationship stress, or work deadlines can overwhelm any sleep intervention.

7. **Room environment:** Keep bedroom temperature, noise, and light levels constant. Use blackout curtains and earplugs if needed.

8. **Medications:** Note any changes in medications, supplements (melatonin, magnesium), or over-the-counter sleep aids.

### What a positive result would look like

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

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

**Sleep onset latency decreases by ≥10 minutes** (e.g., from 30 minutes to 20 minutes).

**Daytime sleepiness score drops by ≥3 points** on the Epworth scale.

**Subjective energy/mood ratings increase by ≥2 points** on a 1–10 scale.

**Consistency:** The improvement should be seen in at least 10 out of 14 nights, not just a few good nights.

**Important caveat:** This study found that 44% of medical students have poor sleep, but it did not test any intervention. A positive result in your n=1 experiment would suggest that *your* poor sleep is at least partially due to modifiable factors (e.g., insufficient time in bed, poor sleep hygiene) rather than an underlying medical condition (e.g., insomnia disorder, sleep apnea, depression). If you try these interventions for 4–6 weeks and see no improvement, consider consulting a sleep specialist for a more thorough evaluation.