Stress Management and Resilience Training Among Department of Medicine Faculty: A Pilot Randomized Clinical Trial

**Intervention:** The SMART program — a single 90-minute one-on-one training session delivered by a trained facilitator. The program teaches attention regulation and interpretation skills designed to reduce stress and build resilience. Key components include:

Training to shift attention from threat-focused thinking to novelty and reward-focused attention

Techniques to reframe stressful interpretations of events

Practices for cultivating gratitude, compassion, acceptance, and meaning

Strategies to reduce "thought suppression" (trying not to think about something) and instead use acceptance-based approaches

**Comparator:** Wait-list control group — participants received no intervention during the 8-week study period but were offered the SMART training after the study ended.

**Primary outcome measures (measured at baseline and week 8):**

**Connor-Davidson Resilience Scale (CDRS):** A 25-item scale measuring resilience (ability to bounce back from stress). Scores range from 0–100, higher = more resilient.

**Perceived Stress Scale (PSS):** A 10-item scale measuring how unpredictable, uncontrollable, and overloaded respondents find their lives. Scores range from 0–40, higher = more stress.

**Smith Anxiety Scale (SAS):** Measures state anxiety. Higher scores = more anxiety.

**Linear Analog Self-Assessment Scale (LASA):** A single-item scale measuring overall quality of life on a 0–10 scale, higher = better quality of life.

**Sample size:** 40 physicians randomized (32 completed the study — 16 in each group)

**Population:** Department of Medicine faculty physicians at Mayo Clinic, Rochester, MN (a tertiary care academic medical center)

**Setting:** Single academic medical center in the United States

**Dropout rate:** 8 out of 40 (20%) did not complete the study — 4 from each group

**Demographics:** Not reported in detail (age, gender, years of practice, specialty breakdown not provided in the abstract or available full text)

All four outcome measures were self-report questionnaires administered at two time points: baseline (before randomization) and at 8 weeks (post-intervention for the treatment group, end of wait-list period for controls).

**Connor-Davidson Resilience Scale (CDRS):** 25 items, each rated 0–4. Total score 0–100. Measures resilience as "personal competence, trust in one's instincts, tolerance of negative affect, positive acceptance of change, and spiritual influences." Validated in general population and clinical samples.

**Perceived Stress Scale (PSS):** 10 items, each rated 0–4. Total score 0–40. Measures the degree to which situations in one's life are appraised as stressful. Widely validated, one of the most commonly used stress measures in research.

**Smith Anxiety Scale (SAS):** Measures state anxiety (anxiety "right now" or "recently"). Specific scoring range not provided in the abstract.

**Linear Analog Self-Assessment Scale (LASA):** Single item: "How would you rate your overall quality of life?" on a 0–10 visual analog scale. Validated in oncology populations.

**Study design:** Pilot randomized clinical trial (RCT) with a wait-list control design.

**Randomization:** Participants were randomized to either the SMART intervention group or the wait-list control group. The randomization method is not specified in the abstract — it's unclear if they used computer-generated random numbers, sealed envelopes, or another method. This is a minor methodological weakness.

**Blinding:** This was an open-label trial — no blinding was used. Participants knew whether they received the intervention or were on the wait-list. The outcome measures were self-report questionnaires, so there was no blinded assessor. This is a significant limitation because expectation effects (placebo) can influence self-reported stress, anxiety, and quality of life.

**Duration:** 8 weeks from baseline to follow-up assessment. The intervention itself was a single 90-minute session delivered at the start of the 8-week period. No booster sessions or ongoing support were provided.

**Statistical approach:** The authors used two-sample t-tests to compare the mean change from baseline to week 8 between the intervention and control groups. They report t-statistics, degrees of freedom, and p-values. They do not report effect sizes (like Cohen's d) or confidence intervals for the between-group differences, which limits interpretation of the magnitude of effects.

**What this design can prove:**

The RCT design with a concurrent control group allows causal inference — the observed improvements can be attributed to the SMART intervention rather than to natural recovery, regression to the mean, or the passage of time.

The wait-list control design ensures that all participants eventually receive the intervention, which is ethical and may improve recruitment.

**What this design cannot prove:**

Cannot prove that effects last beyond 8 weeks — no longer-term follow-up was conducted.

Cannot prove that the intervention works in other populations (only Mayo Clinic physicians were studied).

Cannot rule out placebo effects or demand characteristics because there was no blinding and no active control condition (e.g., a "sham" training or attention control).

Cannot determine which specific components of the SMART program are responsible for the effects — the intervention is a "package" of multiple techniques.

Cannot determine if the effects are clinically meaningful versus statistically significant — no established minimal clinically important differences (MCIDs) are reported for these scales in this population.

**Major methodological weaknesses:**

1. **Small sample size:** Only 32 completers (16 per group). This is a pilot study, and the small sample means the results may not be stable or generalizable.

2. **No blinding:** Participants knew their group assignment, which can inflate effects on self-report measures.

3. **No active control:** The wait-list control does not control for attention, expectation, or the simple act of meeting with a facilitator.

4. **Short follow-up:** 8 weeks is too short to know if effects are sustained.

5. **Single site:** All participants were from one academic medical center, limiting generalizability.

6. **No intention-to-treat analysis reported:** The abstract reports results only for the 32 completers. If the 8 dropouts differed systematically, this could bias results.

7. **No correction for multiple comparisons:** Four primary outcomes were tested, increasing the risk of Type I error (false positive).

**Primary outcome: Resilience (CDRS)**

Intervention group: Mean change from baseline = +9.8 points (SD = 9.6) — *improvement*

Control group: Mean change from baseline = -0.8 points (SD = 8.2) — *slight worsening*

Between-group difference: t(30) = 3.18, p = 0.003

**Statistically significant improvement in resilience in the SMART group compared to controls**

**Primary outcome: Perceived Stress (PSS)**

Intervention group: Mean change from baseline = -5.4 points (SD = 8.1) — *reduction in stress*

Control group: Mean change from baseline = +2.2 points (SD = 6.1) — *increase in stress*

Between-group difference: t(30) = -2.76, p = 0.010

**Statistically significant reduction in perceived stress in the SMART group compared to controls**

**Primary outcome: Anxiety (SAS)**

Intervention group: Mean change from baseline = -11.8 points (SD = 12.3) — *reduction in anxiety*

Control group: Mean change from baseline = +2.9 points (SD = 8.9) — *increase in anxiety*

Between-group difference: t(30) = -3.62, p = 0.001

**Statistically significant reduction in anxiety in the SMART group compared to controls**

**Primary outcome: Quality of Life (LASA)**

Intervention group: Mean change from baseline = +0.4 points (SD = 1.4) — *slight improvement*

Control group: Mean change from baseline = -0.6 points (SD = 1.0) — *slight worsening*

Between-group difference: t(30) = 2.29, p = 0.029

**Statistically significant improvement in quality of life in the SMART group compared to controls**

**Summary of all four outcomes:** The intervention group improved on all measures (resilience up, stress down, anxiety down, quality of life up), while the control group either stayed the same or worsened slightly over the 8-week period.

**Resilience (CDRS):** The SMART group improved by about 10 points on a 100-point scale, while controls declined by about 1 point. The net benefit was roughly 11 points. To put this in context: a 10-point change on the CDRS is roughly equivalent to moving from the 50th percentile to the 65th percentile in general population norms. This is a moderate-to-large effect, but without Cohen's d or confidence intervals, the precision is unknown.

**Perceived Stress (PSS):** The SMART group reduced stress by about 5.4 points on a 40-point scale, while controls increased by 2.2 points. The net difference was about 7.6 points. For reference, a 4–5 point change on the PSS is often considered clinically meaningful in intervention studies. The control group getting worse suggests that physician stress may naturally increase over an 8-week period without intervention.

**Anxiety (SAS):** The SMART group reduced anxiety by about 12 points, while controls increased by about 3 points. This is a large net difference of roughly 15 points. However, without knowing the scale range (likely 20–80 or similar), it's hard to translate into everyday terms.

**Quality of Life (LASA):** The SMART group improved by 0.4 points on a 0–10 scale, while controls declined by 0.6 points. The net difference was 1.0 point. A 1-point change on a 0–10 scale is generally considered a small but perceptible improvement — roughly equivalent to moving from "fair" to "good" on a single global rating.

**In plain English:** After a single 90-minute training session, physicians reported feeling about 10% more resilient, 15% less stressed, and noticeably less anxious 8 weeks later, compared to colleagues who received no training. The control group actually got slightly worse over time, suggesting that physician well-being may decline naturally without intervention.

**Acknowledged by authors:**

The authors describe this as a "pilot" study, implying they recognize the need for larger trials with longer follow-up.

They note that wider dissemination and larger clinical trials are warranted.

**Additional limitations a critical reader would note:**

1. **Small sample size (n=32 completers):** With only 16 people per group, the study is underpowered to detect small-to-moderate effects reliably. The results could easily change with a larger sample.

2. **No blinding:** This is the most critical limitation. When people know they received an intervention, they often report improvements simply because they expect to feel better (placebo effect). Without a sham training or active control, we cannot separate the specific effects of SMART from general expectation effects.

3. **No active control group:** The wait-list control does not control for the effects of meeting with a facilitator, receiving attention, or simply taking time to focus on well-being. A better design would compare SMART to another active intervention (e.g., a relaxation training session of equal duration).

4. **Short follow-up (8 weeks):** We don't know if the effects last beyond 8 weeks. Many stress reduction interventions show initial improvements that fade over time without ongoing practice.

5. **Single site, homogeneous population:** All participants were physicians at one academic medical center (Mayo Clinic). Results may not generalize to other settings, professions, or geographic locations.

6. **Self-report measures only:** All outcomes were self-reported. No objective measures of stress (e.g., cortisol, heart rate variability, blood pressure) or behavioral outcomes (e.g., sick days, job performance, patient satisfaction) were collected.

7. **No intention-to-treat analysis:** The abstract reports results only for the 32 completers. If the 8 dropouts had worse outcomes, the results could be biased in favor of the intervention.

8. **Multiple comparisons:** Four primary outcomes were tested without correction (e.g., Bonferroni). With four tests, there's a ~19% chance that at least one significant result is a false positive.

9. **No effect sizes reported:** The authors report p-values but not Cohen's d, eta-squared, or confidence intervals, making it harder to assess the practical significance of the findings.

10. **No information on fidelity:** We don't know if the intervention was delivered consistently across participants, or if there was a manual or protocol to ensure standardization.

11. **No information on concurrent treatments:** Participants may have been using other stress management techniques (exercise, meditation, therapy) during the study, which could confound results.

12. **Industry funding:** While the authors report "no conflict of interest," the study was funded by the Department of Medicine at Mayo Clinic. This is not problematic per se, but it's worth noting that the intervention was developed at the same institution.

For someone running their own n=1 experiment:

### What to test

Test the core components of the SMART program as described in this study:

**Attention training:** Practice shifting your attention from threat-focused thinking (e.g., "What could go wrong?") to novelty and reward-focused thinking (e.g., "What's interesting about this situation? What can I learn?")

**Reframing:** When you notice a stressful interpretation of an event, deliberately generate alternative, less threatening interpretations

**Gratitude practice:** Spend 2–3 minutes daily identifying things you're grateful for

**Acceptance:** Instead of trying to suppress or avoid unpleasant thoughts, practice acknowledging them without judgment and letting them pass

**Meaning-making:** Connect daily tasks to your larger values and sense of purpose

**Dose:** The study used a single 90-minute training session. For a self-experiment, you could:

Option A: A single intensive session (90 minutes) where you learn the techniques, then practice them daily for 8 weeks

Option B: Break it into shorter daily practices (10–15 minutes per day) for 8 weeks

### Minimum meaningful duration

**8 weeks minimum** — this matches the study duration

**Daily practice** of at least 5–10 minutes is likely needed

**Measure at baseline and at 8 weeks** — do not measure weekly, as stress and resilience change slowly

### What to measure (specific metrics)

Use the same scales as the study (all are freely available online):

1. **Resilience:** Connor-Davidson Resilience Scale (CD-RISC) — 25 items, ~10 minutes to complete. Score 0–100. A positive result would be an increase of 10+ points.

2. **Perceived Stress:** Perceived Stress Scale (PSS-10) — 10 items, ~5 minutes. Score 0–40. A positive result would be a decrease of 5+ points.

3. **Anxiety:** Smith Anxiety Scale or a simpler alternative like the Generalized Anxiety Disorder scale (GAD-7) — 7 items, ~3 minutes. Score 0–21. A positive result would be a decrease of 4+ points.

4. **Quality of Life:** Single-item LASA (0–10 scale): "How would you rate your overall quality of life?" A positive result would be an increase of 1+ point.

**Optional objective measures (if you have access):**

Resting heart