Compassion-focused therapy as guided self-help for enhancing public mental health: A randomized controlled trial.

The intervention was a 9-week guided self-help program based on Compassion-Focused Therapy (CFT). Participants received a printed workbook ("Compassion as a Way of Life") divided into nine chapters, each covering a different theme (e.g., understanding compassion, developing the compassionate self, applying compassion to daily life). Each chapter included psychoeducation, experiential exercises (e.g., compassionate imagery, compassionate letter-writing), and daily practices (e.g., 5–10 minutes of compassionate breathing or soothing rhythm breathing). Participants also received three 30-minute telephone coaching sessions (at weeks 1, 4, and 7) from trained psychology graduates, focused on motivation, problem-solving barriers, and reinforcing practice.

The comparator was a waitlist control group — participants assigned to this group received no intervention during the active study period but were offered the CFT program after the 3-month follow-up.

Primary outcome: well-being (emotional, psychological, and social well-being combined). Secondary outcomes: depression, anxiety, stress, self-compassion, self-criticism, self-reassurance, positive affect, negative affect, and gratitude.

242 Dutch adults recruited from the general population via advertisements in local newspapers and social media

Mean age = 52.87 years (SD = 9.99, range approximately 25–75)

74.8% female

Inclusion criteria: age ≥ 18, low to moderate levels of well-being (score ≤ 4 on the Mental Health Continuum-Short Form emotional well-being subscale, or ≤ 4 on ≥ 1 of the psychological or social well-being subscales), access to internet and telephone, fluent in Dutch

Exclusion criteria: current psychological treatment, severe mental illness (psychosis, bipolar disorder, substance dependence), high suicide risk, or currently practicing meditation > 2 hours per week

120 randomized to CFT, 122 to waitlist

82.5% of CFT group completed the post-intervention assessment; 73.3% completed 9-month follow-up

All measures were self-report questionnaires completed online at baseline, post-intervention (3 months after baseline), 3-month follow-up (6 months after baseline), and 9-month follow-up (12 months after baseline).

**Mental Health Continuum-Short Form (MHC-SF):** 14 items measuring emotional, psychological, and social well-being. Total score range 0–70 (higher = better well-being). Primary outcome.

**Hospital Anxiety and Depression Scale (HADS):** 14 items, two subscales (anxiety and depression, each 0–21, higher = worse).

**Perceived Stress Scale (PSS):** 10 items, range 0–40 (higher = more stress).

**Self-Compassion Scale-Short Form (SCS-SF):** 12 items, range 1–5 (higher = more self-compassion).

**Forms of Self-Criticizing/Attacking and Self-Reassurance Scale (FSCRS):** 22 items measuring inadequate self, hated self, and reassured self (higher scores = more of each).

**Positive and Negative Affect Schedule (PANAS):** 20 items, two 10-item subscales (positive affect and negative affect, each range 10–50).

**Gratitude Questionnaire (GQ-6):** 6 items, range 6–42 (higher = more gratitude).

**Design:** Two-arm, parallel-group, randomized controlled trial (RCT) with a waitlist control group. Participants were randomly assigned 1:1 to CFT or waitlist using a computer-generated random sequence (block randomization with varying block sizes, stratified by gender). Allocation was concealed from the researcher who enrolled participants.

**Blinding:** This was an open-label trial — participants knew whether they were in the intervention or waitlist group. The researchers who delivered the coaching calls were not blinded. Outcome assessors (data analysts) were blinded to group assignment. This is a significant limitation because self-report outcomes are susceptible to demand characteristics and expectation effects.

**Duration:**

Intervention period: 9 weeks

Post-intervention assessment: at 3 months (immediately after the 9-week program)

Short-term follow-up: 3 months post-intervention (6 months from baseline)

Long-term follow-up: 9 months post-intervention (12 months from baseline)

**Statistical approach:** Intention-to-treat (ITT) analysis using linear mixed models, which handles missing data under the missing-at-random assumption. Effect sizes reported as Cohen's d (difference in change scores between groups divided by pooled baseline SD). Moderators tested: age, gender, education, baseline well-being, baseline self-compassion, and number of coaching sessions completed.

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

**Can prove:** That the CFT program caused greater improvements in well-being and secondary outcomes compared to no intervention over the same period (causal inference from randomization).

**Cannot prove:** That CFT is superior to an active control (e.g., mindfulness, cognitive-behavioral therapy, or a placebo intervention). The waitlist design controls for the passage of time and regression to the mean, but not for attention, expectation, or the nonspecific effects of receiving a structured program with coaching calls. Participants on a waitlist may also experience disappointment or demoralization, which could inflate the apparent benefit of the intervention.

**Cannot prove:** That the effects are due specifically to compassion-focused techniques rather than general self-help, behavioral activation, or the therapeutic alliance from coaching calls.

**Cannot prove:** Long-term durability beyond 9 months, or effectiveness in clinical populations (the sample had low-to-moderate well-being, not diagnosed mental illness).

**Methodological weaknesses:**

No active control group (e.g., a self-help program without compassion content, or a supportive phone call condition)

No blinding of participants or coaches

Self-report measures only (no behavioral or physiological outcomes)

High dropout: 17.5% of CFT group lost to post-intervention, 26.7% by 9-month follow-up

Waitlist group was not followed beyond 3-month follow-up (they received the intervention after that), so long-term comparisons are within-group only for the CFT arm

Sample was predominantly female, middle-aged, and Dutch — limits generalizability

**Primary outcome — well-being (MHC-SF):**

At post-intervention (3 months): CFT group improved significantly more than waitlist. Cohen's d = 0.51, 95% CI [0.25, 0.77], p < 0.001. This is a medium effect size.

At 3-month follow-up (6 months): effect was slightly smaller but still significant. d = 0.39, 95% CI [0.13, 0.65], p < 0.001.

At 9-month follow-up (12 months): within the CFT group, well-being scores were maintained or slightly improved compared to post-intervention (no comparison with waitlist possible because waitlist group had received the intervention by then).

**Secondary outcomes (at post-intervention, CFT vs. waitlist):**

Depression (HADS): d = -0.35, 95% CI [-0.61, -0.09], p = 0.008 (CFT group improved more)

Anxiety (HADS): d = -0.33, 95% CI [-0.59, -0.07], p = 0.013

Stress (PSS): d = -0.37, 95% CI [-0.63, -0.11], p = 0.005

Self-compassion (SCS-SF): d = 0.57, 95% CI [0.31, 0.83], p < 0.001

Self-criticism (FSCRS inadequate self): d = -0.40, 95% CI [-0.66, -0.14], p = 0.003

Self-reassurance (FSCRS): d = 0.46, 95% CI [0.20, 0.72], p < 0.001

Negative affect (PANAS): d = -0.33, 95% CI [-0.59, -0.07], p = 0.014

Positive affect (PANAS): d = 0.18, 95% CI [-0.08, 0.44], p = 0.17 (NOT significant)

Gratitude (GQ-6): d = 0.38, 95% CI [0.12, 0.64], p = 0.004

**Moderators:** None of the tested variables (age, gender, education, baseline well-being, baseline self-compassion, number of coaching sessions) significantly moderated the treatment effect. This means the program appeared equally effective across different subgroups.

**Within-group changes at 9-month follow-up (CFT group only):**

All improvements were maintained or amplified. For example, well-being scores at 9-month follow-up were higher than at post-intervention (mean MHC-SF: baseline = 42.6, post = 49.1, 9-month = 50.8). This suggests continued benefit or further skill development after the program ended.

The primary effect (d = 0.51) means that the average person in the CFT group improved by about half a standard deviation more than the average person in the waitlist group. In practical terms, this is roughly the difference between someone who "sometimes" feels good about their life and someone who "often" feels good about their life on the MHC-SF items.

The effect on self-compassion was the largest (d = 0.57), suggesting the program successfully taught its core skill.

The effect on positive affect was the smallest and not statistically significant — meaning the program reduced negative emotions more than it boosted positive ones.

To put the well-being improvement in context: the CFT group moved from a mean MHC-SF score of about 42.6 (moderate well-being, roughly the 40th percentile in general population samples) to about 49.1 (high well-being, roughly the 60th percentile). This is a meaningful shift in day-to-day experience.

**Acknowledged by authors:**

No active control group — cannot rule out that effects are due to nonspecific factors (attention, expectation, structured routine)

Self-report measures only — no objective indicators of well-being or behavior change

High dropout rate (17.5% at post-intervention, 26.7% at 9 months) — completers may differ from dropouts

Sample was predominantly female, well-educated, and middle-aged — limits generalizability to men, younger adults, or less educated populations

Waitlist group not followed beyond 3-month follow-up — long-term comparisons are within-group only

**Additional critical notes:**

No measure of adherence to daily practices (e.g., how many minutes per day participants actually practiced compassionate breathing) — so we don't know the dose-response relationship

The coaching calls were brief (three 30-minute calls) but may have provided significant therapeutic alliance effects — a self-help program without any coaching might show smaller effects

The study was conducted in the Netherlands, a country with high social support and mental health literacy — results may not replicate in other cultural contexts

No assessment of adverse effects or worsening of symptoms — CFT can sometimes trigger distress in people with trauma histories (compassion-focused exercises may bring up painful emotions)

The 9-month follow-up data for the CFT group is uncontrolled (no comparison group), so improvements could be due to natural recovery, seasonal effects, or other life changes

For someone running their own n=1 experiment:

**What to test:**

A 9-week self-compassion training program using a structured workbook or online course, combined with brief weekly check-ins (self-administered or with a buddy). The core practices to test: compassionate breathing (5–10 minutes daily), compassionate imagery (visualizing a compassionate other or your compassionate self), and compassionate letter-writing (writing to yourself from a compassionate perspective about a difficulty).

**Minimum meaningful duration:**

9 weeks appears to be the minimum for durable change. The study showed effects at 3 months (immediately post-program) and maintained at 9 months. A shorter program (e.g., 4 weeks) may produce smaller or less lasting effects. If you want to test a shorter version, measure at 4 weeks and again at 8–12 weeks to see if effects decay.

For daily practice: aim for 10–15 minutes per day. The study's workbook recommended 5–10 minutes of formal practice plus informal practices throughout the day.

**What to measure (specific metrics):**

**Primary:** Well-being — use the Mental Health Continuum-Short Form (MHC-SF, 14 items, free online). Score range 0–70. A meaningful improvement would be an increase of 5–7 points (roughly half a standard deviation).

**Secondary:** Self-compassion — Self-Compassion Scale-Short Form (SCS-SF, 12 items, free). Score range 1–5. Expect an increase of 0.3–0.5 points.

**Secondary:** Self-criticism — Forms of Self-Criticizing/Attacking and Self-Reassurance Scale (FSCRS, 22 items, free). Track the "inadequate self" subscale (expect decrease of 2–4 points) and "reassured self" subscale (expect increase of 2–4 points).

**Secondary:** Stress — Perceived Stress Scale (PSS-10, 10 items, free). Score range 0–40. Expect a decrease of 2–4 points.

**Secondary:** Negative affect — Positive and Negative Affect Schedule (PANAS, 20 items, free). Expect negative affect to decrease by 2–4 points; positive affect may not change much.

**Process measure:** Daily practice log — track minutes of formal practice, number of informal practices, and a 1–10 rating of how compassionate you felt that day.

**Key confounds to control for:**

**Expectation effects:** If you believe compassion training will help, you may report improvements regardless. Consider a 2-week baseline period where you measure well-being daily without any intervention, then start the program. This gives you a within-person baseline.

**Life events:** Major stressors (job loss, relationship changes, health issues) can swamp any intervention effect. Log major life events weekly.

**Seasonal effects:** Well-being varies with seasons (winter blues, summer uplift). If you start in January, improvements by March could be partly seasonal. Best to run the experiment across a season-neutral period or note the season.

**Other practices:** If you also start exercising, meditating, or therapy during the 9 weeks, you can't attribute changes to CFT alone. Keep other wellness practices constant.

**Social support:** The study included three coaching calls. If you do this alone, you may get smaller effects. Consider having a friend or online accountability partner check in weekly.

**Sleep and exercise:** Both affect mood and self-compassion. Track sleep quality and exercise minutes as covariates.

**What a positive result would look like:**

MHC-SF score increases by ≥ 5 points from pre- to post-intervention (e.g., from 42 to 47 or higher)

Self-compassion score increases by ≥ 0.3 points

Self-criticism score decreases by ≥ 2 points

Perceived stress decreases by ≥ 3 points

You notice in daily life: less harsh inner voice when you make mistakes, more ability to comfort yourself during difficult moments,