How to empower patients? A systematic review and meta-analysis

Key points

1. 01Overall pooled effect favored intervention over control (SMD 0.43, 95% CI 0.23-0.62, p < 0.0001) but heterogeneity was very high (I squared = 86-88%)
2. 02Individual sessions produced a higher effect estimate (SMD 0.59) than group-based sessions (SMD 0.31)
3. 03Knowledge alone was not sufficient; combining it with goal setting and action planning was associated with success
4. 04Interventions using 2-3 behavioral change techniques had a higher effect estimate (SMD 0.55) than those using only one (SMD 0.21, non-significant)
5. 05Most evidence comes from diabetes populations; findings cannot be reliably generalized to other chronic diseases

How it was conducted

Design

Systematic review and meta-analysis of randomized controlled trials (PRISMA guidelines, PROSPERO CRD42016052416)

Databases

Medline, Cochrane Library, Web of Science, PsycInfo, Scopus; last search March 29, 2017

Included studies

32 RCTs published in English between 2002 and 2017

Meta-analysis

23 of 32 studies included; pooled standardized mean differences using binary random-effects model with 95% CI

Participants

5,649 participants at baseline across all 32 studies; 3,544 in the meta-analysis

Primary outcome

Patient empowerment, measured by validated instruments (most commonly the Diabetes Empowerment Scale or its short form)

What they found

• Overall pooled SMD 0.43 (95% CI 0.23-0.62, p < 0.0001); heterogeneity I squared = 86%, p < 0.00001
• Diabetes subgroup SMD 0.48 (95% CI 0.25-0.70, p = 0.0001); I squared = 88%
• Non-diabetes subgroup SMD 0.12 (95% CI -0.14 to 0.38, p = .38); I squared = 24%, p = .27 (non-significant)
• Individual sessions subgroup SMD 0.59 (95% CI 0.23-0.94, p = .001); I squared = 92%
• Group-based sessions subgroup SMD 0.31 (95% CI 0.17-0.45, p = .0001); I squared = 33%
• Two or three BCTs subgroup SMD 0.55 (95% CI 0.24-0.87, p = .0006); I squared = 87%
• One BCT subgroup SMD 0.21 (95% CI -0.28 to 0.69, p = .40); non-significant
• Best-evidence synthesis (10 higher-quality studies): SMD 0.24 (95% CI -0.02 to 0.50, p = .07); non-significant
• Goal setting applied in 9 successful vs 5 unsuccessful interventions; action planning in 6 successful vs 2 unsuccessful
• Knowledge was the most frequent BCT, appearing in 13 successful and 12 unsuccessful interventions
• 17 of 33 interventions coded as successful; 16 as unsuccessful

Limitations

• Heterogeneity remained very high across all subgroup analyses, so results must be interpreted with caution
• No study achieved low risk of bias for all key domains; blinding of participants and personnel was high or unclear in 28 of 32 studies
• The large majority of included studies focused on diabetes, limiting generalizability to other chronic conditions
• Intervention content was often poorly described in original papers, requiring some behavioral change techniques to be generalized during coding

Why it matters

For patients

Patients with chronic diseases are more likely to feel in control of their health when their care combines practical knowledge with personalized goal setting and action planning, ideally delivered in one-to-one sessions rather than group classes.

For clinicians

Clinicians designing empowerment programs should go beyond education alone and incorporate structured goal setting and action planning within individual consultations, drawing on frameworks such as motivational interviewing and self-determination theory.

For readers

This review confirms that patient empowerment is measurable and improvable through structured interventions, but highlights a critical need for consensus on its definition, better-quality trials outside diabetes, and validated measurement tools.