Table 9 Details of studies reporting negative or null findings for effects of peer support

Chan et al., 2014 [70]

Patients with well controlled diabetes trained to provide telephone support for others with diabetes. Protocol called for 12 calls over 1 year

RCT of standardized care versus standardized care plus peer support.

No significant differences between groups on clinical measures

Standardized care provided high quality clinical care including initial reports on medication adherence, self management, recommendations for physicians and patients; periodic status updates and recommendations to patients [90]

Subgroup Effects As noted in text, among approximately 20% above norms for distress, the combination of peer support and standardized care substantially reduced distress and hospitalizations.

Chen et al., 2010 [42]

Health coaches (medical assistants or health workers) paired with 1st-year residents in language concordant, stable teamlets for patients with hypertension and/or diabetes

Compared to usual care in teams of 2nd- and 3rd-year residents

Improvements in intended process measures (e.g., assessment of LDL, BMI, smoking; setting self-mgmt plan) but not patient clinical indicators (e.g., HbA1c, BP)

Implementation Problems

Not all received health coaching that was applied according to “time and prioritization of patients who were more complicated or needed more assistance” (p. S611).

Methodological Problems

Study designed to show feasibility within clinical setting rather than outcomes.

Non-equivalent controls. Potential contamination: (i) some 2nd and 3rd year residents having participated in pilot testing of health coaching and/or training of 1st year residents in chronic care, (ii) overlap in attendings for 1st and 2nd/3rd year residents, and (iii) nursing staff who provided health coaching also interacting “regularly with all clinic patients as medical assistants and health workers” (p. S613).

Graffy, 2004 [77]

Intervention to increase breastfeeding compared 1) PS and postnatal in-person visits plus phone calls on request vs 2) UC

Individual randomized design among sample of 720 from among 844 eligible mothers.

No differences in self reported % breast feeding initially or at 4 or 6 mos post-partum

Lack of Acceptance

Although initial antenatal contact by PS counselors achieved in 80% of those randomized to receive it, post-natal contact only if initiated by mothers and occurred for only 62% (p. 3)

Hunkeler et al. 2000 [38]

Augmentation of antidepressant therapy compared 1) Nurse telehealth care including medical and emotional support and advice over the phone plus PS involving in-person support and telephone calls by peers recovered from depression to 2) Nurse telehealth alone, to 3) UC

RCT compared Nurse telehealth + PS to Nurse telehealth alone and UC among 302 drawn from 370 eligibles; 68 refused informed consent.

No differences reported on Hamilton Depression Rating Scale or Beck Depression Index or on SF-12 Mental & Physical Composite Scales at 6-week or 6-month follow-up.

Other Sources of Support Control included Nurse telehealth care including medical and emotional support and advice

Lack of Acceptance: Of 62 Randomized to Nurse telehealth + PS, 31 (50%) had one or fewer contacts among whom 20 (32.3%) had no contacts – refused (11) or no contact (9). 13 had 2 contacts, 14 had 3–5, 4 had 9–20. Only 6 had at least 1 face-to-face

Kaplan, 2011 [80]

Tested unmoderated, unstructured Internet peer support for individuals with serious mental illness.

RCT compared

Internet peer support via listserv, via bulletin board, or control. 300 with Schizophrenia Spectrum or Affective Disorder

No differences on measures of recovery, quality of life, empowerment, social support, or distress

Possible Harm of Unmoderated PS Unmoderated listservs may be inappropriate for those with serious mental illness (as in present case) or other highly stressful diseases or conditions.

May, 2006 [74]

Group-based smoking cessation plus support from another group member in person and by telephone vs. group treatment without group member support

630 randomized to 34 grps (14 with peer support, 20 without). 96 excluded for failure to attend visit 2 quit date.

1-week post quit, borderline (p = 0.06) difference in abstinence (need to check) favored peer support (OR = 1.45, 0.92–2.29).

No other differences between groups.

Other Sources of Support All participants, including controls, participated in group program for smoking cessation.

Lack of Acceptance Peer support appears to have been first introduced in visit 2 quit date and moderately embraced by participants: mean = 2.7 phone calls in first week after quit date, dropped to 1.2, 1.1 and 0.7 in following weeks. (p. 240)

Muirhead, 2006 [75]

“Normal breastfeeding support” (involving a community midwife for 10 days, breastfeeding support groups and breastfeeding workshops) vs Normal breastfeeding support plus the assistance of two peer supporters

Of 284 pregnant women recruited through a physician practice, 59 declined and 225 were randomized to conditions.

No differences in self-reported breastfeeding initiation or duration at 10 days, 8 weeks or 16 weeks post-partum.

Other Sources of Support

Hospital midwives helped mothers in both groups to initiate breastfeeding.

With extensive support in “normal breastfeeding support,” peer support challenged to add additional influence.

Implementation Problems Peer supporters had little/no contact in hospital but were available to women after returning home and if peer supporters were informed in time. Mothers still breastfeeding on return home contacted by peer supporters every 2 days or as often as required either by phone or a personal visit up until day 28. If requested by the mother, the same peer supporters provided further support after 28 days until 16 weeks. (pp. 193–194) No data on actual contacts.

“…Women in the peer support group who did not commence breastfeeding or who stopped while still in hospital received no peer support postnatally” but were included in intention-to-treat analyses of outcomes (p. 196).

Lack of Acceptance “…half of the women in the population simply did not want to breastfeed....”

Lack of acceptance among professionals: “The support and cooperation of health professionals is required for peer supporters to function, and some may be unwilling to accept lay people being involved in the care of women” (p. 196).

Methodological Problems Possible social desirability bias of outcomes in that 10-day assessment surveys completed in presence of health visitor and both 8- and 16-week assessments completed in presence of physician or practice nurse. (p. 194)

Nicholas, 2007 [81]

34 family caregivers of technology-assisted children with chronic lung disease were recruited from a patient database and assigned to dyads for information sharing and support.

Non-randomized and no comparison group.

No significant within group differences over time for perceived social support from friends or family, caregiver stress, coping or social isolation (Meaning of Illness Questionnaire, Coping Health Inventory for Parents

Possible Harm of Unmoderated PS Family caregivers were already under substantial stress. May have been unrealistic to expect them to support each other as opposed to receiving support from a trained supporter. Dyads may fit into pattern of lack of effect for unmoderated support among those with highly stressful diseases or conditions.

Paper did identify effects through qualitative study

Palmas 2014 [76]

For adults with diabetes and elevated HbA1c (> 8%), 12-month CHW intervention included one-on-one visits, group visits, and telephone follow-up.

RCT of PS versus enhanced usual care

No significant differences between groups on clinical outcome measures.

Lack of Acceptance “…in over half of the intervention group, the CHWs were not able to deliver any of the planned one-on-one or small group sessions and only able to contact participants by phone” (p. 968). Adjustment for number of contacts led to a borderline (p = 0.054) effect for the CHW intervention (p. 967).

Salzer, 2010 [82]

Unmoderated internet peer support listserv for women recently diagnosed with breast cancer

Random assignment to peer support listserv or Internet-based educational control condition

Data collection at baseline, 4 and 12 months.

Control group showed significantly greater effect on FACT-B (4 and 12 months, ps < .05). No significant differences between groups on MOS Social Support Scale.

Possible Harm of Unmoderated PS Unmoderated listservs may be inappropriate for those with highly stressful physical illness like newly diagnosed cancer (as in present case) or those with other highly stressful diseases or conditions

Simoni, 2007 [78]

Appraisal, spiritual, emotional, and informational adherence-related support vs. UC to improve antiretroviral medication adherence and depressive symptomatology in HIV+ men and women.

Support = 6, semi-monthly group meetings and weekly contact by peer supporters over 3 mos.

136 participants enrolled (71 randomized to peer intervention and 65 to UC). 53% of eligible patients approached declined to participate… [due to] lacking interest, being too busy, transportation difficulties... or being asocial.” (p. 491)

No significant within or between group differences for adherence based on Electronic Drug Monitoring.

Relationship between attendance and lower deppressive Sx at 6 mos

Lack of Acceptance Those assigned to support condition attended average of 2.1 of 6 meetings. 23% attended none, 26% attended 2, and only 17% attended 5 or 6 of 6 peer meetings. (p. 491) Average number telephone contacts for intervention participants was 5.8 (Range = 0 to 17). (p. 492)

Concern for confidentiality may have suppressed participation in group sessions.

Simmons et al. 2015 [67]

Support by trained patients with diabetes in individual, group, or individual plus group modes over 8–12 months

2 × 2 factorial randomised cluster design of individual peer support, group peer support, individual plus group, or usual care

No significant differences between peer support versus usual care in changes on clinical indicators.

Lack of Acceptance “only 61.4% (592/977) of intervention participants attended an actual peer support session.” Implementation may have been compromised by scope: 127 peer support facilitators in group, individual, and combined group and individual arms, 2 × 2 factorial, cluster randomized design with 1299 randomized participants drawn from three counties and “… 62 general practices, a hospital clinic and Diabetes UK members.” One nurse served as the principal study manager.

Subgroup Effects Significant differences favoring group or group plus individual for systolic blood pressure.

Smith et al. 2011 [58]

Peer led groups including presentation and discussion of topics in diabetes management. 9 sessions scheduled over 2 years

Cluster randomized design. Practices assigned to peer-leg groups or standard care

No significant between-group differences in primary (HbA1c blood pressure, cholesterol) or secondary (BMI) outcomes.

Lack of Acceptance Mean of 5 of 9 sessions attended; 18% attended 0

Questionable whether peer support provided, e.g., infrequent meetings (9 over 24 months) that appear to have been focused on discussion of topics in diabetes management; participants discouraged from contacting peer leaders between meetings

Vilhauer, 2010 [79]

Unmoderated online support among women with metastatic breast cancer. Women sent an introductory email with instructions on how to access group and basic ettiquette.

From over >900 mailings and phone calls to oncologists, breast cancer clinics, and support centers, 42 women replied and 31 determined eligible. Nonrandom assignment to three online support groups to restrict group membership to 10 or 11. Compared to wait-list control.

Among controls, significant within group differences in breast cancer related distress (FACT-B breast cancer subscale, p < .04) and in daily activity (ECOG, p < .04) over first month. At 2 months, control group reported higher activity scores (ECOG, p = .02).

Lack of Acceptance 31 from over 900 mailings engaged in online resource.However, 73% retention rate and average participation of 5.69 days/wk. Average 82 min spent reading messages per week and average 69 min spent writing messages per week.

Possible Harm of Unmoderated PS Unmoderated listservs may be inappropriate for those with highly stressful physical illness like metastatic breast cancer (as in present case) or with other highly stressful diseases or conditions