Table 3.

Summary table of findings of PRISMS systematic reviews and their relevance to the meta-review questions

Reference and weighting*; RCTs, n; Participants, n; R-AMSTAR; Date range of included RCTs	Comparison	Relevance to meta-review questions:	Interventions included	Target group(s)	Synthesis	Main results
		What is the impact? Target groups? Which components? Context?
Bailey 2009 [25]** 4 RCTs 617 participants R-AMSTAR 36 RCTs 2000–2008	Culturally orientated programmes vs. usual care or limited/generic education. FU (mode): 12 mo, range 4–12 mo	Impact Target: Ethnic groups	Education, action plans, triggers and avoidance, collaboration with healthcare services. Language-appropriate asthma educators.	Minority groups: Puerto Rican, African-American, Hispanic, Indian sub-continent. Adults and children.	Meta-analysis Narrative analysis	Reduced hospitalisation in children (RR 0.32, 95% CI 0.15–0.70; 1 RCT) but not reported in adults. Improved QoL in adults (WMD 0.25, 95% CI 0.09–0.41; 2 RCTs). 2 of 2 RCTs reported a reduction in A&E visits and hospitalisations: one reported no difference in ‘use of healthcare resources’; 2 of 3 reported improved QoL (adults).
Bernard-Bonnin 1995 [26]** 11 RCTs 1290 participants R-AMSTAR 27 RCTs 1981–1991	Interactive teaching on self-management vs. standard care.	Impact Target: Children	Interactive teaching (one-to-one or group) to support asthma self-management.	Children 1–18 y. Overall severity classified as ‘mild to moderate’.	Meta-analysis Narrative analysis	Reduced hospitalisation (ES 0.06 ± −0.08) and emergency visits (ES 0.14 ± 0.09); 5 RCTs. Children with high baseline numbers of hospitalisations and emergency visits had greatest subsequent reduction in morbidity.
Bhogal 2006 [23]** 4 RCTs 355 participants R-AMSTAR 41 RCTs 1990–2004	Symptom-based written PAAPs vs. peak flow-based PAAP. FU (mode): 3 mo, range 3–24 mo	Target: Children Components: PEF vs. symptom monitoring	Asthma education plus PAAPs for both parents and children. Generally contained 3 steps: often employing ‘traffic lights’. Monitoring varied: either daily or when symptomatic.	Children 6–19 y with mild to severe asthma.	Meta-analysis	Symptom-based PAAPs reduced unscheduled care compared to peak flow-based PAAPs (RR 0.73, 95% CI 0.55–0.99; 4 RCTs). No difference in hospital admissions (RR 1.51, 95% CI 0.35–6.65. Peak flow-based PAAPs reduced the number of symptomatic days/week (MD 0.45 days/week, 95% CI 0.04–0.26; 2 RCTs). No significant difference for adult or child QoL.
Zemek 2008 [24]** 5 RCTs 423 participants R-AMSTAR 41 RCTs 1990–2005	Written PAAPs vs. no PAAP. Symptom-based vs. PEF-based PAAP. FU (mode): 3 mo, range 0.5–24 mo	Impact: Target: Children Components: PAAP	Education for parents and children, plus PAAPs, with 3 steps: often employing ‘traffic lights’. Monitoring varied: either daily or when symptomatic.	School-aged children with mild to severe asthma.	Meta-analysis	A PEF-based PAAP reduced unscheduled care compared to no plan (WMD −0.50, 95% CI −0.83 to −0.17; 1 RCT). A PEF-based PAAP compared to no plan reduced symptom scores (WMD −11.80, 95% CI −18.22 to −5.38) and number of school days missed (WMD −1.03, 95%CI −1.85 to −0.21; 1 RCT).
Boyd 2009 [27]*** 38 RCTs 7843 participants R-AMSTAR 39 RCTs 1985–2007	Education targeting children/parents vs. low intensity education. FU (mode): 12 mo range 4–12 mo	Impact: Target: Children, A&E attendees	Education plus therapy review, self-monitoring, PAAPs, and trigger avoidance. Range of settings and professionals and mode of delivery.	Children 0–18 y who had attended A&E for asthma within the previous 12 mo.	Meta-analysis Subgroup analyses	Education reduced A&E attendances (RR 0.73, 95% CI 0.65–0.81; 17 RCTs), admissions (RR 0.79, 95% CI 0.69–0.92; 18 RCTs) and unscheduled consultations (RR 0.68, 95% CI 0.57–0.81; 7 RCTs). No effect on QoL (WMD 0.13, 95% CI 0.73–0.99; 2 RCTs). Subgroup analyses (type and timing of intervention, timing of outcome assessment or age of participants) did not change findings.
Bussey Smith 2009 [28]* 9 RCTs 957 participants R-AMSTAR 26 RCTs 1986 - 2005	Computerised education vs. traditional self-management FU (mode): 12 mo, range 3–12 mo	Impact: Components: Technology-based interventions	Interactive computerised educational asthma programmes (games tailored to the individual, web-based education, interactive communication devices).	Patients 3–75 y. 7 RCTs in children, 2 in adults; 4 RCTs in urban or inner-city populations.	Narrative analysis	1 of 4 improved hospitalisation, and 1 of 5 reduced unscheduled care. 5 of 9 studies found statistical improvements in asthma symptoms compared to control.
Chang 2010 [29]** 1 RCT 113 participants R-AMSTAR 40 RCT 2010	Education by IHWs vs. education no IHW. FU: 12 mo	Impact: Target: Ethnic groups	Initial clinical consultation, reinforced by home visits from a trained IHW. Personalised, child-friendly, culturally appropriate education materials.	African-American and Hispanic communities. Children 1–17 y; mean ~7 y.	Narrative analysis	There was no effect on hospitalisations (OR 1.58, 95% CI 0.37–6.79) or A&E attendances (OR 0.30, 95% CI −0.17 to 0.77; 1 RCT). Days absent from school were reduced by 21% in the intervention group (95% CI 5–36%; 1 RCT). Carer asthma QoL was not significantly different (MD 0.25, 95% CI −0.39 to 0.89).
Coffman 2009 [30]** 18 asthma RCTs 8077 participants R-AMSTAR 29 RCTs 1987-2007	School-based asthma education vs. usual care.	Impact: Target: Schoolchildren	School-based education on asthma, medication, monitoring, avoiding triggers. Delivered by nurses, health educators, peer counsellors, teachers, ± computer programmes.	Children 4–17 y. Severity: mild to severe, majority were Black or Latino.	Narrative analysis	Unscheduled healthcare was not reported. School absences significantly reduced in 5 of 13 RCTs. Days with symptoms were reduced in 3 of 8 RCTs. Nights with symptoms improved in 1 of 4 RCTs: 1 found improvement in the control group. QoL improved in 4 of 6 RCTs.
Gibson 2002 [31]*** 36 RCTs 6090 participants R-AMSTAR 39 RCTs 1986 –2001	Self-management programmes vs. usual care.	Impact: Components: Regular review Context: LTC care	Education (100%); self-monitoring of symptoms or PEF (92%); regular review by a medical practitioner (67%); PAAP (50%). Subgroup analyses based on these service models.	Adults and children. Range of settings, including hospital, emergency room, outpatients, community setting, general practice.	Meta-analysis Subgroup analysis	Self-management reduced hospitalisations (RR 0.64, 95% CI 0.50–0.82; 12 RCTs), A&E visits (RR 0.82, 95% CI 0.73–0.94; 13 RCTs] and unscheduled consultations (RR 0.68, 95% CI 0.56–0.81; 7 RCTs). Self-management reduced days off work/school (RR 0.79, 95% CI 0.67–0.93; 7 RCTs) and improved QoL (SMD 0.29, 95% CI 0.11–0.47; 6 RCTs). Optimal self-management (supported by a PAAP and regular review) reduced hospitalisations (RR 0.58, 95% CI 0.43–0.77; 9 RCTs), and A&E visits (RR 0.78, 95% CI 0.67–0.91; 9 RCTs).
Gibson 2004 [32]*** 26 RCTs 6090 participants R-AMSTAR 39 RCTs 1987–2002	Different components of written PAAPs vs. usual care.	Components: PAAPs	Complete PAAPs specified when/how to increase treatment (n = 17); incomplete omitted advice on increasing ICS (n = 4); non-specific (n = 5) only had general instructions.	Adults and children. Variety of settings, including hospital, emergency room, outpatients, community setting, general practice.	Action points % predicted vs. % best Treatment advice Non-specific plans	Benefits were found for any number of action points (2 to 4). Both % predicted and % best reduced hospitalisations, but only % personal best reduced A&E visits. PAAPs which included advice on increasing ICS and starting oral steroids reduced hospitalisations and A&E visits. Efficacy of incomplete and non-specific PAAPs was inconclusive.
Moullec 2012 [33]** 18 RCTs 3006 participants R-AMSTAR 27 RCTs 1990–2010	Interventions to improve inhaled steroid adherence vs. usual care. FU (mode): 12 mo, range 0.25–24 mo	Context: LTC care	All studies included self-management; some included components of CCM: decision support, delivery system design, clinical information systems.	Moderate to severe asthma (one RCT included COPD). Aged 35–50 y. Women over-represented.	Meta-analysis	Effect size for adherence to ICS compared by number of components of the CCM in the study: 1 CCM component (n = 13): small ES 0.29 (95% CI 0.16–0.42) 2 CCM components (n = 5): large ES 0.53 (95% CI 0.40–0.66) 3 CCM components (no studies) 4 CCM components (n = 4) very large ES 0.83 (95% CI 0.69–0.98).
Newman 2004 [34]** 18 asthma RCTs (of 63 RCTs) 2004 participants R-AMSTAR 23 RCTs 1997 –2002	Self-management interventions vs. standard care/basic information.	Impact:	Individual/group interventions, focused on symptom monitoring, trigger avoidance and adherence to medication. A few used techniques to address barriers to effective self-management.	Adults with 3 LTCs (including asthma).	Narrative analysis and comparison between interventions	7 of 11 studies reported a reduction in unscheduled healthcare. 6 of 12 studies reported improved QoL. 3 of 8 studies reported reductions in severity of symptoms, all used education and action plans. 8 of 14 reported improved adherence.
Postma 2009 [35]** 7 RCTs 2316 participants R-AMSTAR 23 RCTs 2004–2008	CHWs vs. usual care. FU (mode): 12 mo, range 4–24 mo	Impact: Target: Ethnic groups, children	CHWs from the same community as participants. Education on asthma, lifestyle and trigger avoidance, with resources to reduce allergen exposure.	Children 5–9 y with allergies and low-income. Mainly African-American and Hispanic.	Narrative review	3 of 6 studies reported reduced hospitalisation and reduced unscheduled consultations. 4 of 6 reported reduced A&E attendances ‘Consistent and significant decrease in caregiver-reported asthma symptoms among intervention subjects compared with control subjects in 6 studies.’
Powell 2009 [36]*** 15 RCTs 2460 participants R-AMSTAR 34 RCTs 1990–2001	Self-management vs. physician-reviewed management. Comparison of modified PAAPs.	Components: PAAP, regular review Context: LTC care	Self- vs. physician adjustment of medication (n = 6 studies). PEF vs. symptoms PAAPs (n = 6). Other variations (n = 3).	Adults with asthma recruited from a range of primary, community, A&E and secondary care.	Self- vs. physician management Symptoms vs. PEF-modified PAAPs	Of 6 studies: 4 reported no difference in hospitalisation, 1 reported no difference in A&E visits, 3 reported inconsistent effects on unscheduled consultations. Of 6 studies, 6 reported no difference in hospitalisation, 5 reported inconsistent effects on A&E visits. Omitting regular review (1 RCT) or reducing intensity of education (1 RCT) increased unscheduled consultations. Verbal (vs. written) PAAPs had no effect on hospitalisations or A&E visits (1 RCT).
Ring 2007 [37]*** 14 RCTs 4588 participants R-AMSTAR 35 RCTs 1993– 2005	Interventions encouraging use of PAAPs vs. usual care.	Context: Organisation of care	Interventions promoting PAAP ownership or use. Diverse interventions (educational, prompting, asthma clinics, asthma management systems, quality improvement).	Adults or children with moderate to severe asthma; some post-exacerbation.	Narrative analysis	4 of 5 studies of education, 1 of 2 studies of telephone consultations, 1 of 2 studies of asthma clinics and 1 of 2 studies of asthma management systems reported increased PAAP ownership. 1 study of self-management education, 1 of 2 studies of telephone consultations and 1 of 2 studies of asthma management systems increased understanding/use of PAAPs.
Tapp 2007 [38]*** 13 RCTs 2157 participants R-AMSTAR 39 RCTs 1979–2009	Asthma education at A&E visit vs. usual care. FU (mode): 6 mo, range 2–18 mo	Impact: Target: Post A&E attendance	Asthma education provided by asthma or A&E nurses within a week of A&E visit included PAAPs, triggers, monitoring, inhalers and medication.	Adults recruited during A&E attendance.	Meta-analysis Narrative analysis	The intervention reduced hospital admissions (RR 0.50, 95% CI 0.27–0.91; 5 RCTs), A&E visits (RR 0.66, 95% CI 0.41–1.07; 8 RCTs). Effect on QoL (2 RCTs) was inconsistent. There was no effect on days off work/school.
Toelle 2004 [39]** 7 RCTs 967 participants R-AMSTAR 38 RCTs 1990– 2001	Written PAAP vs. no plan. Symptom vs. PEF-based PAAP. FU (mode): 12 mo, range 6–12 mo	Components: PAAP	Peak flow-based written PAAP or symptom-based written PAAP delivered in primary or tertiary care.	Adults 28–45 y and children in 1 RCT.	Meta-analysis Subgroup analysis	Unscheduled healthcare: assessed in 1 RCT, not reported by systematic review. No difference between symptom and peak flow-based PAAPs in hospitalisations (RR 1.17, 95% CI 0.31–4.43; 3 RCTs) or A&E attendances (RR 1.17, 95% CI 0.31–4.43; 3 RCTs). Symptom-based PAAPs were more effective at reducing unscheduled consultations (RR 1.34, 95% CI 1.01–1.77; 2 RCTs).
Welsh 2011 [40]*** 12 RCTs 2342 participants R-AMSTAR 41 RCTs 1986–2010	Home-based self-management vs. routine care or general education. FU (mode): 12 mo, range 6–24 mo	Impact: Target: Children	Language-appropriate education (asthma, triggers, medication, inhalers, self-management with PAAPs). Also homework, technology devices, 24-hour hotline.	Children (mostly <12 y) recruited from recent healthcare visit. Mainly ethnic and/or deprived communities in USA.	Meta-analysis Narrative analysis	No difference between groups in mean number of A&E visits (MD 0.04, 95% CI −0.20 to 0.27; 2 RCTs). 2 of 5 studies reported hospitalisation: one found a reduction and one an increase in the intervention group. Effect on A&E visits (6 RCTs) was inconsistent. Overall no effect on QoL was found in 5 studies.
Bravata 2009 [41]*** 63 RCTs 13,476 participants R-AMSTAR 40 RCTs 1966–2006	Self-management QI vs. other QI strategies.	Impact: Target: Children	Self-monitoring or self-management. Patient/caregiver education. Provider education. Organisational change and interventions with multiple QI strategies.	Children <18 y.	Meta-analysis	Interventions targeting parents/caregivers reduced hospitalisation rates by 1.2% per year (95% CI 0.1–2.4; n = 5). Self-management intervention studies improved symptom-free days by 2.8% (95% CI 0.6–5.0), which equalled 0.8 days per month (n = 7); and reduced monthly school absenteeism by 0.4% (95% CI 0–0.7), which equalled 0.1 day per month (n = 16). Longer duration of intervention increased the effect on school absences.
Denford 2014 [43]*** 38 RCTs 7883 participants R-AMSTAR 36 RCTs 1993–2000	Asthma self-care vs. usual/less intensive intervention. FU (mode): 12 mo, range 3–18 mo	Impact: Components: Behaviour change	Commonest behavioural change techniques including: self-monitoring (n = 30), instruction (n = 27), goal-setting (n = 26) and inhaler technique (n = 24).	Adults ≥18 y with a diagnosis of asthma.	Meta-analysis	Intervention group participants had reduced asthma symptoms (SMD −0.38, 95% CI −0.52 to 0.24; 27 RCTs) and unscheduled healthcare use (OR 0.71, 95% CI 0.56–0.9; 23 RCTs). Increased adherence to preventative medication compared to control (OR 2.55, 95% CI 2.11–3.10; 16 RCTs).
de Jongh 2012 [42]** 1 asthma RCT (of 4) 16 participants R-AMSTAR 35 RCTs 1993–2009	Mobile phone messaging for self- management vs. usual care. FU: range 4–12 mo	Components: Mobile phone messaging	Self-management interventions delivered by mobile phone messaging.	Participants of all ages, gender or ethnicity. Included any LTC (one asthma study).	Narrative synthesis	In the single asthma study, there were fewer admissions (2 vs. 7) but more unscheduled consultations (21 vs. 15) in the intervention group compared to the usual care group. The pooled asthma symptom score showed a significant difference between groups, favouring the intervention group (MD −0.36, 95% CI −0.56 to −0.17).
Kirk 2012 [44]** 10 asthma RCTs 2195 participants R-AMSTAR 23 RCTs 1995–2010	Self-care support vs. usual care. FU (mode): 12 mo, range 3–24 mo	Impact: Target: Children	Interventions aiming to help children take control of and manage their condition, promote their capacity for self-care and/or improve their health.	Children ≤18 y with a LTC: asthma (10 RCTs), cystic fibrosis (2) or diabetes (1).	Narrative synthesis	Of 8 RCTs, 2 reported fewer asthma admissions, 5 reported fewer A&E attendances and 2 of 3 reported fewer unscheduled consultations. Control improved in 5 of 8 RCTs. Qol improved in 2 of 5 RCTs.
Marcano Belisario 2013 [45]** 2 RCTs 408 participants R-AMSTAR 39 RCTs 2000–2013	Self-management apps vs. traditional self-management. FU: 6 mo	Components: Smartphone Apps	Self-management support interventions provided by smartphone app.	Adults with clinician-diagnosed asthma.	Narrative synthesis	Of 2 RCTs, 2 reported no difference in hospital admissions; 1 reported fewer A&E attendances compared to control; 1 found no difference in unscheduled GP consultations or out of hours consultations, but reduced primary care nurse consultations; 1 reported no difference in MD in Asthma Control Questionnaire scores between the intervention and control group at 6 months; 1 found improved QoL in the intervention group.
Press 2012 [46]*** 5 RCTs (of 15 studies) 1459 participants R-AMSTAR 34 RCTs 1950–2010	Interventions targeted at ethnic minority groups vs. usual care. FU (mode): 6 mo, range 0.25–32 mo	Impact: Target: Ethnic groups	Interventions targeting ethnic populations in US. 15 were education-based, 9 were system-level interventions, 5 were culturally tailored and community-based, 10 were hospital-based.	Adults ≥18 y. Ethnic minority groups: African-Americans (10 studies, Latinos (4 studies).	Narrative synthesis	An education intervention reduced A&E attendance in 2 of 4 RCTs and hospital admissions in 2 of 3 RCTs. Symptoms were not reduced in any of the 3 RCTs that measured control. QoL was improved in 3 of 4 RCTs that used an asthma-related QoL outcome.
Stinson 2009 [47]* 4 asthma RCTs (of 9 studies) 826 asthma participants R-AMSTAR 28 RCTs 1993–2008	Internet-based self-management vs. usual care. FU (mode): 12 mo, range 3–12 mo	Target: Children Components: Internet-based	Any Internet-based or enabled self-management intervention.	Children 6–12 y or adolescents 13–18 y with LTCs: asthma (4 RCTs), pain (1), encopresis (1), brain injury (1) or obesity (1).	Narrative synthesis	1 RCT reported no difference in hospitalisations compared to control, 1 RCT reported significant reductions in A&E visits and 1 of 2 RCTs showed fewer unscheduled consultations. 4 out of 4 reported significant improvement in a measure of control. 1 of 4 asthma RCTs reported a significant benefit on QoL.

Abbreviations: A&E accident and emergency, CCM chronic care model, CHW community health workers, CI confidence interval, COPD chronic obstructive pulmonary disease, ES effect size, FU follow-up, ICS inhaled corticosteroid, IHW indigenous healthcare workers, LTC long-term condition, MD mean difference, mo months, OR odds ratio, PAAP personalised asthma action plan, PEF peak expiratory flow, QI quality improvement, QoL quality of life, RR risk ratio, SMD standardised mean difference, WMD weighted mean difference, y years