Background
Low back pain
Why it was important to do this review
Methods
Aims and objectives
Design
Study eligibility criteria
Search strategy
Risk of bias assessment of included studies
Data extraction
Data synthesis
Results
Study flow
Study types
Study outcomes
Types of implementation interventions tested
Study | Intervention and control | Outcome measures | Results | Overall intervention success | ||
---|---|---|---|---|---|---|
Dey et al., 2004 [29] | IG: Educational outreach visit supported with PEM & access to back pain clinic for non-responders CG: No intervention/usual care | Primary outcome measure(s) not stated. | Only one in five outcomes was statistical significantly different between the IG and CG: Referral to PT or an education programme | |||
1. Radiograph request rate | 1. 5.1 % IG v 13.7 % CG ρ = 0.62 | |||||
2. Sickness certification rate | 2. 17.7 % IG v 19.2 % CG ρ = 0.74 | |||||
3. Prescription of opioids or muscle relaxants (% patients) | 3. 18.7 % IG v 18.7 % CG ρ = 0.99 | |||||
4. Referral rate to secondary care | 4. 3.4 % IG v 2.3 % CG ρ = 0.12 | |||||
5. Referral rate to PT or an education programme | 5. 26 % IG v 13.8 % CG ρ = 0.01* | |||||
Engers et al., 2004 [46] | IG: 2 h educational workshop involving role-play, focusing on the psychosocial model of LBP backed up with PEM CG: No intervention/usual care | Primary outcome measures: | 6 weeks: | 52 weeks: | 6 weeks: No statistically significant differences at 6 weeks in the direction of the IG. | 52 weeks: Only one in five outcomes was statistically significantly different at 52 weeks: Consulting an HCP other than GP |
1. Median pain intensity over last 7 days | 1. 2 IG v 1 CG ρ = 0.06 | 1. 1 IG v 1 CG ρ = 0.43 | ||||
2. Functional status: a. Median Roland Morris score | 2a. 7.5 IG v 5 CG ρ = 0.06 | 2a. 2 IG v 2 CG ρ = 0.91 | ||||
2. Functional status: b. Median days not functional in last 6 weeks | 2b. 7 IG v 2 CG ρ = 0.03 | 2b. 0 IG v 0 CG ρ = 0.37 | ||||
3. Median days of work absenteeism in the last 6 weeks | 3. 7 IG v 3 CG ρ = 0.01 | 3. 0 IG v 0 CG ρ = 0.66 | ||||
4. Healthcare use in the previous 6 weeks. | 4. 24 % return to GP and 40 % consulting other HCP IG v 17 % & 42 % CG ρ = 0.05 and 0.65 | 4. 0 % & 36 % IG v 1 % & 50 % CG ρ = 0.24 and 0.01* | ||||
Engers et al., 2005 [27] | IG: 2 h educational workshop involving role-play, focusing on the psychosocial of LBP backed up with PEM CG: No intervention/usual care | Primary outcome measure(s) not stated | Only a lower referral rate to a therapist achieved statistical significance | |||
1. Referrals to a therapist | 1. 19 patients or 36 % of patients IG v 25 or 76 % CG OR: 0.2; CI: 0.1 to 0.6; ρ ≤ 0.05* but actual value not stated | |||||
2. Prescription of pain medication on a time-contingent basis | 2. 19 or 62 % IG v 10 or 71 % CG OR: 0.7; CI: 0.6 to 6.3; ρ not stated | |||||
3. Prescription of paracetamol | 3. 7 or 23 % IG v 1 or 7 % CG OR: 4.8; CI: 0.1 to 181; ρ not stated | |||||
4. Prescription of NSAIDs | 4. 19 or 61 % IG v 10 or 71 % CG OR and CI not stated; ρ not stated | |||||
5. Adequacy of patient education rated across nine measures | 5. Explained that no specific cause could be detected: 33 or 64 % IG v 22 or 67 % CG OR: 0.9; CI: 0.3 to 2.8; ρ not stated Explained that back pain will ease by itself: 36 or 69 % IG v 23 or 70 % CG OR: 0.9; CI: 0.3 to 3.1; ρ not stated Explained that there is no harm: 44 or 85 % IG v 20 or 61 % CG OR: 3; CI: 0.6 to 16.3; ρ not stated Explained that it is better to accept the pain: 29 or 56 % IG v 17 or 52 % CG OR: 1.7; CI: 0.2 to 13.3; ρ not stated Handed out an information pamphlet: 11 or 21 % IG v 4 or 12 % CG OR: 1.6; CI: 0.4 to 7.8; ρ not stated Advised to stay active: 42 or 81 % IG v 16 or 49 % CG OR: 1.7; CI: 0.3 to 9.0; ρ not stated Advised to gradually increase activity: 45 or 87 % IG v 9 or 58 % CG OR: 3.3; CI: 0.7 to 17.0; ρ not stated Advised which activities to increase when: 11 or 21 % IG v 5 or 16 % CG OR: 2.1; CI: 0.7 to 11.5; ρ not stated Advised to stop activity when in pain: 9 or 17 % IG v 7 or 21 % CG OR: 2.1; CI: 0.7 to 11.5; ρ not stated | |||||
Matowe et al., 2002 [25] | Postal dissemination of guidelines. No CG as the study was a TIS | Primary outcome measure: Lumbar spine radiography request rate. | Mean request rate reduced by 7.7 from 147.8 in the first month with a 0.4 reduction trend over 13 months CI: −24.7 to 40.2; ρ not stated | The reported change did not reach statistical significance. | ||
Stevenson et al., 2006 [30] | IG: 5-h education session delivered by a local opinion leader CG: Standard in-service session on knee pathologies | Primary outcome measure: Change in PTs’ clinical practice measured using a standardised discharge summary questionnaire assessing time spent using modalities targeted for NSLBP | Number of patients/% of patients with the modality recorded as being used | Only changes in the advice to increase activity levels and attempting to change patient attitudes/beliefs about pain achieved statistical significance. 4 out of 6 primary outcome measures showed no significant difference between groups | ||
1. Advice about work situation | 1. 42 or 37 % IG v 15 or 35 % CG OR: 1.1; CI: 0.5 to 2.5; ρ not stated | |||||
2. Advice on return to normal duties | 2. 34 or 30 % IG v 13 or 30 % CG OR: 1.1; CI: 0.4 to 3.0; ρ not stated | |||||
3. Advice to increase activity levels | 3. 36 or 32 % IG v 7 or 16 % CG OR: 2.1: CI: 0.7 to 6.8; ρ ≤ 0.05* but actual value not stated | |||||
4. Encouraging early return to work | 4. 5 or 4 % IG v 1 or 2 % CG OR: 1.6; CI: 0.1 to 23.1; ρ not stated | |||||
5. Encouraging to undertake activities themselves | 5. 16 or 14 % IG v 18 or 42 % CG OR: 0.3; CI: 0.1 to 0.9; ρ not stated | |||||
6. Attempting to change patient attitudes/beliefs about pain | 6. 25 or 22 % IG v 4 or 9 % CG OR: 2.6; CI: 0.7 to 9.5; ρ ≤ 0.05* but actual value not stated |
Study | Intervention and control | Outcome measures | Results | Overall intervention success | |||
---|---|---|---|---|---|---|---|
Bekkering et al., 2005 [26] | IG: 2 x 2.5 hour group training sessions involving lecture and role play 4 weeks apart; postal dissemination of the guidelines and printed educational material CG: Postal dissemination of the guideline and printed educational materials only, | No one primary outcome measure was stated. Outcome measures were adherence to the guidelines as measured against five criteria recorded in the patient record: | In numbers of patients and % of patients: | Successful in altering practice as recorded in patient records. Statistically significant differences reported. | |||
1. Limiting number of treatment sessions | 1. 32 or 27% IG v 14 or 13% CG. OR 2.39; CI 1.12 to 5.12; p ≤ 0.05* but actual value not stated | ||||||
2. Setting functional goals | 2. 188 or 79% IG v 180 or 71% CG. OR 1.99; CI 1.06 to 3.72; p ≤ 0.05* but actual value not stated | ||||||
3. Using mainly active interventions | 3. 183 or 77% IG v 154 or 60% CG. OR 2.79; CI 1.19 to 6.55; p ≤ 0.05* but actual value not stated | ||||||
4. Giving adequate information | 4. 229 or 96% IG v 221 or 87% CG. OR 3.5; CI 1.35 to 9.55; p ≤ 0.05* but actual value not stated | ||||||
5. Making all 4 above recommendations. | 5. 96 or 42% IG v 75 or 30% CG. OR 2.05; CI 1.15 to 3.65; p ≤ 0.05* but actual value not stated | ||||||
Bekkering et al., 2005 [45] | As Bekkering et al., 2005 [45] above. | Primary outcome measures: | At 6 weeks: | At 12 weeks: | At 26 weeks: | At 52 weeks: | Unsuccessful. There were no statistically significant differences between the groups. ρ values not stated |
1. QBPDS score | 1. 24 IG v 23.5 CG | 1. 20 IG v 17.5 CG | 1. 16 IG v 11 CG | 1. 17 IG v 13 CG | |||
2. Pain rated on a numerical scale | 2. 3 IG v 3 CG | 2. 2 IG v 2 CG | 2. 2 IG v 1 CG | 2. 2 IG v 1 CG | |||
3. Sick leave in number of days off in the past 6 weeks | 3. 25.5 IG v 22.8 CG | 3. 9.6 IG v 9.8 CG | 3. 8.8 IG v 6.8 CG | 3. 9.8 IG v 5 CG |
Study | Intervention and control | Outcome measures | Results | Overall intervention success | |
---|---|---|---|---|---|
Becker et al., 2008 [44] | GPE: 3 MFE sessions with feedback for GPs MC: 3 MFE sessions with feedback for GPs and motivational counselling training for practice nurses CG: Postal dissemination of guidelines only | Primary outcome measure: | Primary outcome measure at 6 months: | Primary outcome measure at 12 months: | Partially successful. There was a statistically significant difference between the MC and CG at 6 months on the HFAQ but not at 12 months Days in pain at 6 months in the MC and GPE were also statistically less than the CG, with the GPE maintaining that difference at 12 months Quality of life score was statistically significantly higher in the MC than the CG at 12 months |
Mean score on the HFAQ | 72.94 GPE v 73.94 MC v 70.29 CG GPE v CG: CI: −0.704 to 6.007; ρ = 0.12 MC v CG: CI: 0.32 to 6.979; ρ = 0.032* | 72.96 GPE v 74.64 MC v 71.56 CG GPE v CG: CI: −2.224 to 5.017; ρ = 0.446 MC v CG: CI: −0.47 to 6.697; ρ = 0.088 | |||
Secondary outcome measures: | Secondary outcome measures at 6 months: | Secondary outcome measures at 12 months: | |||
1. Mean score on the FQPA | 1. 36.47 GPE v 36.29 v 33.51 CG GPE v CG: CI: −1.628 to 7.545; ρ = 0.203 MC v CG: CI: −1.784 to 7.347; ρ = 0.23 | 1. 46.43 GPE v 45.40 MC v 42.88 CG GPE v CG: CI: −1.452 to 8.543; ρ = 0.202 MC v CG: CI: −2.476 to 7.495; ρ = 0.396 | |||
2. Mean days in pain over the last 6 months. | 2. 63.34 GPE v 62.91 MC v 80.78 CG GPE v CG: CI: −26.833 to −6.034; ρ = 0.002* MC v CG: CI: −28.183 to −7.553; ρ = 0.001* | 2. 58.48 GPE v 61.57 MC v 71.32 CG GPE v CG: CI: −23.382 to −2.296; ρ = 0.018* MC v CG: CI: −20.198 to −0.689; ρ = 0.067 | |||
3. Mean days of sick leave in the previous 6 months | 3.12.99 GPE v 13.05 MC v 14.34 CG GPE v CG: CI: −5.972 to 3.287; ρ = 0.569 MC v CG: CI: −5.905 to 3.331; ρ = 0.584 | 3. 6.16 GPE v 6.46 MC v 9.27 CG GPE v CG: CI: −8.582 to 2.358; ρ = 0.256 MC v CG: CI: −8.463 to 2.837; ρ = 0.32 | |||
4. Quality of life score (Euro Qol) | 4. 66.59 GPE v 67.53 MC v 66.85 CG GPE v CG: CI: −2.864 to 2.355; ρ = 0.847 MC v CG: CI: −1.924 to 3.302; ρ = 0.602 | 4. 68.46 GPE v 70.37 MC v 67.65 CG GPE v CG: CI: −1.736 to 3.344; ρ = 0.535 MC v CG: CI: 0.185 to 5.26; ρ = 0.036* | |||
5. FABQ | 5. Not expressed as a result | 5. Not expressed as a result | |||
Bishop & Wing, 2006 [28] | IG1: Copy of the guidelines appropriate to management at that time frame sent at assessment; 0–4 weeks; 5–12 weeks and 12 weeks+ IG2: As the IG1 but with patient educational materials at the same stages. CG: No guidelines sent | Primary outcome measure not specified Outcome measures: | Results in percentage of patients: | Largely unsuccessful. The only statistically significant differences between groups were: recommending aerobic exercise in the IG2 v CG and bed rest=/<4 days recommended in the IG1 v CG. | |
At assessment: | At assessment: | ||||
History of initiating event | History of initiating event recorded: 87 % IG1 v 85 % IG2 v 89 % CG ρ values not stated | ||||
Prior history of a similar symptoms | Prior history of similar symptoms recorded: 30 % IG1 v 27 % IG2 v 24 % CG ρ values not stated | ||||
Neurological examination | Neurological examination recorded: 63 % IG1 v 71 % IG2 v 63 % CG ρ values not stated | ||||
Regional back examination | Regional back examination performed recorded: 93 % IG1 v 93 % IG2 v 91 % CG ρ values not stated | ||||
Reference to presence or absence of red flags | Reference to presence or absence of red flags recorded: 4 % IG1 v 4 % IG2 v 5 % GC ρ values not stated | ||||
0-4 weeks: | 0-4 weeks: | ||||
Exercise and reassurance given | Education and reassurance recorded as given: 10 % IG1 v 6 % IG2 v 7 % CG ρ values not stated | ||||
Aerobic exercise promoted | Aerobic exercise recommended recorded: 38 % IG1 v 53 % IG2 v 43 % CG IG2 v CG; ρ = 0.05* | ||||
Non-narcotic medication prescription | Appropriate medication prescribed: 85 % IG1 v 81 % IG2 v 77 % CG IG1 v CG ρ = 0.14; IG2 v CG ρ = 0.08* | ||||
Physical therapy modalities usage | Not reported ρ values not stated | ||||
Spinal mobilisation usage | Spinal manipulation usage: 2.5 % IG1 v 5 % IG2 v 6 % CG ρ values not stated | ||||
Bed rest of 4 days or less recommended | Best rest greater than 4 days recommended: 10 % IG1 v 18 % IG2 v17 % GC IG1 v CG ρ = 0.05*; IG2 v CG ρ value not stated | ||||
5-12 weeks: | 5-12 weeks: | ||||
Work conditioning programmes utilised | Not reported ρ values not stated | ||||
5-12 weeks discordant: | 5-12 weeks discordant: | ||||
Physical therapy modalities usage. | Continued use of PT: 41 % IG1 v 42 % IG2 v 43 % CG ρ values not stated Continued use of spinal manipulation: 3 % IG1 v 3 % IG2 v 3 % CG ρ values not stated | ||||
12 weeks + concordant: | 12 weeks + concordant: | ||||
Return to full or modified work. | Not reported ρ values not stated | ||||
12 weeks + discordant: | 12 weeks + discordant: | ||||
Continued passive therapy or spinal manipulation; recycling through treatments; use of programmes that had previously failed. | Not reported ρ values not stated | ||||
Goldberg et al., 2001 [31] | IG: Surgeon study group meetings; use of local opinion leaders; GP education sessions; printed educational materials; audit; patient educational materials; financial data analysis meetings. CG: Usual care. | Primary outcome measure: Lumbar spine surgery rate per 100,000 adults. | Net reduction in the IG of 20.9 surgeries per 100,000 adults or 8.9 % v CG; ρ = 0.01* | Successful. The difference between the IG and the CG was statistically significant | |
Kerry et al., 2000 [22] | IG: Guidelines and a covering letter posted at baseline; revised guidelines at 9 months with feedback on referral rates over previous 6 months CG: No guidelines sent | Primary outcome measure: Radiograph request rates. | 15 % reduction in the IG compared to 5 % increase in the CG. CI: 3 to 37; ρ ≤ 0.05* but actual value not stated | Successful. The difference between the IG and the CG was statistically significant | |
Schectman et al., 2003 [23] | E&F: 90-min education session, a copy of the guideline, audit report of patient care during prior year. Follow up at 6 months with a further audit report E&F&PE: As above but with addition of patient educational materials (printed and audio-visual) PE: Patient educational materials only CG: No intervention | Primary outcome measure not specified. Outcome measures: | Unsuccessful. There were no statistically significant differences between groups. | ||
1. Lumbar spine radiograph request rates | 1. 19 % E&F v 18 % CG ρ values not stated | ||||
2. CT/MRI request rates | 2. 5.6 E&F v 7.1 CG ρ values not stated | ||||
3. Sub-speciality referral rates | 3. 8.6 E&F to 7.1 CG ρ values not stated | ||||
4. PT referral rates | 4.10 E&F to 13 in CG No data reported for PE or E&F&PE ρ values not stated | ||||
Winkens et al., 1995 [47] | IG: Regular feedback (x 5 over 2 yrs 7 months) on audit of quantity and quality of diagnostic test referrals CG: No feedback. | Primary outcome measures: | Partially successful. No statistically significant differences between groups in lumbar spine request rate. There was a statistically significant difference between non-rational requests for lumbar spine radiographs pre and post intervention. ρ = 0.004* | ||
1. Radiograph request rate | 1. Not specified but IG before intervention 1128, 1212 at 1 year and 1127 at 2 years. ρ values not stated | ||||
2. Rate of non-rational requests. | 2. % total per Dr: Pre-intervention: 1.92 and 1.38 IG v 1.38 and 1.54 CG 1st data collection point: 0.96 IG v 1.39 CG 2nd DCP: 1.44 IG v 1.22 CG 3rd DCP: IG 0.91 v 1.65 CG 4th DCP: 1.04 IG v 1.24 CG 5th DCP: 1.21 IG v 1.67 CG |
Study | Intervention and control | Outcome measures | Results | Success | ||
---|---|---|---|---|---|---|
Eccles et al., 2001 [24] | IG1: A&F: − audit and feedback IG2: ER - educational reminders IG3: A&F + ER – both interventions. All IGs also received a copy of guidelines CG: Copy of the guidelines only | Primary outcome measure: Radiograph request rate. | −1.53 absolute reduction per 1000 patients ER (95 % CI −2.5 to −0.57) v CG ρ values not stated −0.70 A&F (95 % CI −1.3 to 0.9) v CG ρ values not stated No increased effect A&E + ER. | Successful. Statistically significant differences in lumbar spine radiograph request rate between groups | ||
Ramsey et al., 2003 [48] | IG: Educational reminders CG: Copy of guidelines only | Primary outcome measure: Monthly radiograph request rate for 12 months | Practice mean per month: 1.76 IG v 2.38 CG ρ values not stated 0.64 relative risk (95 % CI 0.43 to 0.96) IG ρ = 0.029* | Successful. Statistically significant differences between groups and no decay of effect over 12 months. | ||
Baker et al., 1987 [50] | IG: Special radiographic requisition form CG: None as a ITS | Primary outcome measure: Radiographic request rate. | Reduction of radiograph request rate from 1443 to 759 ρ values not stated | Successful. A 47 % reduction in the first year maintained for next 3 years | ||
McGuirk et al., 2001 [32] | IG: Special evidence based clinics staffed by motivated practitioners CG: Usual care | Primary outcome measures: | At 3 months: | At 6 months: | At 12 months: | Partially successful. The differences in the VAS at 3 and 12 months were statistically significantly different. The difference between groups in the SF-36 physical functioning at 12 months was statistically ignificantly different. None of these differences in the SF-36 social functioning or physical role were statistically significantly different between groups. The differences in the SF-36 bodily pain at 3 and 6 months were statistically significantly different between groups. |
1. Pain VAS | 2 IG v10 CG ρ = 0.001* | 3 IG v 4 CG ρ = 0.21 | 2 IG v 9 CG ρ = 0.042* | |||
2. SF-36 physical functioning | 1.02 IG v 1.00 CG ρ = 0.364 | 1.04 IG v 1.04 CG ρ = 0.197 | 1.07 IG v 0.91 CG ρ = 0.006* | |||
2. SF-36 social functioning | 1.14 IG v 1.13 CG ρ = 0.853 | 1.15 IG v 1.13 CG ρ = 0.269 | 1.15 IG v 1.15 CG ρ = 0.888 | |||
2. SF-36 physical role | 1.17 IG v 1.12 CG ρ = 0.939 | 1.17 IG v 1.12 CG ρ = 0.35 | 1.18 IG v 1.12 CG ρ = 0.782 | |||
2. SF-36 Bodily pain | 0.93 IG v 0.79 CG ρ = 0.027* | 1.01 IG v 0.90 CG ρ = 0.018* | 1.02 IG v 0.90 CG ρ = 0.123 |
Theoretical models underpinning the implementation intervention
Number and types of healthcare practitioners studied
Effects of the implementation interventions
Studies that used ‘one-off’ or single implementation interventions
Studies testing short-term implementation interventions with no ongoing implementation effort
Studies that tested ongoing implementation interventions
Studies that tested ongoing, intermittently reinforced, implementation interventions
Studies that tested consistently ongoing implementation interventions
Risk of bias
Risk of bias in RCTs, non-randomised controlled trials and controlled before-after studies
Risk of bias category | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Was the allocation sequence adequately generated? | U | L | L | L | L | L | L | L | L | U | L | H | U | H |
Was the allocation adequately concealed? | L | L | L | L | L | L | U | L | L | U | U | H | U | U |
Were baseline outcome measurements similar? | L | U | U | L | U | L | H | L | L | H | H | L | H | H |
Were baseline characteristics similar? | L | U | U | L | L | L | L | L | L | H | H | L | H | L |
Were incomplete outcome data adequately addressed? | L | L | L | L | H | H | L | H | H | L | U | H | H | H |
Was knowledge of the allocated intervention adequately prevented during study? | L | L | L | H | L | U | U | L | L | L | U | U | U | U |
Was the study adequately protected against contamination? | L | L | L | L | L | L | L | L | L | L | L | L | L | U |
Was the study free from selective outcome reporting? | L | L | L | L | L | L | L | L | L | L | L | L | L | L |
Was the study free from other risk of bias? | L | L | L | L | L | L | L | H | H | L | L | L | L | L |