From: Statistical design and analysis in trials of proportionate interventions: a systematic review
First author | Intervention | Tailoring variable and decision rules (response unless otherwise stated) | Primary outcome | Statistical analysis | Analysis of stages |
---|---|---|---|---|---|
Ell [54] | Stepped care, three steps: (1) based on patient preference, patients start PST or antidepressant medication, 8 weeks, (2) a different antidepressant medication or the addition of antidepressant medication or PST, 4 weeks, (3) considered for additional PST, augmentation of low-dose trazodone for insomnia and referral to speciality mental health care | 50% SCL-20 reduction | Depression remission was assessed by SCL-20 < 0.5 or PHQ-9 < 5 | Logistic regression model used to compare the odds of achieving clinically meaningful improvement between treatment groups | No |
Van’t Veer-Tazelaar [55] | Stepped care, four steps: (1) watchful waiting, (2) bibliotherapy, (3) PST, (4) antidepressant medication; stages were in 3-month cycles | CES-D < 16 | MINI/DSM-IV diagnostic status of depressive and anxiety disorders | Incremental effectiveness computed as the difference in the probability of a disorder-free period between groups | No |
Braamse [56] | Stepped care, two steps: (1) internet-based self-help programme, (2) contracting, individual face-to-face counselling, medication or referral to other services | PHQ-9 ≤ 10 and/or HADS < 8 and/or STAI < 40 | Psychological distress using HADS and physical role function using EORTC-QLQ-C30 | ANOVA | No |
Patel [57] | Stepped care, four steps: (1) psychoeducation, (2) antidepressants, (3) interpersonal psychotherapy in addition to antidepressants or an alternative to antidepressants for those who did not respond to them, (4) referral to psychiatrist | Varying | ICD-10 diagnosis | Chi-squared and t-tests; mixed-effects models for longitudinal data | No |
Gilliam [25] | Stepped care, two steps: (1) short therapist sessions and bibliotherapy, (2) longer therapist-directed sessions | Y-BOCS reduction ≥5 points plus a post-treatment score of ≤13 | Y-BOCS total score and the clinician’s CGI severity rating | Repeated measures ANOVA | No |
Kay-Lambkin [58] | Stepped care, four steps: (1) brief integrated CBT/MI intervention, one session, (2) four CBT/MI sessions, (3) four CBT/MI sessions, (4) four CBT/MI sessions | Varying | Depression and methamphetamine use | Small sample size, so no statistical analyses | No |
Richter [18] | Stepped care, six steps: incremental therapy included the following add-on therapies at 4-week intervals: aliskiren 150–300 mg once daily, hydrochlorothiazide 12.5–25 mg once daily and finally amlodipine 5–10 mg once daily, as needed | Meet the target blood pressure at 4-week intervals | Estimated cumulative probability of patients achieving blood pressure target | Probability of reaching the blood pressure target, assessed by estimating control rates of patients who reached target per visit using life-table survivor estimates at each visit; summaries presented of change in blood pressure per treatment step | Yes |
Mitchell [16] | Stepped care, three steps: (1) therapist-assisted self-help for 18 weeks, (2) fluoxetine until 1-year follow-up, (3) full CBT for 6 months | 70% or more reduction in frequency of purging by the end of Session 6 | Recovery (no binge eating or purging behaviours in the past 28 days); remission (no longer meeting DSM-IV criteria) | ANOVA with the site × treatment interaction | No |
Seekles [59] | Stepped care, four steps: (1) watchful waiting, 4 weeks, (2) guided self-help, (3) five short face-to-face PST sessions, (4) pharmacotherapy and/or specialised mental health care | IDS < 14 and HADS < 8 and WSAS < 6 | IDS and HADS | t-tests to compare scores between two groups | No |
Tolin [23] | Stepped care, two steps: (1) bibliotherapy, 6 weeks, (2) therapist-directed ERP sessions | Y-BOCS ≥5 and ≤13 | Y-BOCS and cost | Mixed-effects model | No |
van der Leeden [26] | Stepped care, four steps: (1) randomised to group or individual CBT sessions for children and parents, (2) five manual-based PCTA sessions, (3) additional five PCTA sessions | Children diagnosed with an anxiety disorder or who scored below the cut-off of the MASC | Change in proportion of children with any DSM-IV anxiety disorder | Percentages of children free of any anxiety disorder after each treatment phase and by intervention, e.g. intervention 1 only, 1 and 2, 1–3 and all combined; mixed-effects models for changes on the continuous variables | Yes |
Apil [60] | Stepped care, four steps: (1) watchful waiting, 6 weeks, (2) bibliotherapy self-help booklet, 6 weeks, (3) 12 individual CBT weekly sessions, (4) referral to physician or psychotherapist for any indicated treatment | CES-D ≤16 | Incidence of new depressive episode | Feasibility evaluated descriptively; chi-squared test used to test if selective drop-out biased results of incidence of a new depressive episode | No |
Karp [61] | Stepped care, two steps: (1) 6 weeks open treatment with venlafaxine xr 150 mg/day and supportive management, (2) 14 weeks in which non-responders are randomised to high-dose venlafaxine xr (up to 300 mg/day) with PST for depression and pain or high-dose venlafaxine xr and continued supportive management | PHQ-9 of ≤5 for 2 weeks and at least 30% improvement in the average numeric rating scale for pain | Univariate pain and depression response and both observed and self-report disability | Number needed to treat between two interventions; repeated measures mixed-effect models for self-reported and observed physical disability between the two interventions across time | No |
Dozeman [62] | Stepped care, four steps: (1) watchful waiting, 3 months, (2) activity scheduling, 3 months, (3) life review and consultation with GP, 3 months, (4) consultation with GP to discuss further treatment, 3 months | Improvement of ≥5 points on CES-D | Incidence of major depressive disorder or anxiety disorder using MINI | Incidence rate ratio using an unadjusted and adjusted Poisson regression analysis of MINI/DSM-IV depressive and anxiety cumulative incidence (1 = developed a disorder and 0 = remained disorder-free) on the treatment indicator | No |
Nordin [24] | Stepped care, two steps: (1) low-intensity stress-management intervention given to all patients, (2a) more intensive group stress management treatment, (2b) more intensive individual stress management treatment | Decrease in stress-related symptoms measured by IES or HADS from clinical levels to normal results | Subjective distress (intrusion and avoidance) assessed by IES | Repeated measures ANOVA (continuous variables) and chi-squared test (categorical variables) | No |
Jakicic [11] | Stepped care, six steps: (1) monthly group intervention session plus weekly mailed lessons and submission of self-monitoring diaries, (2) continue step 1 plus 10-minute monthly telephone contact, (3) step 2 plus second 10-minute telephone contact each month, (4) step 3 plus 1 individual in-person intervention contact per month, (5) step 4 plus meal replacement shakes and bars provided to replace one meal and one snack per day, (6) step 5 plus replace one telephone contact with second individual session per month; modified based on weight-loss achievement at 3-month intervals | Weight-loss goals 5% at 3 months, 7% at 6 months, 10% at 9 months, and remained at 10% at 12, 15 and 18 months | Change in weight over 18 months | t-test to compare mean weight loss between groups; mixed-effects models for longitudinal data | No |
Pommer [63] | Stepped care, three steps: (1) four sessions of extensive psycho-education, (2) a course on coping with depression and/or anxiety, 10 consultations, (3) coaching (six booster sessions on top of step 2) complemented with optional antidepressant and/or anxiolytic medication | PHQ-9 < 7 and/or GAD-7 < 8 | PHQ-9, GAD-7 and MINI | Chi-squared and t-tests; mixed-effects models for longitudinal data | No |
Lamb [17] | Stepped care, two steps: (1) Whiplash Book advice or active management advice, (2a) single session of physiotherapist advice or (2b) up to six sessions of physiotherapy | Non-response if persistent symptoms 3 weeks after emergency department attendance (WAD grades I–III) | Neck Disability Index | Mixed models to account for clustering effects from NHS trusts and therapists in step 2 | Yes |
Krebber [27] | Stepped care, four steps: (1) watchful waiting, 2 weeks, (2) guided self-help via internet or booklet, 5 weeks, plus six phone or email coaching sessions, (3) PST administered by a specialised nurse, (4) specialised psychological intervention or antidepressant medication chosen in cooperation between patient and care co-ordinator | HADS-A or HADS-D ≤ 7 | HADS | Repeated measures ANOVA (continuous outcomes); generalised estimating equations used to evaluate longitudinal changes | No |
Borsari [21] | Stepped care, two steps: (1) brief advice session, (2a) brief motivational intervention, (2b) assessment only | Non-response if student has heavy episodic drinking ≥4 and/or alcohol-related consequences ≥5 in the past month they were randomised to receive step 2 or control (assessment only) | Heavy episodic drinking and peak blood alcohol content | Comparison of outcomes at 3, 6 and 9 months between those assigned to (2a) or (2b) using generalised estimating equations for longitudinal data | Yes |
Watson [13] | Stepped care, three steps: (1) behavioural change counselling, one session, (2) motivational enhancement therapy, three sessions, (3) local specialist alcohol services | Three-item AUDIT-C <5 | Average drinks per day | Linear mixed model, to account for variation in GP practice and allocated therapist | No |
Oosterbaan [29] | Stepped care, two steps: (1) self-help course, (2) CBT in combination with antidepressant medication | CGI-S < 3 | % of patients responding to and remitting after treatment measured using CGI-S | Logistic mixed-effects models; analysis after steps 1 and 2 | Yes |
van Dijk [64] | Stepped care, four steps: (1) watchful waiting, (2) guided self-help, (3) PST, (4) referral to GP | PHQ-9 ≥ 6 | Cumulative incidence of DSM-IV major depressive disorder using MINI | Logistic mixed-effects models | No |
Arving [65] | Stepped care, two steps: (1) low-intensity stress management consisting of two counselling sessions over 6 weeks, (2) more intensive stress-management treatment consisting of 4–7 sessions | IES and HADS score at 6-week assessment not clinically significant | Avoidance and intrusions | Repeated measures ANOVA (continuous variables) and chi-squared test (categorical variables) | No |
Mattsson [28] | Stepped care, two steps: (1) self-help material, chat forum and FAQ section, (2) CBT | HADS subscale <7 at 1, 4 or 7 months after inclusion | HADS, 20% change as clinically relevant | Repeated measures ANOVA to compare intervention and control group regarding anxiety, depression, post-traumatic stress and health-related QoL | No |
Carels [12] | Stepped care, three steps: (1) group-based behavioural weight-loss programme, 6 weeks, (2a) behavioural weight-loss programme, 6 weeks or (2b) self-help, (3a) behavioural weight-loss programme, 6 weeks or (3b) self-help | Meet the 3% weight-loss target | % weight loss | Repeated measures ANOVA (continuous variables) and chi-squared test (categorical variables) to compare differences between treatment groups at the end of each stage and the end of the whole intervention | Yes |
van der Aa [66] | Stepped care, four steps: (1) watchful waiting, (2) guided self-help, (3) PST, (4) referral to GP | CES-D < 16 or HADS-A < 7 | MINI | Survival analysis and mixed-effects model | No |
Muntingh [67] | Stepped care, four steps: (1) guided self-help, (2) CBT, six sessions, (3) antidepressant medication prescribed by GP, (4) optimisation of medication in primary care or referral to secondary care | 50% reduction in BAI score and BAI ≤ 11 | BAI score | Difference in gain BAI gain scores from baseline; inverse probability weighting used, accounts for variation in receiving treatment | No |
Hamall [19] | Stepped care, three steps: (1) family resilience and well-being fact sheet, (2) family resilience and well-being activity booklet, (3) family resilience information support group or waiting list control | Step 2: parents eligible if have a child attending one of four selected outpatient clinics at the paediatric hospital. Step 3: eligible if K10 ≥ 15 | Parental well-being (K10); family functioning (McMasters Family Assessment Device); social connectedness (Medical Outcomes Study Social Support Survey); family beliefs | Descriptive statistics used for step 1. ANOVA for effect of booklet intervention for all participants in step 2 and sustained change tested using a repeated measures mixed-effects model for the participants who did not move into step 3. ANOVA to examine additional effect of the information support group relative to waiting list control group | Yes |
Gureje [68] | Stepped care, three steps: (1a) eight weekly psychoeducation and PST sessions, (1b) eight weekly psychoeducation and PST sessions plus doctor’s advice on treatment, (2a) four monthly psychoeducation and weekly PST sessions, (2b) eight weekly psychoeducation and PST sessions, (2c) consult doctor plus eight weekly psychoeducation and PST sessions, (3a) four monthly psychoeducation and weekly PST sessions, (3b) consult doctor plus eight weekly psychoeducation and PST sessions | Step 1: (1a) if PHQ-9 = 11–14, (1b) if PHQ-9 ≥ 18. Step 2: (2a) PHQ-9 < 11, (2b) PHQ-9 = 11–17, (2c) PHQ-9 ≥ 18. Step 3: (3a) PHQ-9 < 11, (3b) PHQ-9 ≥ 11 | Recovery from depression at 12 months as shown by PHQ-9 ≤ 6 | Mixed-effects regression model | No |
Stoop [69] | Stepped care, three steps: (1) four weekly psychoeducation individual meetings, (2) 10 weekly individual meetings covering the coping with depression/anxiety course, (3) advice to meet GP to discuss optional medication and six booster sessions during 6 months; followed by monitoring of symptoms of depression or anxiety if remission | PHQ-9 < 7 and/or GAD-7 < 8 | Symptoms of anxiety and depression after 12 months of intervention and 6 months post-intervention | ANCOVA and clinical significance in terms of effect size | No |
Stam [20] | Risk-factor-guided intervention including: (1) medication adjustment if three or more prescribed fall-risk-increasing drugs, (2) stepped care if anxiety disorder or depression, (3) exercise therapy if impaired functional mobility; those eligible for more than one intervention start them at the same time. Stepped care, four steps: (1) watchful waiting, 6 weeks, (2) guided self-help treatment, 6 weeks, (3) PST maximum six sessions, (4) referral to GP | GAD-7 < 10, PHQ-9 < 10, or positive PHQ-PD score | Dizziness-related impairment, assessed using the Dizziness Handicap Inventory | Mixed-effects models for longitudinal data to compare intervention and control groups, regardless of number of interventions; separate subgroup analyses for three groups that received one of three interventions | No |
Lock [15] | Adaptive intervention, intensive family coaching, consisting of FBT/IPC: four sessions of FBT plus three sessions of IPC | Weight gain ≥2.3 kg after FBT, proceed to IPC | Retentions and treatment use, suitability and expectancy, clinical outcomes, changes in parental self-efficacy | Feasibility and acceptability compared across the randomised groups (FBT versus FBT/IPC) using chi-squared test and t-test | No |
Schuurhuizen [70] | Targeted selection by a nurse (HADS ≥ 13 or Lastmeter ≥ 5), enhanced care (treatment managed by a trained nurse) and stepped care. Stepped care, four steps: (1) watchful waiting, 3 weeks, (2) guided self-help programme, 5–7 weeks, maximum six sessions in 10 weeks, (3) face-to-face PST, (4) psychotherapy, medication or a referral to other services (e.g. social work) | HADS < 13 | Psychological distress measured by HADS | ANCOVA for difference between groups; time patients entered stepped care and the response to treatment (progression or not) are accounted for via a covariate | No |
Haug [71] | Stepped care, three steps: (1) short psychoeducation, (2) 10 weeks’ internet-based self-help programme, (3) 12 weeks of individual CBT | Two out of three of the following criteria: (1) loss of primary diagnosis (SCID-I), (2) CSR ≤ 3 and reduced by at least two points, (3) for panic disorder, BSQ ≤ 2.5, and for seasonal affective disorder, SPS ≤ 25 | CSR, a 0–8 severity rating of the primary anxiety diagnosis | Multiple regression analyses enhanced with the full information; maximum likelihood estimation of missing data | No |
Salloum [72] | Stepped care, two steps: (1) three therapist-led sessions, 11 parent–child meetings at home over 6 weeks using a workbook, weekly brief phone support, online psychoeducation information and video demonstrations, (2) nine trauma-focussed CBT sessions | PTS ≤ 3, or TSCYC-PTS ≤ 39 and an CGI-I rating of 3, 2 or 1 | TSCYC-PTS | Linear mixed-effects models (continuous outcomes); generalised linear mixed-effects models (non-continuous outcome) for longitudinal data | No |
Painter [73] | Stepped care, five steps: (1) watchful waiting, (2) depression care team treatment suggestions (counselling or pharmacotherapy, considering participant preference), (3) pharmacotherapy suggestions after review of treatment history, (4) combination pharmacotherapy and speciality mental health counselling, (5) referral to speciality mental health | Non-response defined on five different measures: antidepressant adherence, counselling non-adherence, report of severe adverse effects, increase in PHQ-9 from baseline by ≥5 or <50% decrease from enrolment PHQ-9 | Quality-adjusted life years and percentage of participants with depression treatment response | Generalised linear models to calculate predicted expenditure for each participant to determine incremental cost; logistic regression models to compare the odds of achieving clinically meaningful improvement (SCL-20 improved by ≥50%) between groups | No |