Modelling treatment-effect heterogeneity in randomized controlled trials of complex interventions (psychological treatments)

We describe instrumental variable (IV) methods for the estimation of the 'dose'-response effects of psychological interventions in randomized controlled trials in which there is variability in the patients' adherence to the allocated therapy (that is, variability in the actual number of sessions of therapy attended) and also variability in the strength of the therapeutic alliance between patients and their therapists. The effect of the therapy on outcome is assumed to be a function of both the number of sessions attended and the strength of the therapeutic alliance, with no intervention effects in the absence of any sessions attended (an exclusion restriction) and the effect of the strength of the alliance being represented by a multiplicative term (interaction) in the treatment-effect model. The IV methods that are described allow for: (a) hidden confounding between sessions, alliance and outcome; (b) measurement errors in the alliance; and (c) that alliance is only measured in those receiving treatment. Three two-stage estimation procedures are illustrated, and their equivalence demonstrated, through Monte Carlo simulation and analysis of the results of an actual trial.     

Recommendations for reporting randomized controlled trials of herbal interventions: Explanation and elaboration

Controlled trials that use randomized allocation are the best tool to control for bias and confounding in trials testing clinical interventions. Investigators must be sure to include information that is required by the reader to judge the validity and implications of the findings in the reports of these trials. In part, complete reporting of trials will allow clinicians to modify their clinical practice to reflect current evidence toward the improvement of clinical outcomes. The consolidated standards of reporting trials (CONSORT) statement was developed to assist investigators, authors, reviewers, and editors on the necessary information to be included in reports of controlled clinical trials. The CONSORT statement is applicable to any intervention, including herbal medicinal products. Controlled trials of herbal interventions do not adequately report the information suggested in CONSORT. Recently, reporting recommendations were developed in which several CONSORT items were elaborated to become relevant and complete for randomized controlled trials of herbal medicines. We expect that these recommendations will lead to more complete and accurate reporting of herbal trials. We wrote this explanatory document to outline the rationale for each recommendation and to assist authors in using them by providing the CONSORT items and the associated elaboration, together with examples of good reporting and empirical evidence, where available, for each. These recommendations for the reporting of herbal medicinal products presented here are open to revision as more evidence accumulates and critical comments are collected.     

Estimating the population impact of preventive interventions from randomized trials

Growing concern about the limited generalizability of trials of preventive interventions has led to several proposals concerning the design, reporting, and interpretation of such trials. This paper presents an epidemiologic framework that highlights three key determinants of population impact of many prevention programs: the proportion of the population at risk who would be candidates for a generic intervention in routine use, the proportion of those candidates who are actually intervened on through a specific program, and the reduction in incidence produced by that program among recipients. It then describes how the design of a prevention trial relates to estimating these quantities. Implications of the framework include the following: (1) reach is an attribute of a program, whereas external validity is an attribute of a trial, and the two should not be conflated; (2) specification of a defined target population at risk is essential in the long run and merits greater emphasis in the planning and interpretation of prevention trials; (3) with due attention to sampling frame and sampling method, the process of subject recruitment for a trial can yield key information about quantities that are important for assessing its potential population impact; and (4) exclusions during subject recruitment can be conceptually separated into intervention-driven, program-driven, and trial-design-driven exclusions, which have quite different implications for trial interpretation and for estimating population impact of the intervention studied.     

Evolving early (pre-dementia) Alzheimer’s disease trials: Suit the outcomes to the population and study design

Assuming that some cases of Alzheimer’s disease (AD) could be prevented or delayed, prevention trials will be developed for this neurodegenerative condition. Initially, stakeholders will have to agree about the definition of prevention—true primary prevention, meaning the prevention of AD neuropathological changes; the prevention of clinical signs and symptoms that often augur AD; or preventing the progression of signs and symptoms to full-blown dementia. True primary prevention trials will have to rely completely upon neuroimaging or biomarker outcomes that reflect AD pathology. On the other hand, trials designed to prevent signs and symptoms of dementia will require researchers to agree on the phenomenology that would constitute an unequivocal endpoint: cognitive worsening on one or more measure compared to a normative group; development of Mild cognitive impairment (MCI); or development of Alzheimer’s dementia. Prevention trials utilizing any of these outcomes in the general public will be large, will have to utilize low risk public health interventions, and might therefore have only a small impact (treatment effect size), especially if the studies are too short or the study populations are too diverse. An alternative to interventions aimed at the general public would be any attempt to prevent signs and symptoms of dementia in individuals thought to be at an increased risk for clinical dementia. These trials could try to reduce the development of signs and symptoms of dementia in cognitively normal subjects, or they could try to prevent progression from some form of Mild Cognitive Impairment to AD, or they could have the more subtle goal of reducing the accumulation of subclinical deficits in MCI subjects. If the populations for these trials are limited to individuals who have abnormal laboratory and neuroimaging studies associated with AD neuropathology, the results will not generalize to biomarker-negative, at risk individuals, who are likely to constitute the majority of any clinically relevant study population. Outcome measures for each study design will depend upon the characteristics of the study.     

The design versus the analysis of observational studies for causal effects: parallels with the design of randomized trials

For estimating causal effects of treatments, randomized experiments are generally considered the gold standard. Nevertheless, they are often infeasible to conduct for a variety of reasons, such as ethical concerns, excessive expense, or timeliness. Consequently, much of our knowledge of causal effects must come from non-randomized observational studies. This article will advocate the position that observational studies can and should be designed to approximate randomized experiments as closely as possible. In particular, observational studies should be designed using only background information to create subgroups of similar treated and control units, where 'similar' here refers to their distributions of background variables. Of great importance, this activity should be conducted without any access to any outcome data, thereby assuring the objectivity of the design. In many situations, this objective creation of subgroups of similar treated and control units, which are balanced with respect to covariates, can be accomplished using propensity score methods. The theoretical perspective underlying this position will be presented followed by a particular application in the context of the US tobacco litigation. This application uses propensity score methods to create subgroups of treated units (male current smokers) and control units (male never smokers) who are at least as similar with respect to their distributions of observed background characteristics as if they had been randomized. The collection of these subgroups then 'approximate' a randomized block experiment with respect to the observed covariates.     

Preventing injuries through interventions for problem drinking: a systematic review of randomized controlled trials.

To assess the effect of treatment of problem drinking on injury risk, we conducted a systematic review of randomized controlled trials by searching 12 computerized databases, cross-checking bibliographies, and contacting authors and governmental agencies. We identified 19 trials of interventions for problem drinking that measured injury outcomes. Treatment for problem drinking was associated with reduced suicide attempts, domestic violence, falls, drinking-related injuries, and injury hospitalizations and deaths, with reductions ranging from 27 to 65%. Interventions among convicted drunk drivers reduced motor vehicle crashes and injuries. The precision of all the point estimates was low, however. We did not combine the results quantitatively, because the interventions, patient populations, and outcomes were so heterogeneous. The results suggest that treatment for problem drinking may reduce injuries and their antecedents. Because injuries account for much of the morbidity and mortality from problem drinking, further studies are warranted to confirm these effects.     

The utility of randomized controlled trials of social interventions: An examination of two trials of HIV prevention

This paper examines the utility and actual usage of evidence from trials of social interventions. It focuses on two case studies of trials of HIV prevention funded by the UK's NHS, one in a genitourinary medicine clinic, and another in a voluntary agency. The research concludes that a variety of contextual factors meant that the trials did not produce very useful findings. The interventions chosen were not always ideal candidates for evaluation. Although process evaluation did occur this was not reported in such a way as to enable consideration of the potential generalizability of the interventions. Although neither trial was led solely by academic researchers, certain stakeholders were excluded from the design and undertaking of both trials. Although findings from both trials were used in informing (or supporting) decisions about the planning and implementation of services, this did not occur in the manner that proponents of evidence-based practice might hope for.     

The feasibility of using a double blind experimental cross-over design to study interventions for sick building syndrome.

Methodological problems have limited scientific investigation of the causes of and solutions for sick building syndrome. The feasibility of using an experimental double blind cross-over study to resolve many of these methodological problems was assessed in a pilot study. The experimental intervention was to vary the amount of outdoor air from 10 cubic feet per minute per person (cfmpp) to 20 cfmpp or 50 cfmpp by central manipulation of the building heating, ventilation and air-conditioning (HVAC) system. Over 6 consecutive study weeks, 2 trials of rates were administered in random order. Study subjects and investigators of the study were blinded to intervention sequence. Unblinding, office environment rating and symptom occurrence were measured weekly. Of 305 eligible workers, 254 participated. Problems were encountered in delivering the lowest dose of ventilation due to building leakage. The prevalence of symptoms diminished steadily over the 6 study weeks, time trends which could be controlled by recommended design modifications. Blinding to the intervention was successfully maintained. Weekly non-response did not introduce a response bias but reduced the number of subjects available for analysis by one-third for each trial. We conclude that this design, with certain modifications, is feasible to evaluate many proposed interventions for sick building syndrome.     

A lesson learnt: the importance of modelling in randomized controlled trials for complex interventions in primary care

BACKGROUND: The Randomised Controlled Trial (RCT) is recognised as the 'gold standard' in quantitative research. However RCTs testing health care interventions can be difficult to design and implement. Health care interventions are often complex in themselves and are always applied in complex settings. Such interventions require a process of careful 'modelling' to maximize the chances of successful trials that will add to knowledge. OBJECTIVES: To describe the terms 'complex' and 'modelling' as used in the setting of randomised controlled trials of complex interventions. To give a practical example of an RCT involving a complex intervention applied in a health care setting to illustrate how this might take place in practice. METHODS: We describe an RCT designed and conducted by the authors. We then use our trial as an example to illustrate how complex interventions such as ours might benefit from modelling during the design of the intervention and the setting within which the intervention is to be tested. RESULTS: Our project was designed and tested before current guidance on complex interventions was published; our RCT was therefore not 'modelled' but was based on the outcome of a single quantitative pilot study. As part of our study we ran a parallel qualitative study, which highlighted several areas of complexity both in our intervention, and in the setting within which we applied it. In this paper we show how modelling might have allowed us to recognise these complexities at an early stage and might therefore have resulted in a study more likely to have demonstrated useful outcomes. CONCLUSION: Careful modelling of complex interventions is an essential step in designing trials of innovations in health care and health care services. Such a process ensures that interventions fit with and reflect the complexities of the settings within which interventions will be applied, and should ensure that the outcomes chosen are those most appropriate to demonstrate any benefits or risks.     

Screening for type 2 diabetes in a high-risk population: study design and feasibility of a population-based randomized controlled trial

ABSTRACT: BACKGROUND: We describe the design and present the results of the first year of a population-based study of screening for type 2 diabetes in individuals at high risk of developing the disease. High risk is defined as having abdominal obesity. METHODS: Between 2006 and 2007, 79,142 inhabitants of two Dutch municipalities aged 40--74 years were approached to participate in screening. Eligible participants had a self-reported waist circumference of [GREATER-THAN OR EQUAL TO]80 cm for women and [GREATER-THAN OR EQUAL TO]94 cm for men, and no known pre-existing diabetes. Of the respondents (n = 20,578; response rate 26%), 16,135 were abdominally obese. In total, 10,609 individuals gave written informed consent for participation and were randomized into either the screening (n = 5305) or the control arm (n = 5304). Participants in the screening arm were invited to have their fasting plasma glucose (FPG) measured and were referred to their general practitioner (GP) if it was [GREATER-THAN OR EQUAL TO]6.1 mmol/L. In addition, blood lipids were determined in the screening arm, because abdominal obesity is often associated with cardiovascular risk factors. Participants in both arms received written healthy lifestyle information. Between-group differences were analyzed with Chi-square tests (categorical variables) and unpaired t-tests (continuous variables). RESULTS: The screening attendance rate was 84.1%. Attending screening was associated with age at randomization (OR = 1.03, 95% CI 1.02-1.04), being married (OR = 1.57, 95%CI 1.33-1.83) and not-smoking currently (OR = 0.52, 95%CI 0.44-0.62). Of the individuals screened, 5.6% had hyperglycemia, and a further 11.6% had an estimated absolute cardiovascular disease risk of 5% or higher, according to the Systematic Coronary Risk Evaluation risk model. These participants were referred to their GP. CONCLUSIONS: Self-reported home-assessed waist circumference could feasibly detect persons at high risk of hyperglycemia or cardiovascular disease. Continuation of the large-scale RCT is warranted to test the hypothesis that targeted population-based screening for type 2 diabetes leads to a significant reduction in cardiovascular morbidity and mortality.Trial registrationISRCTN75983009.     

The Promoting Effective Advance Care for Elders (PEACE) randomized pilot study: theoretical framework and study design

Practice guidelines are available for hospice and palliative medicine specialists and geriatricians. However, these guidelines do not adequately address the needs of patients who straddle the 2 specialties: homebound chronically ill patients. The purpose of this article is to describe the theoretical basis for the Promoting Effective Advance Care for Elders (PEACE) randomized pilot study. PEACE is an ongoing 2-group randomized pilot study (n=80) to test an in-home interdisciplinary care management intervention that combines palliative care approaches to symptom management, psychosocial and emotional support, and advance care planning with geriatric medicine approaches to optimizing function and addressing polypharmacy. The population comprises new enrollees into PASSPORT, Ohio's community-based, long-term care Medicaid waiver program. All PASSPORT enrollees have geriatric/palliative care crossover needs because they are nursing home eligible. The intervention is based on Wagner's Chronic Care Model and includes comprehensive interdisciplinary care management for these low-income frail elders with chronic illnesses, uses evidence-based protocols, emphasizes patient activation, and integrates with community-based long-term care and other community agencies. Our model, with its standardized, evidence-based medical and psychosocial intervention protocols, will transport easily to other sites that are interested in optimizing outcomes for community-based, chronically ill older adults.     

Evaluation of community-wide interventions: the ecologic case-referent study design

In a setting of long-standing, community-wide and generally accepted prevention activities like youth health care services in The Netherlands, evaluative research in the form of experimental studies is hardly possible. Furthermore, as most interventions will bear fruit only after several years and the effects are often described in rather vague terms, even non-experimental study designs are fraught with possible difficulties. Although a study design using aggregate data is generally considered inferior or incomplete, in many cases, especially in health services research, this approach can be the only one feasible to evaluate the effectiveness of preventive programmes and interventions. In this article we present the ecologic case-referent design as a potentially expedient and valid method for estimating the ecologic effect of a population-wide intervention on the outcome rate in those populations. In this case-referent design, many variables are measured at the individual level, whereas the main exposure variable is measured at an aggregate or ecologic level. Using recently published studies as an example, the advantages and drawbacks of the design are discussed using the randomised controlled trial design as the referent study design.     

Members of research ethics committees accepted a modification of the randomized consent design

BACKGROUND AND OBJECTIVE: The use of randomized consent designs has been subject of methodologic and ethical controversy. In most Western countries, research ethics committees make the decision as to whether a randomized consent design can be applied. The purpose of the study is to assess to what extent a randomized consent design and a modification of this design is accepted by research ethics committees, in terms of ethics, health law, and methodology. METHODS: A postal survey was conducted among members of research ethics committees in the United Kingdom, and in The Netherlands, with professional competence in ethics, (health) law, methodology, or clinical practice. RESULTS: In both the UK and in The Netherlands, the modified randomized consent design appears to be statistically significantly more acceptable than the randomized consent design, with respect to ethical and judicial aspects. The overall rejection rate of the randomized consent design was 66% in the UK and 59% in The Netherlands. However, the modified randomized consent design was rejected by 47 and 41% in the two countries, respectively. CONCLUSION: the modified randomized consent design appears to be more acceptable than the randomized consent design. To increase consistency in the way research ethics committees handle study protocols, a discussion about the use of randomized consent designs appears necessary.     

The design of behavioural interventions labelled as patient‐mediated: A scoping review

Objective

Patient‐mediated interventions (PMIs) directed at patients and/or physicians improve patient or provider behaviour and patient outcomes. However, what constitutes a PMI is not clear. This study described interventions explicitly labelled as “patient‐mediated” in primary research.

Methods

MEDLINE, EMBASE, Allied and Complementary Medicine, PsychINFO, HealthSTAR, Social Work Abstracts, CINAHL and Cochrane Library were searched from inception on 1 January 2017 for English language studies that developed or evaluated behavioural interventions referred to as “patient‐mediated” or “patient mediated” in the full text. Screening and data extraction were independently duplicated. Data were extracted and summarized on study and intervention characteristics. Interventions were categorized as 1 of 4 PMI pathways.

Results

Eight studies (4 randomized controlled trials, 1 observational study and 3 qualitative studies) were included. No studies explicitly defined PMI, and few PMIs were described in terms of content and format. Although 3 studies employed physician interventions, only patient interventions were considered PMIs. One study achieved positive improvement in patient behaviour.

Conclusions

Research is needed to generate consensus on the PMI concept, employ theory when designing or evaluating PMIs, establish the effectiveness of different types of PMIs, and understand when and how to employ PMIs alone or combined with other interventions.