|Timestamp||Name||Principles or Qs|
|1/21/2013 15:27:25||David Mowat||1. Must be prioritized according to value which it will add to the health of the population.|
2. Value is best expressed in terms of the maintenabce and improvement in the health status of the population and reduction in disparities rather than clinical outcomes.
3. Research must be translatable into improved population health.
4. Practitioners and policy makers (decision makers) should be involved in setting priorities.
5. the entire value chain, including development and delivery, as well as discovery should be addressed.
|1/21/2013 15:27:28||Paul Ritvo||- to what degree is randomization to no-treatment necessary at all? Can't we randomize to alternatives that subjects are willing to accept?|
- prioritization about what risk reducing interventions are most important? The smaller the array of interventions, the more do-able the designs.
- establishment of ethical obligation for populations to participate in research when they receive governmentally - subsidized prepaid health care
- integration of genetic data (susceptibility data) with intervention exposure
- using smartphones or other forms of electronic connectivity, immediate adherence monitoring (if possible day-by-day) to better ascertain dose-response relationships in terms of intervention effects
|1/21/2013 15:27:54||Jon Kerner||1) What do we and don't we know from existing prevention research?|
2)How can a synthesis of research implications sections from systematic reviews of prevention intervention studies help us answer what we don't know as well as what we know?
3) What do practitioners and policy makers want to know about what works and what doesn't in prevention?
4) How can we best align what we do and don't know from prevention research evidence with the questions practitioners and policy specialists want answered?
5) How can we realign research funding priorities and academic rewards to support prevention research that answers these aligned questions of what is not known and where the practice and policy communities are looking for answers?
6) How can we tranform our prevention research evidence of what we do know into compelling stories that are more accesible to practitioners and policy makers?
|1/21/2013 15:28:04||Erin||To what extent, if at all, does the intervention impact on.....?|
To what extent, if at all, do individual factors moderate the intervention impact on.....?
To what extent, if at all, do environmental factors or context moderate the intervention impact on....?
What are the experiences or perspectives of intervention participants?
To what extent, if at all, was the intervention implemented?
|1/21/2013 15:31:03||Brenda Wilson||1. Equity concerns and impact should be built into any intervention study designs - e.g. differential impact of interventions.|
2. Consider randomizing a group for ongoing 'efforts' to engage in health behaviour change to see if we can learn more general lessons (i.e. not disease specific) about health promotion.
3. Can intervention studies be a way to empower cohort study particiants? (Passivity -> actvity)
4. Starting point must be full understanding of existing evidence.
5. Can we learn about potential interventions from outside the prevention field (e.g. professional behaviour change trials)?
6. Opportunity costs of doing reseaech - what is the value of the information to be gained over and above the best alternative use of the resources? This may be more important than considering sunk costs.
|1/21/2013 15:31:08||Stuart Edmonds||1. Address a significant health issue|
2. Build upon unique strengths in this field in Ontario
3. Build upon or leverage existing infrastructure which includes cohorts
4. Must have clear KT plan including health economics to aid adoption
5. Approach must be multidisciplinary with all relevant stakeholders involved
6. Significant participant engagement
|1/21/2013 15:31:42||John Ioannidis||1. Selection of individual-level or higher societal-level interventions to be tested|
2. Evaluation of the impact of interventions on societal outcomes, e.g. inequality/equity
3. Consideration for additional randomization options for cohort participants who develop a disease during follow-up, e.g. cancer or cancer survivors
4. What registries and other existing databases can be linked to the Ontario study or other existing cohorts and how reliable, accurate, and complete these data are. If not so, can they be easily improved.
5. Mapping other cohorts and biobanks worldwide that may have similar interest in nested randomization, cross-cohort replication and meta-analyses
6. Consideration of stealth interventions with indirect impact on health outcomes
7. Solicitation of proposed randomization choices, creation of large laundry list and cleaning/streamlining the list.
|1/21/2013 15:31:59||Peter Taylor||1. Social change around health is a broader issue than individual interventions, so develop a conceptual model of the multiple levels that this can occur and do research into how polities have been strategic about fostering desired secular changes.|
2. Can built environment and other contextual variables be creatively translated into experiments (as J.I. wondered)? (What exemplars are there to learn from?)
3. Getting things done in a real world of funders and funding emphases and politics requires focus and strategic thinking, but a research agenda might also build in spaces where people can choose to take time out and reflect and incubate not-yet-realistic approaches. (Otherwise, the focused work perpetuates the constraints on what is realistic.)
4. The rapid research review method could be used by researchers who give themselves time to dig deeper into areas that interest or intrigue them, but they haven't developed expertise.
5. Reliable, replicated environmental effect sizes may be small, but are they more readily translated into interventions than the GWAS effects of SNIPS?
6. An honest account of the formation of OHS (and how it has not included an intervention arm) would help clarify what needs to be done to get an intervention agenda adopted and funded.
|1/21/2013 15:32:29||Sandy Sulsky||1. Specify what is meant by 'intervention'. Are we talking about implementing lifestyle changes that already possess a strong evidence base for benefit, or about investigating possible etiologic relationships with health or disease state? The answer will influence the design of the program, and will define many ethical issues around consent, participant choice , etc. |
2. Determine what level of needs assessment will be completed prior to undertaking the design of the program. Which stakeholders will have a voice in choosing key health outcomes and types of interventions will be selected?
3. Select the target population in which the intervention will take place. Specifying the target carefully will help define the relevant comparison groups as well as determine which modes of data collection might be appropriate (e.g., computer based, oral, printed, etc. )
4. Define the domains of key outcomes of concern. Should the intervention focus on prevention of cancers, is heart disease a valid target? What about poverty and violence?
5. What is a reasonable maximum duration for the intervention and its evaluation? Is there a reasonable hope of continued funding if the program is initially successful?
6. How will success if the intervention be defined? Is the metric strictly effectiveness against the targeted health outcome, or must there be a cost-benefit analysis planned for, as well? Is clinical effectiveness define able?
|1/21/2013 15:33:19||Nancy||Determine the stakeholders for the research: General public? Policy makers? Practitioners?|
How might social networking contribute to the research effectiveness?
Why haven't our efforts in prevention interventions been very effective to date? (What are we missing?)
How might we increase physical activity in traveling between home and workplace? At sedentary workplaces?
|1/21/2013 15:33:53||Todd||1. clearly defined research question with background evidence|
2. is there sufficient population/subjects to asses the effectiveness of an intervention (power)
3. study design required to address the research question
5. feasible -- funding/cost - is there funding available (by existing grants, already acquired, etc) to properly address
6. predicted benefit > years of life, reduced risk, etc
|1/21/2013 15:36:33||What do we hope to accomplish by population-based intervention research? What is/are the primary outcome(s) of interest? This will inform priorities and decision-making. |
A principled approach to establishing priorities for intervention research is a first step. These principles help ensure transparency and consistency in decision-making.
Engaging stakeholders from diverse perspectives (e.g., clinicians, policy makers, the "ologies") can only improve the proposed interventions.
How will we predict public acceptance of any proposed intervention?
|1/21/2013 15:36:54||Pat Smith||Six principles that should be addressed by an initiative in population-based intervention research:|
1. Geoffrey Rose principles of the strategies of prevention (high risk vs. population approach) should be considered and all that they entail in terms of the strengths and limitations
a. E.g., intervening with high risk individuals, although it only offers interventions to those who really need it (answering the question “why me, why now?”) and thus can be cost-effective, the approach is palliative (as there will always be more to fill their spot), it is not known how individuals will respond, and it is often not the best approach as asking people to change their behaviour is often asking them to be different than everyone else in their social world (it is difficult to step out of our culture)
2. Consideration of “who is doing what to whom where and how “ using an organizing framework such as the classic “cube” diagram from tobacco control
3. Theory as it pertains to planning is important—e.g., frameworks such as the Precede-Proceed can help planners find a starting point to think about various levels of stakeholders, what is known at various levels of assessment and thus what might be necessary to measure or include in interventions, and what the natural intermediate outcomes might be
4. Be mindful of where we put the solution—we tend to be an individualist society focused on the stereotypic Marlborough Man image of rugged individualism and pull yourself up by the bootstraps—so much talk about interventions puts a focus on individuals to make the behaviour changes/reduce risk factors—but we live in a society of heavy social influence through various types of social media, not the least of which is incessant advertising that advocates we do exactly the opposite of what we are supposed to do…eat more chocolates or fast food, smoke more, drink more, etc.
5. consider stepped care models of interventions
6. consider systems level interventions to create sustainability—as the Surgeon General said about tobacco (poorly paraphrased): if we implemented what we know about tobacco interventions, we could eradicate lung cancer (the number one cause of cancer mortality)
|1/21/2013 15:37:25||Jamie Brehaut||Six principles|
1) There needs to be involvement of a range of disciplinary perspectives to develop these complex interventions;
2) There needs to be more clarity in the theories underlying interventions. Even if it isn't explicit, interventions are based on assumptions on how the intervention should work. Those models should be made explicit
3) There should be detailed reviews of interventions in light of theory-relevant intervention components/constructs. Which theory-relevant intervention descriptions are not generally described in the reports? Which ones are related to effectiveness of the interventions?
4) Need to engage the relevant stakeholders from the outset, and target interventions that match stakeholder priorities
|1/21/2013 15:38:42||Laura||1. Is the intervention feasible - both for study and general population. Will participants/patients comply or participate? Are policy makers interested?|
2. Do the potential benefits outweigh any possible harms? Including the harms that may result from negative findings
3. Will there be sufficient difference between exposed and non-exposed (i.e., non-contaminated control group)?
4. What is the appropriate timing of exposure (critical period, duration of intervention needed, improving uptake)?
5. What data need to be captured for the measurement of outcomes and other important factors?
6. How to encourage study participation, and reduce attrition among both the intervention and control group?
|1/21/2013 15:40:53||Reed||1. The intervention(s) should address multiple levels of aggregation (e.g., individual, family, community)|
2. The intervention(s) should address more than one causal pathway, ideal both distal and proximal, both biological and social and environmental.
3. The intervention(s) should have built into it a concrete way of taking of advantage of positive synergistic/recursive feedback loops, proposing (perhaps from a theoretical framework) why the intervention has the potential to make a major sustaining change to the web of causation.
4. The intervention should be designed in an open fashion, so that it can evolve and be evaluated over time, transforming itself to adapt to the changing environment and unanticipated externalities. The intervention should have from the start a built-in evaluation component that draws from the knowledge of the study participants, ground level staff, and decision makers to adapt the interventions as it develops.
|1/21/2013 15:41:34||Baskerville||1. Engage key stakeholders and decision makers first to help formulate the research agenda. Helps with receptivity in the long run. |
2. Diversity of perspectives is critical. Ensure a balance of disciplines and interests.
3. Match the research interests to the key health behaviour challenges ( eg. Adoption of healthy lifestyles)
4. In seeking evidence-truth, RCTs are the pinnacle of science but are also subject to bias - do not neglect good quality, well executed OBS studies and quasi experimental designs.
5. Encourage collaboration across sectors and disciplines to help garner the infrastructure and support needed for large nested RCT cohorts.
6. A focus on implementation science as compared to discovery of new interventions is needed. This focus needs to include new implementation science methods such as process evaluation and qualitative methods.
7. Gain an understanding of why the past efforts to provide research evidence to help decision makers have failed and where are the exemplars and why.
|1/21/2013 15:42:40||Julian Little||1. Status of evidence on which to base interventions to investigate. Resolving point that many consider enough evidence has accrued to move forward to test vs. lack of replication of many exposure-outcome relationships. Recognise that moving target and that intervention will yield different evidence.|
2. Consideration of context. Is this about differences in confounding structure? Why would randomization not fully address this? Motive for testing in multiple populations, but how these selected should not be opportunistic
3. All interventions complex. Thefore crucial to appreciate multiple levels and design trials that address this. Could one have factorial design of RHS of area-based interventions and individual-level interventions?
4. Crucial importance of public engagement (citizen science) at all levels of research - adds to rights and responsibilities
5. Multiple health outcomes, including mental health. In view of difficulty of getting info on long term outcomes, consifderation of surrogate endpoints.
6. Question of stratifying intensity of intervention, e.g. by family history
|1/21/2013 15:44:42||Michael Wolfson||• John Ioannidis “correlation globe” is a really neat construct|
• But it raises the question, e.g. in the case of “diet”, that an intervention on a specific food item (e.g. tomatoes or chocolate or coffee) will perturb many of the nodes on the correlation globe
• Thus, proper analysis of the “intervention” will require dense data collection on the exposome – e.g. not just tomato consumption, but a wide range of dietary characteristics
• If this is not done, there will be either or both correlated unobserved confounders, and significant measurement error, in turn leading to under-estimation of the relevant effect sizes
• There is a tension between the observations in genetics and many other domains of epidemiologic study that the world is best characterized by myriad small effects, on the one hand, and the observation that there are broad patterns – e.g. the SES gradient in mortality, or the recent “Seven More Years” study by PHO, where there are clearly observable large effects. The reconciliation of this tension may be in noting that the small effects are related to single SNPs or single molecules, or single dietary item (e.g. tomatoes), while the large effects relate to broad composites – e.g. SES, diet, stress, physical activity.
• If reality really is myriad small effects when thinking at the level of detailed factors, and much larger effects when at the level of complexes of factors, then data analysis alone, whether from RCTs or observational studies, is insufficient. As in astronomy and many other fields of science, diverse empirical observations have to be coherently integrated / woven together in a theoretical / conceptual, and ultimately a quantitative simulation model.
• Observational studies really need to be distinguished by whether they are convenience samples or truly population representative samples. Statistical inference from convenience samples, e.g. of etiologic effect sizes, can be biased due to unobserved confounders.
|1/21/2013 15:48:16||Moe||1- Cancer Prevention as the most important population health risk need special attention.|
2-How to engage different research categories under one umbrella? i.e. how to engage biologists, clinicians, epidemiologists and statisticians in preventive research projects in a structural way?
3- are epidemiologists advocate on behalf of the general populations the way they should?
4- how to ensure that the new designs of epidemiology are implemented in real world?
5-is prevention being emphasized in the medial as much as treatment? if not who is responsible? what is the solution?
6- the need for electronic data bases for the development of strong and reliable evidence.
|1/21/2013 15:56:40||H. Richardson||1. What proportion of population-based intervention should continue to evolve around chronic disease etiology?|
2. Should we focus more attention on individual interventions, structural / systems-level interventions or multi-level interventions?
3. What priority should research in gene-environment interactions have in population based intervention research?
4. How best can we integrate knowledge about gene-environment interactions into Public Health Promotion?
5. Duration of lifestyle interventions and surrogate markers / intermediate markers of chronic disease outcomes should continue to be important design considerations in disease etiology research
6. Future primary prevention intervention research will need to be multi-disciplinary with increased attention to different behavioural / psychosocial theories
7. We need to think of new and radical approaches to intervene and improve the health and well-being of the population
|1/21/2013 17:10:59||Tim||Social return on investment (SROI)|
Uptake / ease of uptake
The issue of direct/indirect benefits, including the unintentional consequences
The capturing if context