Written and maintained by CASRAI Editorial Board
Published
On this page
The ORBIT model (Obesity-Related Behavioral Intervention Trials) is a phased framework for developing behavioral treatments for chronic diseases before they are tested in a large efficacy trial. It was published in 2015 by the ORBIT Consortium and divides early work into Phase Ia (Define), Phase Ib (Refine), Phase IIa (Proof-of-Concept) and Phase IIb (Pilot Testing), with a feasibility pilot inside Phase IIb. Each phase ends at a pre-set milestone, and a treatment moves forward only when it reaches it.
This page covers the behavioral-treatment framework used in health psychology and behavioral medicine. It is not about atomic orbit models in chemistry or physics, nor about orbital mechanics.
What is the ORBIT model?
ORBIT was proposed in Czajkowski et al. (2015), From ideas to efficacy: The ORBIT model for developing behavioral treatments for chronic diseases, in Health Psychology. The paper was written by a consortium supported by several US National Institutes of Health institutes and offices (NHLBI, NCI, NIDDK, NICHD and OBSSR). The National Cancer Institute’s Behavioral Research Program describes the phases as mirroring those of the drug development process and as useful for behavioral risk factors related to cancer.
The model is deliberately limited. It covers only the early, pre-efficacy phases, up to the point where a treatment is ready for a Phase III efficacy trial, and it is described as not prescriptive: it does not say what a treatment should contain, only how to organise the evidence that a treatment is ready to advance. Its terminology is borrowed from drug development so that funders and clinicians used to that vocabulary can follow the logic.
Start with a Significant Clinical Question
Before Phase Ia the paper asks investigators to articulate a Significant Clinical Question, stated with the precision of a basic-science hypothesis. Examples include a condition with no current treatment, or an effective treatment that patients adhere to poorly. Naming the problem early lets investigators think ahead to a Phase III trial, to its endpoints, and to how large a behavioral change the treatment must produce to matter.
The ORBIT phases: Ia, Ib, IIa, IIb and the feasibility pilot
The data table below sets out each phase side by side. The notes here explain how to read it.
Phase Ia: Define
The goal is to define the basic elements of the treatment: a hypothesised pathway from treatment to clinical outcome, the candidate components, the targets and the milestones. Typical evidence comes from systematic reviews of guidelines, meta-analyses, experiments, time-series and N-of-1 designs, observational studies and formative qualitative work such as focus groups and interviews. The phase ends when candidate components, methods and milestones are defined.
Phase Ib: Refine
The goal is strength and efficiency: which components act independently or interactively, what dose or duration is needed, how the treatment is delivered and whether it needs tailoring. Factorial and fractional factorial designs and adaptive designs are the typical tools. The phase ends when the investigators are satisfied that the treatment package is complete and the protocol is fixed and ready for testing.
Phase IIa: Proof-of-Concept
Investigators prepare a treatment manual and fidelity-monitoring procedures, then test whether the fixed protocol produces a clinically significant signal on the behavioral target. The typical design is quasi-experimental, treatment-only and within-subjects (pre-post). The paper says the sample can be small, because the aim is clinical rather than statistical benefit, and that sample size calculations are unnecessary at this stage.
Phase IIb: Pilot Testing
Here the question becomes whether the clinically significant signal stands out from noise: whether the benefit exceeds the effects of time and non-specific factors. The design is a randomised test of the fixed protocol with a systematic study of the control group. The milestone is a clinically significant signal over the control group and evidence that the control condition is appropriate.
Feasibility pilot (within Phase IIb)
A second kind of IIb study asks whether the planned efficacy-trial protocol is logistically feasible. It prepares a manual of operations and estimates screening-to-enrollment ratios, dropout or crossover rates and response rates. A pilot study of this kind yields only an unstable estimate of the effect size for the Phase III primary endpoint, so the paper advises weighing it alongside effect sizes from similar past trials when sizing the efficacy trial.
Clinically significant change, milestones and moving backward
The central ORBIT rule is that milestones are set in advance and defined in clinical terms, not only statistical ones. The paper states that specifying them a priori prevents moving the target after the data are in. Thresholds can come from epidemiological, experimental or clinical studies, for instance a cutpoint where biomedical risk rises; the paper gives 3 to 5 percent weight loss in adults with cardiovascular risk factors as a change that can be deemed clinically significant.
| Situation after a phase | What the model says |
|---|---|
| Milestone met at Phase IIa | Clinically significant change on the behavioral risk factor is the essential milestone for moving to Phase IIb. |
| Ambitious criterion approached but not reached | Qualitative optimisation could justify moving forward to Phase IIb. |
| Criterion well short, not approached | Return to Phase I, or abandon the treatment. |
| Suboptimal result, for example in a new population | Return to an earlier phase for refinement; the paper’s summary gives a move back to Phase Ib as an example. |
Backward movement is therefore part of the design, not a failure of it. The model’s value lies in stopping weak treatments cheaply, before a large randomised trial is committed.
ORBIT vs drug development phases
ORBIT borrows the numbering of drug development, but the phases do not map one to one. The paper describes the drug sequence as preclinical studies, a Phase I search for the maximally tolerated dose, a Phase II test of biologic activity at that dose, a pilot for feasibility, and Phase III tests on clinically meaningful endpoints leading to approval. ORBIT differs in three ways described in the paper: it is limited to pre-efficacy development, it applies across a broad range of chronic diseases, and it splits Phases I and II into subphases (Ia/Ib and IIa/IIb). The paper also notes that behavioral treatment development is not regulated or industry-sponsored in the way drug development is. It says ORBIT could be combined with frameworks, such as the NIDA Stage Model, that cover steps beyond pre-efficacy.
For a behavioral model that explains why people change behavior, rather than how to stage a trial, see the COM-B model of behaviour change: it can supply the hypothesised pathway that Phase Ia asks you to build.
Worked illustration: moving a treatment through the phases
Illustrative composite, not a real study. The numbers below are invented to show the logic; the thresholds are the ones the ORBIT paper uses in its own diabetes-prevention example. A team wants a lifestyle treatment to prevent diabetes through weight loss.
- Significant Clinical Question. Can a behavioral treatment help adults at risk of diabetes reach a clinically meaningful weight loss?
- Phase Ia. Prior trials support a biomedical target of 5 percent weight loss, with behavioral targets of 150 minutes per week of moderately vigorous physical activity and a 5 percent reduction in percent fat. In the paper’s example, N-of-1 time series showed participants met these targets but did not sustain them until internet-delivered tailored feedback was added.
- Phase Ib. A fractional factorial design tests daily self-monitoring and tailored feedback alone and together. In the paper’s example the combination was more potent than either alone for the 5 percent target, so both go into the fixed protocol.
- Phase IIa. A small treatment-only pre-post study sets its clinical milestone before enrolment. For a participant weighing 90 kg at baseline, 5 percent is 90 × 0.05 = 4.5 kg, so the participant must weigh 85.5 kg or less to meet it. The team judges the result by how many participants reach that clinical line, not by a significance test.
- Go/no-go. If most participants reach the milestone, the team proceeds. If the result approaches it, qualitative work on why some participants fell short may justify moving on. If it is nowhere near, the team returns to Phase I.
- Phase IIb. A randomised pilot compares the fixed protocol with a control condition and studies the control group itself. A feasibility pilot then estimates recruitment and dropout for the efficacy trial.
The 90 kg figure is the only arithmetic in the example: 90 × 0.05 = 4.5, and 90 − 4.5 = 85.5.
Assumptions and common mistakes
- Treating ORBIT as a recipe. The paper calls the model not prescriptive. It does not define a treatment’s content, targets or designs for you.
- Setting thresholds after seeing the data. A priori milestones are the point of the framework. A threshold chosen afterwards is the moving target the paper warns about.
- Relying on statistical significance alone. The Phase IIa criterion is a clinically significant change on the behavioral target, and a small, accessible sample is acceptable for that purpose.
- Sizing the efficacy trial from one pilot. A feasibility pilot gives an unstable effect size estimate. The paper recommends considering it alongside effect sizes from similar trials.
- Using ORBIT for later phases. It stops before efficacy testing. Phase III and beyond need other frameworks.
- Assuming it only fits weight loss. The name reflects its origin in obesity-related trials, but the paper presents it for behavioral treatments for chronic diseases generally.
How to cite and report the ORBIT model
In a protocol, grant or paper, name the phase you are in and the milestone you set for it, and cite the source paper. An example of how that can read: This study is a Phase IIa proof-of-concept study under the ORBIT model (Czajkowski et al., 2015); the pre-specified clinically significant milestone was a 5% reduction in body weight. The reference in APA 7th edition format:
Czajkowski, S. M., Powell, L. H., Adler, N., Naar-King, S., Reynolds, K. D., Hunter, C. M., Laraia, B., Olster, D. H., Perna, F. M., Peterson, J. C., Epel, E., Boyington, J. E., & Charlson, M. E. (2015). From ideas to efficacy: The ORBIT model for developing behavioral treatments for chronic diseases. Health Psychology, 34(10), 971–982. https://doi.org/10.1037/hea0000161
State the milestone, its source and the date you fixed it, so reviewers can see it was set a priori.
Frequently asked questions
What does ORBIT stand for?
Obesity-Related Behavioral Intervention Trials, per the National Cancer Institute’s Behavioral Research Program, which describes it as a model that mirrors drug development phases.
How many phases does the ORBIT model have?
Four numbered pre-efficacy phases: Ia (Define), Ib (Refine), IIa (Proof-of-Concept) and IIb (Pilot Testing), plus a feasibility pilot that is a type of IIb study. Phase III efficacy trials sit outside the model.
Does ORBIT require a sample size calculation?
Not in Phase IIa; the paper says calculations are unnecessary there because the aim is clinical benefit. It gives no numeric sample sizes for any phase.
Can a treatment go back to an earlier phase?
Yes. The model allows a return to an earlier phase for refinement after suboptimal results, and a treatment that falls well short of its milestone can be sent back to Phase I or abandoned.
Is ORBIT only for obesity research?
No. The acronym reflects its origin, but the source paper is titled for behavioral treatments for chronic diseases generally.
Sources
- Czajkowski, S. M., et al. (2015). From ideas to efficacy: The ORBIT model for developing behavioral treatments for chronic diseases. Health Psychology, 34(10), 971–982. PMC4522392; DOI 10.1037/hea0000161. Read 2026-11-15.
- National Cancer Institute, Behavioral Research Program. Translational research (ORBIT). cancercontrol.cancer.gov. Read 2026-11-15.
- Crossref record for DOI 10.1037/hea0000161 (author list, volume, issue, pages). Read 2026-11-15.
Data: ORBIT model phases: goal, typical designs and exit milestone
Side-by-side summary of the phases of the ORBIT model for developing behavioral treatments (Czajkowski et al., 2015), compiled from the phase table and text of the source paper: central goal, typical designs and the milestone that ends each phase.
Period covered: 2015.
| Phase | Central goal | Typical designs | Exit milestone |
|---|---|---|---|
| Ia Define | Define basic elements | Systematic reviews, meta-analyses, experiments, time series, N-of-1, observational and formative qualitative studies | Candidate treatment components, methods and milestones are defined |
| Ib Refine | Refine for strength and efficiency | Factorial and fractional factorial, adaptive designs, experiments, time series, formative qualitative work | Treatment package complete and fixed protocol ready for testing |
| IIa Proof-of-Concept | Clinically significant signal | Quasi-experimental, treatment-only, within-subjects pre-post test of a fixed protocol | Clinically significant impact on the behavioral target |
| IIb Pilot Testing | Clinically significant signal over noise | Randomized test of a fixed protocol with systematic study of the control group | Clinically significant signal over the control group; control group shown to be appropriate |
| IIb Feasibility Pilot | Feasibility of efficacy trial protocol | Qualitative and quantitative feasibility and acceptability studies | Trial protocol shown to be logistically feasible; estimates obtained for designing the efficacy trial |
Sources
- From ideas to efficacy: The ORBIT model for developing behavioral treatments for chronic diseases (Health Psychology, 2015) (accessed 2026-11-15)
Compiled by CASRAI from the sources above. Free to reuse under CC BY 4.0 with attribution.







