PRECEDE-PROCEED Model: The 8 Phases Explained

The PRECEDE-PROCEED model is an eight-phase framework for planning and evaluating health and social-behavior programmes, developed by Lawrence W. Green and later formalised with Marshall W. Kreuter. Its defining feature is sequencing: the model runs backwards from a community-defined quality-of-life outcome to the specific behaviours, environmental conditions, and administrative resources an intervention has to change to reach it, then runs forward again through implementation and evaluation to check whether it actually did.

That backwards-then-forwards structure is what separates PRECEDE-PROCEED from a framework like RE-AIM, which is evaluation-only. PRECEDE-PROCEED’s first four phases exist to produce the diagnostic case for an intervention design before anything is built; its last four evaluate the specific things those first four phases said mattered. A programme can use RE-AIM to report on an intervention that already exists. It cannot use RE-AIM to decide what that intervention should target in the first place — that is what PRECEDE-PROCEED’s front half is for.

Terminology. PRECEDE is an acronym for Predisposing, Reinforcing, and Enabling Constructs in Ecological Diagnosis and Evaluation — the four assessment phases described below. PROCEED stands for Policy, Regulatory, and Organizational Constructs in Educational and Environmental Development — the implementation and evaluation phases that follow. The model has gone through several editions since its first publication in 1980, with the PROCEED half added in the second edition and the ecological framing strengthened in later ones; the eight-phase structure below reflects the version most widely taught and cited today.

The Eight Phases, and What Each One Actually Produces

Each phase of PRECEDE-PROCEED is easy to describe as a diagnostic question, but the question alone doesn’t tell you what to do with the answer. What matters for actually running the model is the concrete planning artifact each phase is supposed to hand to the next one — the document, list, or dataset that phase 2 can’t start without phase 1’s output, and so on down the chain.

Phase 1: Social Assessment

Question: What quality-of-life problem does the target population itself say it has? Method: participatory data collection directly from the population — surveys, focus groups, community forums, existing social indicators (unemployment, crime, absenteeism, housing instability). Artifact produced: a written statement of the community-prioritized social problem(s) and quality-of-life goals. This is the reference point every later phase has to trace back to — Phase 8’s outcome evaluation measures against exactly this document, not against whatever health metric turned out to be convenient to collect.

Phase 2: Epidemiological, Behavioral, and Environmental Assessment

Question: Which specific health problems contribute to the Phase 1 social problem, and which behaviors and environmental conditions cause those health problems? Method: vital statistics, morbidity and mortality data, surveillance records, and epidemiological literature, decomposed into what people do (behavioral) versus what surrounds them outside individual control (environmental). Artifact produced: a ranked list of health determinants plus a set of specific, measurable behavioral and environmental objectives — who will do what, by how much, by when — that the intervention is accountable for moving.

Phase 3: Educational and Ecological Assessment

Question: Of the behavioral and environmental factors named in Phase 2, which ones can an educational or environmental intervention actually change, and why do they currently look the way they do? Method: sorting causes into three categories — predisposing factors (existing knowledge, attitudes, beliefs, values, perceived self-efficacy: the motivational precursors to a behavior), reinforcing factors (the rewards, social support, or feedback a person gets after performing the behavior, from peers, family, or providers, that determine whether it repeats), and enabling factors (skills, resources, accessibility, and structural conditions that make a behavior possible, not just desired — clinic hours, cost, transportation, service availability) — then rating each factor on importance and changeability. Artifact produced: a predisposing/reinforcing/enabling factor matrix, ranked by importance × changeability, that becomes the direct target list for intervention design. This is the phase that gives PRECEDE its name, and the one most under-resourced programmes shortcut by guessing at factors instead of assessing them.

Phase 4: Administrative and Policy Assessment

Question: Given the target list from Phase 3, what can the implementing organization actually build, and what will help or block it? Method: an audit of available budget, staff time and skill, facility and equipment resources, plus existing policies, regulations, or organizational rules — including those inside partner organizations the programme depends on. Artifact produced: a resourced, sequenced implementation plan — the intervention design, budget, staffing assignment, and timeline — plus a documented list of policy or administrative barriers that have to clear before Phase 5 can start. This is where PROCEED’s own name-giving constructs (Policy, Regulatory, Organizational) get resolved into a go/no-go on the intervention as designed.

Phase 5: Implementation

Question: Does the Phase 4 plan actually run as designed? Method: delivery of the intervention itself — this is not a diagnostic phase, it is the programme happening. Artifact produced: the intervention as actually delivered, documented through attendance and enrollment logs, delivery records, and any deviations from the Phase 4 plan noted as they occur — because Phase 6 has to account for those deviations, not the plan as originally written.

Phase 6: Process Evaluation

Question: Did Phase 5 happen the way Phase 4 planned? Method: measuring reach (who actually enrolled versus who was targeted), dose delivered versus dose planned, and fidelity to the intervention design. Artifact produced: a fidelity and reach report that tells you, if Phase 7 or 8 disappoints, whether the problem was the intervention’s design or its delivery — a distinction the later evaluation phases cannot make on their own.

Phase 7: Impact Evaluation

Question: Did the intermediate objectives set in Phases 2 and 3 actually move? Method: pre/post or comparison-group measurement of the specific predisposing, reinforcing, and enabling factors, and the behavioral and environmental targets, named earlier in the model. Artifact produced: a direct test of whether the programme’s underlying causal logic — change these factors, and behavior changes — actually held up in this population.

Phase 8: Outcome Evaluation

Question: Did the health status and quality-of-life indicators the community named in Phase 1 actually change? Method: outcome measurement against the Phase 1 social-diagnosis goals, on the timescale those goals genuinely require — often years, not the length of a single grant or funding cycle. Artifact produced: the closed loop back to Phase 1: evidence of whether the whole eight-phase chain, from social diagnosis through delivery, actually reached the outcome the population said it wanted.

A Worked Illustrative Example

Illustrative example — the scenario below is a composite constructed to show what each phase’s artifact looks like in practice. It does not describe a real, named programme, institution, or dataset, and no figure in it should be cited as an actual outcome.

A county health department is asked to address rising youth vaping rates. Phase 1 community forums surface that residents’ actual concern is broader — declining adolescent wellbeing and school engagement, of which vaping is one visible symptom. Phase 2 epidemiological review narrows this to two behavioral targets (vaping initiation among 12-14 year-olds, and product availability near schools) and one environmental target (retailer compliance with age-verification requirements). Phase 3 identifies predisposing factors (perceived low harm, peer normalization), reinforcing factors (social-media content that rewards use with attention), and enabling factors (low-cost disposable products sold within walking distance of two middle schools) — ranked so that retailer proximity and perceived-harm messaging score highest on importance × changeability. Phase 4 finds the health department has grant funding for a school-based curriculum but no legal authority over retailer licensing, so the administrative plan pairs a curriculum intervention (within the department’s control) with a referral to the county’s licensing board (outside it, flagged as a dependency). Phases 5 through 8 then deliver the curriculum, check it reached the targeted schools at the planned dose, measure whether perceived-harm attitudes shifted, and — over a longer follow-up window — check whether initiation rates actually declined.

PRECEDE-PROCEED vs. RE-AIM: Planning-and-Evaluation vs. Evaluation-Only

RE-AIM—Reach, Effectiveness, Adoption, Implementation, Maintenance—is a dimension-based framework for evaluating a public-health intervention’s real-world impact. It has no equivalent to PRECEDE-PROCEED’s Phases 1 through 4: RE-AIM does not diagnose a social problem, rank health determinants, sort predisposing/reinforcing/enabling factors, or produce an administrative feasibility assessment. It assumes an intervention already exists and asks how well it reaches its intended population, works under real-world conditions, gets adopted by delivery settings, is implemented with fidelity, and is maintained over time.

That makes the two frameworks complementary rather than competing. PRECEDE-PROCEED’s front half is a program-planning tool: it tells a team what to build and why, grounded in a documented diagnostic chain from community-stated problem to specific behavioral targets. RE-AIM is a reporting tool: it gives an already-built intervention a standardized vocabulary for describing its external validity, most often used when publishing dissemination or translational results. A team can use PRECEDE-PROCEED Phases 1 through 4 to design an intervention, deliver it, and then report on it using RE-AIM’s five dimensions instead of (or alongside) PRECEDE-PROCEED’s own Phases 6 through 8 — the two are not mutually exclusive, and some published programme evaluations use both.

Frequently Asked Questions

What does PRECEDE stand for?

Predisposing, Reinforcing, and Enabling Constructs in Ecological Diagnosis and Evaluation — the label for the model’s four front-end assessment phases.

What does PROCEED stand for?

Policy, Regulatory, and Organizational Constructs in Educational and Environmental Development — the label for the implementation and evaluation phases that follow the PRECEDE assessment.

Who developed the PRECEDE-PROCEED model?

Lawrence W. Green developed the original PRECEDE framework, first published in 1980; Marshall W. Kreuter collaborated on later editions that added the PROCEED implementation-and-evaluation half.

Is PRECEDE-PROCEED a theory or a planning framework?

A planning and evaluation framework, not a behavioral theory in itself. It’s a structure for applying behavioral, epidemiological, and organizational evidence to a specific programme decision — it doesn’t propose a new causal mechanism the way a theory like the Health Belief Model does, and it’s commonly used alongside such theories rather than instead of them.

Do you have to complete all eight phases?

Formally, yes, to call the work a full PRECEDE-PROCEED application — the phases are sequential and each depends on the previous one’s artifact. In practice, many published applications compress or abbreviate the four assessment phases when strong existing data already covers them (a well-documented population, for instance), but skipping Phase 3’s factor-ranking step specifically is the most common reason a downstream intervention targets the wrong thing.

How is PRECEDE-PROCEED different from a logic model?

A logic model is a general-purpose diagram of inputs, activities, outputs, and outcomes; it documents a programme’s intended causal chain but doesn’t specify a method for deriving which factors that chain should target. PRECEDE-PROCEED’s Phases 1 through 4 produce a logic model’s content with a documented derivation behind it — the ranked predisposing/reinforcing/enabling matrix from Phase 3, specifically, is the part a generic logic model template has no equivalent step for.

The PRECEDE-PROCEED Model: What Each of the 8 Phases Produces, and How It Differs From RE-AIM

Ask about The PRECEDE-PROCEED Model: What Each of the 8 Phases Produces, and How It Differs From RE-AIM