Published  |  Last updated

Quick Answer: The PRECEDE-PROCEED model plans a health program backward. Four PRECEDE phases move from the quality-of-life result a community wants, to the behaviors and environments behind it, to the predisposing, reinforcing and enabling factors, to the administrative and policy conditions. Four PROCEED phases then implement the program and evaluate its process, impact and outcome.1 In a 2025 meta-analysis of 14 studies (2,478 adults with type 2 diabetes), HbA1c was 0.41 to 0.63 percentage points lower with programs built on the model, though results varied widely between studies.2

PRECEDE-PROCEED Model for Health Program Planning: The Eight Phases

The model was designed to plan and evaluate health promotion and other public health programs, and it can support a one-time intervention or a long-running program.3 Its central idea is to begin with the outcome, not the activity: diagnose from the desired end result backward to what needs to be done.1 The phase names and questions in the table follow the Community Tool Box;1 the last column, the typical output of each phase, is Editverse’s wording. The Rural Health Information Hub groups the same work under social, epidemiological, educational and ecological assessment, and program and policy development.3

Swipe sideways to see every column.

PhaseQuestion to answerTypical output of the phase
1. Desired resultWhat outcome does the community want?A ranked list of community priorities and a quality-of-life goal
2. Issues and determinantsWhich health issues and which behaviors or environmental conditions stand in the way, and which can an intervention change?Priority behaviors and environmental targets, each with a measurable objective
3. Predisposing, enabling and reinforcing factorsWhat makes the priority behaviors more or less likely?A list of factors to act on, sorted into the three types
4. Administrative and policy factorsWhich organizational conditions and policies decide what you can deliver?A feasible intervention design, budget and approvals
5. ImplementationIs the intervention running as designed?A delivered program with a work plan
6. Process evaluationAre you doing the things you planned to do?Reach, dose and fidelity records
7. Impact evaluationDid the targeted behaviors or environmental factors change?Pre and post measures on the Phase 2 and 3 targets
8. Outcome evaluationDid the Phase 1 result improve?Change in the outcome and quality-of-life indicators

Phase labels vary between texts. A 2022 systematic review lists the same eight as social assessment, epidemiological assessment, educational/ecological assessment, administrative/policy assessment and intervention planning, implementation, process evaluation, impact evaluation and performance evaluation.4 The sequence is the same, so use whichever labels your funder or journal expects.

PRECEDE vs PROCEED: What Each Half Does

PRECEDE is Phases 1 to 4, the diagnosis that decides what the program should do. PROCEED is Phases 5 to 8, which carry out the intervention and then evaluate it, working back to the result chosen in Phase 1.1 The Community Tool Box lists the assumptions behind the model:1

  • Health-promoting behaviors are often voluntary, so the process should be participatory and involve all stakeholders from the beginning.
  • Health is a community issue, shaped by community attitudes, environment and history.
  • Health sits within the larger context of quality of life.
  • A change process should focus first on the outcome, not on the activity.

A major reason to use the model, in the Community Tool Box’s words, is that it is a logic model: it gives a structure for planning, so the intervention is a carefully planned, coherent whole.1

Phases 1 and 2: Community Assessment, From Desired Result to Priority Behaviors

Phase 1 asks the community what result it wants. The Community Tool Box describes collecting this through interviews, questionnaires and focus groups, then involving the whole community in prioritizing the outcomes.1 The model is built to be participatory, so stakeholders are involved from the start rather than consulted at the end.1

Phase 2 turns the result into targets. You list the issues that cause or influence the outcome, including supports and barriers, and then keep those that matter most and that an intervention can influence.1 A factor you cannot change, such as the global economy in the Community Tool Box’s example, may be important and still not be a target.

Phase 3: Predisposing, Reinforcing and Enabling Factors

Phase 3 is the educational and ecological assessment.4 For each priority behavior you ask three questions, and the answers become the content of the intervention.1

Swipe sideways to see every column.

Factor typeWhat it coversExample to look for
PredisposingKnowledge, attitudes, beliefs, values and confidencePatients with diabetes who do not know what their HbA1c target means
EnablingAvailability of resources, accessibility of services, laws and policies, issue-related skillsNo clinic appointment within reach; no skills for reading food labels
ReinforcingLargely the influence of significant others in the social environmentFamily meals that work against the dietary plan

The first column and its content follow the Community Tool Box.1 The examples are written by Editverse to show the sorting, not taken from a study.

Phases 4 and 5: Intervention Design and Implementation

Phase 4 checks the administrative and policy side. The Community Tool Box lists organizational structure, procedures, culture and resources, together with internal policies, funder requirements, regulations, laws and unstated community norms.1 Phase 5 is then the delivery of the intervention you designed in Phases 3 and 4.1

Education was the usual intervention type in published studies. Of 26 experimental or quasi-experimental studies in one review, 25 used educational programs as the main intervention, and 20 used more than two teaching methods such as lectures, group education, one-on-one teaching, online delivery or distributed materials.4 That fits the model’s educational diagnosis, but the model does not require an educational intervention; it asks you to match the activity to the factors you found.

Phases 6 to 8: How to Evaluate the Program

Phases 6, 7 and 8 are three different questions. Process evaluation checks whether you are doing what you planned. Impact evaluation checks whether the targeted behaviors or environmental factors changed. Outcome evaluation checks whether the Phase 1 result is being reached.1 The Rural Health Information Hub phrases these as process, short-term (knowledge, skills, attitudes), intermediate (behavior) and long-term evaluation (incidence or prevalence).3

Formal evaluation of all three phases, with controlled studies, can be beyond a small organization. The Community Tool Box says many users probably do not have the resources for it and suggests that keeping and examining simple records is a workable start.1

A Worked Planning Example (Editverse Illustration)

The example below is hypothetical and contains no real data. A district clinic wants to reduce diabetes complications and uses the eight phases as a checklist.

Swipe sideways to see every column.

PhaseWhat the clinic does
1Runs a short survey and two focus groups; the community ranks “fewer diabetes complications” first.
2Lists contributing factors; selects regular blood glucose checking and diet as the behaviors, and clinic access as the environmental target. Sets measurable objectives for each.
3Finds that many patients do not know their target (predisposing), the nearest check-up is far away (enabling), and family meals work against the plan (reinforcing).
4Confirms staffing, a room, a budget line and data-protection rules; designs group education sessions plus family sessions and a reminder system.
5Delivers the sessions and records attendance.
6Reviews attendance, dose and fidelity every month; adjusts timing.
7Measures knowledge, checking behavior and diet at baseline and follow-up.
8Compares HbA1c and complication indicators with the baseline and, where possible, with a comparison group.

What the Evidence Shows: PRECEDE-PROCEED Programs and HbA1c in Type 2 Diabetes

Does a program built on the model change a health result? We use an open-access meta-analysis as a real case. Tang and colleagues searched nine databases up to 2 June 2024 and included 14 randomized or quasi-experimental studies with 2,478 adults with type 2 diabetes. Eleven were randomized trials, six were from China, six from Iran, one from Spain and one from the United States, and all were published between 2011 and 2023.2 The article is published under CC BY 4.0, so we redrew its pooled results.

Forest-style plot of pooled HbA1c differences with PRECEDE-PROCEED programs: 1 month -0.41 (95% CI -0.58 to -0.24), 3 months -0.50 (-0.67 to -0.33), 6 months -0.63 (-0.93 to -0.33), in percentage points
Pooled change in HbA1c, percentage points, with 95% CI. Redrawn by Editverse from Tang et al. 2025 (CC BY 4.0); computed in R.

Swipe sideways to see every column.

Follow-upStudiesPooled difference in HbA1c (95% CI)I²z (computed)
1 month2−0.41 (−0.58 to −0.24)0%−4.7
3 months5−0.50 (−0.67 to −0.33)54%−5.8
6 months6−0.63 (−0.93 to −0.33)97%−4.1

What the data show. In each group of studies, HbA1c was lower in the program groups than in controls, and none of the three intervals reaches zero. The authors also report that program groups scored higher on total self-management (standardized difference 2.53, 95% CI 1.14 to 3.91) with I² of 98%.2 At 6 months, removing one low-baseline study changed the pooled difference to −0.71 (−0.95 to −0.47) and I² to 75%.2 Editverse recomputed standard errors and z values from the published intervals; they are consistent with the p-values the paper gives for 3 and 6 months.

What they do not show. The three rows pool different studies (2, 5 and 6), so the table is not one group of patients followed over time and does not prove that the effect grows. Heterogeneity is high at 6 months. The authors note that self-management was assessed with self-reported measures, which introduces a possibility of reporting bias. In 13 of the 14 studies the interventions were mainly educational programs, compared with standard education, so the review cannot separate the model itself from the education delivered with it.2 Editverse’s reading is that the figure supports a modest, consistent direction of effect in programs planned with the model, not a precise size.

How Much of the Model Do Published Studies Use?

A second open-access review (Kim and colleagues, CC BY 4.0) pooled 26 experimental or quasi-experimental studies. We counted its Table 1: 16 studies (61.5%) used the model for planning, implementation and evaluation, 2 used it for planning and implementation, and 8 (30.8%) used it for planning only.4 The authors had already reported these counts; the figure splits them by design.

Bar chart of 26 PRECEDE-PROCEED studies: 16 used planning, implementation and evaluation (6 randomized, 10 quasi-experimental), 2 planning and implementation, 8 planning only
Number of studies by phases applied and design. Counted by Editverse from Table 1 of Kim et al. 2022 (CC BY 4.0); computed in R.

In the same review, knowledge was the outcome that could be pooled. Seven randomized trials gave a standardized difference of 3.64 (95% CI 1.95 to 5.33) and four quasi-experimental studies gave 1.68 (0.75 to 2.61), with I² of 98% and 95%. The funnel plot was asymmetric, and a trim-and-fill adjustment lowered the pooled value from 2.58 to 1.73 (95% CI 0.63 to 2.83).4 Differences that large with that much heterogeneity are better read as direction than as size. A third review of 54 studies reported that 51 had effective outcomes and that all included health education.5 Because reviews draw on published studies, the true picture may be less favorable.

How we computed our figures

  • Figure 1. The pooled differences, confidence intervals, study counts and I² values were typed from the Results text of Tang et al. For each interval, the standard error is (upper − lower) / 3.92, z is the estimate divided by the standard error, and the two-sided p-value comes from the normal distribution. R gave z of −4.7, −5.8 and −4.1.
  • Figure 2. Each of the 26 rows of Kim et al. Table 1 was classified by design (12 randomized, 14 quasi-experimental) and by phases applied, then cross-tabulated in R. The totals match the 16, 2 and 8 reported in the paper.
  • Sources. License checked in the PubMed Central XML of both papers (CC BY 4.0). Nothing was taken from the third review except its published summary numbers.

Fitting the Model to Your Resources

The Community Tool Box says the model says much less about the content of an intervention than about its process, which leaves room to adapt methods to your community. It also advises considering whether every part of the model suits your setting.1 The Rural Health Information Hub makes the same point: check that all parts fit the program and the resources available, and expect plan components to change over time.3 A sensible small-scale path is to complete Phases 1 to 4 properly, keep simple process records, and add a comparison group or a formal impact study only where the budget allows.

A Ten-Minute Planning Check

  1. 1Phase 1. Can you state the result in one sentence, and did community members choose it?
  2. 2Phase 2. For each target behavior or environmental condition, is it something your program can change?
  3. 3Phase 3. Does every planned activity answer at least one predisposing, enabling or reinforcing factor?
  4. 4Phase 4. Do you have the staff, budget and approvals the design needs?
  5. 5Phases 6 to 8. For each phase, which measure will you use and when will you collect it?

Need the Analysis Run or Checked? Consult Editverse

Editverse subject-matter experts work to COPE, ICMJE and GPP 2022 guidance, with institutional and country-specific ethics followed.

Editverse logo

Editverse

Study Design Consultation

Planning a program evaluation? Settle the design before you collect data.

You send: your research question and any draft plan or protocol.

You get: guidance on design selection, sample size and power, eligibility criteria, outcome measurement and bias control. Standard turnaround 3 to 5 business days.

When: You are choosing between a trial, a before-and-after design or another comparison.

Study Design Consultation · $400

Plan my design →

Editverse logo

Editverse

Ethics/IRB Support

Community program, human participants, committee application due?

You send: your study details, draft documents and your committee’s requirements.

You get: a prepared protocol, informed consent form, risk-benefit statement and data management plan, with help on committee queries. Standard turnaround 5 to 7 business days.

When: Your evaluation collects data from participants.

Ethics/IRB Application Support · $450

Prepare my application →

Editverse logo

Editverse

Statistical Analysis

Program data collected, but the analysis is not done?

You send: your data, research questions and preferred software.

You get: annotated SPSS, R or STATA files, an output interpretation guide and publication-ready tables. Standard turnaround 4 to 6 business days.

When: Baseline and follow-up data are ready to analyze.

SPSS/R/STATA Statistical Analysis · $400

Analyze my data →

Free Research Tools from Editverse

Frequently Asked Questions

What is the PRECEDE-PROCEED model?

A planning and evaluation framework for health programs. It starts from the quality-of-life result a community wants, works backward through the behaviors, environments and factors that shape it, then implements and evaluates the program in eight phases.1 The Rural Health Information Hub describes it as a structure for assessing health needs and for designing, implementing and evaluating programs to meet them.3

What do PRECEDE and PROCEED stand for?

PRECEDE stands for Predisposing, Reinforcing, and Enabling Constructs in Educational/Environmental Diagnosis and Evaluation. PROCEED stands for Policy, Regulatory, and Organizational Constructs in Educational and Environmental Development.1 Sources differ slightly on whether the first word is Educational or Educational/Environmental.1,3

What are the eight phases of the PRECEDE-PROCEED model?

PRECEDE has four phases: identify the ultimate desired result; identify priority issues and their behavioral and environmental determinants; identify the predisposing, enabling and reinforcing factors; and identify the administrative and policy factors. PROCEED has four: implementation, process evaluation, impact evaluation and outcome evaluation.1

What are predisposing, reinforcing and enabling factors?

Predisposing factors include knowledge, attitudes, beliefs, values and confidence. Enabling factors include the availability of resources, accessibility of services, laws and policies, and issue-related skills. Reinforcing factors are largely the influence of significant others in the social environment.1

How do I use the model for community assessment and intervention design?

Use Phases 1 and 2 for the community assessment: ask the community which result matters most, then which issues and behaviors stand in its way. Use Phases 3 and 4 to design the intervention: match each priority factor to an activity and check the organizational and policy conditions that decide what you can deliver.1

Does the PRECEDE-PROCEED model work?

Published evidence is positive but not clean. In a 2025 meta-analysis of 14 studies in type 2 diabetes, HbA1c was lower in the program groups by 0.41 to 0.63 percentage points, with high heterogeneity at 6 months.2 A 2022 review of 26 experimental or quasi-experimental studies found improved knowledge,4 and a 2025 review reported 51 of 54 studies with effective outcomes.5 These are reviews of studies that were published, and they cannot separate the model from the education delivered with it.

Do I have to use all eight phases?

No. The Community Tool Box calls the model a guide, not a prescription, and says parts can be changed, reordered or dropped to suit your setting.1 In practice many studies use only the planning half: 8 of the 26 studies in one review used the model only for planning.4

How do I evaluate a program planned with the model?

Evaluate in three layers. Process evaluation asks whether you did what you planned; impact evaluation asks whether behaviors and environmental factors changed; outcome evaluation asks whether the result chosen in Phase 1 was reached.1 The Rural Health Information Hub names the same layers as short-term (knowledge, skills, attitudes), intermediate (behavior) and long-term (incidence or prevalence) evaluation.3

References

  1. Community Tool Box, Center for Community Health and Development, University of Kansas. Chapter 2, Section 2: PRECEDE/PROCEED. The eight phases, the premises of the model and how to adapt it (accessed 31 October 2026). ctb.ku.edu/en/table-contents/overview/other-models-promoting-community-health-and-development/preceder-proceder/main
  2. Tang Y, Soh KL, Gan WY, Zhou J, Soh KG. The efficacy of PRECEDE-PROCEED model-based interventions on HbA1c and self-management in type 2 diabetes patients: a systematic review and meta-analysis. BMC Public Health. 2025;25:1980. Open access, CC BY 4.0. doi.org/10.1186/s12889-025-23073-9
  3. Rural Health Information Hub. PRECEDE-PROCEED Model, Rural Health Promotion and Disease Prevention Toolkit. Page last updated 13 June 2026 (accessed 31 October 2026). www.ruralhealthinfo.org/toolkits/health-promotion/2/theories-and-models/precede-proceed
  4. Kim J, Jang J, Kim B, Lee KH. Effect of the PRECEDE-PROCEED model on health programs: a systematic review and meta-analysis. Syst Rev. 2022;11:213. Open access, CC BY 4.0. doi.org/10.1186/s13643-022-02092-2
  5. Yang M, Hu Z, Qu X, Li M, Du W, Fan L. Utilization and effectiveness of the PRECEDE-PROCEED model as a tool in public health interventions: a systematic review. Iran J Public Health. 2025;54(10):2073-2088. doi.org/10.18502/ijph.v54i10.20121

Editverse Research Team | Updated October 2026