Study design and baseline cohort population

The ISFUC is a population-based cohort established through follow-up of participants from the 2021 STEPS survey. The STEPS 2021 survey provides the baseline data, whereas ISFUC constitutes the subsequent follow-up initiated on 4 April 2025. Moreover, participant recruitment and data collection have not been completed and are expected to be finished by 11 June 2026. Preliminary findings on mortality and health outcomes related to non-communicable disease risk factors are expected by late July 2026 and will be limited to descriptive statistics and initially available linked data following basic data cleaning procedures. The baseline cohort for the ISFUC comprises participants recruited in the 2021 STEPS survey, the eighth and most recent STEPS round in Iran [27]. STEPS 2021 was a nationally representative cross-sectional baseline survey undertaken in early 2021 among Iranian adults aged 18 years and older across all 31 provinces. The survey employed a systematic cluster random sampling design, through which 28,821 participants were selected from 3,176 clusters in both rural and urban areas nationwide. In accordance with WHO methodology and national priorities, STEPS 2021 consisted of three sequential steps. The first step involved interviewer-administered questionnaires focusing on sociodemographic characteristics and major metabolic, nutritional, and behavioral risk factors, injury history, health-related quality of life, as well as selected NCD-related medical history. A total of 27,874 participants completed this step. In the second step, 27,745 participants completed standardized physical assessments, encompassing height, weight, waist and hip circumference, and blood pressure. The final step involved 18,119 people aged ≥25, who provided blood and urine samples. The ISFUC study emphasizes on longitudinal assessment over a 5-year follow-up period concerning mortality, healthcare utilization, injuries, reproductive health, fertility-related outcomes, and population growth attitudes.

Follow-up procedures and participant selection

This follow-up study was initiated following ethical approval five years after the establishment of the baseline cohort, with follow-up commencing in April 2025. All participants who completed the first step of the 2021 STEPS survey (N = 27,874) were eligible for inclusion in the present study. From the 27,874 eligible participants who completed Step 1 of the 2021 STEPS survey, 14,010 individuals will be selected using a disproportionate stratified random sampling approach, stratified by province.To preserve national representativeness and ensure adequate provincial-level representation, a minimum sampling threshold will be determined for each province based on the smallest provincial sample proportion in the STEPS 2021 survey. Consequently, smaller provinces will be sampled at relatively higher fractions, whereas more populous provinces will be sampled at lower fractions. All analyses and results in following reports will be weighted based on the 2016 national census, taking into account province, age, sex, and place of residence, in order to ensure the generalizability of the findings to the Iranian population. Follow-up will be conducted through structured telephone interviews by trained staff using provided smartphones with voice recording ability, subsequent to obtaining renewed verbal consent and strictly adhering to confidentiality protocols. Participants will be contacted between 9:00 a.m. and 8:00 p.m. Upon each call, interviewers will introduce themselves, elucidate the study objectives, and inform participants that the interview will last approximately 5–10 minutes. Participants will then be asked to confirm their identities and confirm their availability for participation.

As illustrated in Fig 1, call outcomes are categorized into six predefined groups according to participant responsiveness. A successful call is defined as contact with an eligible participant who consents to participate and completed all sections of the questionnaire. Participants who migrated outside Iran or are confirmed to be deceased by relatives will be documented accordingly, with completion limited to the call-status form. Non-cooperation denotes those who respond the call but declines participation, for whom only call-status information is recorded. An unsuccessful call is defined as a wrong number or a failure to establish contact after three attempts, each made at a different time of day and spaced at least 24 hours apart. In such cases, alternative contact numbers and, subsequently, relatives’ contact numbers will each be attempted. Finally, when participants are temporarily unavailable, a preferred callback time will be arranged and recorded in the call-status form to enable future follow-up. Standardized call-status categories are utilized for each contact number to guarantee uniform reporting and rigorous oversight of response results.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Study flowchart.

Abbreviations: STEPS: STEPwise approach to NCD risk factor surveillance; ISFUC: IRAN-STEPS Follow-Up Cohort.

https://doi.org/10.1371/journal.pone.0356802.g001

Variables and questionnaires

Prior to the initiation of the study and interviewer recruitment, an expert panel discussion was conducted to determine the study questions and their details. At this stage, following a review of the scientific literature and preliminary interviews with health experts and specialists, key indicators were identified and developed to assess major health outcomes, including hospitalization, physician visits, and mortality. Based on these indicators, the initial set of questions and a data collection checklist were developed. After drafting the questions, the questionnaire was evaluated to ensure that the wording and terminology were appropriate and comprehensible for the general population. In addition, an interviewer guide was prepared alongside the data collection checklist to provide necessary clarifications for questions requiring further explanation. All variables and corresponding questions are presented in S1 Table in detail.

Face and content validity

Face validity was assessed qualitatively by a panel of approximately 20 participants, including methodologists, subject-matter experts, and members of the target population. They reviewed the checklist for clarity, simplicity, and comprehensibility of wording and suggested modifications where necessary [33,34]. Content validity was assessed by a panel of 10 experts, comprising methodologists, subject-matter experts, and representatives of the target population, using the Content Validity Ratio (CVR) and the Content Validity Index (CVI). For CVR, experts rated each item on a three-point Likert scale (“essential”, “useful but not essential”, “not essential”) and the CVR was calculated using Lawshe’s method [35]; items with CVR < 0.62 (based on 10 experts) were revised and re-evaluated, and removed if still below the acceptable threshold. CVI was calculated using the Waltz and Bausell approach by rating each item for relevance, clarity, and simplicity on a four-point Likert scale. Item-level CVI (I-CVI) was computed as the proportion of experts assigning a score of 3 or 4; items with I-CVI ≥ 0.79 were retained, 0.70–0.79 were revised, and < 0.70 were removed [33]. Scale-level CVI (S-CVI/Ave) was calculated as the average of I-CVI values across all items, and an S-CVI/Ave ≥ 0.80 was considered acceptable [33].

Data management, quality assurance

Standardized data management and quality control protocols will be implemented to ensure accuracy and completeness. Data will be entered directly into electronic tablets during interviews and securely uploaded to a central server within one week. To minimize the risk of data loss due to hardware malfunctions, communication failures, or unintentional deletion, regular data backups will be conducted systematically, and backup storage devices will be maintained at both field and central levels. In addition, each interviewer will be provided with both a laptop and a mobile device to ensure continuity of data transfer and storage in case of technical problems with one device. Quality control will encompass random audits of audio-recorded interviews compared to entered data, weekly assessments of interviewer performance accompanied by feedback and retraining, random field supervision, and periodic interim analyses to identify and resolve issues. In addition, weekly back-check calls will be conducted for a random sample of completed interviews to assess participant satisfaction and to re-administer selected questionnaire items; responses will be cross-verified against recorded forms, and in cases of discrepancies, the interviewer’s forms will be subjected to more intensive review during the same week. Interviewers will submit weekly activity reports, while the executive manager will engage in daily communication with the field team.

Outcomes

In this cohort study, outcomes were defined a priori based on the structured questionnaire domains and were aligned with the main objectives of assessing health status, healthcare utilization, and fertility-related behaviors and attitudes over time. The primary outcomes included mortality, defined as all-cause death occurring within five years after the baseline assessment; incident hypertension and incident diabetes mellitus were defined as new cases of physician-diagnosed hypertension or diabetes or initiation of antihypertensive or antihyperglycemic drugs reported during follow-up among participants free of the respective condition at baseline. Incidence estimates will therefore be based on self-reported current disease status and medication use at follow-up among participants without the corresponding condition at baseline. Healthcare utilization as a primary outcome was defined as the need for inpatient care within five years after baseline, including any physician or clinic visits, home physician visits, emergency department admissions, or hospitalizations, as well as the frequency of these encounters during the follow-up period.

The secondary outcomes encompassed a broad range of health-related, reproductive, and behavioral indicators assessed longitudinally. These included patterns of outpatient care (type of provider, public versus private sector use, and reasons for visits), injury-related outcomes such as traffic accidents and other unintentional or intentional injuries requiring medical attention, and detailed hospitalization characteristics including causes, type of facility, and duration of long-term admissions. Additional secondary outcomes included reproductive health indicators and fertility-related behaviors and attitudes, with sex-specific measures for women and men, such as pregnancy history, infertility, contraceptive use, utilization of reproductive health services, fertility intentions, and perceptions and use of governmental childbearing support policies.

Statistical analysis plan (SAP)

Data cleaning and validation will be conducted by two independent analysts. Any discrepancies will be resolved through review by a third analyst, and the finalized dataset will be stored securely with restricted access to preserve confidentiality and data integrity.

Descriptive analyses will be used to summarize baseline characteristics and follow-up outcomes. Continuous variables will be expressed as weighted mean (95% CI) or weighted median [95% CI], depending on distributional properties. Categorical variables will be reported weighted percentage (95% CI). The primary time-to-event outcome will be all-cause mortality. Mortality will be analyzed using Cox proportional hazards regression, with time calculated from the baseline STEPS assessment to death. The proportional hazards (PH) assumption for the Cox proportional hazards models will be assessed using Schoenfeld residuals. For common binary outcomes, log-binomial regression or Poisson regression with robust variance estimation will be used to estimate risk ratios (RRs). Logistic regression will be used only if these models fail to converge. Count outcomes, such as number of outpatient visits, hospital admissions, or injury events, will be analyzed using Poisson regression or negative binomial regression in the presence of overdispersion. Ordinal outcomes will be analyzed using ordinal logistic regression, while nominal categorical outcomes with more than two categories will be analyzed using multinomial logistic regression. All regression models will be adjusted for prespecified covariates selected a priori on the basis of the study objectives and available baseline data. These will include age, sex, marital status, province, urban/rural residence, education, occupation, and relevant baseline risk factors, as appropriate for each outcome. Additional covariates may be included based on biological plausibility and prior evidence. Model assumptions will be assessed before final inference. Where necessary, transformations or alternative model specifications will be used.

For missing covariate data, the handling strategy will depend on the proportion and presumed mechanism of missingness. When covariate missingness is between 5% and 10% and considered to be missing at random, complete-case analysis will be performed. However, when missingness exceeds 10%, sensitivity analyses using multiple imputation methods will be conducted. Results obtained from imputed datasets will be compared with complete-case analyses to evaluate the robustness of findings under different missing-data assumptions [36,37].

The final analytical weights for ISFUC will be derived by re-adjusting the original STEPS 2021 sampling weights to reflect the differential sub-sampling probabilities across strata. Consequently, analytical weights will be derived via post-stratification to the 2016 national census. For each stratum will be formed by the cross-classification of province, sex, urban/rural residence, and age category, the weight will be computed as W = N/n, where N is the census population count and n is the number of respondents in that stratum. These weights will adjust for the disproportionate stratified sampling used to select the follow-up cohort and for differential non-response, ensuring that the weighted estimates are representative of the Iranian adult population.

Strengths and limitations

A major strength of ISFUC is the integration of multiple outcome domains that are directly relevant to current public health priorities in Iran. With a large national sample spanning all 31 provinces, it supports the generalizability of the findings to the Iranian population. The ISFUC enables comprehensive linkage of demographic, metabolic, nutritional, behavioral, anthropometric, and biomarker data with mortality and major NCD outcomes, thereby expanding current evidence in the Iranian population. In addition to conventional epidemiological outcomes such as mortality, incident hypertension, and diabetes, the cohort systematically captures outpatient and inpatient healthcare utilization, providing policy-relevant indicators of healthcare demand, access, utilization patterns, and system burden. Therefore, beyond its role as an epidemiological cohort, ISFUC also serves as a valuable health services research platform by enabling assessment of patient interaction with healthcare systems and patterns of healthcare utilization. These measures offer important insights for health system planning, resource allocation, and evaluation of service delivery, particularly in the context of rising NCD prevalence and constrained healthcare resources. In addition, the cohort also enables assessment of traffic- and non-traffic-related injuries at the population level. Furthermore, Iran is currently experiencing an epidemiological and demographic transition characterized by a high burden of non-communicable diseases alongside declining fertility rates and population aging. This “double burden” may have substantial long-term implications for population health, healthcare demand, and socioeconomic development. Thus, as the first nationally representative cohort of its kind, ISFUC provides a unique opportunity not only to evaluate fertility patterns and the population-level impact of recent population growth policies, but also to investigate the potential relationships between NCDs, related risk factors, and fertility outcomes within the STEPS framework. Moreover, the study relied primarily on telephone interviews which provide a cost-efficient approach for follow-up data collection in low-resource settings [29,30]. However, experience with telephone-based cohort studies in low-income countries remains limited worldwide [30].

Notwithstanding these strengths, several limitations inherent to the study design should be acknowledged. First, the follow-up duration may be insufficient to capture the full incidence of diseases with long latency periods such as diabetes complications. Furthermore, these data rely in part on self-reported information, specifically variables such as inpatient care over a 5-year period, may introduce recall bias or misclassification. However, these limitations were partially mitigated through standardized instruments and interviewer training. In addition, the assessment of hypertension and diabetes based on self-report, rather than objective measurements as in the baseline STEPS protocol, may lead to underestimation of their true prevalence. Therefore, findings should be interpreted with caution, particularly when comparing follow-up estimates with baseline STEPS data. Moreover, although loss to follow-up remains a concern, systematic re-contact protocols were implemented to enhance follow-up comprehensiveness. Furthermore, loss to follow-up may exhibit a social pattern, potentially linked to participants’ limited access to communication resources and underlying socioeconomic constraints. Given the telephone-based design of the present cohort, differential follow-up may also arise from the “digital divide,” potentially leading to underrepresentation of marginalized populations with unstable or limited access to telecommunications services [43]. To mitigate this source of bias, multiple contact attempts at different times of day and the use of alternative contact numbers, when available, were implemented to improve participant accessibility and follow-up completeness. In addition, culturally shaped reporting behaviors and social desirability effects may introduce response bias. Consequently, future analyses should interpret findings with caution and consider conducting sensitivity and subgroup analyses to assess the potential impact of differential follow-up, cultural factors, and digital divide–related bias. Taken together, the ISFUC cohort offers a valuable, policy-relevant platform for monitoring emerging public health challenges in Iran, while its methodological limitations underscore the need for interpreting results with caution.