Selective attrition and representativeness in the thyroid cohort study: analysis of fine-needle aspiration biopsy and surgical treatment among patients with thyroid nodules

Abstract

The Ukrainian-American Thyroid Cohort was established for long-term follow-up of individuals exposed to ionizing radiation in childhood as a result of the Chornobyl nuclear accident. A prospective cohort approach allows us to establish causal relationships between radioiodine exposure and the development of thyroid pathology. At the same time, uneven participation or selective attrition of cohort participants during follow-up may be a source of systematic bias, thereby reducing the validity of risk epidemiological estimates for occurrence and progression of nodular and neoplastic thyroid lesions. This study aimed to assess participation of individuals with newly detected thyroid nodules in undergoing fine-needle aspiration biopsy (FNAB) and surgical treatment, and to determine the possibility of selective bias associated with their attrition during repeat examinations. Material and methods. The study included data from the second, third, and fourth screening cycles (2001–2008), during which clinical and ultrasound examinations of the thyroid gland were performed by inpatient and outpatient mobile teams based at the State Institution «V.P. Komisarenko Institute of Endocrinology and Metabolism of the National Academy of Medical Sciences of Ukraine» and regional medical facilities. Participants with thyroid nodules ≥10 mm or smaller with suspicious ultrasound features were referred for FNAB. If there were cytological signs of malignancy or clinical indications, surgical intervention was recommended. Study endpoints included either thyroid surgery
or the last available ultrasound examination with complete data. Participation rates for FNAB and surgery among individuals who were shown these procedures was compared with the sociodemographic and dosimetric characteristics of those who did not show up. Statistical analyses were performed using JASP. Results. Across the three screening cycles, newly diagnosed thyroid nodules were found in 1109 participants of thyroid cohort; TABP was shown in 871 individual (78.5%) according to the specified criteria. Among them, 806 participants (92.5%) arrived for FNAB, repeat assessment demonstrated the absence of indications for the procedure in 14.8% of cases. FNAB was actually performed in 687 individuals (78.9% of those recommended). Many participants agreed to the procedure following repeated contact or logistical support. Based on cytological findings, thyroid surgery was recommended for 187 individuals (27.2% of those who underwent FNAB). The overall proportion
of participants who adhered to the medical recommendation and presented for surgery was 90.9%. In some cases, indications for surgery were not confirmed upon additional examination, while other participants underwent surgical treatment at different institutions. In total, 152 individuals (81.3% of those recommended) received surgical treatment. Sociodemographic analysis revealed no significant differences in age, sex, or type of settlement between those who underwent FNAB or surgical intervention and those who did not; the only exception was the distribution by place of residence at the time of the Chornobyl accident among individuals referred for FNAB. Thyroid radiation dose was not associated with participation in FNAB or surgery. These findings confirm the absence ofdose-related or socially caused attrition and support the ontinued representativeness of the cohort for risk estimation. Conclusions. The obtained data demonstrated a high level of patient participation
in both FNAB and surgical treatment, reflecting the effective organization of the study and strong participant trust. A comparison of baseline characteristics between individuals who underwent the recommended diagnostic and therapeutic procedures and those for whom these procedures were indicated revealed no substantial differences. This confirms the representativeness of the sample and indicates the absence of pronounced selective attrition, thereby ensuring the high reliability and validity of risk estimates for thyroid nodular disease development.