Analysis of Risk and Protective Factor Surveillance for Noncommunicable Diseases Using A MultiMode Data Collection Approach

Abstract Telephone surveys have faced challenges, including declining response rates and increasing costs, prompting the need for alternative data collection methods. Multimode data collection has emerged as a viable approach, producing high-quality data while reducing costs and improving response rates. This study evaluated the feasibility of a multimode data collection approach for the surveillance of risk and protective factors for noncommunicable diseases in Brazil, assessing its impact on response rates, costs, and data quality compared to a telephone only survey. In the multimode study individuals were invited via Short Message Service to participate in an online survey, and non-respondents, after reminders, were subsequently contacted by telephone. Data from the multimode study were compared with telephone-only interviews conducted using the Vigitel 2023 methodology. Design weights were applied to adjust for differential nonresponse, and final calibration weights aligned the sample’s sociodemographic composition with the target population through the raking method. The sociodemographic composition of the samples was compared. The prevalence of 23 health indicators was compared between the samples. Method effect was defined as the percentage variation in indicator prevalence between the multimode and telephone-only, reflecting differences attributable to data collection mode. Performance indicators, including eligibility rate, response and refusal rates, average interview cost, and interview duration, were also assessed. The multimode study showed sociodemographic similarity to the target population. Among 23 health indicators, only one differed between surveys (weighted comparison). Three indicators (passive smoking at home, e-cigarette use, and negative self-rated health) showed high method effect values, and one (obesity) was negatively associated with the multimode study (Prevalence Ratio: 0.68, 95 percent CI: 0.56–0.82). These findings indicate that multimode data collection is a viable option for the surveillance of risk and protective factors for noncommunicable diseases, yielding prevalence estimates comparable to those from telephone-only surveys, with greater operational speed and slightly lower cost.