Does Internet Data Collection Improve Cohort Retention?

RESEARCH QUESTIONS: In longitudinal research, will participant-initiated internet data entry improve 1) participant retention and 2) data quality? RESEARCH CONTEXT: While longitudinal research may be the best tool for examining the course of human health and illness over the lifespan, the burden of participation over many years often results in loss of participants to follow-up. Such attrition reduces statistical power and threatens the internal validity and generalizability of inferences drawn from longitudinal data (Shadish et al., 2002). Because internet data collection is convenient and simplifies complex questionnaires, this method may reduce participant burden and reduce attrition. While internet data collection is already regarded as a viable tool for participant recruitment (Smith et al., 2007) and usually results in data of equivalent or better quality (e.g., factor loadings, reliabilities) than traditional paper-and-pencil methods (Vergnaud et al., 2011), internet data collection as a tool to increase participant retention within an existing longitudinal cohort has not been evaluated previously. The goal of the current project is to determine whether internet data collection increases participant retention and data quality within the APrON (Alberta Pregnancy Outcomes and Nutrition) study.

APrON is an ongoing longitudinal cohort study conducted in Calgary and Edmonton, involving more than 2000 women, 1200 of their partners, and their 2000 children. It is designed to answer questions about the impact of maternal nutrient intake and status during pregnancy on: a) maternal mental health, b) birth outcomes, and c) child health and development. The data collected from Albertans participating in health research (such as APrON) is a provincial treasure, an investment that needs to be maximized with easy, modern technology to reduce participant burden and promote better long-term follow-up. Although software is available for single questionnaire data entry (e.g., RedCap), the innovative technology the investigators plan to introduce and evaluate is a made-in-Alberta solution that automates and simplifies the collection of complex short- and long-term longitudinal data - a process that other software cannot easily achieve. This technology holds strong potential for broad application to all longitudinal health research.

Currently, APrON collects data using traditional paper questionnaires mailed out at regular intervals throughout pregnancy and as children develop over the first 3 years. At some time points, participants are also asked to attend a clinic where biological samples and physical measures are taken. The questionnaires are long and complex (whether a question needs to be answered is dependent upon responses to previous questions) and have been identified as a major time burden by study participants. In a recent survey of 1763 APrON participants, almost 65%, indicated they would prefer to complete study questionnaires via the internet - most of the remainder had no preference. Completion rate for paper questionnaires is approximately 85% when participants come to a study clinic for an in-person visit. However, when a clinic visit does not accompany the mailed questionnaires (questionnaires are completed at home and returned by postage paid envelope) response rates are ~ 65%.

RESEARCH PLAN/METHODOLOGY: Electronic Patient Reported Outcomes (ePRO) is a web-based software tool developed by the Women \& Children's Health Research Institute (WCHRI) at the University of Alberta. Pilot studies have been performed at the University of Alberta, however, this study will be the first use of ePRO in a study population of this size. Questionnaires are built using Checkbox Survey Server. Automatic email invitations to participants are based on the individualized anchor dates and schedules (e.g., the questionnaire for children aged 12 months is linked to the individual child's date of birth). Participants receive links and instructions when a survey is due for completion. Reminders are automatically sent if a questionnaire is not completed within appropriate timelines (e.g., after 2 weeks). Study staff can track late and incomplete questionnaires, re-send invitations, and manually enter data when information is obtained via a paper questionnaire or telephone interview. Importantly, the system facilitates linking of data from different questionnaires and different time points.

The investigators propose a randomized trial of data collection methods. Participants will be allocated using a permuted block randomization of equal block sizes from APrON participants in Calgary to one of two groups: (1) continue with paper-based data collection; and (2) use ePRO for future data collection tasks.

IMPACT/RELEVANCE: In 2008, Alberta invested significant research funds to establish pregnancy and birth cohorts that will help develop knowledge needed to improve the health of mothers and infants. Answering health questions about developmental effects requires continued contact and follow-up with the participants every 3-6 months, which is a significant challenge, and one which most international cohorts are struggling to accomplish with paper questionnaires. What the investigators propose in this grant is to evaluate a novel approach to complex longitudinal research that enables data collection that is convenient, less burdensome, and efficient for participants. The investigators expect that the ePRO system will: 1) increase completion rate by tracking participant progress in the study, notifying them when it is time for another round of questionnaires, administering questionnaires in an efficient manner (skip non-relevant questions), and sending reminders about incomplete questionnaires, and 2) improve data quality by alerting participants to missing items.