Widening access to online health education for lung cancer: a feasibility study
Julia Mueller, Alan Davies, Simon Harper, Caroline Jay, Chris Todd · 2016 · Proceedings of the 13th International Web for All Conference (W4A) · doi:10.1145/2899475.2899495
Summary
This paper presents a feasibility study of "When2Go," a web-based lung cancer symptom appraisal tool designed to reach populations typically excluded from online health information: older adults with low education levels, low health literacy, and psychological barriers such as fear, embarrassment, and stigma associated with lung cancer. The tool was developed at the University of Manchester using a user-centric approach with patient and public engagement groups and Think Aloud usability testing. Content was written at Flesch-Kincaid grade 8 reading level. The tool personalizes information by asking users about their symptoms (type, duration, severity), age, and smoking status, then generating recommendations based on UK clinical guidelines. Theory-based components drawn from the Theory of Planned Behavior include quotes from health professionals targeting specific psychological barriers—for example, addressing the common belief that seeking help would waste doctors' time. The study used a four-arm randomized design: theory-based and personalized (TP), theory-based only (T), personalized only (P), and usual care (UC, based on a typical lung cancer charity website). Built with Python Flask and SQLAlchemy, the tool was recruited through Google AdWords, social media, mailing lists, and referral websites.
Key findings
In 8 weeks, 1,898 users viewed the site and 97 completed the study. Critically, 97.9% reported symptoms warranting medical help-seeking, and 34% had education levels below school leaving qualification—demonstrating the tool successfully reached the high-risk, low-education target population. The tool achieved a 2.9% AdWords click-through rate (above the 2% average), and 25.7% of visitors arrived via organic Google search, indicating good findability for symptom-related searches. The theory-based, personalized condition (TP) produced significantly higher help-seeking intention than the personalized-only condition (M=5.58 vs M=3.5, p=0.01, r=0.46 medium-large effect). The condition lacking theory-based components showed the lowest intention scores, suggesting that psychological theory—particularly addressing fear and embarrassment through health professional quotes—is a key driver of behavior change. Personalized theory-based pages had dramatically lower attrition (3.9%) than static theory-based pages (30.8%, p=0.01). Education level was significantly negatively correlated with help-seeking intention (r=-0.49, p<0.01), meaning users with lower education reported higher intention—potentially indicating the tool is particularly effective for its target audience. However, differential dropout across conditions (the shorter UC condition retained more users) revealed that tool length must be equalized in future trials.
Relevance
This paper demonstrates how web accessibility extends beyond technical compliance to encompass health literacy, reading level, psychological barriers, and the digital divide experienced by older adults with low education. For accessibility practitioners, the findings highlight that personalization and theory-driven content design can significantly improve engagement and outcomes for populations traditionally excluded from online health information. The dramatic difference in attrition between personalized and static content (3.9% vs 30.8%) offers evidence that personalization is not just a nice-to-have but essential for retaining users with low health literacy. The study also illustrates how the Web can serve as an equalizer—reaching people who might not seek face-to-face health advice due to embarrassment or stigma. Limitations include the small sample sizes per condition (15-42 participants), interim nature of the results, and the confound between condition length and attrition. The study is UK-specific due to its basis in NHS clinical guidelines.
Tags: health literacy · healthcare accessibility · plain language · personalization · aging · digital inclusion · online health information · patient empowerment