Semantic analysis of social media messages of patients with neovascular age-related macular degeneration and diabetic macular edema by open Internet sources — a study of patients' opinions in real clinical practice

Abstract

Purpose: to analyze social media messages of patients with neovascular age-related macular degeneration (nAMD) and diabetic retinopathy (DR), or their careers in order to investigates the patients’ opinion in the condition of real clinical routine. Material and methods. Real-life anonymized stories of patients from Russian-language open Internet sources (forums, social networks in Russia) were processed by artificial intelligence techniques: the technologies of automated analysis of unstructured natural language texts, including semantic technologies. In these messages, patients and their careers (mainly, family members) openly and in an ‘uncensored’ way share their experience in diagnostics and treatment while looking for a second opinion or supporting each other. They use general social networks as well as specific disease-related forums or Q&A portals. We identified 73 098 DR/nAMD-related posts, including 13 138 posts by 844 DR patients and 358 posts by 212 nAMD patients. The posts were analyzed in several steps with the technologies of automated analysis of unstructured natural language texts including semantic technologies aimed at processing large volumes of data. The semantic analysis of texts dealt with the whole meaning rather than individual keywords. Results. We obtained information on the patients’ characteristics and treatment plans of retinal diseases in real practice but also on the patients’ attitude to their condition, diagnostic and curative procedures, their needs and difficulties experienced during treatment. The nAMD and DR patients have a low level of Internet activity and poor awareness of these diseases as compared with the patients suffering from non-ophthalmological diseases with lower prevalence (breast cancer, multiple sclerosis, etc.) or other ophthalmological disorders. Most of the content for DR was produced by the patients’ relatives (82.6 % of messages), and for nAMD — by the patients themselves (65 %). The key item for DR patients was diabetic microvascular manifestations (over 42 000 posts discussed ‘diabetic foot’ and only 681 ‘diabetic retinopathy’). Quality of life (QoL) was shown to be significantly affected with inability to work as a major burden for 30 % of nAMD patients, and diabetes-associated comorbidities as a key factor compromising QoL in 20 % of DR patients. In nAMD patients, the average time-to-diagnosis after disease manifestation was 1 year (35 % patients reported 1–2 months), in DR, over a half of the messages mentioned 1–2 years. The key reasons for visiting the clinics included in-depth eye exams (OCT mentioned by 59 % of nAMD patients) and treatment (24.1 %). Only 33.2 % of nAMD patients and 7 % of DR patients noted that they received anti-VEGFs. Treatment unaffordability is one of the key barriers. The patients lack clear understanding of the prognosis and effective treatment options. Conclusion. The study revealed low activity and awareness of nAMD and DR patients with regard to their diseases. This justifies the need of increasing computer literacy and awareness of effective treatment options and efficacy criteria not only in patients, but also in their younger relatives. The results confirm that, among the studied group of retinal patients, vision-related quality of life is compromised. We need to change at least several aspects of nAMD and DR patients’ management: reducing the time to diagnosis, prescribing effective treatment options and increasing the availability of these options.