Reliability of Cancer Treatment Information on the Internet: Observational Study (Preprint)

Abstract

BACKGROUND Finding the correct medical information in a flood of information from the internet is a significant issue for patients with cancer. OBJECTIVE We investigated the reliability of the information on cancer treatment methods available on the internet based on an evaluation by medical oncologists, medical students, and cancer survivors. METHODS Using Google and Yahoo as the search engines, we carried out the information search using 2 keywords, “cancer treatment” and “cancer cure,” and the top 20 information sites were identified. A similar search was conducted on 5 types of cancer. The reliability of the information presented was rated on a 3-level scale (A, B, or C). Level A referred to reliable sites (providing information complying with the clinical practice guidelines for various types of cancer), Level B included sites not falling under either Level A or Level C, and Level C comprised dangerous or harmful sites (providing information on treatment not approved by the regulatory authority in Japan and bombastic advertisements without any relevant clinical evidence). The evaluation was conducted by medical oncologists, medical students, and cancer survivors. The consistency of the information reliability level rating between the medical students or cancer survivors with that of the medical oncologists was assessed by using the kappa value. RESULTS A total of 247 sites were evaluated for reliability. The ratings provided by the medical students’ group were as follows: Level A, 12.1% (30/247); Level B, 56.3% (139/247); and Level C, 31.6% (78/247). The ratings provided by the cancer survivors’ group were as follows: Level A, 16.8% (41/244); Level B, 44.7% (109/244); and Level C, 38.5% (94/244). The ratings provided by the oncologists’ group were as follows: Level A, 10.1% (25/247); Level B, 51.4% (127/247); and Level C, 38.5% (95/247). The intergroup rating consistency between the medical students’ group and oncologists’ group was 87.4% (216/247, kappa=0.77) and that between the cancer survivors’ group and oncologists’ group was 76.2% (186/244, kappa=0.61). CONCLUSIONS Of the investigated sites providing information on cancer treatment on the internet, the percentage of sites that seemed to provide harmful information was much higher than that of sites providing reliable information. The reliability level rating was highly consistent between the medical students’ group and the medical oncologists’ group and also between the cancer survivors’ group and the medical oncologists’ group.