|Chiu, Chung-jung - Jean Mayer Human Nutrition Research Center On Aging At Tufts University|
|Mitchell, Paul - University Of Sydney|
|Klein, Ronald - University Of Wisconsin|
|Klein, Barbara - University Of Wisconsin|
|Chang, Min-lee - Jean Mayer Human Nutrition Research Center On Aging At Tufts University|
|Gensler, Gary - Emmes Corporation|
|Taylor, Allen - Jean Mayer Human Nutrition Research Center On Aging At Tufts University|
Submitted to: Ophthalmology: Journal of The American Academy of Ophthalmology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/14/2013
Publication Date: 3/7/2014
Citation: Chiu, C., Mitchell, P., Klein, R., Klein, B.E., Chang, M., Gensler, G., Taylor, A. 2014. A risk score for the prediction of advanced age-related macular degeneration: Development and validation in 2 prospective cohorts. Ophthalmology: Journal of The American Academy of Ophthalmology. DOI:10.1016/j.ophtha.2014.01.016.
Interpretive Summary: Age-related macular degeneration (AMD) is a blinding disease. It accounts for over 50% of cases of legal blindness in the US. The disease also significantly reduces quality of life and consumes over 50% of eye care costs in the Medicare budget. As Americans are ageing rapidly, AMD has become a major personal and public health issue. Unfortunately, the disease often provides few warnings to alert patients to seek treatment. Therefore, early identification and close follow-up of those at high risk are important to delay the disease progressing into blindness. In this study, we developed a prediction model for both general and eye doctors to predict patients' risk of developing advanced AMD. The prediction algorithm summarizes a patient's personal and eye information into a risk score. An application program (App) for iPhone/iPad use was also developed to make the algorithm easy to use. This practical tool may be useful to clinicians to inform patients of the risk of developing advanced AMD and to guide treatment plans to prevent the disease from occurring.
Technical Abstract: We aimed to develop an eye specific model which used readily available information to predict risk for advanced age-related macular degeneration (AMD). We used the Age-Related Eye Disease Study (AREDS) as our training dataset, which consisted of the 4,507 participants (contributing 1,185 affected vs. 6,992 unaffected eyes during mean 6-y follow-up) recruited from 11 eye clinic across the US. Employing Bayes’ theorem in a logistic model, we used 8 baseline predictors (corresponding to MRj): age, sex, education level, race, smoking status, and presence of pigment abnormality, soft drusen, and maximum drusen size, to devise a macular risk scoring system (MRSS). The MRSS was then tested in the Blue Mountains Eye Study (BMES) cohort consisting of 2,169 residents (69 affected vs, 3,694 unaffected eyes during 10-y follow-up) recruited from the great Sydney area. We assessed the performance of the MRSS by calculating sensitivity, specificity, and the area under the receiver operating characteristic (ROC) curve (i.e. c-index). The internal validated c-index AREDS (0.88, 95% CI:to 0.89) and the external validated c-indexBMES (0.91, 95% CI:0.88 to 0.95) suggested an excellent performance of the MRSS [(=-6.26-ln (disease odds by eye in a target population)+ sigmaMRj)]. Our analysis also suggested the feasibility of using “indidence by person” to estimate “disease odds by eye” in the MRSS. Accordingly, the practical form of the MRSS can be expressed as: [(-6.26-ln (incidence by person in a target population/2)+ sigmaMRj)]. The MRSS was successfully developed and validated to provide satisfactory accuracy and generalizability.