ARE “ COMBO DEBITS ” NECESSARY? THE CASE OF DIABETES AND SMOKING
Steven J. Rigatti, MD Medical Consultant CRL Glastonbury, CT steven.rigatti@crlcorp.com
Introduction At Clinical Reference Laboratories, we often receive interesting questions from our clients and partners. One recent such question was regarding the often- used practice of adding “combo debits” to the ratings of individual life insurance policies on those who smoke when they have a disease which is exacerbated by smoking or which accelerates the development of disease complications. For example, smoking is well known to promote and accelerate the development of heart disease. So, in many underwriting guidelines, an additional debit load is recommended for smokers with heart disease. Similarly, in diabetes, smoking is known to accelerate both the micro-vascular (e.g., nephropathy, neuropathy, retinopathy) and macro- vascular (e.g., coronary disease, peripheral vascular disease, etc.) complications of diabetes, and so combo debits are often recommended for diabetic smokers. The question we seek to answer in this article is: “Does the available data provide evidence for or against this practice?” Methods To answer this question, the CRL data is narrowed to include only those ages 20-80 who have answered “yes” or “no” to questions about their history of dia - betes and their use of tobacco products. A “smoker” is defined as anyone answering “yes” to either of the tobacco questions (cigarette or non-cigarette use) or who tests positive for cotinine at a level of 200 ng/ml. A “diabetic” is defined as anyone answering “yes” to the question about their history of diabetes or who has a hemoglobin A1c level of 7.0% or higher. Note that the recommended diagnostic threshold for hemoglobin A1c is 6.5%, but the higher level is used to help reduce false-positives and because the of- ficial guidelines say that there should be at least two such tests. After this filtering, the data contain over 15 million records with 83,685 deaths and an aver -
Executive Summary It is a common practice in life insurance to rate diabetic smokers with debits for their diabetes and “ combo ” debits for smok- ing. In this article, data from Clinical Reference Labs is used to determine if this common practice is justified. The results show that, generally, the mortality risk of diabetic smokers, while higher than smoking non-diabetics, is covered by an appropriate rating for diabetes without the ap- plication of additional combo debits. age exposure time of 9.1 years. The determination of death was made using the most recent Social Security Death Master File. The analytic strategy is three-fold. First, Cox models will be used to determine the hazard ratios (HRs) for each group, male and female. For this first model, a single variable will be constructed indicating the presence or absence of diabetes and smoking together and indicated as “neither,” “diabetes only,” “smoking only” or “both.” If the HR for “both” is higher than the product of the HRs for “diabetes only” times “smoking only,” then this would suggest that combo debits are necessary. If it is less than or equal to this product, then it would suggest that the combo debits are not necessary. Note that all Cox models are controlled for age. In the second analysis, Cox models will be used as well, but this time two variables will be used to classify applicants, one for smoking (“yes”/“no”) and one for diabetes (“yes”/“no”). Then, an interaction term will be added. While it is not necessary to understand the math behind it, the hazard ratio produced by this term has the following properties. If it is greater than 1 (and statistically significant), it would suggest that the risk is greater than that produced by smoking and diabetes alone, and that combo debits are appropriate. If it is
ON THE RISK vol.40 n.3 (2024)
70
Powered by FlippingBook