How can MyClearview be accurate and informative with just a few questions? Our goal is to get a good sense of likely cost from as few questions as pos

Transcription

1 What is MyClearview? MyClearview is a web application that employs mathematical models and algorithms to help individual consumers and their families identify their best choice among the medical insurance plans available to them. The choice might be from a short menu of plans offered by the individual s employer, from a medium-sized menu on a private exchange, or from a huge menu of plans on a public exchange or a private marketplace for individuals. In any of these scenarios MyClearview sorts through the complexity of premiums, employee contributions, deductibles, co-payments, etc. to recommend the plan that best fits the individual s circumstances. How does MyClearview account for individual circumstances? MyClearview first predicts whether the individual (plus spouse and dependents as applicable) will be high, medium or low users of covered medical services during the future plan year. The user makes a high-level characterization of his/her propensity to consume and then clicks all that apply from a list of anticipated events and a list of anticipated conditions. Most individuals complete the questions in less than five minutes. An algorithm then quantifies the anticipated utilization in terms of a continuous function of percentile that asymptotically approaches zero and one hundred at either end. The assigned percentile will correspond to a predicted amount of consumption for each benefit design considered, adjusted for demand elasticity as a function of benefit richness, and given the typical exponential distribution of claimants. How does MyClearview predict out-of-pocket cost for each benefit design? MyClearview considers not only the way that total claim cost is distributed from lowest percentile to highest percentile, but also the way costs are variably distributed by type-of-service as a function of overall percentile. For example, a 10th percentile user might typically spend 80% of his medical dollars on office visits and drugs while a 95th percentile user might typically spend 90% of his medical dollars on hospital services. So the cost sharing features for office visits and drugs will be relatively more important drivers of value for the low utilizer and relatively less important drivers for the high utilizer (for whom the hospital benefits and out-of-pocket maximum will be big drivers). Two different benefit designs could each have an overall actuarial value of 80% (i.e. both are gold plans) but plan A could have higher cost-sharing for office visits and drugs while B has greater cost-sharing for hospitalization and perhaps a higher OOP-max. At any given percentile level, MyClearview will predict different out-of pocket costs for plan A vs. B (A higher than B for the 10th percentile user and B higher than A for the 95th percentile user). MyClearview separately considers the cost-sharing features of each type of service and meticulously accounts for the applicability of deductibles and maxima to each type of service. 1

2 How can MyClearview be accurate and informative with just a few questions? Our goal is to get a good sense of likely cost from as few questions as possible. From a user perspective less is better because tedium deters use. The conditions and events lists were designed to be as succinct as possible while still being comprehensive. We sought to cover the waterfront in terms of organ systems and (with carefully-considered exceptions) to cover most of the content of a traditional individual underwriting form, which is what carriers used for years to predict expected claims cost. So the conditions list asks broadly about significant disorders of the cardiovascular system, the respiratory system, the immune system and the nervous system. The list is rounded out by three items that cross organ systems: cancer and the two others (requiring multiple office visits and advanced diagnostics). The events list covers the reproductive / genitourinary system and endocrine systems specifically and crosschecks with surgery, inpatient stay, multiple drugs and expensive drugs. What did we intentionally leave out? From the standpoint of organ systems we ask nothing specific about the integumentary system (skin), musculoskeletal system or gastrointestinal systems. Our concern here was that they would invite affirmative responses for minor costdrivers: sprains, strains, rashes, reflux, etc. The costly conditions associated with those systems are either unpredictable (e.g. broken bones) or covered by cancer, immune system or services. With respect to traditional underwriting questions, we left off specific mention of back pain, depression and arthritis. These conditions are incredibly prevalent and the way carriers used them for underwriting was to predict worst case cost, not expected cost. The great majority of people with those conditions are low-to-moderate utilizers. So we think it s more appropriate to let surgery, hospitalization and multiple/expensive drugs identify the likely high-cost people. And don t forget that underlying the conditions and events is the propensity-to-consume question as a starting point. Generally speaking, individuals know better than the insurance underwriter whether or not they will be high utilizers. To the extent that something big is otherwise missing, MyClearview users can simply signal high use through the first question. Finally, why don t we ask for age and sex? Both of course correlate with higher or lower utilization and cost. But they don t operate independently from conditions and events. For example, women in the commercially-insured age category cost on average a bit more than men. That s mostly driven by reproduction which we ask about specifically. And of course likely cost increases with age, but it does not increase independently of the propensity to consume, conditions and events we ask about. That is to say, the older person is more likely to give affirmative responses to our questions. But we don t want to assume that an older person who 2

3 self-reports low propensity to consume and foresees no events or conditions should map to a higher percentile of cost/utilization. A 60 year old can be a 10th percentile user just like a 30 year old can be, and we don t automatically want to assume otherwise. What is the basis of the MyClearview value ranking? Simply stated, medical coverage provides benefits at a cost. The benefits (credits in accounting terms) are quantified as the value of covered services consumed, as well as any employer donation to an H.S.A. account if applicable. The value of services consumed is after discounts, i.e. at the level of typical carrier allowances. The costs (debits in accounting terms) are premium and out-of-pocket cost. For purposes of MyClearview calculations, premium is not necessarily the full carrier premium, but instead is the net premium to the individual after employer contribution, government subsidy, etc. The value of a particular option is simply the sum of the four parameters: services, H.S.A. donation, premium and cost, where services and H.S.A donation are positive numbers and premium and cost are negative numbers. The best value is the option with the highest sum of benefits and costs, and the value ranking simply reflects the ordinal positions of the sums. Note that for projected low utilizers the sum of benefits and costs might be negative. In this case coverage serves the traditional function of catastrophic protection, but best value is still the highest (or least negative) sum. Also note the important distinction between an H.S.A. donation and an available H.R.A. The former is a guaranteed transfer of funds regardless of utilization. In contrast, the H.R.A. acts to attenuate out-of-pocket costs in cases of higher utilization. So MyClearview evaluates H.R.A. availability as a benefit design enhancement that lowers the OOP cost in cases of higher utilization, not as a direct credit. What makes MyClearview different from other individual decision-support tools? MyClearview takes less of the user s time by an order of magnitude (e.g. 3-5 minutes vs. 30 minutes or more). Most decision-support tools operate as conceptually-simple out-of-pocket calculators, multiplying anticipated units of service by average unit costs, and then aggregating. Because of the conceptual construct, the user has to convey to the system specific numbers of expected office visits, drugs, surgeries, etc. This can be tedious and time consuming, tends to undervalue risk-pooling for the unpredictable, and does not account for demand elasticity (which the average user tends not to consider). Further, many decision-support tools ask for a great deal of information about personal income, wealth and risk aversion. These questions feel both tedious and intrusive to many users. Worse, the 3 answers are applied to make subjective judgments about people s un-suitability for leaner plans, in particular high-deductible plans, and in the case of some private exchanges to push the sale of ancillary products. MyClearview departs from the usual in that it asks no financial questions. We believe that what is best for 3

4 everyone, from the low to the highly compensated, is the plan that is likely to provide the most services for the lowest total cost. Is a person living paycheck-to-paycheck worse off with the possibility of an unscheduled $3,000 expense or with the certainty of $3,000 extra in payroll deduction? We think that to presume the answer is at best un-scientific and at worst actively detrimental. The tendency of most people is to over-insure, and our objective analysis will in many cases tend to counteract that propensity. Implementation of an effective employee health plan decision support tool will support many employer health benefit strategies, especially those encouraging consumerism, employing defined contribution, increased choice and a sensitivity to the needs and interests across generations and other employee segments. But MyClearview s clear purpose is to help individuals pick the health plan that bests fits their personal circumstances, resulting in the highest value coverage. 4

5 About Clearview Logix, LLC Clearview Logix was founded by Larry Colley, M.D. and Tim O Shea, experienced health insurance executives and founders of a successful midmarket employee benefits consulting firm (Dominion Benefits). The consulting firm s growth to thousands of clients was propelled by patented employer decision support analytics, which brought sophisticated analysis to midmarket employers (typically 50-1,500 employees). Larry Colley, M.D. was the primary inventor of the BENEFIT SELECT suite of employer tools, which has received one patent and another patent pending. Larry is the chief product development officer of Clearview Logix and is the architect and developer of the applications and algorithms behind MyClearview. Larry is a graduate of Yale University and the University of Virginia School of Medicine and has deep experience in Health Services and Outcomes Research. He was also a Medical Director for successful Blue Cross and Blue Shield organizations earlier in his career. Tim is the chief operating officer of Clearview Logix, working with clients to ensure productive and successful outcomes. Previously, Tim was the managing principal of Dominion Benefits and, following acquisition, the managing principal for Digital Benefit Advisors in Virginia. Tim is a graduate of the University of Virginia as well as its Darden School of Business. Early in his career, Tim held various executive positions with Trigon BCBS and was later regional CEO for UnitedHealthcare. Allen Hatzimanolis is the chief technology officer of Clearview Logix, responsible for transforming the firm s proprietary analytical models into scalable web applications for use by employers, exchanges, and licensees. Previously, Allen was the managing partner of Evenspring s health care and analytics practice, which was acquired by Clearview Logix in October Allen is a graduate of the University of Virginia and a member of the CFA Institute. 5