Rule Sets in Engine-Driven Underwriting

Hank George, FALU, CLU, FLMI

The age of underwriting engines – defined as straight-through processing of some portion of an insurer’s new business – is at hand.

Even in the United States, where life underwriting is relatively more complex for a variety of reasons, we know now that almost 20% of carriers have an engine in place and the majority of the rest are either in process of acquiring/building an engine or at least studying their options in this regard.

There are many issues associated with underwriting engines, asevidenced by the 92-question survey undertaken last year by my company and the UK consulting firm SelectX. It was consciousness-raising for me to see how much I did not understand or fully appreciate about underwriting engines after drafting and executing survey, and then studying its findings.

Companies embracing engine technology are obliged to consider most if not all of these issues if they hope to realize their long term strategic objectives.

One core consideration is endeavoring to maximize the capacity of an engine to make as many decisions as possible in a manner consistent with sound risk appraisal. The fulcrum upon which this is balanced is the adequacy of the rule sets used to triage cases.

At one of my study group meetings last year, a member company reported that it was necessary to dedicate four experienced underwriters to this daunting task. Moreover, it took them no fewer than 12 months to manufacture the rule sets and other elements needed to optimize the functionality of their engine.

No doubt many companies flirting with the idea of acquiring or building an underwriting engine will underestimate the length and breadth of this commitment. This will be due in part to exaggerating the adequacy of generic rule sets provided by some engine manufacturers.

Writing rule sets will be misperceived by some companies as a rather facile undertaking.

After having fashioned rule sets for both reinsurer and direct company clients, it became all too apparent to me that this task is anything but facile. Indeed, it is likely that only a small share of even the most experienced underwriters will attain a comfort zone in this regard.

The output of underwriting engines depends entirely on the caliber of the rule sets driving its actions. If key elements are absent, or if their relative significant is either understated or exaggerated, there could be potentially ominous consequences going right to the insurer’s bottom line.

Clearly, the best way to gather risk histories to populate rule sets is the teleinterview. This is because teleinterviews allow us to ask the essential questions about a given risk history which directly correlate with rule set content.

Traditional risk-history taking by producers, paramedical technicians and medical examiners typically fails to elicit the specific information needed to design high-functioning rule sets. Nevertheless, the survey showed us that many engine users do not as yet rely upon teleinterviews in this context.

A major aspect of medical rule sets is the impact of clinical interventions to manage prevalent chronic conditions amenable to engine triage, such as hypertension and hyperlipidemia.

Unfortunately, writing competent rule sets for pharmaceuticals is perhaps the most challenging aspect among all medical risk components. Medical manuals are notoriously inadequate in this regard and most underwriters are not well grounded in the relative risk implications of taking drug A versus drug B.

For any given medical condition suitable for engine-driven triage, there are a minimum number of questions that must be asked and answered. This is necessary to effectively distinguish between risks which can be auto-approved versus those which inherently require further assessment by an underwriter.

For example, consider depression, one of the more common histories we see in underwriting. For life insurance at least, a substantial share of depression cases can be approved as applied for.

In a fully-underwritten context, these are the questions I believe need to be addressed and then factored into rule sets:

  • When did depression symptoms commence and how long did the applicant have these symptoms? Has there been more than one episode and if so, when was the most recent episode and what was its duration?
  • Are depression symptoms currently present or is the applicant in remission?
  • Were there any RED FLAG features, such as suicidal ideation or attempts?
  • Was the applicant given a specific diagnosis and if so, what is that diagnosis?
  • Has the applicant been treated medically for this depression? If so what drug(s) did he take or is he currently taking? If he has had a change in Rx within the recent past, what medication was he taking prior to that change?
  • Has the applicant had any other forms of treatment (psychotherapy, electroconvulsive therapy, etc.) and if so, what were those treatments and when was the last time they were taken?
  • Has the applicant ever been hospitalized as an inpatient for depression? If yes, when and for how long?
  • Has the applicant ever been seen in an emergency department for symptoms related to depression? If so, when and in what specific context?
  • Is the applicant using any other remedies, including alternative and complementary interventions? If so, what are those remedies?
  • Has the applicant been referred to a mental health care professional? If so, has he seen that person?
  • Has the applicant been advised to have any tests which have not as yet been completed? If so, what are those tests and why were they not done?
  • Has the applicant been advised to have treatment that has not yet been taken or completed? If yes, what is that treatment and why was it not undertaken?
  • Has the depression had any significant adverse occupational implications and if so what was their nature and extent?

One could, of course, cut back on some of these questions. However, if too much is left off the table, the capacity of the engine to make as many decisions as possible could be greatly compromised.

How many underwriters would feel comfortable being asked to create rule sets based on potential prevalent answers to these questions, such that most cases eligible for approval as applied for are acted upon and the remainder directed to an underwriter for scrutiny?

Bottom line: To create rule sets that maximize the impact of their underwriting engine, insurers need to make certain they ask the necessary risk-related questions. The paradigm for triaging on the basis of the answers must be as consistent as possible with what an experienced underwriter would do in determining insurability.

In order to assist insurers in making the best decisions on all matters related to engine procurement and utilization, my company
and SelectX have teamed up to prepare a comprehensive 90+ page underwriting engine white paper.

We believe this paper is essential reading for all companies contemplating, acquiring or configuring an underwriting engine…and judging from responses we have received from reinsurers, service providers and insurers, this conclusion appears to be grounded in reality!

Our white paper is available for purchase. The price is only $2500 (US).

For more information, please contact Esther at or at 414.328.9010.