U.S. patent application number 11/269282 was filed with the patent office on 2006-03-23 for health cost calculator/flexible spending account calculator.
Invention is credited to Michael Schoenbaum, Mark Spranca.
Application Number | 20060064332 11/269282 |
Document ID | / |
Family ID | 27498383 |
Filed Date | 2006-03-23 |
United States Patent
Application |
20060064332 |
Kind Code |
A1 |
Schoenbaum; Michael ; et
al. |
March 23, 2006 |
Health cost calculator/flexible spending account calculator
Abstract
A method and system of providing comparative cost information
for health insurance plans. Claims files are generated for the
reference population of real historical patients for each the
plans. Information is presented to users on the distribution of
out-of-pocket costs for health care that users are likely to incur
in the coming year, based on the parameters of health plans,
information on the user and his/her household, and the actual
health care use and costs for a reference population comparable to
the users. Information is presented to users on optimal
contributions to their flexible spending account for health care in
the coming year and solving a dynamic numerical model based on
users' objective function; solutions are based on the parameters of
health plans, information on the user and his/her household, and
the actual health care use and costs for a reference population
comparable to the user.
Inventors: |
Schoenbaum; Michael;
(Bethesda, MD) ; Spranca; Mark; (Venice,
CA) |
Correspondence
Address: |
M. John Carson;FULBRIGHT & JAWORSKI L.L.P.
Forty-First Floor
555 South Flower Street
Los Angeles
CA
90071
US
|
Family ID: |
27498383 |
Appl. No.: |
11/269282 |
Filed: |
November 7, 2005 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
09841756 |
Apr 24, 2001 |
|
|
|
11269282 |
Nov 7, 2005 |
|
|
|
60200495 |
Apr 25, 2000 |
|
|
|
60219909 |
Jul 21, 2000 |
|
|
|
60223205 |
Aug 4, 2000 |
|
|
|
Current U.S.
Class: |
705/4 ; 705/2;
705/36T |
Current CPC
Class: |
G06Q 10/10 20130101;
G06Q 40/08 20130101; G06Q 40/10 20130101; G06Q 40/02 20130101 |
Class at
Publication: |
705/004 ;
705/002 |
International
Class: |
G06Q 40/00 20060101
G06Q040/00 |
Claims
1. A method of providing comparative cost information for health
insurance plans comprising the steps of: incorporating parameters
of health plans; assembling recent health care use and cost data
for a reference population; acquiring personal and health
information from users on themselves and their household members;
estimating the distribution of out-of-pocket costs the user and
his/her household is likely to face in the coming year in various
health plans, based on the experience of comparable households in
the reference population; and outputting premium and estimated
out-of-pocket expenses.
2. The method as in claim 1 of providing comparative cost
information for health insurance plans further comprising the step
of: providing costs for a "worst-case" scenario.
3. The method as in claim 2 of providing comparative cost
information for health insurance plans further comprising the steps
of: outputting out-of-pocket costs for an individual for particular
medical conditions and health events with "in-network" treatment;
and outputting out-of-pocket costs for an individual for particular
medical conditions and health events with "out of-network"
treatment.
4. The method as in claim 3 of providing comparative cost
information for health insurance plans further comprising the steps
of: outputting cost data based upon various factors including all
"in-network" treatment and all "out-of-network" treatment;
outputting cost data based upon various factors including expected
very high, high, moderate, low and no usage of health care; and
including said parameters of health plans such as individual and
household deductibles, coinsurance and co-payments, individual and
household stop-loss provisions, and services covered and not
covered.
5. A computer-based method of providing comparative cost
information for health insurance plans comprising the steps of:
processing data with a central processing unit; storing data on a
mass storage device; storing data and commands in volatile memory;
incorporating parameters of health plans; assembling recent claims
cost data for each health plan; generating claims-level files for a
reference population; acquiring user's household members' health
and personal information; comparing household members health and
personal information to reference population; outputting cost data
for household members based on said comparison; outputting cost
data based upon various factors including expected very high, high,
moderate, low and no usage of health care; outputting cost data
based upon various factors including all "in-network" treatment and
all "out-of-network" treatment; including said parameters of health
plans such as individual and household deductibles, coinsurance and
co-payments, individual and household stop-loss provisions, and
services covered and not covered; outputting premium and estimated
out-of-pocket expenses; providing costs for a "worst-case"
scenario; outputting out-of-pocket costs for an individual for
particular medical conditions and health events with "in-network"
treatment; and outputting out-of-pocket costs for an individual for
particular medical conditions and health events with
"out-of-network" treatment.
6. A computer-based method of providing comparative cost
information for health insurance plans comprising the steps of:
processing data with a central processing unit; storing data on a
mass storage device; storing data and commands in volatile memory;
incorporating parameters of health plans into stored data;
assembling recent claims cost data for each health plan and
entering them into data storage; generating claims-level files for
a reference population from historical claims data and entering
said files into data storage; acquiring user's household members'
health and personal information and incorporating that information
into temporary storage; comparing household members health and
personal information to reference population by processing the data
with a central processing unit; outputting, onto display device,
cost data for household members based on said comparison, and
providing for optional printed output; outputting, onto display
device, cost data based upon various factors including expected
very high, high, moderate, low and no usage of health care;
outputting, onto display device, cost data based upon various
factors including all "in-network" treatment and all
"out-of-network" treatment and providing for optional printed
output; including, for display on a display device, said parameters
of health plans such as individual and household deductibles,
coinsurance and co-payments, individual and household stop-loss
provisions, and services covered and not covered and providing for
optional printed output; outputting, on a display device, premium
and estimated out-of-pocket expenses and providing for optional
printed output; providing costs for a "worst-case" scenario, for
display on a display device and providing for optional printed
output; outputting, on a display device, out-of-pocket costs for an
individual for particular medical conditions and health events with
"in-network" treatment and providing for optional printed output;
and outputting, on a display device, out-of-pocket costs for an
individual for particular medical conditions and health events with
"out-of-network" treatment and providing for optional printed
output.
7-13. (canceled)
14. A system for providing comparative cost information for health
insurance plans comprising:; at least one computer comprising a
central processing unit, a data entry device, and volatile memory
for performing calculations; a mass storage for storing parameters
of health plans and recent claims cost data for health plans, for a
reference population; claims-level files for said reference
population which have been generated by calculations of said
computer applied to said recent claims cost data; user's household
members' health and personal information entered into computer by
said data entry device; comparison cost data for household members,
output by computer, said data calculated for household members
health and personal information, compared to said reference
population; and output premium and estimated out-of-pocket expense
cost data for household members, output by computer, said data
calculated for household members health and personal
information.
15. The system as in claim 14 further comprising: an output
calculated by the computer from costs for a "worst-case" scenario
from said personal data and said reference population data.
16. The system as in claim 15 further comprising: output of
out-of-pocket costs for an individual for particular medical
conditions and health events with "in-network" treatment, as
calculated by the computer from personal, reference population and
health plan parameter data; and output of out-of-pocket costs for
an individual for particular medical conditions and health events
with "out of-network" treatment, as calculated by the computer from
personal, reference population and health plan parameter data.
17. The system as in claim 16 further comprising: output cost data
based upon various factors including "all in-network" treatment and
"all out-of-network" treatment, as calculated by the computer from
personal, reference population and health plan parameter data;
output cost data based upon various factors including expected very
high, high, moderate, low and no usage of health care, as
calculated by the computer from personal, reference population and
health plan parameter data; output cost data based upon said
parameters of health plans such as individual and household
deductibles, coinsurance and co-payments, individual and household
stop-loss provisions, and services covered and not covered, as
calculated by the computer from personal, reference population and
health plan parameter data.
18. A computer-based system of providing comparative cost
information for health insurance plans comprising: a central
processing unit for processing data; a mass storage device for
storing data; volatile memory for storing data and commands; a data
entry device for said mass storage; an output viewing device; a
print output device parameters of health plans stored on said mass
storage device; recent claims cost data for each health plan stored
on aid mass storage device; claims-level files for a reference
population calculated by said central processing unit from data on
said mass storage device; user's household members' health and
personal information provided for storage on said mass storage
device by entering data utilizing said data entry device; household
members health and personal information compared to reference
population by said central processing unit and output by said
central processor to output viewing device; cost data for household
members output to output viewing device, said cost data calculated
by said central processing unit based on said comparison and upon
various factors including expected very high, high, moderate, low
and no usage of health care, and including "all in-network"
treatment and "all out-of-network" treatment; parameters of health
plans such as individual and household deductibles, coinsurance and
co-payments, individual and household stop-loss provisions, and
services covered and not covered stored on mass storage device and
used for cost data calculations by said central processing unit;
wherein viewable output options for viewing on viewing device
include premium and estimated out-of-pocket expenses, costs for a
"worst-case" scenario, out-of-pocket costs for an individual for
particular medical conditions and health events with "in-network"
treatment, and out-of-pocket costs for an individual for particular
medical conditions and health events with "out-of-network"
treatment.
19-25. (canceled)
Description
[0001] This application is a divisional of application Ser. No.
09/841,756 filed Apr. 24, 1002, which claims the benefit of
Provisional Applications 60/200,495 filed on Apr. 25, 2000;
60/219,909 filed on Jul. 21, 2000; and 60/223,205 filed on Aug. 4,
2000.
FIELD OF THE INVENTION
[0002] This invention relates to estimating and calculating health
care costs for individuals and to optimal estimation for flexible
spending account amounts to set aside for pre-tax savings.
BACKGROUND
[0003] A number of methods and systems have been developed to look
at the costs associated with the service delivered by the
physicians. The purpose is to compare physicians or hospitals in
terms of those groups which are rendering the most efficient
treatment per dollar of cost. This type of analysis is made from
the standpoint of the managers of the managed care systems. Thus in
U.S. Pat. Nos. 5,724,379, 5,924,073 and 5,953,704, the analysis is
oriented toward what should be the cost of the treatment, and for
what should a medical provider actually bill. The point of view is
not from the point of view of an actual user of Medicare, i.e., one
receiving treatment, but from the point of view of the managed care
providers. Systems and methods for analyzing medical claim
histories and billing patterns have been devised. For example, see
U.S. Pat. No. 5,557,514.
[0004] Another type of health costs estimation program is that
which is from the physician's point of view. For example, a
physician might want to decide whether to stay with a
fee-for-services or change to a capitation system where the
physician is paid an amount for treatment per patient. This type of
cost estimating system is, for example, shown in U.S. Pat. No.
5,918,208.
[0005] It would be desirable, in addition to having computer based
cost analysis programs for health care managers and for physicians,
to also have a computer based cost analysis program from the point
of view of the user, i.e., the consumer, of the medical services.
The insured person, as the one who ultimately pays for the medical
services, should be provided with information on which to make
choices about which insurance program would be most suitable for
him or her or the household.
[0006] Today there are many different types of health insurance
plans, including HMOs (Health Maintenance Organizations), PPOs
(Preferred Provider Organizations), POSs (Point of Service) and
FFSs (Fee For Service). Within these categories, there are many
different specific plans, each with different benefit designs,
costs, and other characteristics. Consumers who can choose between
two or more insurance plans thus face a complex choice. Such an
individual would do well to have some guidance as to what or how
that individual may optimally provide coverage for him- or herself.
In the event there is a household, further guidance would be useful
for deciding on the household coverage, since most people cover
their dependents and themselves in the same insurance plan.
Ideally, the user of a medical care guidance system would be
provided with comparisons and contrasts of different health plans
as to the likely distribution of out-of-pocket costs that an
individual or household would incur in each plan in the coming
year, and with respect to other plan characteristics. In
particular, people choose health insurance for future periods, such
as the coming year, yet they do not know how much health care they
will use in this future period. For example, a person might be a
high user or low user of medical services. There can also be cases
where episodes of illness occur. It would be desirable to know what
is the likelihood of the illness episodes continuing to happen and
what would be the effect on the health costs according to the
coverage chosen by the individual. It is highly desirable to have a
health cost calculator, which can calculate, over a variety of
health or medical situations, what the likely distribution of
future medical costs to an individual or to a family household. An
additional desirable feature would be to include historical patient
information. This would allow prediction, by a statistical
comparison of similar individuals and households, of an
individual's or household's statistically predicted cost results,
from their choice of medical insurance plan, and to provide
probabilities of certain types of illnesses and the resulting costs
of such, including out of pocket costs.
[0007] Moreover, if a consumer of medical services had a good cost
calculation of that consumer's likely distribution of future
medical costs, that consumer would be in a position to estimate how
much money he or she might want to allocate to a flexible spending
account (FSA). A consumer might want to have "enough" set aside in
the FSA, since that amount would not be taxed. The best "enough"
would exactly match the "out-of-pocket" amount spent on medical
costs. That way, maximum tax benefit would be obtained and no money
would be left "unused" at the end of the year, since that money is
not carried over to next year's FSA, but is lost if not spent. The
FSA amount decision is made up-front, at the beginning of the year.
Consequently, some guidance from a computer-based analysis and
prediction program would be useful. The consumer is likely to find
such a method for producing optimal estimates of the amount to be
aside for the FSA, for the year, to be most desirable.
SUMMARY OF THE INVENTION
[0008] The invention comprises a method and system of providing
comparative cost information for health insurance plans and
episodes of health care need (e.g., illness, injury). The method
and system includes using parameters of the health plans, including
yearly or monthly premiums, coverage rules, copayments,
coinsurance, stop loss provisions, benefit limits and other
details. It also includes acquiring personal and health information
from users on their household members (including themselves). It
also includes assembling recent data (preferably no more than three
years old) on health care use and costs for a large reference
population, to be used as a basis for actuarial analyses. The
method and system identifies individuals or households in the
reference population who are comparable to the user or his/her
household, respectively. The method and system uses the health plan
parameters and the data from the comparable members of the
reference population to perform actuarial estimates of the total
annual out-of-pocket costs for health care for particular users or
their households if they enrolled in each of the various health
plans, or actuarial estimates of the annual out-of-pocket health
care costs associated with certain episodes of health care use
(e.g., illness, injury) for particular users or their households if
they enrolled in each of the various health plans. The method and
system then outputs the premiums and estimated out-of-pocket costs
to the user.
[0009] This invention also comprises a method for calculating
optimal flexible savings account contributions. The method and
system includes using parameters of the health plans, including
yearly or monthly premiums, coverage rules, copayments,
coinsurance, stop loss provisions, benefit limits and other
details. It also includes acquiring personal and health information
from users on their household members (including themselves),
including information on their degree of risk aversion. It also
includes assembling recent data (preferably no more than three
years old) on health care use and costs for a large reference
population, to be used as a basis for actuarial analyses. It also
involves formulating a dynamic numerical model based on a user's
objective function; formulating a user's utility function and a
health transition equation; calibrating the health transition
equation with historical claims data linked to the user's health
status; solving the numerical model by numerical calculation
methods with assigned exogenous parameters and with test values for
the preference parameters; estimating preference parameters using
parameter values which correspond to solutions of the dynamic
program which are close to observed historical expenditures of
like-situated members of a given health plan, to input regarding
the user's risk aversion, and to the user's income. Then the
dynamic programming model is solved by numerical calculation
methods for optimal flexible spending account contribution for a
particular user in one or more particular health plans (or no
health insurance), with assigned exogenous parameters and with
estimated values for the preference parameters. The system and
method then outputs the calculated optimal contributions to the
user.
BRIEF DESCRIPTION OF THE DRAWINGS
[0010] For a more complete understanding of the present invention,
and the advantages thereof, reference is now made to the following
descriptions taken in conjunction with the accompanying drawings,
in which:
[0011] FIG. 1 a shows screen to provide a short summary of the
tool;
[0012] FIG. 2 shows a screen which presents a of what the Health
Cost Calculator (HCC) does;
[0013] FIG. 3 shows a screen which allows the user to begin
personalizing their HCC experience;
[0014] FIG. 4 shows a screen where the user verifies the
information entered on the screen of FIG. 3.
[0015] FIG. 5 shows a screen where the user enters health
information about him/herself and each other household member;
[0016] FIG. 6 shows a screen which presents overview information
about plan costs and benefits;
[0017] FIG. 7 shows a screen which has a glossary;
[0018] FIG. 8 shows a screen which illustrates basic plan cost
structures:
[0019] FIG. 9 shows a screen which shows covered services and
explains how out-of-pocket costs are dependent upon different plan
services;
[0020] FIG. 10 shows a screen which explains how out-of-pockets
costs are dependent upon whether in-network or out-of-network
providers are used;
[0021] FIG. 11 shows a screen which has a tabular presentation of
the monthly premiums;
[0022] FIG. 12 shows a screen which is an intermediary introductory
screen;
[0023] FIG. 13 shows a screen which illustrates plan benefits
tables;
[0024] FIG. 14 shows a screen to encourage the user to begin
thinking about how much health care they'll need;
[0025] FIG. 15 shows a screen which asks the user for estimated
health care use;
[0026] FIG. 16 shows a screen which has five levels of health care
use;
[0027] FIG. 17 shows a screen presenting in-network, yearly, out-of
pocket costs for each plan;
[0028] FIG. 18 shows a screen presenting out-of-network, yearly,
out-of pocket costs for each plan;
[0029] FIG. 19 shows a screen which presents a worst-case
scenario;
[0030] FIG. 20 shows a screen which reminds the user of open
enrollment dates;
[0031] FIG. 21 shows a screen which allows the user to select
particular conditions and events;
[0032] FIG. 22 shows a screen which looks at cost by condition;
[0033] FIG. 23 shows a screen which provides access to a series of
screens showing out-of-network costs; and
[0034] FIG. 24 shows a screen which shows out-of-network costs for
a person.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
[0035] The following description is of the best mode presently
contemplated for carrying out the invention. This description is
not to be taken in a limiting sense, but is merely made for the
purpose of describing the general principles of the invention. The
scope of the invention should be determined with reference to the
claims.
[0036] In everything that follows, health insurance choices may
include insurance for medical care, prescription drugs, and other
medical products and services; and other health-related insurance,
such as, but not limited to, mental health insurance, dental
insurance, vision insurance, and chiropractic insurance.
[0037] HCC--Health Cost Calculator A first embodiment of the
invention is an embodiment of the health cost calculator (HCC),
which comprises a computer based tool that is designed to help
consumers compare alternative health insurance choices available to
them, by illustrating and comparing how much they would spend
out-of-pocket for health insurance and medical care in the coming
year in various plans. The methods used in the health cost
calculator can be used to show the out-of-pocket costs of any of
number and type of health insurance plans, including, for example,
fee for service (FFS), PPO, HMO, POS, and so on. The health cost
calculator can be implemented to run on a single computer or
through a computer network including Internet, Intranet, or
Extranet. Users must be able to enter data into the tool using an
input device such as a keyboard, mouse trackball or a touch-screen
and must be able to view the output on a monitor. The HCC is
browser independent and operating system independent and platform
independent.
[0038] This first embodiment of the invention comprises a method of
providing comparative cost information for health insurance plans.
The method involves processing data with a central processing unit;
storing data on a mass storage device; storing data and commands in
volatile memory; incorporating parameters of health plans;
assembling recent data on health care use and costs for a large
reference population; acquiring personal and health information
from users on their household members (including themselves);
comparing household members' health and personal information to the
reference population; outputting cost data for household members
based on said comparison; outputting cost data based upon various
factors, including expected very high, high, moderate, low and no
usage of health care; outputting cost data based upon various
factors, including all "in-network" treatment, all "out-of-network"
treatment, or a user-specified mix of in-network and out-of-network
treatment; including the parameters of health plans such as
individual and household deductibles, coinsurance and co-payments,
individual and household stop-loss provisions, and services covered
and not covered; outputting premium and estimated out-of-pocket
expenses; providing costs for a "worst-case" scenario; outputting
out-of-pocket costs for an individual for particular medical
conditions and health events with "in-network" treatment, and
outputting out-of-pocket costs for an individual for particular
medical conditions and health events with "out-of-network"
treatment.
[0039] FSAC: Another embodiment of the invention also comprises a
computer-based method for calculating optimal flexible savings
account (FSA) contributions comprising the steps of processing data
and performing numerical calculations with a central processing
unit; storing data and computer programs on a mass storage device;
storing data and commands in volatile memory; formulating a dynamic
numerical model based on a user's objective function; formulating a
user's utility function and a health transition equation;
calibrating the health transition equation with historical claims
data linked to the user's health status; solving the numerical
model by numerical calculation methods with assigned exogenous
parameters and with test values for the preference parameters;
estimating preference parameters using parameter values which
correspond to solutions of the dynamic program which are close to
observed historical expenditures of like-situated members of a
given health plan, to input regarding the user's risk aversion, and
to the user's income; solving the dynamic programming model by
numerical calculation methods for optimal flexible spending account
contributions for a particular user in one or more particular
health plans (or no health insurance), with assigned exogenous
parameters and with estimated values for the preference parameters;
and outputting the calculated optimal contributions to the
user.
[0040] A third embodiment comprises the Health Cost Calculator
feeding its data into the required data fields of the Flexible
Spending Account Calculator, in a seamless manner.
[0041] HCC--Health Cost Calculator: Returning to the detailed
description of the first embodiment, the Health Cost Calculator
(HCC) is a computer-based tool that is designed to help consumers
compare alternative health insurance choices available to them, by
illustrating and comparing how much they would spend out-of-pocket
for health insurance and medical care in the coming year in various
plans.
[0042] For each application of the HCC, one identifies a set of
health insurance plans about which the HCC will provide
information, and then obtains the benefit design for each such
plan. "Benefit design" includes, but is not limited to, the monthly
premium for different household types; the cost structure of the
plan, including deductibles, coinsurance, co-payments, and
stop-loss provisions; covered services; pre-authorization and
referral rules; and description of the provider network and rules
for out-of-network care, if applicable. Plan designs may be
obtained from an employer, based upon what that employer offers, or
it may be based upon plans from insurance broker, or from the
directly from the insurers, if the user is selecting from these
plans rather than from an employer's offerings.
[0043] Also, for each application of the HCC, one obtains data on
the actual health service use of a reference population that is
comparable to the particular set of users for which the HCC is
intended. Specifically, one obtains relatively recent data
(ideally, no more than three years old) including the age and sex
of each person in the reference population; the medical conditions
of each person; and codes identifying members of the same
household. In addition, the data include information on each health
service used by each person in the reference population over a
period of twelve continuous months. Also included are the date the
service was provided; the type of service, described using CPT-4
procedure codes or analogous descriptors (for prescription drugs,
one obtains the drug name, strength, dosage, and volume dispensed).
The primary medical diagnosis associated with the service is
included, described using ICD-9 diagnosis codes or analogous
descriptors. Services covered in these data include medical care,
prescription drugs, and behavioral health care. The data described
in this section are referred to as the "claims data."
[0044] For instance, if the HCC is to be implemented for the active
employees of a particular firm, one obtains claims data on a large
privately insured population ("private" as opposed to public
insurance such as Medicaid). If the firm permits employees to cover
dependents in the employer-sponsored health insurance plans, these
data include employed individuals and their dependents. The person
eligible for insurance coverage, his/her covered spouse if any),
and his/her covered. dependents (if any) are referred to as a
"household." The reference population includes information on the
health service use of people at all the ages of the particular set
of users for which the HCC is intended.
[0045] For each application of the HCC, one obtains estimates of
the price of specific health care services that would be faced by
the particular set of users for which the HCC is intended.
Specifically, one needs relatively recent data (ideally, no more
than three years old) on the price paid for the full range of
health care services that are generally covered by private health
insurance plans, including medical care, prescription drugs, and
behavioral health care. Ideally, the claims data on health service
use that is obtained would include price information for each
service listed in the data (in general, this would be true of
claims data). One mainly uses the actual amount paid to the
provider for each service, e.g. the sum of insurance payments,
patient payments, and any adjustments from other sources that may
apply; one refers to this as the "cost" of each service. However,
one also obtains the prices that consumers would be charged if they
did not have any health insurance coverage; one refers to this as
the "billed charge" for each service. If the claims data do not
include price, one obtains price data from another source, such as
the publicly available price schedule for services paid for by
Medicare (for outpatient and inpatient care); or average wholesale
prices for prescription drugs.
[0046] If the price data one obtains come from a previous year, one
obtains data on the rate of medical inflation from the year the
price data to date. These data are publicly available from the
Federal government. Prices for all health care services are
inflated, using a separate inflation factor for prescription drugs
and all other medical services, respectively.
[0047] Using the data on "cost" and "billed charges," two
additional measures are created for each health care service used
by each individuals in the claims data: the cost of the service,
and the billed charge for the service. Specifically, if the claims
data already contains these prices, one uses the claims data
directly. Otherwise, one assigns each service the respective price,
based on the type of service and, if germane, the number of units
provided. One refers to the resulting data file as the
"claims-level" data file.
[0048] Examples follow the description of the Flexible Spending
Plan Calculator and illustrate both the Health Cost Calculator and
the Flexible Spending Plan Calculator.
[0049] FSAC--Flexible Spending Account Calculator: The Flexible
Spending Account Calculator (FSAC) is a computer-based tool that is
designed to help consumers decide how much money to contribute to
their flexible spending account for health care. Flexible spending
accounts permit consumers to pay for health care using pre-tax
dollars; however, in general consumers must decide how much money
to contribute to their account at the beginning of a calendar year,
and they lose any money that they do not spend by the end of that
year. Consumers' decisions about the optimal amount to contribute
are considerably complicated by two factors: uncertainty regarding
the incidence of medical expenditures over the course of the coming
benefit year and the loss of any unspent money in the FSA at the
end of the year.
[0050] The methodology underlying the second embodiment as the FSAC
represents an addition to the first embodiment, the Health Cost
Calculator methodology (HCC).
[0051] The FSAC can be implemented together with the HCC or as a
stand-alone tool for consumers. As a stand-alone tool, the FSAC
methodology incorporates many of the methodological steps of the
HCC but skips the presentation of much or all of the HCC content to
consumers.
[0052] For any application of the FSAC, one collects a range of
data necessary for implementing the tool. These include all the
data described in the HCC description, plus some additional
data.
[0053] Identify relevant health insurance plans: For each
application of the FSAC, one identifies a set of health insurance
plans about which the tool will provide information, and then
obtains the benefit design for each such plan. "Benefit design"
includes, but is not limited to, the monthly premium for different
household types; the cost structure of the plan, including
deductibles, coinsurance, co-payments, and stop-loss provisions;
covered services; pre-authorization and referral rules; and
description of the provider network and rules for out-of-network
care, if applicable. If plan designs change over time, one needs
the design that will apply when users' insurance choices take
effect.
[0054] For instance, if the FSAC is to be implemented for the
employees of a particular firm, one collects the plan designs for
the health insurance plans sponsored by the firm and available to
employees. The tool can provide information for a single health
insurance plan if the firm only offers one, or on a subset of the
plans, sponsored by the firm. However, it is most useful if it
covers all plans sponsored by the firm. If the FSAC is to be
implemented for users considering or having chosen personal or
family health insurance through a broker or directly from insurers,
one collects the plan designs for health insurance plans available
to such users.
[0055] Health care data base for a suitable reference population:
For each application of the FSAC, one obtains data on the actual
health service use of a reference population that is comparable to
the particular set of users for which the FSAC is intended.
Specifically, one obtains relatively recent data (ideally, no more
than three years old) including the age and sex of each person in
the reference population; the medical conditions of each person;
and codes identifying members of the same household. In addition,
the data include information on each health service used by each
person in the reference population over a period of 24 continuous
months. Also included is the date the service was provided. The
type of service, described using CPT-4 procedure codes or analogous
descriptors (for prescription drugs, one obtains the drug name,
strength, dosage, and volume dispensed) is part of the data. The
primary medical diagnosis associated with the service, described
using ICD-9 diagnosis codes or analogous descriptors is included.
Services covered in these data include medical care, prescription
drugs, and behavioral health care. One refers to the data described
in this section as the "claims data." For instance, if the FSAC is
to be implemented for the active employees of a particular firm,
one obtains claims data on a large privately insured population
("private" as opposed to public insurance such as Medicaid). If the
firm permits employees to cover dependents, in the
employer-sponsored health insurance plans, the data included also
references employed individuals and their dependents. One refers to
the person eligible for insurance coverage, his/her covered spouse
(if any), and his/her covered dependents (if any) as a "household."
The reference population includes information on the health service
use of people at all the ages of the particular set of users for
which the FSAC is intended.
[0056] Obtain suitable data on health care prices: For each
application of the FSAC, one obtains estimates of the price of
specific health care services that would be faced by the particular
set of users for which the FSAC is intended. Specifically, one
needs relatively recent data (ideally, no more than three years
old)on the price paid for the full range of health care services
that are generally covered by private health insurance plans,
including medical care, prescription drugs and behavioral health
care.
[0057] Ideally, the historical claims data on health service use
that one obtains would include price information for each service
listed in the data (in general, this would be true of claims data).
One mainly uses the actual amount paid to the provider for each
service, e.g. the sum of insurance payments, patient payments, and
any adjustments from other sources that may apply; one refers to
this as the "cost" of each service. However, one also obtains the
prices that consumers would be charged if they did not have any
health insurance coverage; one refers to this as the "billed
charge" for each service. If the claims data do not include price,
one obtains price data from another source, such as the publicly
available price schedule for services paid for by Medicare (for
outpatient and inpatient care); or average wholesale prices for
prescription drugs.
[0058] If the price data one obtains come from a previous year, one
obtains data on the rate of medical inflation from the year of the
price data to date. These data are publicly available from the
Federal government. One inflates prices for all health care
services, using a separate inflation factor for prescription drugs
and all other medical services, respectively.
[0059] Creation of claims-level data file for reference population:
Using the data on "cost" and "billed charges," one creates two
additional measures for each health care service used by each
individuals in the claims data: the cost of the service, and the
billed charge for the service. Specifically, if the claims data
already contains these prices, one uses the claims data directly.
Otherwise, one assigns each service the respective price, based on
the type of service and, if germane, the number of units provided.
One refer to the resulting data file as the "claims-level" data
file.
[0060] Calculating Optimal FSA Contributions: A model of optimal
contributions to a flexible savings account: The main incentive
that employees have for contributing to a flexible savings account
(FSA) is the ability to spend pre-tax dollars on medical care.
However, the optimal amount to contribute is considerably
complicated by two factors: (1) uncertainty regarding the incidence
of medical expenditures over the course of the coming benefit year
and (2) loss of any unspent money in the FSA at the end of the
year. The purpose of this section is to describe the model that the
FSA Calculator will use to derive its suggestions.
[0061] This FSAC includes a novel method for calculating optimal
FSA contributions. Previous authors have developed methods (i.e.,
Nunnikhoven 1992; Auster and Sennetti, 1994; Cuddington 1999).
However, each of these methods suffers from a number of limitations
that are addressed here. Most notably, each of these methods
neglects the fact that consumers who have not spent down their FSA
can increase medical expenditures toward the end of a benefit year,
considerably mitigating the risk of losing left-over funds at the
end of the year. In other words these models assume that consumers
spend on medical care only if they suffer a health shock. In the
model used here, consumers can improve health status by spending
more on medical care even if they do not receive health shocks.
This assumption is more realistic given that healthy individuals
regularly consume medical care services in the form of medical
exams or other preventive care services. In addition, the other
methods assume a one-to-one correspondence between the health
shocks and medical care expenditures. This restriction on consumer
behavior implicitly assumes a completely inelastic demand for
health care. In the model used here, consumers are responsive to
the marginal price of medical care expenditures and are allowed to
reduce demand for medical care in the face of higher marginal
prices. Finally, other methods assume a very simplistic form of
utility function. In particular, they assume that individuals are
risk neutral and maximize the expected value of resources available
for consumption. The utility function used here is more
realistic.
[0062] Consumers' objective function: This method uses a general
framework in which consumers derive utility from health status and
consumption of goods and services. In this framework, one can
calculate optimal FSA contributions for consumers that depend upon
their degree of risk-aversion. Combining this method with real data
on expenditure patterns over the course of the year, one can
construct more realistic estimates of optimal FSA contributions.
The model starts with consumers choose FSA contribution, G, at the
beginning of the benefit year and at the same time plan consumption
of medical care and other goods and services for the year. The
consumption plan for medical care specifies medical care use for
every possible contingency the consumer might face. The only source
of uncertainty in the model is the incidence of health shock over
the course of the year. Therefore the consumers' consumption plan
specifies use of medical care and other goods for every possible
health shock a consumer might face. In the general framework for
this model, consumers choose consumption and FSA contributions to
maximize their expected utility: G , { m , c max } = - .infin.
.infin. .times. EU = .intg. - .infin. .infin. .times. U .function.
( h , c ) .times. f .function. ( , .theta. ) .times. d .times.
.times. .times. ( 1 ) ##EQU1## where, [0063] G represents the FSA
contribution; [0064] {m.sub..epsilon.,
c.sub..epsilon.}.sub..epsilon.=-.infin..sup..infin. represents the
consumption plan for every possible health shock .epsilon.; [0065]
U(h, c) represents the utility of the consumer from health status h
and consumption of non-medical goods c; and [0066] f(.epsilon.,
.theta.) is the probability density function of the distribution
for health shocks, [0067] where .theta. parameterizes the
distribution of health shocks and will depend on the
characteristics of the consumer. [0068] 1. The instantaneous
utility function U (h, c) encodes information about the preferences
that the consumer has over health states and expenditures on other
consumption goods (which is proxied by income). Medical care
expenditures do not enter directly since the model assumes that
people care only about the improvement in health that can be bought
with medical expenditures, not the expenditures themselves, except
to the extent that medical expenditures prevent expenditures on
other goods. The utility function also encodes information
regarding the extent of the consumer's aversion to monetary and
health risk. This is made explicity using the following
instantaneous utility function: U .function. ( h , c ) = { ( ( 1 -
.delta. ) .times. h .rho. + ( .delta. ) .times. c .rho. ) 1 / .rho.
if .times. .times. h .gtoreq. h min 0 if .times. .times. h < h
min ( 2 ) ##EQU2## Where, .rho.<1, .delta..epsilon.[0,1] and
h.sub.min are the parameters of the utility function. With a
utility function like (5), the consumer dies if health falls below
h.sub.min and has a CES utility function if health is greater than
or equal to h min 1 1 - .rho. ##EQU3## is the elasticity of
substitution between health and other consumption goods and .delta.
specifies the relative importance of consumption in the utility
function. Consumers are risk neutral if .rho. equals one and risk
averse otherwise.
[0069] Health transition equation and the distribution of health
shocks. The purpose of expenditures on medical care is to improve
health, and indeed, in order to calculate a solution to this
problem a health transition equation must be estimated:
h=f(h.sub.0, m, .epsilon.; .eta.) (3) where .epsilon..sub.1
represents shocks to health in period t. A new health shocks
arrives each period, and it may either improve or diminish the
health status of the consumer. While consumers do not know in
advance the exact health shocks that they will receive over the
course of the benefit year, they presumably have some information
regarding the probability distribution from which the health shocks
are drawn: .epsilon..about.F(.epsilon.;.theta.) t=1 . . . 12 (4)
where F(.) is the cumulative density function of the distribution
of shocks, and .theta. parameterizes that distribution. One of the
objects needed by the optimal FSA calculator is .theta., the method
for obtaining this object is described below. In the calculation of
the optimal FSA, the model assumes that .epsilon. is normally
distributed, which seems reasonable since a negative health shock
is being modeled. This functional form assumption for F is flexible
enough to reflect some well-known facts about health shocks (for
example, shocks are more likely and more severe for older
people).
[0070] The health transition function depends in a predictable way
on prior health status, medical expenditures, and shocks. Health
status is sticky from period to period; that is, better health in
the previous period implies better health this period. More medical
expenditure may improve health in the following period. And, health
shocks have a negative effect on health (though this is just a
normalization, since draws of c that are less than zero imply a
positive effect on health). The particular functional form used for
the health transition function is described below. For now, .eta.
represents the parameters associated with this functional form.
Assumptions regarding this functional form must meet the basic
requirements for a health transition equation, as long as .eta. is
suitably constrained: .differential. h .differential. h 0 .gtoreq.
0 ; .differential. h .differential. m .gtoreq. 0 ; .differential. h
.differential. < 0 ( 5 ) ##EQU4##
[0071] Tax advantage from FSA contributions: The budget constraint
is nonlinear, and reflects both the costs and benefits of
contributing to an FSA. The main cost, of course, is that each
dollar contributed into an FSA is deducted from income at the
beginning of the year, hence cannot be spent on consumption of
non-medical goods and services. On the other hand, taxes are saved
on each dollar of contribution to the FSA. The non-linearity in the
asset transition function reflects the fact that the marginal price
of medical care to the consumer is zero for all medical expenses
below FSA contribution and one for all medical expenses in excess
of the FSA contribution. Thus, the asset transition equation is
given by: I=G+.tau.(I-G)+c if mp.ltoreq.G (6a)
I=G+.tau.(I-G)+c+(mp-G) if mp>G (6b) Where, .tau. is the
marginal tax rate; p is the relative price of medical care; and I
is the yearly income.
[0072] Summarizing the model Table 1 summarizes the pieces of the
model. Of course, the optimal solution for {m.sub..epsilon.} and
{c.sub..epsilon.} cannot be a fixed quantity known at the beginning
of the benefit year, since there is uncertainty regarding health
events over the course of the year. Thus, an optimal solution will
consist of a plan (or policy function) that sets optimal medical
and other expenditures as a function of the state of the world. On
the other hand, since G must be chosen at the start of the program,
its optimal value can only depend on data that are known then (such
as the distribution of future medical shocks, F
(.epsilon.;.theta.), and income, I). TABLE-US-00001 TABLE 1 Summary
of Model Model Object Specification and Description Objective
Equations (1) and (2) Function Choice G-Optimal FSA contribution
Variables {m.sub..epsilon.}-A plan for optimal medical care
expenditure based upon an observed health shock, .epsilon.
{c.sub..epsilon.}-A plan for optimal expenditures on other goods
based upon an observed health shock, .epsilon. State I-Income
Variables h-Health status .epsilon.-Health shock Transition
h-Equation (3) Equations I-Equation (6a) and (6b) Exogenous
.tau.-Marginal tax rate Parameters .theta.-Health shock
distributional parameter .eta.-Health transition function
parameters p-Price of medical care h.sub.min-Health status, below
which the consumer dies Preference 1-.delta.-Weight on health in
the utility function Parameters .delta.-Weight on other consumption
in the utility function (1-.rho.).sup.-1-Elasticity of
substitution
[0073] Integrating the FSAC--(Flexible Spending Account Calculator)
with the HCC--Health Cost Calculator: In performing its
calculations, the FSAC and HCC both use information that the user
provides to classify the user into household types that are based
upon the size, demographic characteristics, and health
characteristics of the user and his/her households; this
information is described in greater detail in the description of
the HCC. Then, the calculator estimates the distribution of medical
expenditures in the following year conditional on the user's
household type. For the FSAC, we will use the same assignment of
users into household types, described above and in the HCC
description, and the dynamic program described in Table 1 is solved
conditional on the information pertinent to each user's household
type. In some applications of the FSAC, household types may be
subdivided into categories of health care use, as described in the
description of the HCC, and optimal FSA contributions are
calculated for these smaller groups.
[0074] In the following sections, the methods for deriving
estimates for each of the parameters in the dynamic program are
described. These estimates will naturally be specific to each
household type. The solution methodology for the dynamic program as
a whole given values for all these parameters is also described.
This solution methodology is necessary to derive the suggested
optimal FSA contribution for users in each household type.
[0075] Calibrating the health transition equation,
h=f(h.sub.0,m,.epsilon.): The strategy for calibrating the vector
of parameters, .eta., in the health transition equation is to use
two years of linked health insurance claims records on a reference
population; these data have the same characteristics as those
described in the description of the HCC, with the addition that
they include data on each individual and household for two
consecutive years. These data contain extensive information on
medical expenditures by families in each of two consecutive
years.
[0076] The main idea underlying the estimated parameters of the
health transition equation with these data is that medical
expenditures can serve as a proxy for the health status of people
in the user's household type. Those household types with high
medical expenditures in a given year presumably have users who are
on average less healthy than household types with low medical
expenditures. Thus, medical expenditures will serve two different
roles in the calculations: (1) an input into the health transition
equation, where increased medical expenditures should improve
health (or else there would be no reason to spend money on health
care in this model), and (2) as a proxy for health status, where
high medical expenditures signifies poor health.
[0077] In order to resolve these two apparently contradictory roles
of medical expenditures, the health transition equation will be
specified as follows:
h=.eta..sub.0+.eta..sub.1h.sub.0+.eta..sub.2m+.epsilon. (11)
[0078] In order to derive estimates of .eta..sub.0 and .eta..sub.1,
within each household type medical expenditures in the more recent
year are regressed on expenditures in the previous year. Let the
estimates from this linear regression be n.sub.0 and n.sub.1,
corresponding to the constant term and the coefficient on prior
year medical expenditures respectively. This regression will
produce parameter estimates for a yearly health transition
equation.
[0079] With these two parameters in hand, the only remaining
unknown parameter in the health transition equation left to be
calibrated is .eta..sub.2, which is the coefficient on m.sub.t.
Rather than derive a single value for this parameter, which is a
controversial one in the literature, the model can use several
different values within well-accepted bounds. The lower bound for
.eta..sub.2 is 0. If it were any less than zero, then medical
expenditures would implausibly have a negative effect on future
health. The upper bound for .eta..sub.2 is the size of effect of
previous period health, .eta..sub.1h.sub.0. This must be an upper
bound, or else it would be possible to reverse aging simply by
spending money on medical care. In general, the dynamic program is
solved three times, once with a value of .eta..sub.2 toward the low
end of the bounds, once with a value in the middle, once with a
value toward the high end, and the FSAC reports the optimal value
of the FSA to the user under each of the three assumptions. Some
applications of the FSAC may use more or fewer different values of
.eta..sub.2.
[0080] Calibrating the distribution of health shocks,
F(.epsilon.,.theta.): One of the outputs of the health transition
regression is a distribution of residuals, e, which by construction
of an ordinary least squares estimation procedure will have zero
mean. e=h-(n.sub.0-n.sub.1h.sub.0) (15) This distribution of
residuals is fit to a normal distribution with parameters .eta.
[0081] Measuring income, I, and the marginal tax rate, .tau.:
On-line users are asked to enter their yearly pre-tax family income
levels. Based on this information, an estimate the marginal tax
rate of the user is made, assuming standard deductions to calculate
state and federal income taxes. In addition, the calculation of the
marginal tax rate will take into account the contribution to the
marginal tax rate of FICA and Medicare taxes.
[0082] Solving the dynamic program: The objective of the FSAC is to
obtain the optimal FSA contribution based on the parameters of the
utility function, budget constraint and the health shock
distribution, and to show changes in the optimal FSA contribution
in response to changes in the these parameters. Since there is no
analytical solution for the consumers' utility maximization
problem, the optimal FSA contribution will be estimated using a
multi-step numerical approach. In the first step the optimal
consumption of medical care and other consumption goods for each
possible health shock are calculated, taking the amount of the FSA
contribution as given. Inputting the optimal consumption plan for
each possible health shock into the utility function gives the
maximum utility attainable in each state of the world. In the
second step the expected maximum utility is calculated by
multiplying the maximum utility under each health shock by its
probability of occurrence and adding over all health shocks. These
two steps are repeated for each possible FSA contribution to
calculate the maximum expected utility for each FSA contribution.
Finally the FSA contribution with the highest expected maximum
utility is chosen as the optimal FSA contribution. Table 2 gives a
schematic view of the above solution methodology. Programming is
carried out in C++ or Fortran, or another readily available
computer language. TABLE-US-00002 TABLE 2 Algorithm to Solve the
Program Step1 Given .times. .times. G , calculate .times. .times.
max .times. .times. U .times. .times. for .times. .times. each
.times. .times. possible .times. .times. health .times. .times.
shock U max .function. ( G , ) = Max [ U .function. ( h , c )
.times. s . t . { h = h 0 + m + I = G + .tau. .function. ( I - G )
+ c + ( mp - G ) .times. .times. if .times. .times. mp > G I = G
+ .tau. .function. ( I - G ) + c .times. .times. if .times. .times.
mp .ltoreq. G ] ##EQU5## Step2 Calculate .times. .times. expected
.times. .times. value .times. .times. of .times. .times. the
.times. .times. Maximum .times. .times. Utility .times. .times.
given .times. .times. G E .function. [ U max .function. ( G ) ] =
.times. U max .function. ( G , ) .times. f .function. ( ; .theta. )
##EQU6## Step3 Repeat Steps 1 and 2 for all possible values of G
Step4 Choose .times. .times. G .times. .times. with .times. .times.
the .times. .times. highest .times. .times. Expected .times.
.times. Maximum .times. .times. Utility G Opt = Arg .times. .times.
max .times. { E .function. [ U max .function. ( G ) ] }
##EQU7##
[0083] Estimating the preference parameters: The methods discussed
in the previous section give a quick method to solve the program
for each consumer if one knows the parameters of the objects in the
model. However, up to now, only the calibration of the exogenous
parameters in the model has been described, not the preference
parameters. The main object of the next step in the estimation is
to obtain consistent estimates of these parameters (which are
collected into a vector called .mu. for convenience). The
preference parameters are estimated using an iterative strategy.
For a given value of .mu., the optimal medical care expenditure
paths for each consumer. In this model, .mu..sup.(j) represents the
value of .mu. in the j.sup.th iteration, X.sub.i*[.mu..sup.(j)] is
the optimal values obtained from solving the dynamic program, and
let X.sub.i is the actual values of these outcomes for consumer i.
In particular, the FSAC uses data on actual medical expenditures
from the health insurance claims data on the reference population.
The basis for estimating the preference parameters is to pick them
such that the difference between the observed medical expenditures
and the predicted medical expenditures (from the model) are as
close as possible. The value of .mu. in the next
iteration--.mu..sup.(j+1)--are calculated using a least squares
distance function: S .function. [ .mu. ] = i = 1 N .times. .times.
( X i - X i * .function. [ .mu. ] ) .times. W ' .times. W
.function. ( X i - X i * .function. [ .mu. ] ) ( 18 ) ##EQU8##
Where, W is a weighting matrix. The ultimate goal of the analysis
is to minimize the distance function with respect to .mu.. Using
standard hill climbing methods, .mu..sup.(j+1) can be calculated
based on the derivatives of S[.mu..sup.(j)]. The model will have
converged to the estimate of .mu. when the norm of the first
derivative of S is sufficiently close to zero.
[0084] Calculating optimal FSA contributions: Finally, with all the
exogenous and the preference parameters estimated, the dynamic
program, using the methodology described, is solved one more time.
This solution combines the data on typical medical expenditures
(which are also used to construct the HCC, in applications where
the HCC and FSAC are implemented together) and the reported data
from the user. It represents a customized number or numbers that,
while specific to the user, draws from the experience of people who
are similar to the user.
[0085] REPORTING THE OPTIMAL FSA CONTRIBUTIONS Generate optimal FSA
contributions in each health insurance plan, based on households
similar to the user's. To generate optimal FSA contributions for a
particular user of the FSAC, one utilizes the user's household
profile and the additional information collected. Specifically, one
extracts from the household-level data file on the reference
population all the households with the same household profile as
the user. In general, one then calculates the mean optimal FSA
contribution among the households of this type, in each health plan
about which the FSAC is providing information and that is available
to the user, and for each of the three values of .eta..sub.2. In
some applications of the FSAC one also calculates the mean optimal
FSA contribution among the households. This is done within the
respective categories of health care use, as already described, in
each health plan about which the FSAC is providing information and
that is available to the user, and for each of the three values of
.eta..sub.2.
[0086] The solution derived from the dynamic program represents the
optimal FSA contribution under the assumption that all
out-of-pocket spending for health care covered by the FSA is
non-discretionary. To estimate the overall optimum for a particular
user, one assumes that the user will incur the discretionary
expenses, with certainty, and this amount is added to the optimal
FSA contribution. In the event that the estimated optimum for a
particular user and combination of health plan, category of health
care use (if applicable), and value of .eta..sub.2 exceeds the
maximum legal FSA contribution, the optimal contribution will be
set as the maximum legal FSA contribution.
[0087] The plan designs of many insurance plans make different
provisions for services provided "in network" vs. "out of network."
In general, therefore, one calculates the optimal FSA contribution
in the respective categories of health care use and for the various
values of .eta..sub.2 under two scenarios: (1) all health care is
received in-network, and (2) all health care is received
out-of-network; or other scenarios as appropriate. If a particular
plan has more tiers (e.g. a POS plan), more corresponding scenarios
are calculated.
EXAMPLE 1
HCC
[0088] Introductory Screen (FIG. 1) The first page (FIG. 1) of the
HCC is designed to attract the user's attention, provide a short
summary 32 of the tool, and to motivate the user to begin using the
tool by clicking on the "Step 1" button 34.
[0089] This screen is comprised of three main sections: navigation
(left section) 36, header (upper right section) 38, and content
(lower right section) 40.
[0090] The navigation area 36 is comprised of an image file that
displays several numbered "steps" 42 to be followed to navigate
through the HCC from beginning to end. This image file 42, an
"active region programmed with numbered "hyperlinks" enables the
user to navigate to different parts of the HCC simply by clicking
on the appropriate active region, i.e., hyperlink, of the
navigation graphic.
[0091] The header 38 is comprised of an image file that displays
the title 46 of the tool and a photo montage design 48. The photos
contain images related to health care (such as doctors, patients,
prescription bottles, etc.) 48 and an image of a calculator 50.
[0092] The content area is comprised of text and may include active
hypertext "links" (underlined text) 52 which allow the user to
navigate to other screens. Both are written in a common Web
formatting language. The end of the content area is delineated by a
horizontal line 54 across the bottom of the screen (a simple image
file), and an active hypertext "link" (underlined text) 56 that
allows the user to email questions or comments.
[0093] Introduction (FIG. 2) This page, FIG. 2, presents a more
thorough description 58 of what the HCC does and how it can benefit
the user in choosing a health plan. It also explains the goals 60
that are served by the HCC in that it calculates an estimated
yearly out-of-pocket health care cost and defines which plan would
best suit the user and his or her family based on this information
62.
[0094] This screen is comprised of three main sections: navigation
(left section) 36, header (upper right section) 38, and content
(lower right section) 40.
[0095] The navigation area 36 is comprised of an image file that
displays a logo 64 and several numbered "steps" 42 ("active" region
of numbered hyperlinks) to be followed to navigate through the HCC
from beginning to end.
[0096] The header 38 is comprised of an image file that displays
the title of the screen 38 that the user is currently viewing. This
title corresponds to the "step" 42 in the navigation graphic that
the user selected. For example, the header image file corresponding
to "Step 1" 38 displays the word "Introduction" 38, which is also
displayed in the navigation image file 66.
[0097] The content area is comprised of text and is similar to FIG.
1 content in style.
[0098] USER ENTERS PERSONAL INFORMATION: User enters number and
type of household members ( FIG. 3). This screen (FIG. 3) allows
the user to begin personalizing their HCC experience by defining
the number 68, sex 70, and age 72 of the people that may be covered
by their chosen plan. Sex 70, age 72, spouse 78 and how many
children 80 are entered and submitted 82 on one screen. As an
acknowledgement of the user's sense of privacy, there is also a
reminder 74 that the tool will not retain or use any of the
information provided.
[0099] This screen (FIG. 3) is comprised of three main sections:
navigation (left section) 36, header (upper right section) 38, and
content (lower right section) 40. The navigation area 36 is as
described in FIG. 2.
[0100] The header 38 is comprised of an image file that displays
the title 38 of the screen that the user is currently viewing. This
title 38 corresponds to the "step" 76 in the navigation graphic 42
that the user selected.
[0101] The content area (FIG. 3) 40 is comprised of text and a
data-collection table. Both are written in a common Web formatting
language. The table cells contain selection buttons 78 and empty
text fields 80 to collect user input. The content area 40 also
contains a button 82 for the user to submit the information entered
in the input areas. A warning message box appears on screen if the
user attempts to navigate off this screen without providing enough
input. FIG. 4 asks for confirmation of data; if correct 84, one is
moved to the next step, if not 86, one is returned 88 to the
beginning of the step to provide correct information. The end of
the content area is delineated by a horizontal line 54 across the
bottom of the screen (a simple image file), and an active hypertext
"link" (underlined text) that allows the user to email questions or
comments.
[0102] User verifies sex and age of each household member (FIG. 4).
By verifying the sex and age of the family members covered under
the user's health plan, this step further personalizes the
information that will be presented to the user regarding health
plan costs on subsequent pages. This screen (FIG. 4) is comprised
of three main sections: navigation (left section) 36, header (upper
right section) 48, and content (lower right section) 40. This is as
described for FIG. 3.
[0103] The content area 40 (FIG. 4) is comprised of text and a
dynamically-generated data-presentation table 92. Both are written
in a common Web formatting language. Table headers 94 are
dynamically generated based on user input from using client/server
communications technology to access a database on the Web server
and display headers 94 customized to the user (See also FIG. 11).
The table cells contain selection buttons 84, 86 and empty text
fields 80 (FIG. 3) to collect user input. The content area also
contains a button for the user to submit the information entered in
the input areas 82. A warning message box appears on screen if the
user attempts to navigate off this screen without providing enough
input. The end of the content area is delineated by a horizontal
line 54 across the bottom of the screen (a simple image file), and
an active hypertext "link" (underlined text) 56 that allows the
user to email questions or comments.
[0104] User enters medical conditions of each household member
(FIG. 5) In this step, the user enters health information about
him/herself and each other household member who will be covered
under this health insurance coverage. Specifically, for each
person, the user views a list of common medical conditions (e.g.
hypertension, diabetes) and events (e.g. pregnancy) that might
occur in the coming year; the user then chooses all such conditions
or events that apply for each household member. This exercise helps
the user to begin considering what their family's usage may be
based on the presence of these conditions.
[0105] The particular medical conditions and events from which
users can choose depend on the particular application of the HCC.
In some applications, the HCC may not permit users to specify
certain medical conditions or events depending on the population of
users for which the HCC is being developed, and on the quality and
size of the claims data on the reference population. In some
applications, users may be able to asked to provide additional
detail about particular conditions/events, such as whether the
condition is chronic or newly diagnosed.
[0106] This screen (FIG. 5) is comprised of three main sections:
navigation (left section) 36, header (upper right section) 38, and
content (lower right section) 40. Navigation 36 and header 38 are
similar to the description of FIG. 3.
[0107] The content area 40 (FIG. 5) is comprised of text and a
dynamically generated data-collection table. Both are written in a
common Web formatting language. Table headers are dynamically
generated based on user input using client/server communications
technology to access a database on the Web server and display
headers customized to the user. The table cells 96 contain
selection buttons 98 and empty text fields 100 to collect user
input. The content area also contains a button 102 for the user to
submit the information entered in the input areas. A warning
message box appears on screen if the user attempts to navigate off
this screen without providing enough input. The end of the content
area is delineated by a horizontal line 54 across the bottom of the
screen (a simple image file), and an active hypertext "link"
(underlined text) 56 that allows the user to email questions or
comments.
[0108] Tool constructs household profile for user. For each user,
one takes the information provided in the previous screens, FIG. 3
through FIG. 5, to construct a profile of the user's household. One
includes the number of people the user is considering covering by
the health insurance plans he/she is considering; and the age, sex,
and health characteristics of these people.
[0109] Specifically, the data about the user's household goes to a
common gateway interface (CGI) program, which sends it to a data
processing program such as the STATISTICAL ANALYSIS SYSTEM.RTM. or
ORACLE.RTM.. The data processing program registers all the data and
constructs a household profile for the user.
[0110] For example, one can imagine the following household
profile: a user indicates that he/she plans to cover him/herself,
his/her spouse/partner, and one child. The user indicates that she
is female and 45 years old, that the spouse is male and 50 years
old. and that the child is female and 15 years old. The user also
indicates that neither she nor her daughter has any current medical
conditions, and that the spouse/partner has current
hypertension.
Then the household profile looks like this:
[0111] User: female, age 45, no current medical conditions
[0112] Spouse: male, age 50, current hypertension
[0113] Child: female, age 15, no current medical conditions
[0114] The degree of detail at which household profiles are defined
depends on the particular application of the HCC, and on the
quality and size of the claims data on the reference population
available for a particular application of the HCC. In some
applications, for instance, one may group households by age (e.g.
households where the female adult is age 45-54), or in other ways,
with the goal of increasing the number of households in the claims
data that have the same household profile as particular users', in
order to make actuarial analyses more precise. In some applications
of the HCC, one can also incorporate multivariate regression
techniques in the actuarial analyses, to control for various
personal and health characteristics of households in the reference
population.
[0115] PRESENT INFORMATION ABOUT PLAN COST AND BENEFITS: Overview
(FIG. 6) To educate the user about the two primary types of costs,
explanations of premiums and out-of-pocket costs are presented,
with access provided to a glossary (FIG. 7) that defines common
out-of-pocket terms. It also explains the three factors that make
out-of-pocket costs more difficult to compare across different
health plans, and links the user to three corresponding drill-down
pages that describe these factors.
[0116] This screen (FIG. 6) is comprised of three main sections:
navigation (left section) 36, header (upper right section) 38, and
content (lower right section) 40. Navigation 36 and header 38 are
similar to the description of FIG. 3.
[0117] The content area 40 (FIG. 6) is comprised of text and may
include active hypertext "links" (underlined text) 102, 104, 106,
108, 110, 112 which allow the user to navigate to other screens.
Both are written in a common Web formatting language. The end of
the content area is delineated by a horizontal line 54 across the
bottom of the screen (a simple image file), and an active hypertext
"link" (underlined text) 56 that allows the user to email questions
or comments.
[0118] Glossary (FIG. 7) This glossary page defines common terms
that will assist the user in understanding their health costs. It
can be reached from hypertext links 102, 104 and 106 of FIG. 6.
[0119] This screen (FIG. 7) is comprised of three main sections:
navigation (left section) 36, header (upper right section) 38, and
content (lower right section) 40. Navigation 36 and header 38 are
similar to the description of FIG. 3.
[0120] The content area 40 (FIG. 7) is comprised of text, written
in a common Web formatting language. The end of the content area is
delineated by a horizontal line 54 across the bottom of the screen
(a simple image file), and an active hypertext "link" (underlined
text) 56 that allows the user to email questions or comments.
[0121] Basic plan cost structures (FIG. 8): This page explains how
out-of-pocket costs depend on how often household members use
health care services. This screen (FIG. 8) is comprised of three
main sections: navigation (left section) 36, header (upper right
section) 38, and content (lower right section) 40. Navigation 36
and header 38 are similar to the description of FIG. 3.
[0122] The content area (FIG. 8) may include active hypertext
"links" (underlined text) which allow the user to navigate to other
screens. Both are written in a common Web formatting language. The
end of the content area is delineated by a horizontal line 54
across the bottom of the screen (a simple image file), and an
active hypertext "link" (underlined text) 56 that allows the user
to email questions or comments.
[0123] Covered services (FIG. 9): This page explains how
out-of-pocket costs are dependent upon how different plans cover
services and how this information should be factored into the
user's health plan choice. This screen (FIG. 9) is comprised of
three main sections: navigation (left section) 36, header (upper
right section) 38, and content (lower right section) 40. Navigation
36 and header 38 are similar to the description of FIG. 3. The
content area (FIG. 9) is comprised of text, written in a common Web
formatting language. The end of the content area is delineated by a
horizontal line 54 across the bottom of the screen (a simple image
file), and an active hypertext "link" (underlined text) 56 that
allows the user to email questions or comments.
[0124] In-network and out-of-network use (FIG. 10) This page
explains how out-of-pockets costs are dependent upon whether
in-network or out-of-network providers are used.
[0125] This screen (FIG. 10) is comprised of three main sections:
navigation (left section) 36, header (upper right section) 38, and
content (lower right section) 40. Navigation 36 and header 38 are
similar to the description of FIG. 3.
[0126] The content area 40 is comprised of text, written in a
common Web formatting language. The end of the content area is
delineated by a horizontal line 54 across the bottom of the screen
(a simple image file), and an active hypertext "link" (underlined
text) 56 that allows the user to email questions or comments.
[0127] Health insurance premiums for available plans (FIG. 11) This
table presents the monthly premiums 114 that would be paid by the
user under each health care plan about which the particular
application of the HCC is providing information. The tool uses the
personal information 116 previously collected from prior screens to
highlight the column of this table that applies to the user and his
or her household.
[0128] This screen (FIG. 11) is comprised of three main sections:
navigation (left section) 36, header (upper right section) 38, and
content (lower right section) 40. Navigation 36 and header 38 are
similar to the description of FIG. 3.
[0129] The content area 40 (FIG. 11) is comprised of text and a
dynamically generated data table. Both are written in a common Web
formatting language. Data presented in the table 118 are
dynamically generated based on user input, using client/server
communications technology to access a database on the Web server
and display data in the appropriate table cell. The end of the
content area is delineated by a horizontal line 54 across the
bottom of the screen (a simple image file), and an active hypertext
"link" (underlined text) 56 that allows the user to email questions
or comments.
[0130] Benefits tables for available plans -introductory screen
(FIG. 12) This screen provides hypertext links to other screens
containing tables that display information about available health
plans.
[0131] This screen (FIG. 12) is comprised of three main sections:
navigation (left section) 36, header (upper right section) 38, and
content (lower right section) 40. Navigation 36 and header 38 are
similar to the description of FIG. 3.
[0132] The content area 40 (FIG. 12) is comprised of text and may
include active hypertext "links" (underlined text) 120 which allow
the user to access benefits tables. Both are written in a common
Web formatting language. The end of the content area is delineated
by a horizontal line 54 across the bottom of the screen (a simple
image file), and an active hypertext "link" (underlined text) 56
that allows the user to email questions or comments.
[0133] Benefits tables for available plans (FIG. 13) This is an
example of the table or tables that a user will be able to access
to compare the benefits offered by the health plans available to
them. This is based on information acquired previously acquired.
The table can be created by a document creation program such as
ADOBE ACROBAT.RTM., which converted the table from a word
processing application into a compressed file format that can be
easily downloaded and viewed from a Web page.
[0134] GENERATE HOUSEHOLD-LEVEL MEASURES OF HEALTH CARE USE AND
SPENDING FOR EACH HOUSEHOLD IN THE REFERENCE POPULATION.
[0135] Estimate out-of-pocket costs for each household in the
reference population. Using the plan design of each health
insurance plan for which the HCC will provide information, and the
claims-level data for the reference population, a calculation is
made to determine what the annual out-of-pocket costs for
households in the reference population would have been had they
paid for services according to the plan design of the respective
health insurance plans.
[0136] Specifically, for each household in the claims data, the
claims are sorted for individual services by household member and
by the date it was provided. Using the plan design, and the price
data already acquired, the health care claims are then processed
chronologically, assigning out-of pocket costs according to the
plan design of the particular health insurance plan being modeled.
One repeats this for each plan for which the HCC will provide
information. This processing takes into account all cost components
of the respective plans, including individual and household
deductibles; coinsurance and co-payments; individual and household
stop-loss provisions; and what services are covered and not
covered. For covered services provided in-network, deductibles and
coinsurance rates are calculated based on the "cost" of the
service; for services that are not covered by a particular plan,
and covered services that are provided out-of-network, deductibles
and coinsurance rates are calculated based on the "billed charge"
for the service.
[0137] One notes that the plan designs of many insurance plans make
different provisions for services provided "in network" (e.g. by a
specific group of health care providers with whom the health
insurance plan has a special contractual relationship) versus "out
of network" (e.g. by any other health care provider). For instance,
Preferred Provider Organizations (PPO) typically have a higher
deductible and higher coinsurance rates for out-of-network services
than for in-network services. Health Maintenance Organizations
(HMO) generally do not provide any coverage for out-of-network
services (unless specifically authorized by the insurer), although
members may use such services if they pay for them entirely
out-of-pocket.
[0138] To reflect the fact that the user and his/her dependents may
receive health care in-or out-of network, one estimates
out-of-pocket costs in each plan under two scenarios: (1) all
health care is received in-network, and (2) all health care is
received out-of-network. In other implementations, one estimates
out-of-pocket costs in each plan under additional scenarios, such
as outpatient care received out-of-network and inpatient care
received in-network; or one solicits an expected mix of in- and
out-of-network services from the user and estimates out-of-pocket
costs under the user-provided mix. For illustration, one can
envision two simple plans:
[0139] Plan A:
[0140] Monthly employee premium for family coverage: $50
[0141] In-Network Use: Each household member pays $10 for each
outpatient physician visit and all associated outpatient services
(e.g. diagnostic tests); $0 for each hospitalization and all
associated inpatient services; $20 for each emergency room visit
and all associated services; and $10 for each prescription that
they fill.
[0142] Out-Of-Network Use: Not covered.
[0143] Plan B:
[0144] Monthly employee premium for family coverage: $100
[0145] In-Network Use: Each household member pays 100% of the cost
of the first $200 of health services in a year (i.e., a $200
individual deductible); and then pays 20% of the cost for all other
health services in that year (i.e., a 20% coinsurance rate), up to
a maximum out-of-pocket payment (excluding the deductible)of $1500
per year (i.e., a $1500 stop loss provision).
[0146] Out-Of-Network Use: Each household member pays 100% of the
cost of the first $400 of health services in a year; and then pays
40% of the cost for all health services in that year, up to a
maximum out-of-pocket payment (excluding the deductible)of $4000
per year.
[0147] Using the hypothetical household profile, as entered for
FIG. 3 through FIG. 5, and using hypothetical data on health
service use and prices, Table 1 a illustrates the process of
calculating out-of-pocket costs for this household in Plans A and
B. All services are assumed to be provided in-network in Table 3a,
while Table 3b illustrates the process of calculating out-of-pocket
costs assuming all services are provided out-of-network.
TABLE-US-00003 TABLE 3a Calculation of Out-of-Pocket Costs.
Assuming All Services In-Network Plan A Plan B Date of Billed
Out-of-Pocket Out-of-Pocket Member Service Description Cost Charge
Cost Cost Comments Employee 15 Jan Outpatient visit $100 $150 $10
$100 15 Jan Prescription drug $50 $90 $10 $50 05 Jan Annual exam
$150 $210 $10 $70 Plan B deductible met Spouse 15-Jan Outpatient
visit $120 $150 $10 $120 15 Jan Prescription drug $200 $275 $10
$104 Plan B deductible met 15 Feb Prescription drug $200 $275 $10
$40 15 Mar $200 $275 $10 $10 $40 17 Mar Outpatient visit $120 $150
$10 $24 15 Apr Prescription drug $200 $275 $10 $40 15 May
Prescription drug $200 $275 $10 $40 15 Jun Prescription drug $200
$275 $10 $40 17 Jun Outpatient visit $120 $150 $10 $24 15 Jul
Prescription drug $200 $275 $10 $40 15 Aug Prescription drug $200
$275 $10 $40 15 Sept Prescription drug $200 $275 $10 $40 15 Oct
Prescription drug $200 $275 $10 $40 15 Nov Prescription drug $200
$275 $10 $40 02 Dec Outpatient visit $150 $10 $30 05 Dec Outpatient
visit $150 $10 $30 10 Dec Outpatient visit $120 $150 $10 $24 15 Dec
Prescription drug $200 $275 $10 $40 16 Dec 3 day $5,000 $9,000 $0
$904 Plan B stop- hospitalization loss met 20 Dec Prescription drug
$100 $125 $10 $0 20 Dec Outpatient visit $200 $260 $10 $0 24 Dec
Outpatient visit $120 $150 $10 $0 Child 03 Mar Emergency room $500
$800 $20 $260 Plan B visit deductible met 03 Mar Prescription drug
$80 $100 $10 $16 03 Mar Prescription drug $35 $45 $10 $7 06 Mar
Outpatient visit $100 $150 $10 $20 15 Mar Outpatient visit $100
$150 $10 $20 15 Aug Annual physical $150 $210 $10 $30 Total $10,865
$15,740 $310 $2,273 Annual Premium $600 $1,200 Total Out- Of-Pocket
Spending $910 $3,473
[0148] TABLE-US-00004 TABLE 3b Calculation of Out-of-Pocket Costs.
Assuming All Services In-Network Plan A Plan B Date of Billed
Out-of-Pocket Out-of-Pocket Member Service Description Cost Charge
Cost Cost Comments Employee 15 Jan Outpatient visit $100 $150 $150
$150 15 Jan Prescription drug $50 $90 $90 $90 05 Jan annual exam
$150 $210 $210 $180 Plan B deductible met Spouse 15-Jan Outpatient
visit $120 $150 $150 $150 15 Jan Prescription drug $200 $275 $275
$260 Plan B deductible met 15 Feb Prescription drug $200 $275 $275
$110 15 Mar $200 $200 $275 $275 $110 17 Mar Outpatient visit $120
$150 $150 $60 15 Apr Prescription drug $200 $275 $275 $110 15 May
Prescription drug $200 $275 $275 $110 15 Jun Prescription drug $200
$275 $275 $110 17 Jun Outpatient visit $120 $150 $150 $60 15 Jul
Prescription drug $200 $275 $275 $110 15 Aug Prescription drug $200
$275 $275 $110 15 Sept Prescription drug $200 $275 $275 $110 15 Oct
Prescription drug $200 $275 $275 $110 15 Nov Prescription drug $200
$275 $275 $110 02 Dec Outpatient visit $150 $200 $200 $80 05 Dec
Outpatient visit $150 $200 $200 $80 10 Dec Outpatient visit $120
$150 $150 $60 15 Dec Prescription drug $200 $275 $275 $110 16 Dec 3
day $6,000 $9,000 $9,000 $2,440 Plan B stop- hospitalization loss
met 20 Dec Prescription drug $100 $125 $125 $0 20 Dec Outpatient
visit $200 $260 $260 $0 24 Dec Outpatient visit $120 $150 $150 $0
Child 03 Mar Emergency room $500 $800 $800 $560 Plan B visit
deductible met 03 Mar Prescription drug $80 $100 $100 $40 03 Mar
Prescription drug $35 $45 $45 $18 06 Mar Outpatient visit $100 $150
$150 $60 15 Mar Outpatient visit $100 $150 $150 $60 15 Aug Annual
physical $150 $210 $210 $84 Total $10,865 $15,740 $15,740 $5,642
Annual Premium $600 $1,200 Total Out- Of-Pocket Spending $16,340
$6,842
[0149] In settings in which the HCC is to provide information on
one or more Point of Service (POS) plans with three tiers of
benefits, one estimates out-of-pocket costs in POS plans under
three scenarios:
[0150] (1) All health care is provided under Tier 1, which most
commonly requires that services are provided in-network, and that
patients are referred for specialty care by their designated
primary care provider; this corresponds to the "in-network"
scenario for HMOs, PPOs and fee-for-service (FFS) plans.
[0151] (2) All health care is provided under Tier 2, which most
commonly requires that services are provided in-network, but that
patients can self-refer themselves for specialty care.
[0152] (3) All health care is provided under Tier 3 in which
patients most commonly can self-refer to any out-of-network
provider: this corresponds to the "out-of-network" scenario for
HMOs. PPOs and FFS plans. In other implementations, one estimates
out-of-pocket costs in each plan under additional scenarios; for
instance, one solicits an expected mix of service use across tiers
from the user and estimates out-of-pocket costs under the
user-provided mix. In some PPOs and POS plans, there rules for
using "in-network" and "out-of-network" care can be more complex
than illustrated here. For instance, some PPO plans have one set of
rules for out-of-network physician visits and another for
out-of-network hospital admissions. The particular way in which one
applies these rules in the "out-of-network" scenario (PPOs), and
"Tier 2" and "Tier 3" scenarios (POS plans) will depend on the
details of the particular plans, and on the particular application
of the HCC.
[0153] This step thus creates the following measures for each
household -in the reference population: what the household would
have spent out-of-pocket for health care in each health insurance
plan for which the particular application of the HCC provides
information. It also indicates under two (for HMOs, PPOs, and FFS
plans) or three (for POS plans) scenarios what the household would
have spent out-of-pocket if they had no health insurance and had
been required to pay the billed charges for all services.
[0154] Divide the reference population into categories of health
care use. For each household in the reference population, one
calculates the total of health care used during the year covered by
the claims data. Specifically, one prices each health care service
using its cost. One then adds up the cost of all services for each
member of a household to obtain the household total. For instance,
the total health care use of the hypothetical household illustrated
in Table 3 over that year is $10,865 (the sum of the "cost"
column). One then sorts all households in the reference population
by the total value of health services used during the year.
[0155] Next, one stratifies the households in the reference
population into a small number of categories of health care use, by
the total value of health services used during the year. The number
of categories, and the range of total values of health services
that define the categories, can vary depending on the particular
group of users for which the HCC is being implemented.
[0156] For instance, one could use five categories of health care
use, defined as follows: "No" use: No member of the household used
any health care during the year covered by the claims data.
[0157] "Low" use: The total value of health services used during
the year ranged from $1 to $1000 during the year covered by the
claims data.
[0158] "Moderate" use: The total value of health services used
during the year ranged from $1000 to $3000.
[0159] "High" use: The total value of health services used during
the year ranged from $3000 to $10000.
[0160] "Very High" use: The total value of health services used
during the year was more than $10000.
[0161] Each household in the reference population thus gets
assigned to one category of health care use. A specific
implementation of the HCC could use more or fewer categories, and
could define the categories differently.
[0162] This step thus creates the following measures for each
household in the reference population: (1) total cost of the health
care used by the household during the continuous 12 month period
covered by the data and (2) the category of health care use to
which the household belongs. That is, chosen from among the
categories used for a particular application of the HCC.
[0163] Calculate the use of specific types of health care for each
household in the reference population. For each household in the
reference population, one calculates the number of units of
specific types of health care used by the household in the year
covered by the claims data. The specific types of health care will
vary depending on population for which the particular application
of the HCC is being developed.
[0164] As an example, one calculates the number of total outpatient
visits, including visits to physicians, physician's assistants,
nurses, psychologists, mental health social workers, and other
providers of ambulatory care. One calculates the number of
emergency room visits and inpatient admissions, including any
overnight stay in a hospital, nursing home, or other inpatient
medical facility; and prescriptions used by the members of the
household during the year. In other applications, one might use
more or fewer types of health care, and/or different types.
[0165] Create household-level data file for reference population.
From the claims-level data file, one creates a household-level data
file for the reference population. This file contains the
demographic and health status information on each member of the
household, and all of the household measures created. One refers to
this as the "household-level" data file.
[0166] CALCULATE MEAN COST OF SPECIFIC CATEGORIES OF HEALTH CARE
USE. Using the claims-level data file, the mean unit cost of
specific types of health care, corresponding to the categories
discussed, ("no use", "low", "moderate", "high" and "very high") is
calculated. As an example, one may work with four categories of
health care: outpatient visits and associated services; emergency
room visits and associated services; inpatient admissions and
associated services; and prescriptions. These categories of use are
defined as follows: (1) outpatient visit and associated services:
one counts each outpatient visit, and ancillary services provided
on the same day as the visit, as one unit of outpatient care; (2)
one prices these using the corresponding cost. If there is more
than one visit and ancillary services on a given day, so that it is
unclear with which visit the ancillary services were associated,
one assigns them all to the first visit listed in the data on that
date.
[0167] Emergency room (ER) visit and associated services: One
counts each emergency room visit, and ancillary services provided
on the same day as the visit, as one unit of emergency room care.
One prices these using the corresponding cost. If there is more
than one ER visit and ancillary services on a given day, so that it
is unclear with which visit the ancillary services were associated,
one assigns them to the first visit listed in the data on that
date.
[0168] Inpatient admissions and associated services: One prices all
services associated with an inpatient admission by their associated
cost, and sum these to obtain the cost of each inpatient
admission.
[0169] Prescriptions: One counts each prescription in the
claims-level data, priced using the corresponding cost.
[0170] For each type of health care used in the particular
application of the HCC, one then calculate the mean cost per unit
of care. Depending on the particular application of the HCC, one
calculates the mean cost per unit across the whole reference
population, or for households with the same household profile as
the user's.
[0171] For illustration, one could imagine that the average cost of
an outpatient visit and associated services in the claims-level
data on the reference population is $200; of an emergency room
visit and associated services, $700; of an inpatient admission and
associated services, $8000; and of a prescription, $150.
[0172] PROVIDE INFORMATION ON LEVELS OF HEALTH CARE USE: Overview (
FIG. 14) To encourage the user to begin thinking more clearly about
how much care they'll need, three questions 122, 124, 126 are posed
that will help the user to estimate their projected utilization.
These questions focus on previous use, certain future use, and
possible future use.
[0173] This screen (FIG. 14) is comprised of three main sections:
navigation (left section) 36, header (upper right section) 38, and
content (lower right section) 40. Navigation 36 and header 40 are
similar to the description of FIG. 3.
[0174] The content area (FIG. 14) is comprised of text, written in
a common Web formatting language. The end of the content area is
delineated by a horizontal line 54 across the bottom of the screen
(a simple image file), and an active hypertext "link" (underlined
text) 56 that allows the user to email questions or comments.
[0175] Estimated health care use (FIG. 15) Once the user begins
considering the care they may use, this page further breaks down
this utilization into as small number of categories of health care,
; in this example, one designates four specific categories of use.
These numbers will give a finer estimate of use, and is broken down
by family member to create a higher level of accuracy.
[0176] This screen (FIG. 15) is comprised of three main sections:
navigation (left section) 36, header (upper right section) 38, and
content (lower right section) 40. Navigation 36 and header 38 are
similar to the description of FIG. 3.
[0177] The content area 40 (FIG. 15) is comprised of text and a
dynamically generated data-collection table 128. Both are written
in a common Web formatting language. Table headers 130 are
dynamically generated based on user input using client/server
communications technology to access a database on the Web server
and display headers customized to the user. The table cells 132
contain empty text fields to collect user input. The content area
also contains a button 134 for the user to submit the information
entered in the input areas. The end of the content area 40 is
delineated by a horizontal line across the bottom of the screen (a
simple image file) 54, and an active hypertext "link" (underlined
text) 56 that allows the user to email questions or comments.
[0178] For instance, one can imagine that the hypothetical user,
expects the following utilization patterns in the coming year:
TABLE-US-00005 TABLE 4 Anticipated Health Care Use in Coming Year,
for Hypothetical User Household Outpatient Inpatient member visits
ER visits admissions Prescriptions User 2 0 0 1 Spouse 8 0 1 13
Child 3 1 0 2 Total 13 1 1 16
[0179] Once this information has been entered by the user, it gets
sent to the CGI program and then to the data processing program.
One then calculates the total cost of this pattern of health care.
Specifically, one prices the units in the respective categories of
use by the corresponding average cost already calculated. In this
example, the anticipated health care use described in Table 3 would
cost $11,556. This result is sent back to the CGI program and then
presented to the user.
[0180] Calculation of mean household use of particular types of
health care, within category of health care use. To illustrate
patterns of use of different health care for a particular user of
the HCC, one designates the user's household profile. Specifically,
using the data processing program, one extracts from the
household-level data file on the reference population all the
households with the same household profile as the user. Within each
category of health care use, one then calculates the mean number of
units of specific types of health care. For example, included are
outpatient visits, emergency room visits, inpatient admissions, and
prescriptions, respectively, used during the year covered by the
claims data, among the households in the respective category of
health care use. Of these households with the same household
profile as the user, one also calculates the fraction within each
category of health care use.
[0181] To illustrate, one can imagine the hypothetical user, and
the categories of use described (i.e., "no," "low," "moderate,"
"high," and "very high" use). One extracts from the claims data on
the reference population all the households with the same household
profile as the hypothetical user. Using these households and the
types of health care illustrated, one calculates the following
numbers:
[0182] The mean number of outpatient visits, emergency room visits.
inpatient admissions, and prescriptions used during the year
covered by the claims data, among households with no use and the
fraction of households with no use.
[0183] The mean number of outpatient visits, emergency room visits,
inpatient admissions, and prescriptions used during the year
covered by the claims data, among households with low rise and the
fraction of households with low use.
[0184] The mean number of outpatient visits, emergency room visits,
inpatient admissions, and prescriptions used during the year
covered by the claims data, among households with moderate use and
the fraction of households with moderate use.
[0185] The mean number of outpatient visits, emergency room visits,
inpatient admissions, and prescriptions used during the year
covered by the claims data, among households with high use and the
fraction of households with high use.
[0186] The mean number of outpatient visits, emergency room visits,
inpatient admissions, and prescriptions used during the year
covered by the claims data, among households with very high use and
the fraction of households with very high use.
[0187] Levels of health care use (FIG. 16) Based on the information
provided by the user, this screen presents illustrates different
patterns of health care use, and associated information. In this
example, five different levels of health use 136 are defined for
the user. These levels vary as a function of the average number of
services used each year and total expenditures for these services.
This table is intended to assist users in categorizing themselves
into a level of use to help estimate the costs that may be incurred
during the year. It gives perspective through the column that shows
the percentage of families like theirs who utilize care in each of
the five utilization levels 136.
[0188] This screen (FIG. 16) is comprised of three main sections:
navigation (left section) 36, header (upper right section) 38, and
content (lower right section) 40. Navigation and header are similar
to the description of FIG. 3.
[0189] The content area 40 (FIG. 16) is comprised of text and a
dynamically generated data table. Both are written in a common Web
formatting language. Data table headers and data presented in the
table are dynamically generated based on user input from the steps,
as shown, using client/server communications technology to access a
database on the Web server and display data in the appropriate
table cell. The end of the content area is delineated by a
horizontal line across the bottom of the screen (a simple image
file) 54, and an active hypertext "link" (underlined text) 56 that
allows the user to email questions or comments.
[0190] PROVIDE INFORMATION ON OUT-OF-POCKET COSTS Generate
out-of-pocket cost estimates in each health insurance plan, based
on households similar to the user's.
[0191] To generate out-of-pocket cost estimates for a particular
user of the HCC, one designates the user's household profile.
Specifically, one extracts from the household-level data file on
the reference population all the households with the same household
profile as the user. Within each category of health care use, one
then calculates the mean out of pocket cost among the households in
the respective category of health care use, in each health plan
about which the HCC is providing information and that is available
to the user.
[0192] The plan designs of many insurance plans make different
provisions for services provided "in network" vs. "out of network."
In general, therefore, one calculates the mean out-of pocket cost
in the respective categories of health care use under two
scenarios: all health care is received in-network, and all health
care is received out-of-network. (For some implementations, one
might want to calculate costs under other scenarios of the mix
between in-network and out-of-network care, in addition to or
instead of these two.) To illustrate, one can imagine the
hypothetical user and the categories of use. One extracts from the
claims data on the reference population all the households with the
same household profile. Using these households, one calculates the
following numbers:
[0193] The mean of total out-of-pocket spending for health services
used during the year covered by the claims data, among households
with no use.
[0194] The mean of total out-of-pocket spending for health services
used during the year covered by the claims data, among households
with low use.
[0195] The mean of total out-of-pocket spending for health services
used during the year covered by the claims data, among households
with moderate use.
[0196] The mean of total out-of-pocket spending for health services
used during the year covered by the claims data, among households
with high use.
[0197] The mean of total out-of-pocket spending health services
used during the year covered by the claims data, among households
with very high use.
[0198] One calculates these values for each health plan about which
the particular application of the HCC is providing information. For
each of these plans, one calculates these values once for the
"in-network" scenario and then again for the "out-of-network"
scenario (and/or for whatever alternative scenarios of the mix of
in-network and out-of-network care one presents in the particular
implementation of the tool). For POS plans, one also calculates
these values for a third scenario, representing the case in which
all care is provided in network but patients self-refer to medical
specialists. Finally, one calculates the mean of total
out-of-pocket spending for health services used during the year, at
each level of use, under the scenario of no health insurance
coverage, so that households were required to pay billed charges
for all health care.
[0199] Presenting in-network costs (FIG. 17) This section of the
HCC provides a definition of in-network care and presents yearly
out-of pocket costs for each plan when all care is within the
health plan's network. Based on information provided by the user,
the appropriate utilization field is automatically highlighted 138
by the tool to present dollar estimates of in-network costs for the
user when enrolled in different health plans. The bottom row 140 of
this table illustrates the value of the health plans available by
listing estimates of utilization costs for a user with no health
insurance.
[0200] The cells in the table 142 (FIG. 17) represent the mean
total-of-pocket cost for the population for each level of use in
each of the health plan about which the particular application of
the HCC is providing information, plus the annual premium the use
would pay in each of the respective health plans. If a "no use"
level is presented, the cells thus contain just the applicable
annual premium.
[0201] This screen is comprised of three main sections: navigation
(left section) 36, header (upper right section) 38, and content
(lower right section) 40. Navigation and header are similar to the
description of FIG. 3.
[0202] The content area 40 (FIG. 17) is comprised of text and a
dynamically generated data table. Both are written in a common Web
formatting language. Data presented in the table are dynamically
generated based on user input using client/server communications
technology to access a database on the Web server and display data
in the appropriate table cell. The end of the content area is
delineated by a horizontal line 54 across the bottom of the screen
(a simple image file), and an active hypertext "link" (underlined
text) 56 that allows the user to email questions or comments.
[0203] Presenting out-of-network costs (FIG. 18) This section of
the HCC provides a definition of out-of-network care and presents
yearly out-of-pocket costs 148 for each plan when all care is used
outside the plan's network of providers. Based on information
provided by the user, the appropriate utilization field is
automatically highlighted 144 by the tool that presents dollar
estimates of out-of-network costs for the user when enrolled in
different health plans. The bottom row 146 of this table
illustrates the value of the health plans available by listing
estimates of utilization costs for a user with no health
insurance.
[0204] This screen (FIG. 18) is comprised of three main sections:
navigation (left section), 36 header (upper right section) 38, and
content (lower right section). 40 Navigation and header are similar
to the description of FIG. 3.
[0205] The content area (FIG. 18) 40 is comprised of text and a
dynamically generated data-table. Both are written in a common Web
formatting language. Data presented in the table are dynamically
generated based on user input, using client/server communications
technology to access a database on the Web server and display data
in the appropriate table cell. The end of the content area is
delineated by a horizontal line across the bottom of the screen (a
simple image file) 54, and an active hypertext "link" (underlined
text) 56 that allows the user to e-mail questions or comments.
[0206] In this hypothetical application of the HCC, there are no
POS plans available to users, so that the third scenario is not
illustrated, all use in-network, but self-referral, to medical
specialists in this instance.
[0207] Worst-case scenario (FIG. 19) While the user may have
accurately defined what his or her family's projected care may be
for the year, this often does not include a "worst-case scenario"
which could greatly increase the family's health costs. This screen
(FIG. 19) directs users to consider how much their projected health
care costs would be if they needed significantly more care than
they had expected. This screen's purpose is two-fold in that is
educates the user about the worst-case scenario possibility 150 and
is also a directive to consider these higher utilization costs 152
when choosing a health plan. Highlighting costs under this scenario
is particularly important because this is when insurance coverage
is most valuable.
[0208] This screen (FIG. 19) is comprised of three main sections:
navigation (left section) 36, header (upper right section) 38, and
content (lower right section) 40. Navigation 36 and header 38 are
similar to the description of FIG. 3.
[0209] The content area 40 (FIG. 19) is comprised of text, written
in a common Web formatting language. The end of the content area 40
is delineated by a horizontal line 54 across the bottom of the
screen (a simple image file) 54, and an active hypertext "link"
(underlined text) 56 that allows the user to email questions or
comments.
[0210] Concluding screen (FIG. 20): The final screen reminds the
user of open enrollment dates 154 and provides a link 156 to the
company's Human Resources website to access health plan
information. This final screen also contains a button 158 that
enables the user to erase all information entered into the
calculator to secure the user's privacy after using the tool.
[0211] This screen (FIG. 20) is comprised of three main sections:
navigation (left section) 36, header (upper right section) 38, and
content (lower right section) 40. Navigation 36 and header 38 are
similar to the description of FIG. 3.
[0212] The content area 40 (FIG. 20) is comprised of text and may
include active hypertext "links" (underlined text) which allow the
user to navigate to other screens. Both are written in a common Web
formatting language. The content area also contains a button 158
for the user to delete the personal information entered into prior
data-collection screens. The end of the content area 40 is
delineated by a horizontal line 54 across the bottom of the screen
(a simple image file), and an active hypertext "link" (underlined
text) 56 that allows the user to email questions or comments.
[0213] UPDATING THE HCC PERIODICALLY: Update information on health
insurance plans. As the health insurance plans available to a
particular population of users for which the HCC has been
implemented change, one collects updated information on the
available plans and their respective benefit designs. One then
updates that particular application of the HCC for the new set of
plan choices.
[0214] Update data on health care use for a reference population.
Because patterns of health care use change over time, one updates
the claims data on the reference population periodically, with the
goal of using claims data that are three or fewer years old.
[0215] Update data on health care prices. Because health care costs
change over time (and because new services and products are
introduced), the data on health care costs and billed charges are
updated periodically, with the goal of using data that are three or
fewer years old. In addition, each year in which the HCC is
implemented, current inflation factors are used for adjusting
prices from previous years to the current economy.
[0216] SUPPLEMENTAL STEP-PRESENT INFORMATION ON OUT-OF-POCKET COSTS
FOR PARTICULAR MEDICAL CONDITIONS AND HEALTH EVENTS.
[0217] For some applications of the HCC, users are provided with
estimates of the annual out-of pocket health care costs associated
with particular medical conditions in a year. For illustration, one
presents this as a final step in the sample application of the HCC.
However, depending on the application, it will be included there,
elsewhere in the HCC, or not at all.
[0218] Tool constructs individual profile for user and user's
household members. Information provided by the user is used to
construct an individual profile for the user and each member of the
user's household. In an example already used, the user's profile
would be:
[0219] User: female, age 45, married (or living with partner), no
current medical conditions, her spouse would have the following
profile: spouse: male, age 50, married (or living with partner),
current hypertension, and so on for each household member
identified by the user.
[0220] The degree of detail at which individual profiles are
defined depends on the particular application of the HCC, and on
the quality and size of the claims data on the reference population
available for a particular application of the HCC. In some
applications, for instance, one may group households by age
category (e.g. households where the female adult is age 45-54), or
in other ways, with the goal of increasing the number of households
in the claims data that have the same household profile as
particular users', in order to make actuarial analyses more
precise. In some applications of the HCC, one can also incorporate
multivariate regression techniques in the actuarial analyses, to
control for various personal and health characteristics of
households in the reference population.
[0221] Calculate condition-specific out-of-pocket costs for each
health condition and event users can identify (i.e., in FIG. 5),
for each individual in the reference population. For each person in
the reference population, it is determined whether the person had
each of the respective health conditions and events that users can
identify during the year covered by the claims data. Using the plan
design of each health insurance plan for which the HCC will provide
information and the claims-level data for the reference
population,. The following are calculated: the annual out-of-pocket
costs associated with each specific health condition or event that
users can identify, had the person with the condition/event paid
for services according to the plan design of the respective health
insurance plans.
[0222] Specifically, for each person in the claims data identified
as having the particular health condition or experiencing the
particular health event during the year covered by the claims data,
one sorts the health care claims by those attributable to the
condition/event by date. The total cost of those claims is
calculated, with each claim priced based on its "cost". Then the
plan design and the price data are used to process the health care
claims chronologically, assigning out-of-pocket costs according to
the plan design of the particular health insurance plan being
modeled. This is repeated this for each plan for which the HCC will
provide information. Rules for pricing health care claims are the
same as those previously described.
[0223] In plans with individual or household deductibles or
stop-loss provisions, the out-of-pocket cost for a particular claim
depends on the health care used previously during the year by the
individual and/or the household. In processing the health care
claims for particular health conditions/events, out-of-pocket costs
are assigned assuming that the claims attributable to the
particular condition/event are the only care used during the year.
To reflect the fact that health care may be provided in-or
out-of-network, out-of pocket costs are estimated for each
condition/event in each plan under two (HMO, FFS, and PPO plans)or
three (POS plans) scenarios. These scenarios are as defined
previously.
[0224] This step creates measures for each person in the reference
population. First it creates indicators of whether the person had
each of the respective health conditions, second, events that users
can identify during the year covered by the claims data. It also
creates measures for the total cost of health care attributable to
the respective conditions/events. And it creates a measure of what
the individual would have spent out-of-pocket cost for health care
attributable to the respective conditions/events in each health
insurance plan for which the particular application of the HCC
provides information. It covers the situation for two (for HMO,
PPO, and FFS plans) or three (for POS plans) scenarios. Also, it
creates a measure of what the individual would have spent
out-of-pocket if they had no health insurance and had been required
to pay the billed charges for all such health care.
[0225] Creation of person-level data file for reference population.
From the claims-level data file a person-level data file is created
for the reference population. This file contains the demographic
and health status information on each person, and all of the
person-level measures already created. This is referred to as the
"person-level" data file.
[0226] Generation of out-of-pocket cost estimates for each
condition/event specified by the user, based on similar
individuals. To generate out-of-pocket cost estimates for the
particular conditions/events of individual members of the user's
household, the person's individual profile is used. Specifically
extracted from the person-level data file on the reference
population are all the individuals with the same individual profile
as the person, and who had the particular condition/event in the
year covered by the claims data. These people are sorted by the
total cost of care attributable to the condition/event and are
divided into a number of groups based on this cost.
[0227] To illustrate, one can imagine estimating the out-of-pocket
cost associated with hypertension in a 50 year-old married male.
One divides the population in the reference population with
hypertension and the same individual profile according to their
total cost of care attributable to hypertension during the year
covered by the claims data, One also divides this sub-population
into five groups(sub-subpopulations): Lowest quintile (20%) of
costs, Second lowest quintile of costs, Middle quintile of costs,
Second highest quintile of costs, Highest quintile of costs
[0228] Within each group, one calculates the mean of total
out-of-pocket spending for health care attributable to
hypertension. One calculates these values for each health plan
about which the particular application of the HCC is providing
information, and for each of the scenarios of in-network use and
out-of-network use. Finally, one calculates the mean of total
out-of-pocket spending for health care attributable to
hypertension, under the scenario of no health insurance, so that
individuals were required to pay billed charges for this health
care.
[0229] Select particular conditions/events (FIG. 21): This screen
provides access to a series of screens showing in-network costs for
treating each condition, by person. Previously entered personal
information is used to generate a list of hypertext links 160 that
link to cost tables for each person and each condition.
[0230] This screen is (FIG. 21) comprised of three main sections:
navigation (left section) 36, header (upper right section) 38, and
content (lower right section) 40. Navigation 36 and header 38 are
similar to the description of FIG. 3.
[0231] The content area 40 (FIG. 21) is comprised of text and a
dynamically generated table containing active hypertext "links"
(underlined text) 160. Both are written in a common Web formatting
language. Hypertext links 160 presented in the table are
dynamically generated based on user input, using client/server
communications technology to access a database on the Web server
and display data in the appropriate table cell. The end of the
content area is delineated by a horizontal line across the bottom
of the screen (a simple image file) 54, and an active hypertext
"link" (underlined text) 56 that allows the user to email questions
or comments.
[0232] Looking at cost by condition: In-network costs ( FIG. 22)
This screen presents in-network costs 164 for Person 1 and
Condition 1 162. This screen is comprised of three main sections:
navigation (left section) 36, header (upper right section) 38, and
content (lower right section) 40. Navigation 36 and header 38 are
similar to the description of FIG. 3. The content area 40 (FIG. 22)
is comprised of text and a dynamically generated data-table. Both
are written in a common Web formatting language. Data presented in
the table are dynamically generated based on user input, using
client/server communications technology to access a database on the
Web server and display data in the appropriate table cell. The end
of the content area is delineated by a horizontal line across the
bottom of the screen (a simple image file) 54, and an active
hypertext "link" (underlined text) 56 that allows the user email to
questions or comments.
[0233] Select particular conditions/events (FIG. 23) This screen
provides access to a series of screens showing out-of-network costs
for treating each condition, by person. Previously entered personal
information is used to generate a list of hypertext links 166 that
link to cost tables for each person 168 and each condition 170.
[0234] This screen (FIG. 23) is comprised of three main sections:
navigation (left section) 36, header (upper right section) 38, and
content (lower right section) 40. Navigation 36 and header 38 are
similar to the description of FIG. 3.
[0235] The content area 40 is comprised of text and a dynamically
generated table containing active hypertext "links" (underlined
text) 166. Both are written in a common Web formatting language.
Hypertext links 166 presented in the table are dynamically
generated based on user input, using client/server communications
technology to access a database on the Web server and display data
in the appropriate table cell. The end of the content area 40 is
delineated by a horizontal line across the bottom of the screen (a
simple image file) 54, and an active hypertext "link" (underlined
text) 56 that allows the user to email questions or comments.
[0236] Looking at cost by condition: Out-of-network costs (FIG. 24)
This screen shows out-of-network costs for person 1 with condition
1 174 It is accessed by a hyperlink from FIG. 24. Previously
entered personal information has been used to generate a list of
hypertext links that link to cost tables for each person and each
condition.
[0237] This screen (FIG. 24) is comprised of three main sections:
navigation (left section), 36 header (upper right section) 38 and
content (lower right section) 40. Navigation 36 and header 38 are
similar to the description of FIG. 3.
[0238] The content area 24 (FIG. 24) is comprised of text and a
dynamically generated table containing active hypertext "links"
(underlined text) 172. Both are written in a common Web formatting
language. Hypertext links presented in the table are dynamically
generated based on user input, using client/server communications
technology to access a database on the Web server and display data
in the appropriate table cell. The end of the content area 40 is
delineated by a horizontal line 54 across the bottom of the screen
(a simple image file), and an active hypertext "link" (underlined
text) 56 that allows the user to email questions or comments.
EXAMPLE II
Flexible Spending Account Calculator: Introductory Screen (FIG. 1),
of the HCC Description
[0239] The first page of the FSAC is designed to attract the user's
attention, provide a short summary of the tool and to motivate the
user to begin using the tool by clicking on the "Step 1" button.
For the description see FIG. 1. Flexible Spending Account
Calculator and FSAC are substituted for Health Cost Calculator and
HCC when the FSAC is run as a stand alone basis. When the HCC and
the FSAC are run together, a joint annotation may be used, or the
FSAC annotation may be used exclusively. Analogous, as used here,
means that the screens may be identical except as to the
annotation, or that screen correspond with regard to content, e.g.,
purpose of HCC and purpose of FSAC. For further description, see
the description of HCC, FIG. 1.
[0240] Introduction (analogous to FIG. 2, of the HCC description)
This page presents a more thorough description of what the FSAC
does and how it can benefit the user in deciding how much to
contribute to their FSA. For further description, see the
description of HCC (FIG. 2).
[0241] PRESENT INFORMATION ABOUT FSA PURPOSE AND USE: Overview
(analogous to FIG. 6, of the HCC description): Summary information
on the use and purposes of FSAs is provided, along with links to
additional relevant information. For further description, see the
description of HCC, FIG. 6.
[0242] Glossary (analogous to FIG. 7, of the HCC description) This
glossary page defines common terms that will assist the user in
understanding their FSA. For further description, see the
description of HCC, FIG. 7.
[0243] Covered services (Analogous to FIG. 9, of the HCC
description) This page explains the health-related products and
services for which FSA contributions can be used. For further
description, see the description of HCC FIG. 7.
[0244] USER ENTERS PERSONAL INFORMATION: User enters number and
type of household members (analogous to FIG. 3a, of the HCC
description): This page allows each user to begin personalizing
their FSAC experience by defining the number. sex, and age of the
people that may be covered by their chosen plan. As an
acknowledgement of the user's sense of privacy, there is also a
reminder that the tool will not retain or use any of the
information provided. For further description, see the description
of HCC, FIG. 3a, 3b. Alternatively the information is entered on
FIG. 3b and verified on FIG. 4b.
[0245] User enters/verifies sex and age of each household member
(analogous to FIG. 4a of the HCC description) By entering/verifying
the sex and age of the family members covered under the user's
health plan, this step helps personalizes the information that will
be presented to the user regarding FSA contributions on subsequent
pages. For further description, see the description of HCC (FIGS.
4a, 4b).
[0246] User enters medical conditions of each household member
(analogous to FIG. 5, of the HCC description) In this step, the
user enters health information about him/herself and each other
household member who will be covered under their health insurance
coverage. Specifically, for each person, the user views a list of
common medical conditions (e.g., hypertension. diabetes) and events
(e.g. pregnancy) that will or might occur in the coming year; the
user then chooses all such conditions or events that apply for each
household member. This exercise helps the user to begin considering
what their family's usage may be based on the presence of these
conditions. The particular medical conditions and events from which
users can choose depend on the particular application of the FSAC.
In some applications, the FSAC may not permit users to specific any
medical conditions or events, depending on the population of users
for which the FSAC is being developed and on the quality and size
of the claims data on the reference population. In some
applications users may be able to provide additional detail about
particular conditions/events, such as whether the condition is
chronic or newly diagnosed. For further description, see the
description of HCC FIG. 5.
[0247] This screen is comprised of three main sections: navigation
(left section) 36, header (upper right section) 38, and content
(lower right section) 40. The navigation area 36 is comprised of an
image file that displays a logo and several numbered "steps" to be
followed to navigate through the FSAC from beginning to end. This
image file has been programmed to contain various "active" regions.
which enables the user to navigate to different parts of the FSAC
simply by clicking on the appropriate active region of the
navigation graphic. The header 38 is comprised of an image file
that displays the title of the screen that the user is currently
viewing. This title corresponds to the "step" in the navigation
graphic that the user selected.
[0248] The content area 40 is comprised of text and may include a
dynamically generated data-collection table. Both are written in a
common Web formatting language. Table headers are dynamically
generated based on user input. using client/server communications
technology to access a database on the Web server and display
headers customized to the user. The table cells contain selection
buttons and empty text fields to collect user input. The content
area also contains a button for the user to submit the information
entered in the input areas. A warning message box appears on screen
if the user attempts to navigate this screen without providing
enough input. The end of the content area is delineated by a
horizontal line across the bottom of the screen (a simple image
file), and an active hypertext "link" (underlined text)that allows
the user to email questions or comments.
[0249] Tool constructs household profile for user. For each user.
one takes the information provided to construct a profile of the
user's household, including the number of persons the user is
considering covering under the health insurance plans he/she is
considering; and the age, sex, and health characteristics of these
people.
[0250] Specifically, the data about the user's household goes to a
common gateway interface (CGI) program, which sends it to data
processing program such as the STATISTICAL ANALYSIS SYSTEM.RTM. or
ORACLE.RTM.. The data processing program registers all the data and
constructs a household profile for the user.
[0251] For example, one can imagine the following household
profile: a user indicates that he/she plans to cover him/herself,
his/her spouse/partner, and one child. The user indicates that she
is female and 45 years old, that the spouse is male and 50 years
old, and that the child is female and 15 years old. The user
indicates that neither she nor her daughter has any current medical
conditions and that the spouse's partner has current hypertension.
Then the household profile looks like this:
[0252] User: female, age 45, no current medical conditions
[0253] Spouse: male, age 50, current hypertension
[0254] Child: female, age 15, no current medical conditions
[0255] The degree of detail at which household profiles are defined
depends on the particular application of the FSAC, and on the
quality and size of the claims data on the reference population
available for a particular application of the FSAC. In some
applications, for instance. one may group households by age
category (e g. households where the female adult is age 45-54), or
in other ways, with the goal of increasing the number of households
in the claims data that have the same household profile as
particular users', in order to make actuarial analyses more
precise. In some applications of the FSAC, one can also incorporate
multivariate regression techniques in the actuarial analyses, to
control for various personal and health characteristics of
households in the reference population..
[0256] User enters key information for FSAC, analogous to FIGS. 3a,
3b, and analogous to FIGS. 4a, 4b of the HCC description, but with
different content. In this step, a user enters additional
information about him/herself and his/her covered household members
that the FSAC will use in calculating optimal contributions. Unlike
the information described previously, the information provided here
is not required for the HCC but only for the FSAC.
[0257] (1) Risk aversion--Risk aversion is the disutility that
users have from facing risk with respect to their uncertain health
status and uncertain future income. Standard gamble and
time-tradeoff methodologies are used to assess users' risk
aversion; these methods are described in Gold et al., 1996. These
questions provide a measure of how much users would be willing to
pay to avoid health and/or income risk, relative to situations in
which they would not face any such risks.
[0258] (2) State of residence--Users are asked to provide the state
in which they reside. This will be used for estimating marginal tax
rates.
[0259] (3) Household Income--The advantage of contributing money to
an FSA is that medical expenses can be paid with pre-tax income.
This advantage clearly depends on the user's marginal tax rate. At
the extreme, for instance. users whose income is not taxed at all
have no advantage from contributing to an FSA. Therefore, users are
asked to provide their estimated total household income for the
coming year (i.e., the year the FSAC will cover). If users choose
not to provide an income range, FSA contribution recommendations
are not provided. This information is used to estimate their
marginal tax rate, taking into account: the amount of discretionary
medical spending.
[0260] Spending on health care in the coming year can be thought of
as having two components: a certain component representing costs
that the user and the covered members of his/her household will
incur with certainty in the coming year; and an uncertain
component, attributable to unanticipated health shocks that might
occur in the coming year. Claims data, previously described, is
used to assess and model the later component.
[0261] However, users are asked to provide information on the
health care use and costs they expect to incur with certainty in
the coming year. More specifically, users are asked to distinguish
between such spending that is "discretionary" and such spending
that is not "discretionary." "Discretionary" means that it is not
medically necessary; this reflects the fact that FSA contributions
can generally be used to cover health-related products and services
that health insurance plans may not cover but that consumers may
wish to buy. such as prescription sunglasses or orthotics, and
therapies, such as chiropractic care or massage therapy.
Non-discretionary spending that consumers know they will incur in
the coming year could include the out-of-pocket costs associated
with a birth or with the treatment of a chronic medical
condition.
[0262] Tool estimates user's marginal tax rate. Information is
collected on the tax schedules for federal income and payroll
deduction (FICA, Medicare. Social Security) taxes, and for the
income tax in states that are relevant to a particular application
of the FSAC. This information is incorporated into the FSAC, and is
used, together with the information on household income provided,
to estimate users' marginal income tax rates. In some applications
of the FSAC, users may be asked to provide their actual or
estimated marginal tax rate.
[0263] GENERATE HOUSEHOLD-LEVEL MEASURES OF HEALTH CARE USE AND
SPENDING FOR EACH HOUSEHOLD IN THE REFERENCE POPULATION. Estimate
out-of-pocket costs for each household in the reference population.
Using the plan design of each health insurance plan for which the
FSAC will provide information, and the claims-level data for the
reference population, the annual out-of-pocket cost are calculated
for households in the reference population had they paid for
services according to the plan design of the respective health
insurance plans.
[0264] Specifically, for each household in the claims data, the
claims are sorted for individual services by household member and
by the date it was provided. Using the plan design, and the price
data, then the health care claims are processed chronologically,
assigning out-of pocket costs according to the plan design of the
particular health insurance plan being modeled. This is repeated
for each plan for which the FSAC will provide information. This
processing takes into account all cost components of the respective
plans, including individual and household deductibles; coinsurance
and co-payments; individual and household stop-loss provisions; and
what services are covered and not covered. For covered services
provided in-network, deductibles and coinsurance rates are
calculated based on the "cost" of the service. For services that
are not covered by a particular plan. and covered services that are
provided out-of-network, deductibles and coinsurance rates are
calculated based on the "billed charge" for the service.
[0265] Note that the plan designs of many insurance plans make
different provisions for services provided "in-network" (e.g. by a
specific group of health care providers with whom the health
insurance plan has a special contractual relationship), vs. "out of
network" (e.g. by any other health care provider). For instance,
Preferred Provider Organizations (PPO) typically have a higher
deductible and higher coinsurance rates for out-of-network services
than for in-network services. Health Maintenance Organizations
(HMO) generally do not provide any coverage for out-of-network
services (unless specifically authorized by the insurer), although
members may of course use such services if they pay for them
entirely out-of-pocket.
[0266] To reflect the fact that the user and his/her dependents may
receive health care in-or out-of network, estimates of
out-of-pocket costs are made in each plan under two scenarios: all
health care is received in-network, and all health care is received
out-of-network. In other implementations, one estimates
out-of-pocket costs in each plan under additional scenarios, such
as outpatient care received out-of-network and inpatient care
received in-network; or one solicits an expected mix of in- and
out-of-network services from the user and estimates out-of-pocket
costs under the user-provided mix.
[0267] For illustration, one can compare two simple plans:
[0268] Plan A: 4: Monthly employee premium for family coverage:
$50
[0269] In-Network Use:
[0270] Each household member pays $10 for each outpatient physician
visit and all associated outpatient services (e.g. diagnostic
tests), $10 for each hospitalization and all associated inpatient
services; $20 for each emergency room visit and all associated
services; and $10 for each prescription that they fill.
[0271] Out-Of-Network Use:
[0272] Not covered.
[0273] Plan B Monthly employee premium for family coverage:
$100
[0274] In-Network Use
[0275] Each household member pays 100% of the cost of the first
$200 of health services in a year (i.e., a $200 individual
deductible); and then pays 20% of the cost for all other health
services in that year (i.e., a 20% coinsurance rate), up to a
maximum out-of-pocket payment (excluding the deductible) of $1500
per year (i.e., a $1500 stop loss provision).
[0276] Out-Of-Network Use
[0277] Each household member pays 100%of the cost of the first
$1000 of health services in a year; and then pays 40% of the cost
for all health services in that year, up to a maximum out-of-pocket
payment (excluding the deductible)of $4000 per year. Using the
hypothetical household profile described, and hypothetical data on
health service use and prices, Table 3a illustrates the process of
calculating out-of-pocket costs for this household in Plans A and
B, assuming all services are provided in-network.
[0278] Table 3a illustrates the process of calculating
out-of-pocket costs assuming all services are provided
in-network.
[0279] Table 3b illustrates the process of calculating
out-of-pocket costs assuming all services are provided
out-of-network.
[0280] In settings in which the FSAC is to provide information on
one or more Point of Service (POS) plans with three tiers of
benefits, out-of-pocket costs are estimated in POS plans under
three scenarios:
[0281] 1) All health care is provided under Tier 1, which most
commonly requires that services are provided in-network, and that
patients are referred for specialty care by their designated
primary care provider; this corresponds to the "in-network"
scenario for HMOs, PPOs and fee-for-service (FFS) plans.
[0282] 2) All health care is provided under Tier 2, which most
commonly requires that services are provided in-network. but that
patients can self-refer themselves for specialty care.
[0283] 3) All health care is provided under Tier 3 in which
patients most commonly can self-refer to any out-of-network
provider: this corresponds to the "out-of-network" scenario for
HMOs, PPOs and FFS plans. In other implementations, one estimates
out-of-pocket costs in each plan under additional scenarios; for
instance, one solicits an expected mix of service use across tiers
from the user and estimates out-of-pocket costs under the
user-provided mix. In some PPO and POS plans, the rules for using
"in-network" and "out-of-network" care can be more complex than
illustrated here. For instance. some PPO plans have one set of
rules for out-of-network physician visits and another for
out-of-network hospital admissions. The particular way in which the
rules are applied in the "out-of-network" scenario (PPO), and
"Tier: 2" and "Tier 3" scenarios (POS plans) will depend on the
details of the particular plans, and on the particular application
of the FSAC.
[0284] This step thus creates the following measures for each
household in the reference population: (1) what the household would
have spent out-of-pocket for health care in each health insurance
plan for which the particular application of the FSAC provides
information, under two (for HMOs, PPOs, and FFS plans) or three
(for POS plans) scenarios; also, (2) what the household would have
spent out-of-pocket if they had no health insurance and had been
required to pay the billed charges for all services.
[0285] Dividing the reference population into categories of health
care use. For each household in the reference population, the total
of health care used during the year covered by the claims data is
calculated. Specifically, each health care service is priced using
its cost. Then the cost of all services for each member of a
household are added up to obtain the household total. For instance,
the total health care use of the hypothetical household illustrated
in Table 3 over that year is $10,865 (the sum of the "cost"
column). Then all households in the reference population are sorted
by the total value of health services used during the year.
[0286] Next, the households in the reference population are
stratified into a small number of categories of health care use, by
the total value of health services used during the year. The number
of categories, and the range of total values of health services
that define the categories, can vary depending on the particular
group of users for which the FSAC is being implemented.
[0287] For instance, five categories of health care use can be
defined as follows:
[0288] "No" use: No member of the household used any health care
during the year covered by the claims data.
[0289] "Low" use: The total value of health services used during
the year ranged from $1 to $1000 during the year covered by the
claims data.
[0290] "Moderate" use: The total value of health services used
during the year ranged from $1000 to $3000.
[0291] "High" use: The total value of health services used during
the year ranged from $3000 to $10000.
[0292] "Very High" use: The total value of health services used
during the year was more than $10000.
[0293] Each household in the reference population thus gets
assigned to one category of health care use. A specific
implementation of the FSAC could use more or fewer categories, and
could define the categories differently.
[0294] This step thus creates the following measures for each
household in the reference population: (1) total cost of the health
care used by the household during the continuous 12 month period
covered by the data; and (2) the category of health care use to
which the household belongs, among the categories used for a
particular application of the FSAC.
[0295] Creation household-level data file for reference population.
From the claims-level data file, one creates a household-level data
file for the reference population. This file contains the
demographic and health status information on each member of the
household, and all of the household measures, as previously
defined. One refers to this as the "household-level" data file.
[0296] CALCULATING OPTIMAL FSA CONTRIBUTIONS: A model of optimal
contributions to a flexible savings account The main incentive that
employees have for contributing to a flexible savings account (FSA)
is the ability to spend pre-tax dollars on medical care. However,
the optimal amount to contribute is considerably complicated by two
factors: (1) uncertainty regarding the incidence of medical
expenditures over the course of the coming benefit year and (2)
loss of any unspent money in the FSA at the end of the year. The
estimated optimal contributions for different scenarios are then
calculated, based on the user's input and the user's household's
likely distribution (estimated by the FSAC, using the methods we
have described) of out-of-pocket costs for the coming year in
relevant health plans, and considering the degree of risk aversion
indicated by the consumer. The key innovation in the method for
calculating optimal contributions is the recognition that medical
expenditures toward the end of the benefit year can improve health
or be otherwise productive. Previous methods, which assigned zero
weight to medical expenditures at the end of the benefit year,
systematically underestimated optimal FSA contributions.
[0297] Although the present invention and its advantages have been
described in detail, it should be understood that various changes,
substitutions and alterations can be made herein without departing
from the spirit and scope of the invention as defined by the
appended claims. Moreover, the scope of the present application is
not intended to be limited to the particular embodiments of the
process, machine, manufacture, composition of matter, means,
methods and steps described in the specification.
[0298] As one of ordinary skill in the art will readily appreciate
from the disclosure of the present invention, processes, machines,
manufacture, compositions of matter, means, methods, or steps,
presently existing or later to be developed that perform
substantially the same function or achieve substantially the same
result as the corresponding embodiments described herein may be
utilized according to the present invention. Accordingly, the
appended claims are intended to include within their scope such
processes, machines, manufacture, compositions of matter, means,
methods, or steps.
* * * * *