Group Chap 21: Estimating Medical Claim Cost

Question 1

Introduction

Answer 1

1. Insurers produce group rate manual to estimate medical costs
- Outlines pricing adjustments to plan design, demographics, utilization management, group size, area, etc

2. Small groups were formerly defined as 2-50 people, but ACA now defines it as up to 100 employees
- Protecting Affordable Coverage for Employees (PACE) Act - allowed states to choose whether small group is defined as 50 or 100 employees. Most states continued with 50 employees
- Under ACA, estimated cost for insurer’s small group population must be calculated based on entire book of small group business for the insurer (“single-risk pool”)

3. Expected medical costs used for rate manuals. Rate manuals used for premium rate development

Question 2

Data Considerations

1. Factors to consider when estimating claim costs
   a. Appropriate data
   b. Level of detail
   c. Data source

Answer 2

1. Summarizing medical claim data is complex: Multitude of benefit plans, variations in claim coding, varying group characteristics

2. Factors to consider when selecting data sources described below
a. Is the data appropriate for Estimating Claim Costs?
- Source data should reflect characteristics of company’s population and benefits

• Need to be adjusted for demographic mix, benefit levels, provider contracting details, rating period

b. What level of detail is needed?
- Aggregate data sufficient to estimate medical claim costs

• Developing manual rates, insurer’s data for entire line of business used
• If experience rate for a particular group, then data for that group is used
• For ACA rating, all groups must be aggregated togethr
• Aggregate methods supplemented with more detailed analyses, as insurers ability to handle large volumes of data continues to imrpove
• Employee choice between plan options, causes need for more detailed analysis - consumer-driven products are becoming more common and complicate matters further
• The more stable and the fewer differences between experience period and the rating period characteristics, the more likely that aggregate data is appropriate

c. Data Sources
- Best source is company’s own experience, if credible. Will reflect company’s cost patterns, administration, and risk selection

• Variery of internal sources - from difference systems and different formats
  i. ASOP 5 requires actuary to check reasonability of data and ASOP 23 provides more detail about how data can be used
• Typical sources of Internal Medical Data
  i. Medical Claims System - Captures billed claims, claim eligibility status, allowed amounts and paid amounts
  ii. Pharmacy Benefit Manager (PBM) Data - if a third-party PBM is used
  iii. Premium Billing and Eligibility Data - Used to calculate historical claim costs on per member or per employee basis and also to assess historical mix of business
  iv. Provider Contract System- Keep track of discounts and reimbursements to providers
• Sometimes internal data is not appropraite for projecting future costs
  i. where changes are occurring in benefit design, type of group, provider contracts, utilization management, underwriting practices, or claim administration
  ii. Where information not at detailed claim level
• When using external data, must understand the characteristics of the population that may affect the data
• Public sources of data; CMS, Bureau of Labor Statistics, or purchased from industry or actuarial consulting firms

Question 3

Collecting Data for Manual Rates

Variables to be considered in Collection Process
1. Financial Information
2. Experience Period
3. Exposure Basis
4. Frequency of Data Collection

Answer 3

1. Collected in Level of detail consistent with rating methodology

2. Following variables should be considered in colelction process
a. Financial Information
- Define type of claim dollars to be collected
i. If plans or reimbursement are changing, collect paid claim amounts as well as billed and allowed claim dollars

ii. Difference between paid and allowed reflects insured cost sharing

iii. Difference between allowed and billed reflects negotiated provider arrangements
- Define treatment of reinsurance recoveries and COB recoveries

b. Experience Period
- Twelve-month period ensures a complete seasonal cycle (claim levels vary over the year)
i. Less elective care services used in Nov and Dec - care often postponed during the holidays
ii. Claims in the earlier months of the year lower because benefits do not apply until after the deductible is paid
iii. Non-calendar year deductibles add another level of complexity

• Collecting data for a two or three year time period should be considered - helps identify variances between years
• Claims may be analyzed on incurred basis (dates of service) or a paid basis (date claims are received by providers)
• Capture corresponding exposure information
• Incurred basis claims should be “completed”
  i. By collecting data with payment dates extending far enough beyond incurral period, or
  ii. Estimate completion factors to apply to project ultimate claims

c. Exposure Basis
- Monthly counts of insured over the experience period
i. No. of employees
ii. No. of contracts
iii. No. of subscribers
iv. No of covered persos (members)
- First three bases subject to mis-estiamation when family composition changes
- Third basis requires exposure for spouses and each child
- In some cases, spouse and children counts are estimated to calculate costs (PMPM)

d. Frequency of Data Collection
- For annual basis, results can be used to study trend and impact of cost containment programs, plan design changes, or utilization management
- Less detailed study can evaluate overall medical costs and trends

Question 4

Normalizing Data for Important Rating Variables

1. Age and Gender
2. Geographic Area
3. Group characteristics
4. Benefit Plan
5. Utilization Management Program
6. Provider Reimbursement Arrangements
7. Other Risk Adjusters

Answer 4

1. Age and Gender
a. Variation by age and gender can be substantial

b. Relationships do not change significantly year-to-year, so performacne age/gender studies infrequently

c. May adjust historical costs to a “standard population”
- Alternatively, may treat population underlying the experience as the standard population

d. Separate age/gender factors appropriate
- For major service categories
- Can develop separate age/gender factors for high deductible plans
- Typically 5-year or 10-year age bands

e. in the past, manual rate was adjusted to reflect actual age/gender composition of group being rated
- ACA reform eliminates gender-specific rates and restricts premium variation due to age factor 3:1 for individuals and small groups

2. Geographic Area
a. Claim costs by geographic vary widely

b. Companies lack volume to study their claim cost at zip code, or metropolitan statistical area

c. Use area factors from competitor or actuarial consultant, then monitor experience

d. Might focus on fine level of service, while keeping area broad

e. For details at more granular level of geography, might focus only on total claim costs

f. Under ACA, small groups can vary by geographic area
- Area factors allowed to vary by provider contract differences, not by morbidity of area (morbidity differences can only be captured in the overall risk pool)
- E.g. Area A expected to have 10% higher costs than Area B. Then actually has 20% higher costs in a population with worse health status. Only 10% differential is allowed in the area factor under ACA small group rules

3. Benefit Plan
a. Claim costs adjusted to reflect different plan from average plan included in claim data

b. Managed care variations include copayments, dolalr maximum for specific benefit, covered services, combinations of copayments deductibles and coinsurance

c. Induced Utilization or Induced Demand - different benefit plans have different utilization based on degree of cost sharing

d. Higher deductible acts as a deterrent to seeking medical care. People less likely to submit claims under deductible

e. Practical to adjust all experience data to reflect richest benefit plan
- Process uses allowed charge data, and adjusting utilization to reflect richer benefit plan

f. HDHPs utilization may vary based on how much money from account is available

g. Under ACA, insurers can’t reflect plan selection factors since that is reated to morbidity variance (risk must be spread across entire risk pool - similar to considerations for geographic factors)

4. Group Characteristics
a. Manual rate should represent “average group”
- Historical experience must be adjusted to reflect average group characteristics

b. Industries with above average costs involve physical labor of those where employees highly aware of available benefits

c. As the size of group increaes, the impact of individuals with serious conditions is dampened

d. ACA restricts allowable rating variables for individuals and small groups (and certain large groups)
- Allowable: family status (individual vs. family), geographic area, age (3:1 variation permitted), tobacco use (1.5:1 variation permitted)
- Industry and group size no longer allowed

e. Tendency for smaller employers to provide coverage in a way that selects against insurers
- Claim cost studies on blocks of small cases often made on durational basis

f. With HIPAA’s restrictions, first year claims much closer to ultimate
- Additionally, number of years to reach ultimate claim levels is less
- ACA erodes durational claims pattern further, since it restricts premium variances and underwriting

5. Utilization Management Program
a. Focus of prospective UM on necessity of proposed treatment and appropriateness of setting

b. Prior to care, preauthorization where proposed treatment approved before it is delivered

c. During care, UM involves monitoring patient progress

d. Retrospective, UM ensures reported care was actually delivered or to assess whether care was necessary

e. UM also includes disease management, focused on care to patients with chronic disease

f. Source data should be adjusted for changes in UM programs

6. Provider Reimbursement Arrangements
a. Variety of forms (Discounts from billed charges, per diems, case rates, Medicare DRG-based payment levels, capitation)
b. If arrangements changed during experience period, adjust the experience

7. Other Risk Adjusters
a. Age, gender, geographic, and industry factors used historically in health insurance
b. More refined risk adjusters emerging based on claim, diagnosis, encounter data, and pharmacy claims

Question 5

Projecting Experience Period Costs to Rating Period

Answer 5

1. Claim costs trended from experience base period to the rating period
2. Trend includes changes in average unit cost per service, utilization, medical practice, mix of services, provider reimbursement arrangements, new technologies, drugs, plan design
3. Secular trend: percentage change only from factors that affect first dollar 100% benefits
   - Components of secular trend are 1. changes in utilization and 2. average unit cost per service
4. Deductibles and copayments create leveraging which increases impact secular trend has on claim costs
5. Historical changes
   - Examined by comparing costs for all services combined or by individual types of service
   - Historical trends separated further between changes in utilization and unit costs
6. Twelve month trends indicate annual trends, while three month pattern a leading indicator of trend direction
7. When negotiated arrangements include scheduled reimbursements
   - Secular trend on unit costs estiamted by exaining changes in these arrangements
   - Arrangements based on discounts off billed charges are less useful in estimateing trend
8. Managed Care backlash has caused increases in MC trends

Question 6

Methods of Adjusting Manual Base Rates

Manual rate must be adjusted to reflect demographics of the group, geographic area, group size, industry, tobacco use
- ACA limits allowable rating variables in small group market

Benefit adjustments for traditional insurance calculated very differently than MC products

1. Claim Probability Distribution
2. Actuarial Cost Models
3. Using Cost Model to Estimate Impact of Benefit Plan Changes

Answer 6

1. Claim Probability Distributions (CPD)
a. To estimate impact of deductibles, coinsurance, OOP maximums, and annual maximums on claim costs

b. Used where reimbursement and benefit design are primarily a deductible and coinsurance (e.g. HDHPs)

c. To construct CPD
- Annual claims are summarized on a per member basis
- Members are then grouped into cost ranges to calculate average annual claims and frequency

d. This method useful for pricing comprehensive or major medical, where there is deductible, coinsurance, and OOP limit

2. Actuarial Cost Models
a. Develops claim cost for each detailed type of service category

b. used by HMOs, PPOs, POS and other plans utilizing copayments

c. Most useful where copayments and limits apply to specific services, rather than to all services combined

d. Gross benefit cost quoted as PMPM calculated as
- One-twelfth the product of Annual frequencies and Average allowed charges
- Reduced by value of any copayments
- Total net benefit cost PMPM is the derived

e. Provider arrangements should be recognized

f. A weighted average of arrangements may be appropriate

g. Assumptions and methods must be chosen consistent with provisions and contract arrangements, which may vary by service category

3. Using Cost Models to Estimate the Impact of Benefit Plan Changes
a. Actuary can evaluate changes in benefit design using these models

b. Common for benefit design to influence insured’s behavior

c. Insurers often measure cost changes net of copayments

d. If office visit copayment increased, office visit utilization will decrease and likely decrease number of presciptions written

e. For POS and PPO plans, depending on efficiency of in-network providers, and depdning on reimbursement levels, the net cost to plan sponsor of in-network might actually be lower that out-of-network

f. Contractual arrangement with providers can impact cost changes due to benefit deisgn change
- E.g. If a certain benefit is covered under capitation, a change in copay will not have any rate impact unless capitation payment can be adjusted

Question 7

Small Group Considerations

1. General Considerations
2. Evoluation of Underwriting and Rating
3. Experience Conisderations
4. Underwriting the Small Employer
5. Rating Structure and Paramerer
6. Risk Adjustment Programs

Answer 7

1. General Considerations
a. Small group originally meant 2-50 employees (under HIPAA)

b. Under ACA, small group means up to 100 employees
- Also must consider part-time, temporary and seasonal employees in this count
- PACE Act allows states to choose 50 or 100 employee definition

c. Many small groups don’t offer medical plans to employees or have high cost sharing, simply due to cost considerations and relatively high eates of employee turnover

d. Higher potential for antiselection in small market as only some employers will provide this medical coverage, and those are often the ones that know they need it

e. Under ACA, group of 50 or mroe are required to offer coverage or pay a penalty fee per employee
- Small groups under 25 are offered temporary tax credits for providing coverage

f. With ACA features, employees now have other ways to get health insurance
- 35% of groups with less than 50 employees offered health covered in 2013 (pre ACA), and then dropped to 30% in more recent years
- Tax incentives haven’t been a significant enticement for small employees to offer coverage

2. Evolution of Underwriting and Rating
a. Under HIPAA, small groups had to offer medical on guaranteed acceptance and renewal basis
- Could consider group health status, size and industry (with certain limits for these factors)
- Early underwriting based on evaluting the business entity and examining health status and other features of the individuals

b. ACA Changes
- Essential health benefit requirements and standardized plans (metallic plans)
- Elimination of a number of rating variables; Elimination of health status rating - eliminated need for medical underwriting
- Further limitations on allowable rating factors

c. Rating changes results in higher subsidization by the healthier, younger groups to the older, less healthy groups

d. Groups that were likely to experience higher premiums with new ACA plans waited to adopt these plans when possible
- Grandfathered plans and transitional plans were allowed to continue, leading to potential disparity in pre-ACA and ACA plan experience
- ACA expereince more likely comprised of groups adversely selecting against the pool

e. Some small groups will exit the market due to higher costs

3. Experience Considerations for Insurers for Small Group Rating
a. Insurer can’t decline or rate groups for various coverage

b. Financial viability
- Is it a legitimate business being insured? How long have they been in business?

c. Industry/Occupation
- Type of business and duties is related to expected claim costs

d. Group Size
- Affects expected claim levels and per capita acquisition and maintenance expenses

e. Workers Compensation
- In states where WC is not required, there will be increased risks and may have to cover some of these expenses

f. Participation
- ACA does not allow participation minimum limits (formerly in place to help fight adverse selection)

g. Employer Contributions
- ACA does not allow requirements on this (formerly helped as higher employer contribution rates resulted in higher particiption from employees)

h. Prior Coverage
- Want to understand motives of why a group is buying coverage, particularly if they didn’t it before

4. Underwriting the Small Employer
a. Allowable Criteria for Underwriting Small Groups
- Verification that entity is a licensed employer in the state
- Participation and contribution requirements outside of open enrollment period
- Requirement that employees live, work or reside in service area of plan’s network
- Employee eligibility requirements (hours, eligibility class, etc)
- Enforcement of employer restrictions on coverage for late entrants (waiting periods)

5. Rating Parameters and Structures
a. Allowable Rating Parameters for Small Groups Under ACA
- Age - Maximum 3:1 factor (for ages 21-64)
- Geographic Area - Allowed to rate by geographic zones with rate variation by expected claim cost, provider arrangement, marketing expense differences, variance in loss ratio minimums by state
- Benefit Plan - Rates can vary by plan design
- Managed Care and Negotiated Discounts - Vary by care managemetn and cost variance for network arrangemetns
- Family Composition - Calculate average enrollee premium per person above age 21 and average below 21 and then just sum rates based on number of members above and below 21 (limited to 3 children under 21)
- Tobacco Use - Maximum load of 50% (i.e. ratio of 1.5:1)

6. Risk Adjustment Program
a. Helps normalize the risk of gurarnteed issue and new rating restrictions of the ACA
b. Formula accounts for demographic risk and health status risk
c. Program intended to be “zero-sum” - higher risk insurers get payments and lower risk insurers pay into the pool

Question 8

Data Sources

Answer 8

1. Federal Government Publications
* National Ambulatory Medical Care Survey (NAMCS)
* National Hospital Care Survey (NHCS)
* National Health Interview Survey (NHIS)
* Consumer Price Index (CPI)
* Health, United States Annual Report
* Public Health Service Sources – CDC, Center for Infectious Diseases, Center for Prevention Services, etc
* Health Resources and Services Administration
* Alcohol, Drug Abuse, and Mental Health Administration
* National Institutes of Health
* Centers for Medicaid and Medicare Services
* Bureau of the Census
* Bureau of Labor Statistics

2. Other External Data Sources and Publications
* Society of Actuaries Research
* State Health Data Organizations
* HMO and PPO Data Sources
* National Committee for Quality Assurance
* Hospital, Medical and Other Periodicals and Sources
* Actuarial Consulting and Analytics Firms
* Other Organizations