Gi Ch. 34 Medical Claim Cost Trend Analysis

Front 1

 The Market Force of Trend means the average Claim Cost trend from all types of private payers of health benefits.
 Fee-for-service trend > Market Force of Trend
 Capitation trend < Market Force of Trend

Back 1

Blank

Front 2

How to Analyze Trend

Back 2

 Traditional way: split it into
 Utilization trend
 Cost-per-service trend
 According to this chapter, the two components are:
 the Market Force of Trend (macroeconomic factors)
 Company-specific factors (microeconomic factors).

Front 3

ECONOMIC ISSUES IN TREND ANALYSIS
(“Components of the Market Force of Trend”)
(“Factors Affecting the Level of Consumption of Health Care”)

These are 3 ways of asking for the Macroeconomic Factors.

Back 3

 Demographic changes (aging)
 Wealth
 Inflation
 physician supply
 cost-shifting
 proportion of specialty physicians
 Managed Care’s effects

Front 4

FACTORS THAT AFFECT A COMPANY’S OBSERVED TREND
These are the Microeconomic Factors.

Back 4

 Antiselection
 Makes a plan’s trend > underlying cost trend
 ph’s optimize their financial situation
 disproportionately high claim costs within a given demographic cell
 Antiselection does not include actual changes in the demographic mix.

 Cost-Shifting
 Reduced or negotiated payments by HMO’s
 causes one-time decrease in cpy’s trend
 followed by a recovery due to fee raises (negation)
 unless growth occurs in the line of business subject to discount.

 Risk Shifting
 immunizes the MCO to the market force of trend
 but providers raise fees to negate this.
 one-time effect, again.

 Managed Care Initiatives (Interventions)
 UM implementation causes one-time drop in trend.

 Benefit Design and Leveraging
 policies with deductibles, and stop loss policies, cause leveraging.
 causes observed trend to be > market force of trend

 Type of Service
 Inpatient vs. outpatient
 Pharmacy trends lower than medical

 Random Fluctuations
 significant for small blocks
 use linear regression to remove the random component.

See numerical examples in chapter notes.

Front 5

EXTERNAL SOURCES OF INFORMATION – PUBLIC MEASURES OF MEDICAL COST TRENDS

Back 5

 Medicare Reports
Advantages: history of Medicare claim costs
Disadv: eligibility / legislative changes distort the data

 National Health Expenditure portion of the GDP
Advantages: None
Disadv: slow publication – out of date.

 HCI (Health Cost Index) and Medical CPI
Advantages: Up to date; widely used
Disadvs: Underweights hosp/phys; overweights vision/dental
Based on billed charges – discounts not reflected.

 Trend Surveys (Compiled by consultants)
Advantages: Easy to obtain
Disadv: Statistical rigor varies
purpose varies

Front 6

Considerations in using External Data Sources:

Back 6

 source of the data
 objective of the data
 differences between the different external sources

Front 7

Advantages of Comparing Claims to an External Benchmark

Back 7

 notice company-specific anomalies faster
 Shows the trend being experienced by competitors

Front 8

THE DIFFERENT APPROACHES TO ANALYZING CLAIM COST TRENDS
Techniques not using external data (non-causal models)

Back 8

 “Historical Averages and Graphs”
1. Graph cpy’s and competitors’ data
2. Use judgment to project trends.

Disadvantages:
 Subjective
 Projected trend depends heavily on recent results
 Leads to the typical 6-year underwriting cycle.

 “Actuarial Model”
This is the method we used in SN 202
1. Split PMPM costs into U and C
2. split by type of service, type of contract, etc.
3. Project forward based on “known impacts”

Advantage:
 Greater level of detail
Disadvantages:
 Result is still mostly based on historical experience.
 Judgment is still used

 Linear Regression (Applied to Historical Experience)
Advantage:
 removes random fluctuations
Disadvantage:
 also removes seasonality.

 Time Series – ARIMA models
 Only good for short-term projections.

Front 9

A technique using external data (Causal Model)
(Promoted by this chapter)

Back 9

1. Gather leading macroeconomic indicators
2. Assume these indicators are a proxy for the market force of trend.
3. Collect historical data.
4. Determine the correlation between historic data and the indicators
 by graphing.
5. Use linear regression to predict costs = f(indicators).

Recommendations:
 Include the following variables:
 HCI
 Managed Care interventions
 Seasonality
 Deviation Years

Front 10

DIFFICULTIES IN MEASURING CLAIM COST TREND
(Common Problems In Trend Analysis)

Back 10

Practical difficulties
 Lack of data, models, staff
 Competitive considerations limit ability raise rates even if justified

Data difficulties
 Separating components:
 Price trends & Utilization
 Macroeconomic & microeconomic factors
 observed trend and Market Force of Trend
 due to Antisel, Cost-shifting, Leveraging, Risk-transfer, etc.
 one-time vs. ongoing trends (catastrophic claims)
 Pulse outliers; level shifts
 real vs. random fluctuation

 immature claims data
 Changes in claim payment speeds distort reserves
 Changing Demographics
 Changing data systems and products
 terminations / new sales make data incomparable
 Choosing the length to analyze:
 A 12-month moving-average is stable, but less responsive;
 A 3-month average is responsive, but volatile.

Front 11

THE IMPACT OF MANAGED CARE ON THE FORCE OF TREND
(Changes in Trend Analysis under Managed Care)

The MCO must consider:

Back 11

 Competitivity
 One-time anomalies, such as:
 catastrophic claims
 enrollment shifts
 interventions; e.g.
 UM program implementation
 stricter claims adjudication & processing changes
 change in underwriting
 Must separate capitations from Ffs payments

Front 12

LONG-TERM MEDICAL TREND ANALYSIS

Back 12

Needed for: employers’ retiree benefits, Medicare, LTC
Considerations in analyzing long-term trend: COLA guarantees, historical trends vs. GDP, aging of the population

Done.