Concept Architecture
Concept
Theoretically, Pharmacy-Based Risk Score is a quantitative risk adjustment measure that estimates an individual's expected healthcare expenditure, morbidity or healthcare utilisation using prescribed medication data rather than, or in addition to, diagnostic information. The concept is founded on the observation that patterns of pharmaceutical use reflect underlying disease burden, disease severity and chronic health conditions. Pharmacy-based risk scores are widely used in health insurance, provider payment and population health management to improve prediction of healthcare costs when diagnostic information is incomplete or unavailable.
Mathematically, a Pharmacy-Based Risk Score is represented as a weighted combination of pharmacy-derived risk indicators and other recognised explanatory variables. Medication classes are mapped to predefined pharmacy risk groups, and regression coefficients estimated from historical healthcare data quantify the contribution of each medication category to expected healthcare expenditure or utilisation. The resulting weighted sum forms an individual's overall risk score.
In practice, pharmacy-based risk scores are calculated from pharmacy claims, prescription dispensing records or electronic prescribing systems using recognised classification systems such as Pharmacy-Based Cost Groups (PCGs), RxGroups or RxRisk models. Medication records are classified into therapeutic categories, weighted according to published algorithms and combined with demographic variables where appropriate. The resulting risk scores are used in risk adjustment, capitation payments, expenditure forecasting and healthcare performance assessment.
Purpose
Used to estimate patient morbidity from medication use, predict healthcare expenditure, support risk-adjusted payment systems, improve population health management, forecast healthcare utilisation and compensate for incomplete diagnostic information.
Mathematical Formulae
Primary Formula
Risk Score? = ?? + ? ??P??
where:
- P?? = pharmacy-based risk indicator or medication category for individual i
- ?? = estimated coefficient associated with medication category j
- ?? = intercept
Supporting Formulae
Expected Cost? = ?? + ??X?? + ??X?? + ? + ??X??
Residual:
e? = Y? ? ??
Related Mathematical Methods
- Pharmacy-based risk adjustment
- Multiple linear regression
- Generalised linear models
- Case-mix adjustment
- Expected value modelling
- Predictive modelling
Example
A pharmacy-based risk model assigns the following coefficients:
- Intercept = 0.45
- Insulin therapy = 0.70
- Heart failure medication = 0.80
- Chronic obstructive pulmonary disease medication = 0.40
A patient prescribed insulin and heart failure medication has:
Risk Score = 0.45 + 0.70 + 0.80
= 1.95
A risk score of 1.95 indicates substantially higher expected healthcare expenditure than the reference population with a score of 1.00.
Excel Implementation
| Function | Example Formula | Health Economics Application |
|---|---|---|
| SUMPRODUCT | =SUMPRODUCT(B2:F2,$J$2:$J$6)+$J$1 | Calculates pharmacy-based risk scores from medication category weights. |
| XLOOKUP | =XLOOKUP(A2,Drug_Table[Drug_Code],Drug_Table[Risk_Weight]) | Maps medication codes to recognised pharmacy risk weights. |
| IF | =IF(B2="Yes",0.70,0) | Applies medication-specific coefficients when indicator variables are used. |
| LINEST | =LINEST(C2:C1000,D2:H1000,TRUE,TRUE) | Estimates regression coefficients during model development. |
VBA (Optional)
A VBA routine can automatically classify pharmacy claims into recognised medication risk groups, calculate pharmacy-based risk scores and generate risk-adjusted payment files.
Sources
- Fishman PA, Goodman MJ, Hornbrook MC, Meenan RT, Bachman DJ, O'Keeffe Rosetti MC. Risk Adjustment Using Automated Ambulatory Pharmacy Data: The RxRisk Model.
- Sloan KL, Sales AE, Liu CF, et al. Construction and Characteristics of the RxRisk-V Risk Adjustment Instrument.
- van de Ven WPMM, Ellis RP. Risk Adjustment in Competitive Health Plan Markets.
- Iezzoni LI. Risk Adjustment for Measuring Health Care Outcomes.
- ISPOR Good Practice Reports.
Related Concepts (2)
Library
Publications
1
Uncertainty and the Welfare Economics of Medical Care — Kenneth J. Arrow, Vol. 53, No. 5 ed., 1963 (American Economic Review)
The founding paper of health economics as a discipline, analysing how uncertainty, asymmetric information, trust and the special features of medical markets prevent them from behaving like ordinary competitive markets — the intellectual origin of the entire field.
Journal ArticleView source →
Frequently Asked Questions (6)
What is a pharmacy-based risk score?
A numerical estimate of expected future healthcare costs calculated using prescription drug utilisation patterns as an indicator of health status.
Source: Ellis 2008
What does a pharmacy-based risk score use to gauge health status?
A pharmacy-based risk score is a numerical estimate of expected future healthcare costs that uses a person's prescription drug patterns as a marker of their health. Because the medicines someone takes reveal the conditions they are being treated for, drug use serves as an indirect signal of how sick they are and how costly they are likely to be. This makes it a practical tool where drug data are easier to gather than full diagnoses, though it differs from a diagnosis-based risk score, which reads health directly from recorded conditions. Reading health from medicines to forecast cost is what it does. Ellis (2008) sets out such scores.
Source: Ellis 2008
How is a pharmacy-based risk score calculated?
A pharmacy-based risk score is calculated using prescription drug utilisation patterns as an indicator of health status, so the drugs an individual uses inform an estimate of their expected future healthcare costs. So a pharmacy-based risk score is calculated from drug use, which is why drug patterns matter, since they indicate health, and a pharmacy-based risk score is calculated by using an individual's prescription drug utilisation as an indicator of health status to estimate their expected future healthcare costs.
Source: Ellis 2008
Why are prescription drug patterns used in a pharmacy-based risk score?
Prescription drug patterns are used in a pharmacy-based risk score because the drugs an individual uses can indicate their health conditions, so drug utilisation serves as an indicator of health status for predicting expected costs. So drug patterns indicate health for the score, which is why they are used, since medications reflect conditions, and prescription drug patterns are used in a pharmacy-based risk score as an indicator of health status, since the drugs taken point to conditions that predict expected future healthcare costs.
Source: Ellis 2008
What does a pharmacy-based risk score estimate?
A pharmacy-based risk score estimates expected future healthcare costs, so it gives a numerical prediction of an individual's likely costs, based on their prescription drug utilisation patterns as an indicator of health. So a pharmacy-based risk score estimates expected costs, which is why it is used in risk adjustment, since predicting costs supports adjustment, and a pharmacy-based risk score estimates an individual's expected future healthcare costs as a number derived from their prescription drug utilisation patterns.
Source: Ellis 2008
How does a pharmacy-based risk score differ from a diagnosis-based risk score?
A pharmacy-based risk score differs from a diagnosis-based risk score in the indicator used: a pharmacy-based score uses prescription drug utilisation patterns as an indicator of health, while a diagnosis-based score uses documented medical diagnoses. So they differ in the health indicator, which is why they are distinguished, since one uses drug use and the other diagnoses, and a pharmacy-based risk score estimates expected costs from drug utilisation patterns, whereas a diagnosis-based risk score uses documented diagnoses alongside demographics.
Source: Ellis 2008
Trust Record
Verified by Dr Darrin Baines
British health economist
Professional identity: darrinbaines.org
Verification date: 16 Jan 2026
Content version: 1.0.0
Canonical Identity
- Term code
- HS-HP-HI-130
Stable URI · Machine-readable · Resolvable · CC BY 4.0