MII IG PRO
2027.0.0-ballot.1 - Germany

This page is part of the MII Implementation Guide PRO (v2027.0.0-ballot.1: 2027 Ballot 5) based on FHIR (HL7® FHIR® Standard) R4. This is the current published version. For a full list of available versions, see the Directory of published versions

Cross-Instrument Mappings

Overview

Cross-instrument mapping enables the translation of scores between different PRO instruments that measure the same construct. This is essential for harmonizing data from different sources and ensuring comparability of study results.

Related pages:

Mapping Methods

1. Equipercentile Linking

The most common method is based on the assumption that persons with the same percentile rank on different instruments exhibit the same level of the measured construct.

2. IRT-Based Calibration

Item Response Theory enables the placement of different instruments on a common metric through co-calibration. Details on IRT methodology see Domain-Based Scoring.

3. Regression-Based Prediction

Linear or non-linear regression models for predicting scores of one instrument based on another.

Depression Domain: Comprehensive Mapping Table

The following figure shows validated mappings between PROMIS Depression T-Scores and eight other established depression scales:

Depression Scale Mappings

Figure 1: Translations of PROMIS T-Scores to other scales – Comprehensive mapping table for the depression domain

Interpretation Guide for the Mapping Table

The table shows for each PROMIS T-Score (horizontal axis, 30-90):

  • Upper row: Corresponding raw score of the respective instrument
  • Lower row (in parentheses): 95% confidence interval

Example reading:

  • A PROMIS T-Score of 60 corresponds to:
    • PHQ-9: Score of 10 (95% CI: 7.55-8.91)
    • BDI-II: Score of 20 (95% CI: 21.22-26.01)
    • HADS: Score of 15 (95% CI: 14.74-15.75)

Supported Instruments in the Depression Mapping

  1. BDI-II (Beck Depression Inventory-II)
    • Range: 0-63
    • Cut-offs: Minimal (0-13), Mild (14-19), Moderate (20-28), Severe (29-63)
    • See also: BDI-II in PRO Library
  2. CES-D (Center for Epidemiologic Studies Depression Scale)
  3. EPDS (Edinburgh Postnatal Depression Scale)
  4. HADS (Hospital Anxiety and Depression Scale - Depression Subscale)
  5. K6 (Kessler Psychological Distress Scale)
  6. PHQ-2 (Patient Health Questionnaire-2)
    • Range: 0-6
    • Screening cut-off: >=3
  7. PHQ-8 (Patient Health Questionnaire-8)
    • Range: 0-24
    • Cut-offs similar to PHQ-9
  8. PHQ-9 (Patient Health Questionnaire-9)
    • Range: 0-27
    • Cut-offs: Minimal (0-4), Mild (5-9), Moderate (10-14), Moderately Severe (15-19), Severe (20-27)
    • See also: PHQ-9 in PRO Library

Practical Application

Use Case 1: Study Harmonization

A multi-center study uses different instruments:

  • Center A: PHQ-9
  • Center B: BDI-II
  • Center C: HADS

Solution: All scores are mapped to PROMIS T-Scores for unified analysis.

Use Case 2: Instrument Change in Longitudinal Studies

A patient was initially assessed with BDI-II (Score: 25), follow-up with PHQ-9:

  1. BDI-II Score 25 maps to PROMIS T-Score ~62
  2. Expected PHQ-9 score at the same severity level: ~11-12

Use Case 3: Meta-Analyses

Systematic reviews can compare effect sizes across studies using different instruments through transformation to a common PROMIS metric.

Important Limitations

1. Consider Confidence Intervals

Confidence intervals indicate the uncertainty of mappings. For critical clinical decisions, these intervals should be taken into account.

2. Population Specificity

Mappings were developed and validated in specific populations. Generalizability to other populations (e.g., different cultures, age groups) should be verified.

3. Conceptual Differences

Despite measuring the same construct, instruments may emphasize different aspects:

  • PHQ-9: DSM-5 criteria-based
  • BDI-II: Emphasizes cognitive symptoms
  • HADS: Excludes somatic symptoms

4. Floor and Ceiling Effects

At the extremes of the scales, mapping accuracy may decrease, especially with:

  • Very low scores (floor effect)
  • Very high scores (ceiling effect)

FHIR Implementation

Observation with Mapping Documentation

// FSH
Instance: Depression-Score-Mapped
InstanceOf: Observation
* code = LOINC#77861-3 "PROMIS Depression T-score"
* valueQuantity = 60 '[T-score]'
* derivedFrom = Reference(PHQ9-Response)
* method = SCT#702663005 "Equipercentile equating"
* note.text = "Mapped from PHQ-9 raw score of 10 (95% CI: 7.55-8.91)"

Quality Assurance

Validation Requirements

  1. Concurrent Validity: Correlation between mapped scores >=0.8
  2. Classification Accuracy: Agreement of severity categories >=85%
  3. Test-Retest Reliability: ICC >=0.75 for mapped scores

Documentation Standards

For every mapping, document:

  • Source instrument and raw score
  • Target score and confidence interval
  • Mapping method and version
  • Population reference

Future Developments

Planned Mappings (2026-2027)

  • Anxiety Domain: GAD-7 to/from PROMIS Anxiety to/from HADS-A
  • Fatigue Domain: FACIT-F to/from PROMIS Fatigue to/from FSS
  • Pain Domain: BPI to/from PROMIS Pain to/from NRS

Methodological Improvements

  • Machine learning-based mappings
  • Individual precision estimates
  • Dynamic, population-specific mappings
  • Real-time mapping updates based on new data

Resources and Tools

Online Calculators

Scientific Publications

  • Choi et al. (2014): "Establishing a common metric for depressive symptoms"
  • Schalet et al. (2015): "Linking scores with patient-reported health outcome instruments"

Summary

Cross-instrument mappings are a powerful tool for harmonizing PRO data. The depression domain exemplifies how various established instruments can be mapped onto a common PROMIS metric. Despite inherent limitations, these mappings enable better data integration, comparability, and continuous patient care across different settings.