Skip to main content

Table 2 Performance of Claims-Based Surrogates of Frailty Against Clinical Frailty Assessment in the Health and Retirement Study-Medicare Data

From: Accuracy of diagnosis and health service codes in identifying frailty in Medicare data

Claims-Based Surrogates of Frailtya

Sample Size (%)

Clinical FI Mean (SD)

Clinical Frailty Phenotype

Sensitivity

Specificity

PPV

NPV

Hospital beds and associated supplies

44 (1.4%)

0.49 (0.14)

0.06

0.99

0.73

0.82

Pressure ulcer

39 (1.3%)

0.43 (0.17)

0.04

0.99

0.59

0.82

Cachexia

< 10 (< 0.3%)

0.43 (0.11)

0.01

> 0.99

0.75

0.81

Wheelchairs, components, and accessories

63 (2.6%)

0.41 (0.15)

0.09

0.99

0.63

0.82

Adult failure to thrive

< 10 (< 0.3%)

0.36 (0.17)

0.01

> 0.99

0.57

0.81

Muscle weakness

164 (5.3%)

0.34 (0.16)

0.12

0.96

0.44

0.82

Accessories for oxygen delivery devices

149 (4.8%)

0.34 (0.16)

0.11

0.97

0.44

0.82

Debility

50 (1.6%)

0.34 (0.15)

0.04

0.99

0.48

0.81

Difficulty in walking

135 (4.4%)

0.32 (0.15)

0.10

0.97

0.44

0.82

Walking aids and attachments

130 (4.2%)

0.32 (0.14)

0.10

0.97

0.47

0.82

History of fall

56 (1.8%)

0.31 (0.16)

0.04

0.99

0.41

0.81

Transportation services including ambulance

300 (9.7%)

0.31 (0.16)

0.24

0.93

0.46

0.84

Abnormality of gait

231 (7.5%)

0.31 (0.15)

0.18

0.95

0.45

0.83

Anorexia

15 (0.5%)

0.30 (0.22)

0.01

> 0.99

0.33

0.81

Abnormal loss of weight and underweight

112 (3.6%)

0.28 (0.16)

0.08

0.97

0.44

0.82

Muscular wasting and disuse atrophy

23 (0.7%)

0.28 (0.12)

0.02

0.99

0.39

0.81

Senility without mention of psychosis

< 10 (< 0.3%)

0.26 (0.21)

< 0.01

> 0.99

0.33

0.81

Malaise and fatigue

740 (23.9%)

0.25 (0.15)

0.38

0.79

0.30

0.84

  1. Abbreviations: FI frailty index, HCPCS Healthcare Common Procedure Coding System, ICD International Classification of Diseases, NPV negative predictive value, PPV positive predictive value, SD standard deviation
  2. aClaims-based surrogates of frailty were defined as occurrence of the respective codes in a 12-month period in any of the inpatient, outpatient, skilled nursing facility, home health, carrier, and durable medical equipment datasets