Skip to main content

Table 2 Representativeness of the different topic models per category

From: Analyzing patient experiences using natural language processing: development and validation of the artificial intelligence patient reported experience measure (AI-PREM)

Question

Positive categories in total

Per topic

Negative categories in total

Per topic

Q1

94.4% (n = 72)

T1: 100% (n = 36)

T2: 88.9% (n = 36)

55.6% (n = 18)

T1: 60% (n = 10)

T2: 50% (n = 8)

Q2

93.3% (n = 75)

T1: 97.1% (n = 35)

T2: 100% (n = 10)

T3: 85% (n = 20)

T4: 90% (n = 10)

71% (n = 31)

T1: 100% (n = 3)

T2: 100% (n = 3)

T3: 83.3% (n = 6)

T4: 100% (n = 3)

T5: 75% (n = 4)

T6: 28.6% (n = 7)

T7: 60% (n = 5)

Q3

98.4% (n = 63)

T1: 100% (n = 43)

T2: 95% (n = 20)

76.9% (n = 39)

T1: 100% (n = 4)

T2: 33.3% (n = 3)

T3: 85.7% (n = 7)

T4: 100% (n = 5)

T5: 66.7% (n = 3)

T6: 77.8% (n = 9)

T7: 62.5% (n = 8)

Q4

100% (n = 65)

T1: 100% (n = 41)

T2: 100% (n = 12)

T3: 100% (n = 12)

86.7% (n = 15)

T1: 100% (n = 5)

T2: 80% (n = 10)

Q5

86.2% (n = 29)

T1: 85.7% (n = 21)

T2: 87.5% (n = 8)

55.5% (n = 20)

T1: 50% (n = 10)

T2: 60% (n = 10)

  1. Representativeness is defined as the number of texts within a certain topic that fit the description of the topic. The percentage is calculated by dividing the texts that fit the description of the topic by the total number of texts within the topic. Q: AI-PREM question. T: automatically extracted topic