Skip to main content
. 2021 Apr 15;21:349. doi: 10.1186/s12913-021-06345-z

Table 5.

Summary of statistical findings for letter content features against GP gradings

Content feature (independent variable) Present in successful letters Present in unsuccessful letters Test Df.* n χ2 value Effect size -phi coefficient P value
Discharging physician (name & role) 86.8% 79.4% χ2 1 375 2.984 −.097 0.084
Reason for admission 99.1% 86.5% χ2 1 375 24.176 −.266 < 0.001
Diagnosis 97.4% 74.5% χ2 1 375 44.386 −.353 < 0.001
Tests/procedures performed 97.0% 74.5% χ2 1 375 41.841 −.343 < 0.001
Investigation results 96.2% 77.3% χ2 1 375 30.194 −.293 < 0.001
Medication names 84.6% 78.7% χ2 1 375 1.719 −.075 0.190
Medication dose & frequency 84.2% 78.7% χ2 1 375 1.437 −.069 0.231
Medication changes 61.5% 48.9% χ2 1 375 5.193 −.123 0.023
Reasons for changes 32.1% 18.4% χ2 1 375 7.606 −.149 0.006
Hospital plan/actions 70.5% 50.4% χ2 1 375 14.475 −.202 < 0.001
Patient concerns/wishes 8.1% 8.5% χ2 1 375 .000 .007 1.000
Information to patient 38.5% 24.8% χ2 1 375 6.764 −.140 0.009
GP plan & actions 69.7% 53.2% χ2 1 375 9.591 −.166 0.002
Acronyms (unexplained) 78.2% 86.5% χ2 1 375 3.482 .103 0.062
Medical jargon (unexplained) 71.8% 84.4% χ2 1 375 7.095 .144 0.008

*Df. degrees of freedom