Author of

• 10984

  +

  P1.22 - Poster Session 1 - Epidemiology, Etiology (ID 166)

  • Event: WCLC 2013
  • Type: Poster Session
  • Track: Prevention & Epidemiology
  • Presentations: 1
  • Moderators:
  • Coordinates: 10/28/2013, 09:30 - 16:30, Exhibit Hall, Ground Level

  • 10995

    +

    P1.22-011 - Assessment of the accuracy and reliability of health related behavioural data obtained from patient-reported surveys (PRS) compared with electronic patient records (EPR) in lung cancer patient population (ID 2948)

    09:30 - 09:30  |  Author(s): C. Brown
    • Abstract

    Background
    Cigarette smoking, alcohol consumption and presence of co morbidities are important factors that affect health status and mortality in patients diagnosed with lung cancer. While the gold standard for presence or absence of co-morbidities is EPR, the gold standard for obtaining accurate data pertaining to health-related behaviours is by PRS. The purpose of this study is to ascertain, whether in the absence of patient self-reported data, health related behavioural data pertaining to cigarette smoking and alcohol consumption abstracted from EPR provides an accurate and reliable surrogate.

    Methods
    731 lung cancer patients completed a PRS pertaining to information on their lifetime tobacco use, alcohol consumption and whether or not they had been diagnosed with certain co-morbid conditions. Relevant smoking, alcohol consumption and co-morbidity data was collected independently from EPR. Kappa coefficient analysis was used to assess the agreement.

    Results
    Results can be seen in Table 1. Ever/never status for smoking showed almost perfect agreement (k=0.95) between PRS and EPR and surpassed all other health behavioural measures and all co-morbidity agreement values. The calculation of pack-years from EPR and PRS showed substantial agreement (k=0.77); However, categorizing the smoking status into current/ former / never, resulted in only moderate agreement (k=0.47). Alcohol ever/ never status agreement was moderate (0.44) with high sensitivity (0.90) but low specificity (0.50). The lung related co-morbidities like emphysema (k=0.41) and chronic bronchitis (k=0.28) showed fair agreement but with substantial missing data through EPR.

    Table 1
    Health Behaviour	N	Missing Data in EPR	Agreement (k)	95% CI (P value)	Se	Sp
    Smoking (E/N)	709	0	0.95	(0.79, 0.89)	0.995	0.94
    Smoking (Pkyrs)*	606	81(11%)	0.77	P<0.0001
    Smoking (C/F/N)**	705	4(0.5%)	0.47	(0.41, 0.51)
    Alcohol (E/N)	575	150(20.5%)	0.44	(0.36, 0.52)	0.9	0.5
    Comorbidity
    Emphysema	589	126(17.2%)	0.41	(0.33, 0.49)	0.41	0.95
    Chronic Bronchitis	601	94(12.8%)	0.28	(0.19, 0.37)	0.39	0.88
    *Spearman correlation coefficient
    **Weighted kappa

    Conclusion
    In the absence of PRS data, EPR provides a reliable surrogate for ever/ never smoking status and moderately reliable for lifetime smoking exposure in this lung cancer population. However current/ former/ never smoking status and ever/ never alcohol status cannot be reliably ascertained from medical records. Missing EPR data related to smoking pack years, alcohol consumption and lung co-morbidities is concerning and suggests more systematic or synoptic reporting by physicians would improve opportunities for research