Influence of Selection Bias in Survey Studies Derived From a Patient-Focused Organization: A Comparison of Response Data From a Single Tertiary Care Center and the Acoustic Neuroma Association.

BACKGROUND

The Acoustic Neuroma Association (ANA) is a national, nonprofit organization, focused on the education and support of patients with vestibular schwannoma (VS). The aim of the present study is to characterize the profile of ANA survey respondents and compare them with non-ANA patients evaluated at a single tertiary academic referral center to investigate the potential influence of selection bias.

METHODS

A prospectively maintained VS quality-of-life (QOL) database, comprised of patients evaluated at the authors' center and members of the ANA, was queried. Demographic variables, patient-reported symptoms and tumor characteristics, as well as patient-reported outcome scores were captured. Health-related QOL was evaluated using the disease-specific Penn Acoustic Neuroma QOL (PANQOL) questionnaire. Multivariable regression models were fitted for PANQOL domain and total scores as well as satisfaction with treatment adjusting for baseline demographics, symptoms, and PANQOL scores.

RESULTS

A total of 1,060 patients (802 [76%] ANA respondents) were analyzed. Overall, ANA patients were slightly younger (mean age: 59 vs 60 yr, p = 0.145), more likely to be women (72 vs 55%, p < 0.001), and had a larger tumor size (overall p < 0.001). Furthermore, a significantly higher proportion of ANA patients were more likely to undergo microsurgery (57 vs 21%) or radiation (21 vs 8%) and less likely to be managed with observation (16 vs 65%, overall p < 0.001). A significantly higher proportion of ANA patients reported hearing loss (95 vs 88%, p < 0.001), tinnitus (80 vs 73%, p = 0.034), dizziness (78 vs 64%, p < 0.001), headache (56 vs 45% p = 0.003), and facial paralysis (37 vs 12%, p < 0.001). On multivariable analysis, ANA respondents exhibited significantly lower PANQOL scores for hearing (OR: 0.47, 95% CI: 0.35-0.64, p < 0.001), balance (OR: 0.51, 95% CI: 0.38-0.70, p < 0.001), pain (OR: 0.63, 95% CI: 0.46-0.86, p = 0.004), facial function (OR: 0.58, 95% CI: 0.42-0.80, p = 0.001), energy (OR: 0.44, 95% CI: 0.32-0.59, p < 0.001), anxiety (OR: 0.54, 95% CI: 0.40-0.74, p < 0.001), general (OR: 0.72, 95% CI: 0.53-0.98, p = 0.03), and total QOL (OR: 0.40, 95% CI: 0.30-0.55, p < 0.001). No statistically significant difference was seen with regard to treatment satisfaction.To determine the true clinical relevance of these differences, the two groups were compared using the minimal clinically important difference (MCID) for each domain. MCID is defined as the smallest difference in score in the domain of interest that patients perceive as important, either beneficial or harmful. The domains for hearing, balance, energy, anxiety, and total QOL reached their respective MCID thresholds, indicating that the ANA cohort has QOL scores that are clinically, perceptually worse for these domains compared to the non-ANA group.

CONCLUSION

These data help delineate some of the inherent limitations and biases associated with survey studies incorporating data from national patient support organizations. The population profile of ANA survey respondents likely differs significantly from the greater population of patients with VS that may be encountered at a tertiary referral center.