Informatively missing quality of life and unmet needs sex data for immigrant and Anglo-Australian cancer patients and survivors

Abstract

Although cancer can seriously affect peoples' sexual well-being, survivors and patients may be reluctant to answer questions about sex. This reluctance may be stronger for immigrants. This study aimed to investigate missing sex data rates and predictors of missingness in two large studies on immigrants and Anglo-Australian controls with cancer and to investigate whether those with missing sex data may have worse sexual outcomes than those with complete data. We carried out two studies aimed at describing the quality of life (QoL) and unmet needs amongst Arabic, Chinese and Greek immigrants versus Anglo-Australians cancer survivors (n=596, recruited from cancer registries) and patients (n=845). Logistic regression was used to model the probability of having missing sex data in either of the questionnaires. We compared the mean of the unmet sex needs responses of those who had missing QoL sex data (but not needs) to those who had completed both, and vice versa. Missing sex data rates were as high as 65%, with immigrants more likely to skip sex items than Anglo-Australians (p=0.02 for registry study, p<0.0001 for hospital study). Women, older participants and participants with more advanced disease had increased odds of missingness. There was evidence that data were informatively missing. Additionally, the questionnaire which stated that the sex questions are optional had higher missing data rates. High missing data rates and informatively missing data can lead to biased results. Using the questionnaires that state that they may skip sex items may lead to an underestimation of sexual problems or an overestimation of quality of life.