Emotional reactions to infertility diagnosis: thematic and natural language processing analyses of the 1000 Dreams survey.

ABSTRACT

RESEARCH QUESTION What are the emotional effects of infertility on patients, partners, or both, and how can qualitative thematic analyses and natural language processing (NLP) help evaluate textual data? DESIGN: A cross-sectional, multi-country survey conducted between March 2019 and May 2019. A total of 1944 patients, partners, or both, from nine countries responded to the open-ended question asking about their initial feelings related to an infertility diagnosis. A mixed-method approach that integrated NLP topic modelling and thematic analyses was used to analyse responses. Sentiment polarity was quantified for each response. Linear regression evaluated the association between patient characteristics and sentiment negativity. RESULTS: Common emotional reactions to infertility diagnoses were sadness, depression, stress, disappointment, anxiety, frustration, confusion and loss of self-confidence. NLP topic modelling found additional reactions, i.e. shared feelings with partners, recollections about causes of infertility and treatment experience. Responses to the open-ended question were brief (median three words) with 71.8% conveying negative sentiments. Some respondent characteristics showed small but significant associations with sentiment negativity, i.e. country (Spain, China and France were more negative than the USA, P < 0.001, P < 0.003 and P < 0.009 respectively), treatment engagement (no treatment was more negative than one or more treatment, P = 0.027) and marital status (missing/other was more negative than divorced, P = 0.003). CONCLUSION: Infertility diagnoses create an emotional burden for patients and partners. The mixed-method approach provides a compelling synergy in support of the validity of these findings and shows potential for these techniques in future research.