Validity and reliability of the Amharic version of supportive care needs survey - short form 34 among cancer patients in Ethiopia.

Supportive care needs survey short form has a total of 34 items that have 5 domains that measure the unmet needs of cancer patients. It is important to validate this tool since there are differences in culture, geographic areas, and clinical care service which influence patients' needs. Therefore, this study aimed to assess the construct validity and reliability of the tool. The study was conducted among 170 cancer patients from April 1st to 30th 2019 in Hawassa hospital, South Ethiopia. Confirmatory factor analysis was done using fit indices. Convergent and discriminant validity was evaluated using average variance extracted and maximum shared variance respectively. Known group validity was checked using the Mann-Whitney U test. The reliability of the instrument was examined using Cronbach's alpha. Domains except for health system and information, and patient care and support maintained convergent and divergent validity. The remaining validity was maintained after removing items that were redundant and double loading. The average variance extracted of domains varied from 0.52-0.81. The Square of correlation between constructs was lower than the average variance extracted for the constructs. The tool had reliability r = 0.932. The root mean square error of approximation was 0.057, comparative fit index 0.954, and the other fit indices were also indicating a good fit. Known groups difference was seen by age and type of treatment taken across the different domains. After the health system and information, and patient care, and support domain validity issues were corrected by removing 8 items, the reduced tool was found to be a valid and reliable tool. The validated tool will be valuable if included in routine cancer care in our clinical settings.