The authors developed a new instrument to measure HRQOL. The BSIqol was designed to be a generic questionnaire and was validated in a sample of cancer patients. The great advantage of the proposed instrument is its likely clinical usefulness since it was answered quickly and without difficulty by the interviewees, even in a population having low educational level where approximately 60% had less than four years of formal schooling and 10% were illiterate.
The BCH is exclusively dedicated to oncology and serves a needy population in terms of financial resources, originated from all corners of Brazil. The development of BSIqol comes to fill a gap in clinical practice observed by the authors of this research, where many patients have difficulty answering instruments that assess “patient-reported outcomes”, even instruments that have been appropriately validated in the country. This is probably because the population of low socioeconomic level treated in the hospital does not reflect the populations subjected to validation studies often conducted around the country.
Considering HRQOL within a multidimensional concept, the instruments proposed for its measurement must assess at least the physical, emotional, functional and social aspects. To this end, the tools developed are mostly long and sometimes complex, or too stressful for some individuals.
The main advantage in measuring HRQOL is the observation of clinical benefits on the treatment from the patient’s own perspective. In oncology, the two most widely used instruments in research are EORTC QLQ-C30 and FACT-G. The EORTC QLQ-C30, comprised of 30 items, and FACT-G, with 28 items, are usually answered within 5 to 10 min. The BSIqol was answered in a median time of <2 min, which was very relevant, considering the low socioeconomic level of the population assessed in this study.
The search for instruments that would measure the HRQOL with few items is not new[17–19]. Some instruments with few items proved to be valid in some specific situations; however, they are still rarely used[20–22]. Data in the literature suggest that short questionnaires have higher response rate (when mailed to patients) and lower rates of unanswered items, compared with long questionnaires[23, 24]. Moreover, short questionnaires would possibly be less of a burden to patients and would facilitate the operational logistics involved in health services.
The reduction in items from longer questionnaires in order to make them shorter has proved to be feasible in previous studies, and the short instruments did not lose their psychometric characteristics[26, 27]. One instrument worth mentioning is the Quick-FLIC, with 11 items, whose total score showed good correlation with the EORTC QLQ-C30 global health, with correlation coefficients ranging from 0.71 to 0.77. The global score of the Quick-FLIC may vary from 0 to 100 and it was more clinically relevant than the scores of individual items. Similarly, the BSIqol global score correlated with the EORTC QLQ-C30 global health with a high correlation coefficient (r = 0.76). As with the Quick-FLIC global score, we believe the global score is the most useful index extracted from BSIqol for future studies and for use in clinical practice.
Some studies suggest that social support may interfere with HRQOL[29, 30]. The BSIqol social domain specifically assessed interpersonal relationships based on the theoretical premise that a good relationship with people considered important to patients would be associated with good social support. The BSIqol social correlated significantly with the social functional domain of the EORTC QLQ-C30, however, the correlation coefficient was found to be low (r = 0.223). This can be explained in that the EORTC QLQ-C30 assesses the social impact on HRQOL caused by cancer or its treatment; the BSIqol, however, upon assessing the interpersonal relationships, does not associate them with the disease or its treatment. The authors believe that problems in interpersonal relationships may impact HRQOL regardless whatever its cause may be. Interestingly enough, we observed that the BSIqol social scale correlated with the financial difficulties subscale and also correlated negatively with the family income of patients. This finding points to the association between problems in interpersonal relationships and financial difficulties. Financial difficulties secondary to cancer are common and may interfere with HRQOL.
The BSIqol emotional showed high correlation coefficient with EORTC QLQ-C30 emotional functioning domain (r = 0.775) and also with the ESAS-emotional (r = −0.593). Similarly, the BSIqol physical scale correlated with the physical functioning (r = 0.669), role functioning (r = 0.715), global symptoms (r = −0.701) and the ESAS-physical (r = −0.699). These correlations point to an adequate validation of the physical and emotional BSIqol domains.
In spite of the adequate correlation coefficients (r > 0.5) displayed by BSIqol functional when correlated with the functional domain of EORTC QLQ-C30, the BSIqol global showed even higher rates. However, in order to validate the BSIqol functional, we found the highest correlation with ECOG-PS when we correlated the functional performance with BSIqol functional (r = −0.714). Furthermore, functional BSIqol scores were significantly higher in patients who maintained work activity, in relation to those inactive at work. In this analysis, the BSIqol functional was the domain most associated with work activity.
In addition to the color concept, the BSIqol uses short and simple sentences, with large letters. Each item is shown on a different page, except for items 5 and 6 (functional domain), shown on a single page because they are interrelated with one another. The authors believe that such “layout” facilitates the understanding of patients, since only one of the respondents reported some difficulty to understand it. The Moorehead-Ardelt Quality of Life Questionnaire is an instrument originally developed to measure HRQOL in obese patients after surgery. This questionnaire uses symbols and colors that facilitate the understanding of patients. It is believed that such peculiarities made the instrument have a wide spread use in certain corners of the world due to its convenience and ease of being understood. In the present study, we chose the colors red, yellow and green, because they are related to traffic signs, possibly being considered as a universal language.
This preliminary study has some limitations. One of them is that the questionnaire responsiveness of was not evaluated. Another limitation is the lack of data regarding the minimum difference in scores that are clinically significant to the patient. While this study is preliminary in nature, further studies are necessary to clarify the mentioned limitations.
We believe that the major contribution posed by this study is to show the feasibility as for the development of very short instruments with good reliability. Furthermore, the characteristics of the instrument (use of colors, one item per page, several words/phrases showing the same domain, etc.) could be employed as a model for future instruments.