Development of a Thai version of the neutropenia subscale for the Functional Assessment of Cancer Therapy–Neutropenia questionnaire

Our team translation consisted of two oncology nurses (the first and the second investigators), one hematologist, four English instructors from the Language Institute and the Faculty of Mass Communication, Chulalongkorn University (one of the instructors is a native English speaker who can read and write Thai fluently), and one professional translator.

The 19-item NS was translated according to the FACIT translation methodology guidelines [13]. The translation process ensured equivalence with the English version and cultural appropriateness. A rigorous multistep translation/back-translation process was employed. First, two independent forward translations from English to Thai were prepared by two native Thai speakers (a senior English instructor and a professional translator). Then, a hematologist formulated a single translated version by combining the two forward translations. A back-translation of the combined version into English was performed by the English instructor, who is a native English speaker. The back-translation was reviewed by people from the FACIT organization and our team (one English instructor and two oncology nurses), and following this, two English instructors proofread the final Thai version. The FACIT staff then formatted the test versions into Word documents. We performed a pilot testing with ten patients having hematological malignancies and receiving chemotherapy. The ten participants were asked to read and answer the Thai version of the questionnaire and were questioned on its conceptual clarity. We also solicited feedback and comments to refine the translation further.

Once this 19-item Thai version of the NS was approved, we obtained a copyright for it from FACIT.org for use in the main study.

To exam psychometric properties of the Thai version NS, 260 patient participants were randomly selected from four hospitals in Thailand: King Chulalongkorn Memorial Hospital, Siriraj Hospital,

Narasual Hospital, and Sapphasitthiprasong Hospital. The participants were aged 18 years or older, diagnosed with lymphoma, acute myeloid leukemia, or multiple myeloma, and physically and mentally capable of completing the self-administered questionnaire in Thai. Those with active infection or progressive disease were excluded. All participants provided written informed consent and completed the questionnaires at the time of expected neutropenia (7–14 days after receiving chemotherapy). This happened either at the hospital, while awaiting follow up with the hematologist, or at home, while awaiting the subsequent cycle of chemotherapy. For those who completed the questionnaires at home the participants returned them by post.

The instruments used in this study were the personal data form, medical record form, and FACT-N. The personal data form recorded age, sex, marital status, and education, while the medical record form recorded details of the disease type, performance status, and types of chemotherapy received. FACT-N consists of the 27 items of the FACT-G questionnaire and 19 items of the NS. The details of FACT-N have already been described.

Demographic characteristics of the participants were analyzed by descriptive statistics. Content and construct validity and reliability were used to confirm the psychometric properties of the Thai NS.

Content validity was evaluated by a panel of 5 experts. The purpose of this method is to examine the relevancy of items for measuring a concept or construct [14]. The panel of experts was composed of one hematologist, and 4 nurses. All of the nurses have a PhD in nursing, and have more than 10 years of experience in work related to cancer patients. The experts were asked to rate the relevancy between the original and the Thai version of each item using three point rating scale: 1 (irrelevant), 2 (questionably relevant), and 3 (relevant). The content validity index (CVI) was computed as the proportion of either 2 or 3 on the scale by all rates. A CVI of 0.78 to 1.0 was considered acceptable to justify relevance between the original English and Thai versions of the NS [15].

Construct validity was examined by confirmatory factor analysis (CFA) using the LISREL statistical software package (version 8.72). Nineteen items that were identified in the analyses of the English version were used in the model [8]. Multiple criteria were employed for evaluating absolute fit indices of the model. These included non-significant P-values on a χ² test, a goodness-of-fit index (GFI) greater than 0.95, a comparative fit index (CFI) > 0.95, and a root mean square error of approximation (RMSEA) of less than 0.08 [16].

The reliability of the Thai NS and its three dimensions (as per the original English version) were examined for internal consistency using Cronbach’s α coefficients. Coefficient values ≥0.70 were considered acceptable reliability coefficients [17].

The patients during pilot testing reported that they understood the translation and that the items in the translated version were easy to score. Among the 19 items, three had feedback comments. Single unmarried participants found it difficult to answer items N7 and N8 (which refer to a spouse or partner); however, they answered these items with a score of 0, which is the correct response. Some participants preferred that the more specific term of “white blood cells” be used instead of “blood cells” in item N1 (which referred to white blood cells); however, they were able to understand the translation. In consensus with the team, the term “blood cells” was retained in order to maintain accuracy of translation from the English to the Thai version. Thus, the newly formulated Thai version of the 19-item NS was approved without further changes following the pilot study.

Among the 260 patients who participated in this study, 57.7% were male, their mean age of all participants was 45.7 years (SD = 16.54). Over half (56.5%) were married, and a major portion (31.2%) had only elementary school education. The majority of the patients had a diagnosis of non-Hodgkin lymphoma (56.5%), a performance status of “some symptoms, but did not require bed rest during awake hours” (60.8%). Cyclophosphamide, hydroxydaunorubicin, Oncovin (vincristine), prednisone or prednisolone (CHOP) or rituximab and CHOP (R-CHOP) was the treatment regimen (38.8%) most commonly encountered (Tables 1 and 2).