• The standard forward-backward process of translating an instrument into one or more target languages can be broken down into the following steps: 1) forward translation, 2) reconciliation meeting, 3) back translation (quality control), 4) pilot-test, and 5) final reconciliation. The aim of each step is to add quality to the preceding step in terms of conceptual equivalence between source and target versions, in order to produce a final version. It is common practice for at least two forward translations to be completed and reconciled because of the subjective nature of translation. However, having more than one forward translation tends to be more important when the questionnaire being translated is complex and less of a symptom measure than is the case with the 9-item LCSS patient form. The methodology suggested by the developers of the LCSS, which includes a single forward translation, is considered to be adequate; however, many of the newer languages have received two. Moreover, two independent backward translations were used to complete this standard process prior to pilot testing.
  
  
• The purposes for conducting a pilot test with a newly-translated questionnaire are as follows:
  1) to assess the comprehensibility of the translation, 2) to test any translation alternatives which have not been resolved by the translators, and 3) to identify any items which may be inappropriate at a conceptual level. Each translation for the LCSS was pilot tested with seven in-country patients with lung cancer, using a face-to-face interview format. Respondents were asked to complete the questionnaire while "thinking aloud" to the interviewer about what they took each item to mean. From this process the interviewer can get a sense of the meaning of the item, any misunderstandings that are occurring, and the ease of response. After the respondent completed this process, he/she was asked a series of six open-ended questions related to the LCSS items and responses. Changes were made when needed by the reconciliation panel.    

Policy #1: Changing Existing Translations
Based on the standard processes and pilot testing (outlined above), LCSS Associates, in conjunction with the outcomes research company, Oxford Outcomes, in England, recommend that the current translations be used as they are, with the rationale that these are most closely equivalent to the well-tested and validated original version. Please note that the original version (in American English) was specifically developed in lay clinical terms to reduce patient burden and has been assessed for readability at an early reading level (U.S. 2nd grade reading level). Additionally, it was built on existing tested instruments that in themselves may have had minor flaws. Minor grammatical and regional language differences to date have not changed the meaning of the words or the conceptual understanding in the patient scale. If, indeed, as the new LCSS translations are used, changes need to be considered, then LCSS Associates and Oxford Outcomes will welcome your suggestions.

Two Translated Versions of the LCSS Patient Scale
The Lung Cancer Symptom Scale (LCSS) has two versions for each of the translated patient scales. The original LCSS uses the words "lung cancer" in the instructions as well as in question #7 (related to overall symptomatic distress). Thus, there is a "lung cancer version."

The use of the term "cancer" raises problems in many countries for two reasons: (1) a literal translation of "cancer" is felt to be shocking and is considered almost a taboo word; and (2) in some settings, patients may not be fully informed about the nature of their cancer, and the malignancy may be referred to as a "lung illness." Thus, there is a "lung illness version" of the LCSS.

It is important that the patient be comfortable in completing the scale, especially in trials using repeated measures. For those patients who have been informed of their condition, to use a vague description of their illness will cause confusion and distraction. Similarly, to use a term that neither the patient nor oncologist has used in discussing the illness will cause distress and distraction. Any confusion or distraction could affect the accuracy of this self-report quality of life measure by patients.

A chart has been provided that delineates which of the two versions was pilot tested in each country and the rationale for this choice (see Table 1). The intention is to provide a document to generate discussion and a decision by consensus of the investigators as to which version should be used in the clinical trial in that country. 

Policy #2: Choice of Translated Version
Quality of Life Research Associates, in conjunction with the outcomes research company, Oxford Outcomes, in England, believe that the difference between the two versions is minimal in terms of their effect on the psychometric properties of reliability and validity. 

While our group would prefer the use of the "lung cancer version," we recognize that there is controversy surrounding this issue. It is not our position to decide what is necessary culturally. But, to ensure scientific rigor, a decision needs to be made for which version to use per country (not per treatment site or per patient). We would recommend that for a particular trial, the same version be used at all sites within a country.

We further recommend that if in doubt as to which version to use, it is best to use the version that was pilot tested in the country of interest. Because of the likelihood that more patients will be educated about their exact disease in the future, it is expected that there will be less use of the lung illness version.

Quality of Life Research Associates and Oxford Outcomes (Revised 09/26/2000)

Table 1. LCSS Versions (Lung Cancer vs. Lung Illness)
YES = version that was pilot tested