This study shows that the SDAT has acceptable to good internal consistency, as well as intra-rater and inter-rater reliability. Criterion-related and concurrent validity were also in the range considered as substantial [30
]. Finally, the presence of at least one severely unmet need significantly predicted the occurrence of a family meeting to define discharge disposition, even when controlling for depressive symptoms. Overall, the psychometrics properties of the SDAT instrument appear good enough to support further investigation of its predictive validity.
These results deserve several comments. First, the assessment of the SDAT's internal consistency (Cronbach's α and inter-item correlations) showed only moderate correlations. This is not surprising given the concept used to develop this instrument. The spirituality construct underlying its development was explicitly multidimensional, including four distinct dimensions. High correlations between items measuring needs related to these four separate dimensions were therefore not expected. From a more technical stand point, high reliability coefficients would also have been surprising as they depend not only on item homogeneity, but also on their number in the scale, a number limited to five in the SDAT. Nevertheless, item-to-total correlations were highly significant, indicating that each item contributes additional specific information. This last appreciation is further supported by results of the factor analysis that clearly identified two main factors. The first factor (loading on Meaning, Transcendence and Identity spiritual needs) could be interpreted as reflecting the patient's intrinsic inner spirituality, while the second factor (loading on Values needs) would reflect the combined balance between the patient's and the health professional's system of values. Overall, these results show that the SDAT has acceptable psychometrics properties that make it a valid and reliable instrument to assess spiritual distress in older patients hospitalized in post-acute rehabilitation.
The second comment is related to the cut-off used to define spiritual distress. This cut-off was determined according to a clinical definition of spiritual distress. This cut-off is debatable and will probably have to be refined according to further sensitivity analyses of the instrument's predictive validity.
The final comment relates to the specific contribution of the present study to the field of spirituality research. Results from this study provide a preliminary estimate of the prevalence of spiritual distress in older patients hospitalized in post-acute rehabilitation. Overall, these results indirectly raise the question whether spiritual distress could be a neglected problem in these patients. Future studies need to investigate in more details the potential influence of spiritual distress on patients' health outcomes and quality of life. This is a necessary step to determine whether specific interventions targeting spiritual distress should be developed and tested in the future.
This study has some limitations. Test-retest assessment would have been more accurate if the SDAT interviews could have been repeated. However, this option was considered inappropriate because two successive interviews investigating intimate concerns were considered too demanding in this vulnerable population. In addition, the current lack of knowledge on the dynamic of spiritual state when undergoing post-acute rehabilitation would have made it difficult to choose a time window both large enough to avoid recall bias in the interviewer and tight enough to minimize the potential effect of numerous factors that could influence these patients' spiritual state. Alternatively, using a different interviewer would have resulted in investigating inter-rater agreement at the same time, a clearly unsatisfying option. Video-taped interviews and measures of intra-rater reliability were therefore preferred. Nevertheless, only four interviews were videotaped and analyzed by three judges and this limited number of cases limits also the statistical power analysis.
Refusal rate (31.8%) is another limitation of this study. Patients showed some reluctance to participate in the study as most considered they had already answered too many questions on admission. However, comparison of the characteristics of participants and refusers did not suggest differences that could have affected results. Finally, additional limitations concerning the sampling bias exist. Subjects suffering from significant cognitive impairment or considered too ill to complete the interview were excluded from the study. This sampling bias might have influenced the results and underestimated the prevalence of spiritual distress in this population of older hospitalized patients.
This study also has clear strengths. The SDAT underwent an extensive validation process. Most instruments currently available to assess spirituality have not undergone such a rigorous and complete validation process [19
]. Previous work showed good face validity and acceptability [23
]. This study further extends documentation of the SDAT properties by showing its reliability and validity to assess spiritual distress in older hospitalized patients.
Another original contribution of this work is to propose a validated instrument based on a semi-structured interview rather than on a set of closed questions. The SDAT is unique in this regard as it offers the possibility to assess spirituality through an approach that is centered on the patient. Rigorous validation of such semi-structured interviews, as reported here, is uncommon.