Responses to the Comments on “Caregiver Burden and Disability in Somatoform Disorder: An Exploratory Study”

We thank Andrade et al. for their interest and critical appraisal of our published article. ¹ They have raised an important issue: the unexpected finding of comparable disability between somatoform disorder and schizophrenia. We agree with the explanations provided by the authors, which we also had pointed out; the study setting would have naturally biased the sample toward a more severely ill sample of patients with somatoform disorder, as is the case with all facility-based studies. However, their assertion that “almost all patients with schizophrenia would receive inpatient care, regardless of illness severity” may not be entirely accurate; indeed, a systematic investigation into hospital admission incidence rates for schizophrenia showed a decline in admission incidence in the current century compared to historical data. Moreover, recent data from the National Mental Health Survey of India (2016) also points to a treatment gap of 84.5% for mental morbidity, indicating that four in five individuals with mental illness do not receive appropriate treatment. ² Further, duration of illness correlates with disability levels in somatoform disorder. ³ Among our samples, the mean (±SD) duration of illness among patients with somatoform disorders (71.07 ± 82.09 months) was greater than in those with schizophrenia (61.07 ± 70.09 months). This could have also contributed to the findings.

We used the Clinical Global Impression-Severity scale to assess the severity of illness among participants. Although the paper did not mention that consecutive patients were recruited (an error of omission), Table 1 showed that severely ill patients were more in the schizophrenia group (42.9%) compared to the somatoform disorders group (28.6%); this difference, however, did not reach statistical significance. Because the study setting is a tertiary care center, the possibility that more severely ill patients were recruited applies to both groups. These findings and the possible reasons were indeed discussed in the paper. With regard to sample size estimation, the actual effect size (ES) for patient disability noted in the sample was 0.07; with this small an ES, a larger sample powered to detect a small ES (0.2) might also have been inadequate. Being employed has shown little correlation with caregiver burden in schizophrenia in prior studies from India and other low- and middle-income countries^4,5; hence, we do not regard these as incongruous findings.

Another issue raised pertains to multiple hypothesis testing and the lack of statistical correction for it. The exploratory nature of the study is indicated in the title itself. Further, this limitation was also acknowledged in the paper. In this context, we would like to point out that exploratory studies have their own place in research, particularly in fields with relatively few data such as the one we studied. Exploratory studies are hypothesis generating and help to accumulate evidence needed to evolve more robustly articulated research questions, support the case for effectiveness studies, and guide full-scale evaluations of interventions. ⁶ For the same reason, they should be more concerned about sensitivity (in this case, detecting variables that may be potentially useful for further testing) than specificity (excluding variables that may eventually prove useless in trials). ⁷ Given these considerations, exploratory studies often require a flexible approach for design and analysis, and multiplicity should not be seen as a problem as long as the “significant” results arising from the exploratory analyses are clearly labeled as exploratory findings, ⁸ which is what we had done. The reader can either take them or leave them. However, we do agree that our conclusions could have been worded more conservatively to highlight these nuances and thank the authors for bringing this to readers’ attention.