Strengths and limitations
Most of the existing studies that compare services use across diagnostic groups rely on chart reviews, which cannot show contacts with multiple health services. Studies of patients who had appointments during the pandemic are limited, since they exclude those who avoid hospitals out of fear of COVID-19. To avoid those limitations, we selected a sample who contacted our clinic before the pandemic (when no pandemic awareness existed) and assessed them about one year later. We used official ICD-10 diagnoses of the patients enrolled in the study, which are more reliable than self-reports of patients on their diagnoses.
Collecting self-report data using online questionnaires is valuable when the sample is the general population. When the target is mental health patients, however, interviews have clear advantages over self-reported data. Some patients may not give reliable information because of their psychopathology (fears, lack of trust, delusions, etc.); clinician’s judgment may be superior to self-reported data in such cases. On the other hand, interviews are time-consuming and costly, since they require specialist/clinician time. Our advantage was that we were able to recruit most of our experienced residents as interviewers. Nevertheless, having 12 interviewers would necessarily increase inter-rater variability, the effect of which we did not analyze. Our sample, on the other hand, was large enough to conduct essential statistical analyses.
Although we used the official ICD-10 codes in the patient charts, we were not able to include all psychiatric diagnoses; and we needed to regroup diagnoses for ease of analysis. We also had to use judgement (a hierarchy) in grouping comorbid cases, which may have resulted in bias (probably a type II bias). The differences observed between the four diagnostic groups would probably be more pronounced, had we compared non-comorbid (i.e., pure) diagnostic groups. We had to drop 68 patients that belonged to the “other diagnosis” category, since this group included 26 drug dependence cases, and several other diagnoses, which were too small to be meaningfully analyzed. Our assessments of psychiatric symptom worsening, COVID-19 related anxiety, PGI or CGI did not specify exact time periods; rather our questions were worded as “during the pandemic…”, which referred to the past 10-12 months.