Vickybot, a Chatbot for Anxiety-Depressive Symptoms and Work-Related Burnout in Primary Care and Health Care Professionals: Development, Feasibility, and Potential Effectiveness Studies.

BACKGROUND: Many people attending primary care (PC) have anxiety-depressive symptoms and work-related burnout compounded by a lack of resources to meet their needs. The COVID-19 pandemic has exacerbated this problem, and digital tools have been proposed as a solution. OBJECTIVE: We aimed to present the development, feasibility, and potential effectiveness of Vickybot, a chatbot aimed at screening, monitoring, and reducing anxiety-depressive symptoms and work-related burnout, and detecting suicide risk in patients from PC and health care workers. METHODS: Healthy controls (HCs) tested Vickybot for reliability. For the simulation study, HCs used Vickybot for 2 weeks to simulate different clinical situations. For feasibility and effectiveness study, people consulting PC or health care workers with mental health problems used Vickybot for 1 month. Self-assessments for anxiety (Generalized Anxiety Disorder 7-item) and depression (Patient Health Questionnaire-9) symptoms and work-related burnout (based on the Maslach Burnout Inventory) were administered at baseline and every 2 weeks. Feasibility was determined from both subjective and objective user-engagement indicators (UEIs). Potential effectiveness was measured using paired 2-tailed t tests or Wilcoxon signed-rank test for changes in self-assessment scores. RESULTS: Overall, 40 HCs tested Vickybot simultaneously, and the data were reliably transmitted and registered. For simulation, 17 HCs (n=13, 76% female; mean age 36.5, SD 9.7 years) received 98.8% of the expected modules. Suicidal alerts were received correctly. For the feasibility and potential effectiveness study, 34 patients (15 from PC and 19 health care workers; 76% [26/34] female; mean age 35.3, SD 10.1 years) completed the first self-assessments, with 100% (34/34) presenting anxiety symptoms, 94% (32/34) depressive symptoms, and 65% (22/34) work-related burnout. In addition, 27% (9/34) of patients completed the second self-assessment after 2 weeks of use. No significant differences were found between the first and second self-assessments for anxiety (t8=1.000; P=.34) or depressive (t8=0.40; P=.70) symptoms. However, work-related burnout scores were moderately reduced (z=-2.07, P=.04, r=0.32). There was a nonsignificant trend toward a greater reduction in anxiety-depressive symptoms and work-related burnout with greater use of the chatbot. Furthermore, 9% (3/34) of patients activated the suicide alert, and the research team promptly intervened with successful outcomes. Vickybot showed high subjective UEI (acceptability, usability, and satisfaction), but low objective UEI (completion, adherence, compliance, and engagement). Vickybot was moderately feasible. CONCLUSIONS: The chatbot was useful in screening for the presence and severity of anxiety and depressive symptoms, and for detecting suicidal risk. Potential effectiveness was shown to reduce work-related burnout but not anxiety or depressive symptoms. Subjective perceptions of use contrasted with low objective-use metrics. Our results are promising but suggest the need to adapt and enhance the smartphone-based solution to improve engagement. A consensus on how to report UEIs and validate digital solutions, particularly for chatbots, is required.

Prognostic factors in Spanish COVID-19 patients: A case series from Barcelona.

BACKGROUND: In addition to the lack of COVID-19 diagnostic tests for the whole Spanish population, the current strategy is to identify the disease early to limit contagion in the community. AIM: To determine clinical factors of a poor prognosis in patients with COVID-19 infection. DESIGN AND SETTING: Descriptive, observational, retrospective study in three primary healthcare centres with an assigned population of 100,000. METHOD: Examination of the medical records of patients with COVID-19 infections confirmed by polymerase chain reaction. Logistic multivariate regression models adjusted for age and sex were constructed to analyse independent predictive factors associated with death, ICU admission and hospitalization. RESULTS: We included 322 patients (mean age 56.7 years, 50% female, 115 (35.7%) aged ≥ 65 years): 123 (38.2) were health workers (doctors, nurses, auxiliaries). Predictors of ICU admission or death were greater age (OR = 1.05; 95%CI = 1.03 to 1.07), male sex (OR = 2.94; 95%CI = 1.55 to 5.82), autoimmune disease (OR = 2.82; 95%CI = 1.00 to 7.84), bilateral pulmonary infiltrates (OR = 2.86; 95%CI = 1.41 to 6.13), elevated lactate-dehydrogenase (OR = 2.85; 95%CI = 1.28 to 6.90), elevated D-dimer (OR = 2.85; 95%CI = 1.22 to 6.98) and elevated C-reactive protein (OR = 2.38; 95%CI = 1.22 to 4.68). Myalgia or arthralgia (OR = 0.31; 95%CI = 0.12 to 0.70) was protective factor against ICU admission and death. Predictors of hospitalization were chills (OR = 5.66; 95%CI = 1.68 to 23.49), fever (OR = 3.33; 95%CI = 1.89 to 5.96), dyspnoea (OR = 2.92; 95%CI = 1.62 to 5.42), depression (OR = 6.06; 95%CI = 1.54 to 40.42), lymphopenia (OR = 3.48; 95%CI = 1.67 to 7.40) and elevated C-reactive protein (OR = 3.27; 95%CI = 1.59 to 7.18). Anosmia (OR = 0.42; 95%CI = 0.19 to 0.90) was the only significant protective factor for hospitalization after adjusting for age and sex. CONCLUSION: Determining the clinical, biological and radiological characteristics of patients with suspected COVID-19 infection will be key to early treatment and isolation and the tracing of contacts.