Health Condition 1: F411- Generalized anxiety disorder Health Condition 2: F410- Panic disorder [episodic paroxysmal anxiety] Health Condition 3: F320- Major depressive disorder, singleepisode, mild Health Condition 4: F321- Major depressive disorder, singleepisode, moderate Health Condition 5: F322- Major depressive disorder, singleepisode, severe without psychotic features Health Condition 6: F329- Major depressive disorder, singleepisode, unspecified Health Condition 7: F328- Other depressive
Conditions
Interventions
Intervention1: Nil: Nil
Sponsors
SEARCH (Society for Education, Action and Research in Community Health)
Eligibility
Inclusion criteria
Inclusion criteria: Inclusion Criteria at the level of the patients is the following 1. Patients of age greater than 18 years presenting as new patients in the OPD 2. Able to and willing to provide informed consent and can communicate in either Hindi or Marathi Inclusion criteria at the level of Users is the following 1. Above 18 years of age with a bachelors degree in any discipline fluency in Marathi and Hindi and able to and willing to provide informed consent Inclusion criteria at the level of Data is the following 1. All records from patients meeting the patient level inclusion criteria and completing the full set of questionnaires PHQ 2 GAD 2 PHQ 9 GAD 7 PADIS without system interruption
Exclusion criteria
Exclusion criteria: Exclusion criteria at the level of the patient is following 1. Patients exhibiting symptoms of mania or psychosis conditions the AI has not been validated on. 2. Patients with severe cognitive impairment or sensory deficits example hearing loss that make them unable to interact with the user operating the chatbot. Exclusion criteria at the level of users is as following 1. Lack of basic digital proficiency never used WhatsApp and not willing not able to learn it in a brief training session 2. Visual motor or cognitive impairments affecting device use 3. Prior extensive involvement in developing the AI tool 4. Prior non compliance risk, examples are History of low adherence to study procedures or clinical guidelines in prior research. 5. Unwillingness to follow the study s data-entry, reporting, or oversight requirements as observed prospectively too 6. Any other condition or circumstance that, in the investigator s judgment, would compromise safe, reliable, or unbiased use of the decision-support system Exclusion criteria at the level of data are as following 1. Any session with incomplete responses skipped questions or premature exit. Any single missing response in a required questionnaire PHQ 2 GAD 2 PHQ 9 GAD 7 PADIS will be excluded as AI requires all the responses to make a decision. 2. Technical errors at the level of software, examples are interface timeouts and message delivery failures. 3. Incorrect information given at any level by the patient or an error in input of data by the user, which is realized by the user/patient during the session. 4. Technical errors due to the user device or connection, examples are low battery, poor network. 5. User not remembering to run the chatbot flow on the patients.
Design outcomes
Primary
| Measure | Time frame |
|---|---|
| Our primary outcomes broadly consist of the following: Accuracy of the mental health care service provider-AI system in comparison to the psychiatrist. This will be analyzed along the 4 broad headings: Screening, Diagnosis, Psychoeducation, and Treatment. The agreement between AI and the clinician will be analysed separately for 4 main headings: i.Screening (Binary) ii.Diagnosis (Categorical) iii.Psychoeducation (Content headings- Binary, Quality ordinal score from 1-10 given by the expert) iv.Pharmacotherapy (at the level of class of drug, specific drug used, and the dose). A composite treatment utility score will be calculated. Timepoint: The output of AI and the psychiatrist will be recorded immediately after the patient encounter. The composite primary outcome will be calculated at the end of the study. | — |
Secondary
| Measure | Time frame |
|---|---|
| We have selected 4 secondary outcome measures: i. Gross disagreement of the user with the AI system: ii. Feasibility factors. Feasibility will be assessed through four domains using validated, widely cited instruments: Workflow integration: NoMAD (Normalization Measure Development) Usability: System Usability Scale (SUS) Acceptability & Appropriateness: Acceptability of Intervention Measure (AIM) + Intervention Appropriateness Measure (IAM) Cognitive workload: NASA Task Load Index (NASA-TLX) Aglobal acceptability question and a global usability question. iii. Learning curve: Time to completion, Error-rate trends, and help requests and interruptions. iv. Safety and errors: Algorithm errors, Malfunctions of software/hardware, User errors. Timepoint: Overall secondary outcomes will be measured at the end of the study. Specific measurement time for some of the parameters are as follows: A. Workflow Integration (NoMAD) - Mid and End of study B. Usability (SUS) - After the first 5-10 patient encounters C. Cognitive Workload (NASA-TLX) - After the first 5-10 patient encounters D. Acceptability and Appropriateness(AIM plus IAM)- Mid-study | — |
Countries
India
Contacts
Public ContactDr Ojas Krishnani
SEARCH(Society for Education, Action and Research in Community Health)
Outcome results
None listed