Table 3:
Challenges and solutions in adapting cascade analysis tool (CAT)
| Major modification domains | mPCAT | PedCAT | FPCAT | MHCAT | CCS-CAT | HCAT | NCAT |
|---|---|---|---|---|---|---|---|
| 1) Time periods for data aggregation and analysis | Problem: Small counts make monthly cascades difficult because of “zero” cells; highly variable monthly counts lead to implausible proportions using cross-sectional approach Solution: Cascade covers 3 month long period to increase overall counts, cascade uses cohort-based population to retain relevant denominators and plausible proportions |
Problem: Clients receiving naloxone have unstructured visit schedule that could result in repeat visits within a month to syringe services; however, naloxone use is infrequent Solution: Naloxone distribution cascade uses one week long period; naloxone use cascade uses 3 month long period |
|||||
| 2) Use of cross-sectional vs. cohort-based populations | |||||||
| 3) Handling low patient volumes | Problem: Small counts in naloxone use cascade make monthly cascades difficult because of “zero” cells Solution: Naloxone use covers 3 month long period to increase overall counts |
||||||
| 4) Analyzing a full catchment area vs. patients presenting for care | Problem: Catchment area population undefined Solution: Cascade begins with population presenting to health center |
Problem: Catchment area population undefined Solution: Cascade begins with population presenting to syringe service programs |
|||||
| 5) Estimating denominators | Problem: Estimating denominator for viral load monitoring not feasible directly Solution: Utilized pseudo-denominators from simple mathematical formulas |
Problem: Clients returning for repeat visits may have previously received service Solution: Include only new clients to avoid repeat testing challenges |
|||||
| 6) Managing appropriate exit and re-entry of groups from cascade | Problem: Heterogeneity between sites in assessing HIV testing eligibility Solution: Included as optional step in cascade to accommodate sites with and without step |
Problem: Not all clients are eligible for HIV testing, which depends on risk profile and recentness of testing Solution: Included appropriate exit for individuals who did not require testing at time point |
Problem: Clients have two streams of services within a similar visit structure (either requiring provision of naloxone or not) and all contribute to final step of cascade of appropriate possession Solution: Cascade includes appropriate exit and re-entry steps in programming to reflect holistic end step in cascade |
||||
| 7) Integrating with electronic health records or mobile platforms | Problem: Access to computers with Excel limited potential users of tool; results challenging to share in intuitive format Solution: Migrated from Excel-based to mobile phone-based platform, with data visualization graphics that can be sent via WhatsApp and other messaging services |
Problem: No existing data tools to link patient records over time; high data collection needs to populate cohorts Solution: Created new mental health modules in electronic medical record; integrated cascade with electronic medical record to auto-populate |
|||||
| 8) Expanding CAT operators to include new HCW cadres | Problem: Clients served outside of a health facility setting by community-based organizations Solution: Cascade filled and operated by community-based organizations |
||||||
| 9) Additional unique challenges | Problem: Age bands to define children and adolescents not reflected in aggregated data sources Solution: Abstract data from paper registers to enable custom age bands; this solution has downside of taking time and resources |
Problem: Cascade steps may differ based on mental health diagnoses and for patients over time Solution: In process of resolving challenge |
Problem: No existing data tools for hypertension; medical management may differ based on co-morbidities Solution: Created new hypertension registers for government; medical management steps left flexible to accommodate changing guidelines and comorbidities |
Problem: Existing data sources exist but are time intensive and bias against service or representative denominator Solution: Introduce new streamlined data tool to populate cascade and separate from intensive questionnaire |