LATAM-RX — Context-Aware Rheumatology Risk Adjustment for Latin America

Abstract

LATAM-RX is a context-aware rheumatology risk-adjustment skill that incorporates regional infectious burden, monitoring constraints, treatment access, insurance formulary availability, and population-specific considerations relevant to Latin American practice. Its purpose is not to redefine disease biology, but to improve clinical realism in decision support under real-world constraints.

Clinical Rationale

Most clinical decision-support tools in rheumatology are developed for high-income healthcare systems with full laboratory access, specialist availability, and broad biologic formularies. Latin American practice operates under different structural constraints: higher TB burden, limited biologic access in public insurance, diagnostic delays, endemic infections requiring pre-treatment screening, and steroid self-medication culture. This skill adjusts risk context for these realities.

Methodology

Four adjustment domains:

1. TB risk modifier: country-level incidence from WHO 2023
2. Access fragility: urban/rural, specialist availability, lab access, diagnostic delay, self-medication
3. Biologic access: IMSS/ISSSTE/private formulary lookup for specific drug
4. Endemic infection flags: Chagas, Strongyloides, Histoplasma, Coccidioides screening

Composite fragility = 35% access + 30% TB + 20% self-medication + 15% specialist availability.

Limitations

• Context-aware heuristic; not a validated epidemiologic risk engine
• Insurance formulary data is approximate and may not reflect current procurement
• TB incidence data from WHO 2023 estimates; local variation exists
• Endemic infection screening recommendations are general guidance
• Designed to improve clinical realism, not to replace local guidelines

References

1. Pons-Estel BA, et al. GLADEL multinational Latin American prospective inception cohort of 1,214 SLE patients. Medicine. 2004;83(1):1-17. DOI: 10.1097/01.md.0000104742.42401.e2
2. Ugarte-Gil MF, et al. Socioeconomic differences in SLE outcomes from GLADEL-PANLAR. Arthritis Care Res. 2022;74(2):203-210. DOI: 10.1002/acr.24447
3. Pelaez-Ballestas I, et al. Epidemiology of rheumatic diseases in Mexico: COPCORD. J Rheumatol Suppl. 2011;86:3-8. DOI: 10.3899/jrheum.100952
4. WHO. Global Tuberculosis Report 2023. Geneva: WHO; 2023.
5. COFEPRIS. Listado de medicamentos biotecnologicos aprobados en Mexico. 2025.

Executable Code

#!/usr/bin/env python3
"""
LATAM-RX — Context-Aware Rheumatology Risk Adjustment for Latin America.

Applies regional modifiers for infectious burden, monitoring constraints,
treatment access, and population-specific considerations relevant to
Latin American rheumatology practice.

Authors: Zamora-Tehozol EA (ORCID:0000-0002-7888-3961), DNAI
License: MIT

References:
- Pons-Estel BA, et al. GLADEL multinational Latin American prospective
  inception cohort of 1,214 patients with systemic lupus erythematosus:
  ethnic and disease heterogeneity among "Hispanics". Medicine. 2004;
  83(1):1-17. DOI:10.1097/01.md.0000104742.42401.e2
- Ugarte-Gil MF, et al. Socioeconomic-related differences in clinical
  outcomes and mortality in SLE patients from the GLADEL-PANLAR
  multinational cohort. Arthritis Care Res. 2022;74(2):203-210.
  DOI:10.1002/acr.24447
- Peláez-Ballestas I, et al. Epidemiology of the rheumatic diseases in
  Mexico: a study of 5 regions based on the COPCORD methodology.
  J Rheumatol Suppl. 2011;86:3-8. DOI:10.3899/jrheum.100952
- WHO. Global Tuberculosis Report 2023. Geneva: World Health Organization; 2023.
- COFEPRIS. Listado de medicamentos biotecnológicos aprobados en México. 2025.
"""

from dataclasses import dataclass, asdict
from typing import Dict, Any, List
import json


# Regional TB incidence per 100,000 (WHO 2023 estimates)
TB_INCIDENCE = {
    "mexico": 23, "brazil": 40, "peru": 123, "argentina": 25,
    "colombia": 25, "chile": 14, "bolivia": 92, "guatemala": 21,
    "venezuela": 28, "ecuador": 47, "default_latam": 35,
}

# Simplified biologic access tiers
BIOLOGIC_ACCESS = {
    "imss": {"rituximab": True, "tocilizumab": True, "adalimumab": True,
             "etanercept": True, "infliximab": True, "abatacept": True,
             "secukinumab": False, "upadacitinib": False, "anifrolumab": False,
             "belimumab": False, "baricitinib": False, "tofacitinib": True},
    "issste": {"rituximab": True, "tocilizumab": True, "adalimumab": True,
               "etanercept": True, "infliximab": True, "abatacept": True,
               "secukinumab": False, "upadacitinib": False, "anifrolumab": False,
               "belimumab": False, "baricitinib": False, "tofacitinib": True},
    "private": {"rituximab": True, "tocilizumab": True, "adalimumab": True,
                "etanercept": True, "infliximab": True, "abatacept": True,
                "secukinumab": True, "upadacitinib": True, "anifrolumab": True,
                "belimumab": True, "baricitinib": True, "tofacitinib": True},
    "no_insurance": {"rituximab": False, "tocilizumab": False, "adalimumab": False,
                     "etanercept": False, "infliximab": False, "abatacept": False,
                     "secukinumab": False, "upadacitinib": False, "anifrolumab": False,
                     "belimumab": False, "baricitinib": False, "tofacitinib": False},
}


@dataclass
class LatAmContext:
    country: str
    insurance_type: str  # imss, issste, private, no_insurance
    urban_rural: str     # urban, semi_urban, rural
    has_rheumatologist: bool
    lab_access: str      # full, basic, minimal
    biologic_needed: str # drug name or "none"
    endemic_infections: List[str]  # chagas, strongyloides, histoplasma, etc.
    diagnostic_delay_months: float
    steroid_self_medication: bool


def clamp(x, lo=0.0, hi=1.0):
    return max(lo, min(hi, x))


def tb_risk_modifier(country: str) -> float:
    inc = TB_INCIDENCE.get(country.lower(), TB_INCIDENCE["default_latam"])
    if inc >= 80: return 1.0
    if inc >= 40: return 0.7
    if inc >= 20: return 0.4
    return 0.2


def access_fragility(ctx: LatAmContext) -> float:
    score = 0.0
    if ctx.urban_rural == "rural": score += 1.5
    elif ctx.urban_rural == "semi_urban": score += 0.7
    if not ctx.has_rheumatologist: score += 1.3
    if ctx.lab_access == "minimal": score += 1.2
    elif ctx.lab_access == "basic": score += 0.5
    if ctx.diagnostic_delay_months >= 24: score += 1.0
    elif ctx.diagnostic_delay_months >= 12: score += 0.5
    if ctx.steroid_self_medication: score += 0.8
    return clamp(score / 5.8)


def biologic_access_flag(ctx: LatAmContext) -> Dict[str, Any]:
    tier = BIOLOGIC_ACCESS.get(ctx.insurance_type, BIOLOGIC_ACCESS["no_insurance"])
    drug = ctx.biologic_needed.lower()
    if drug == "none":
        return {"drug": "none", "available": None, "barrier": None}
    available = tier.get(drug, False)
    barrier = None if available else f"{drug} not in {ctx.insurance_type} formulary"
    return {"drug": drug, "available": available, "barrier": barrier}


def endemic_infection_flags(ctx: LatAmContext) -> List[str]:
    flags = []
    for inf in ctx.endemic_infections:
        inf_lower = inf.lower()
        if "chagas" in inf_lower:
            flags.append("Screen for Chagas disease before immunosuppression (T. cruzi serology)")
        if "strongyloid" in inf_lower:
            flags.append("Screen for Strongyloides before high-dose steroids or biologics (ivermectin pre-treatment)")
        if "histoplasm" in inf_lower:
            flags.append("Consider histoplasma antigen/antibody before TNF inhibitors in endemic area")
        if "coccidioid" in inf_lower:
            flags.append("Coccidioidomycosis risk — serology before biologics in endemic zone")
    return flags


def run_latam_rx(ctx: LatAmContext) -> Dict[str, Any]:
    tb_mod = tb_risk_modifier(ctx.country)
    access = access_fragility(ctx)
    bio_flag = biologic_access_flag(ctx)
    endemic_flags = endemic_infection_flags(ctx)
    
    composite_fragility = round(0.35*access + 0.30*tb_mod + 0.20*(1.0 if ctx.steroid_self_medication else 0.0) + 0.15*(0.0 if ctx.has_rheumatologist else 1.0), 3)
    
    band = "high" if composite_fragility >= 0.60 else ("intermediate" if composite_fragility >= 0.35 else "lower")
    
    return {
        "input": asdict(ctx),
        "tb_risk_modifier": round(tb_mod, 3),
        "access_fragility": round(access, 3),
        "biologic_access": bio_flag,
        "endemic_infection_flags": endemic_flags,
        "composite_fragility": composite_fragility,
        "fragility_band": band,
        "clinical_adjustments": [
            "Reinforce TB screening (IGRA/TST) before any biologic/JAK initiation." if tb_mod >= 0.4 else None,
            "Monitor for steroid complications — self-medication history present." if ctx.steroid_self_medication else None,
            "Limited lab access — adjust monitoring frequency expectations." if ctx.lab_access in ("basic","minimal") else None,
            f"Diagnostic delay of {ctx.diagnostic_delay_months} months — assess for accumulated damage." if ctx.diagnostic_delay_months >= 12 else None,
            "No specialist access — consider telemedicine or referral pathways." if not ctx.has_rheumatologist else None,
        ],
        "limitations": [
            "Context-aware heuristic; not a validated epidemiologic risk engine.",
            "Insurance formulary data is approximate and may not reflect current procurement.",
            "TB incidence data from WHO 2023 estimates; local variation exists.",
            "Endemic infection screening recommendations are general guidance.",
            "Designed to improve clinical realism under resource constraints, not to replace local guidelines."
        ]
    }
    # Clean nulls from adjustments
    


if __name__ == "__main__":
    demo = LatAmContext(
        country="mexico",
        insurance_type="imss",
        urban_rural="semi_urban",
        has_rheumatologist=False,
        lab_access="basic",
        biologic_needed="anifrolumab",
        endemic_infections=["strongyloides", "histoplasma"],
        diagnostic_delay_months=18,
        steroid_self_medication=True,
    )
    print("=" * 70)
    print("LATAM-RX — Context-Aware Rheumatology Risk Adjustment")
    print("Authors: Zamora-Tehozol EA (ORCID:0000-0002-7888-3961), DNAI")
    print("=" * 70)
    result = run_latam_rx(demo)
    # Clean None from adjustments
    result["clinical_adjustments"] = [a for a in result["clinical_adjustments"] if a]
    print(json.dumps(result, indent=2))

Demo Output

======================================================================
LATAM-RX — Context-Aware Rheumatology Risk Adjustment
Authors: Zamora-Tehozol EA (ORCID:0000-0002-7888-3961), DNAI
======================================================================
{
  "input": {
    "country": "mexico",
    "insurance_type": "imss",
    "urban_rural": "semi_urban",
    "has_rheumatologist": false,
    "lab_access": "basic",
    "biologic_needed": "anifrolumab",
    "endemic_infections": [
      "strongyloides",
      "histoplasma"
    ],
    "diagnostic_delay_months": 18,
    "steroid_self_medication": true
  },
  "tb_risk_modifier": 0.4,
  "access_fragility": 0.655,
  "biologic_access": {
    "drug": "anifrolumab",
    "available": false,
    "barrier": "anifrolumab not in imss formulary"
  },
  "endemic_infection_flags": [
    "Screen for Strongyloides before high-dose steroids or biologics (ivermectin pre-treatment)",
    "Consider histoplasma antigen/antibody before TNF inhibitors in endemic area"
  ],
  "composite_fragility": 0.699,
  "fragility_band": "high",
  "clinical_adjustments": [
    "Reinforce TB screening (IGRA/TST) before any biologic/JAK initiation.",
    "Monitor for steroid complications \u2014 self-medication history present.",
    "Limited lab access \u2014 adjust monitoring frequency expectations.",
    "Diagnostic delay of 18 months \u2014 assess for accumulated damage.",
    "No specialist access \u2014 consider telemedicine or referral pathways."
  ],
  "limitations": [
    "Context-aware heuristic; not a validated epidemiologic risk engine.",
    "Insurance formulary data is approximate and may not reflect current procurement.",
    "TB incidence data from WHO 2023 estimates; local variation exists.",
    "Endemic infection screening recommendations are general guidance.",
    "Designed to improve clinical realism under resource constraints, not to replace local guidelines."
  ]
}