A comprehensive analysis of 204 countries and territories using GBD 2021 data reveals a complex picture of progress, persistent inequality, and urgent priorities for health systems worldwide.
Medical treatment saves millions of lives every year — but it also causes harm. Adverse effects of medical treatment (AEMT) encompass everything from medication errors and surgical complications to device failures and misadministration. Until now, no global study had quantified this burden comprehensively across time, geography, and socioeconomic context. A major new study published in the Korean Journal of Internal Medicine, produced by the GBD 2021 Collaborator Network, fills this gap with the most detailed and current longitudinal analysis available.
KEY NUMBERS AT A GLANCE — 2021 GLOBAL ESTIMATES
| 150.44 per 100,000 Global Incidence Rate | 64.19 per 100,000 DALY Rate (2021) | 1.53 per 100,000 Mortality Rate (2021) | –36% 1990–2021 DALY Rate Decline |
KEY FINDINGS
Declining DALY & Mortality Rates — But Progress Is Uneven: Age-standardized DALY rates fell from 106.49 to 64.19 per 100,000 between 1990 and 2021, and mortality rates dropped from 2.40 to 1.53. This represents meaningful progress — yet the gains are concentrated in high-SDI quintiles, while low-income regions lag substantially behind.
Healthcare Overutilization Drives AEMT in High-Income Settings: Older adults in high-SDI countries show the highest prevalence and incidence of AEMT across all age groups. The HAQ Index correlates positively with incidence (r = 0.40, p < 0.0001), pointing to healthcare overuse — rather than poor quality — as the primary mechanism in wealthy nations.
Neonates Bear a Disproportionate Global Burden: DALY rates peak in the early neonatal group at 4,789 per 100,000 — more than 70× the global average. Mortality-incidence ratios (MIRs) in this group are up to 49× higher in low-SDI settings compared to high-SDI regions, and declined far more slowly between 1990 and 2021 (≈12% vs. ≈73% reduction).
Better Healthcare Quality Reduces Fatality, Not Exposure: While HAQ Index is positively correlated with AEMT incidence (r = 0.40), it is strongly and negatively correlated with DALY rates (r = –0.66) and mortality (r = –0.59). Better access and quality do not prevent the occurrence of AEMT, but dramatically improve survival and reduce long-term disability.
Global MIR Declined 33% — Greatest Gains in Neonatal Care: The overall MIR fell by 33.11% globally from 1990 to 2021. The largest reductions occurred in neonatal groups in high- and high-middle-SDI quintiles (up to 72.8%). By contrast, MIR among the oldest populations (95+) has remained nearly unchanged, signalling an unresolved challenge in late-life care.
WHY THIS MATTERS FOR HEALTH MANAGEMENT PROFESSIONALS
For hospital managers, policymakers, and quality officers, this study provides actionable intelligence at a level of granularity previously unavailable. The findings reframe AEMT not merely as a clinical problem but as a systemic and organizational one — shaped by insurance structures, clinician incentive design, primary care density, and safety culture. The multi-level intervention framework proposed by the authors maps directly onto Donabedian’s structure–process–outcome model, positioning this study as a natural anchor point for quality improvement programmes worldwide.
Notably, the data suggest that one-fifth to one-third of patients in high-income settings receive unnecessary healthcare services. Addressing this requires both patient-side reforms (such as cost-sharing adjustments) and clinician-side incentives (equitable compensation structures that do not reward low-value care). At the population level, increasing the ratio of primary care physicians to specialists is flagged as a practical lever for reducing healthcare overuse and its downstream harms.
DIRECTIONS FOR FUTURE RESEARCH
AEMT Subtype Disaggregation: The global analysis necessarily pools all adverse effect categories. Future work using harmonized registries should separate adverse drug events, surgical complications, device-related harm, and medical management failures — enabling subtype-specific policy responses.
Severity & Preventability Stratification: Knowing the magnitude of AEMT is essential; knowing what proportion is preventable and how severe each episode is would allow resources to be allocated far more precisely. A global taxonomy aligned with national trigger-tool frameworks would be transformative.
Healthcare System Archetype Comparisons: Variations in insurance models, reimbursement structures, and care delivery architectures within the same SDI quintile remain unexplored. Comparative analyses across single-payer, multi-payer, and out-of-pocket-dominant systems would clarify which institutional features are protective.
Dedicated Pediatric & Neonatal Cohort Studies: Given the disproportionate DALY burden in under-2 populations and the widespread use of off-label prescribing in neonates, dedicated global studies incorporating gestational age, birth weight, and formulary access variables are urgently needed.
Economic Burden Quantification: Linking AEMT estimates to direct healthcare costs, productivity losses, and catastrophic expenditure data would translate the epidemiological burden into the language of health financing — critical for budget impact analyses and priority-setting.
Reporting Quality & ICD Coding Accuracy: Healthcare providers are often reluctant to code events as AEMT. Studies examining the gap between observed rates and true incidence — using electronic trigger tools or NLP on clinical records — would sharpen global estimates and reduce systematic underreporting bias.
Original Article: GBD 2021 AEMT Collaborators. Global burden of adverse effects of medical treatment from 1990 to 2021: a Global Burden of Disease Study 2021. Korean J Intern Med. 2026;41:350–366. https://doi.org/10.3904/kjim.2025.278