Development of the Susceptibility-Spectrum Discrepancy Index (S2DI): A novel metric for antimicrobial stewardship in hospitalised patients

This study introduces the Susceptibility-Spectrum Discrepancy Index (S2DI), a novel quantitative metric that evaluates the appropriateness of parenteral antimicrobial therapy for bloodstream infections by measuring the gap between an isolate's susceptibility and the spectrum of prescribed agents, revealing that while hospital-level variations persist, antimicrobial stewardship has improved over time in Japanese hospitals.

The Gist

The Big Picture: The "Goldilocks" Problem in Hospitals

Imagine a hospital is like a giant kitchen. When a patient gets a serious infection in their blood (bacteraemia), the doctors are the chefs. Their job is to pick the right "ingredient" (antibiotic) to kill the bad bacteria.

But there's a tricky rule:

• Too weak: If they pick a weak antibiotic, the bacteria survive, and the patient gets sicker.
• Too strong: If they pick a "super-weapon" antibiotic (one that kills almost everything), it's like using a sledgehammer to crack a nut. It might kill the bad bacteria, but it also wipes out the good bacteria in the patient's body and teaches the bad bacteria how to become "super-resistant" (like a villain leveling up). This is Antimicrobial Resistance (AMR).

The goal of Antimicrobial Stewardship is to find the "Goldilocks" antibiotic: not too weak, not too strong, but just right.

The Problem: How Do We Measure "Just Right"?

In a doctor's office, it's easy to spot bad prescribing (e.g., giving antibiotics for a cold virus). But in a hospital, it's much harder. Doctors often start with a "broad-spectrum" antibiotic (the sledgehammer) just in case, because the patient is very sick. Once lab tests come back showing exactly what bacteria is causing the infection, the doctor should switch to a "narrow-spectrum" antibiotic (the precise scalpel). This switch is called de-escalation.

The problem is: How do we know if a hospital is actually doing this switch correctly?
Until now, there wasn't a simple way to measure this across different hospitals. It was like trying to judge a cooking competition without a scoring sheet.

The Solution: The S2DI Score (The "Mismatch Meter")

The authors of this paper invented a new score called the Susceptibility–Spectrum Discrepancy Index (S2DI).

Think of the S2DI as a "Mismatch Meter."

1. Susceptibility (A): How easy is the specific bacteria to kill? (Is it a weak enemy or a tough one?)
2. Spectrum (B): How "wide" is the antibiotic being used? (Is it a tiny laser or a massive explosion?)

The score is calculated by subtracting the "ease of killing" from the "size of the weapon."

• Score of 0: Perfect match. The weapon size fits the enemy perfectly. (Great job!)
• High Score: The weapon is way too big for the enemy. (You're using a sledgehammer on a gnat.)

They calculated this score for patients 7 days after they started treatment. If the score is still high after a week, it means the hospital hasn't switched to the right, smaller antibiotic yet.

What They Did (The Recipe)

The researchers looked at data from 67 hospitals in Japan over 7 years (2017–2023). They focused on two very common "villains":

1. Staphylococcus aureus (Staph)
2. Escherichia coli (E. coli)

They gathered data on over 14,000 patients with blood infections. A team of 10 expert doctors helped create the "scoring rules" for the S2DI, ranking antibiotics from "narrow" to "broad."

What They Found (The Results)

1. The "Mismatch" is Real: Even 7 days after starting treatment, many patients were still on antibiotics that were too strong for their specific bacteria. The average "Mismatch Meter" score wasn't zero.
2. Hospitals are Different: Just like some schools have better reading programs than others, some hospitals are much better at switching to the right antibiotics than others. The study showed a huge gap between the best hospitals and the ones that need help.
3. Things are Getting Better: Over the years (2017 to 2023), the scores improved. Hospitals are slowly learning to use the "scalpel" instead of the "sledgehammer."
4. Who Gets the Best Care? Interestingly, younger patients and women with E. coli infections tended to get better-matched antibiotics. Also, hospitals with a dedicated "Infectious Disease" department did better with Staph infections.
5. No Harm Done: Crucially, patients who got the "perfect match" (low score) didn't die more often than those who got the "too strong" antibiotics. This proves that being precise doesn't hurt the patient; it actually helps the future fight against superbugs.

Why This Matters

This study is like giving every hospital a report card for their antibiotic use.

• Before: Hospitals didn't really know how they compared to their neighbors.
• Now: They have a simple number (the S2DI) that tells them, "Hey, you're using too many heavy-duty antibiotics. Try to narrow it down."

The Bottom Line

The authors created a simple tool to help hospitals stop overusing powerful antibiotics. By measuring the gap between the bacteria's weakness and the antibiotic's strength, they can spot which hospitals need to improve.

The good news? Hospitals are getting better at this every year. The bad news? There is still a lot of room for improvement, and some hospitals are lagging behind. But now, we have a map to help them catch up.

In short: We found a way to measure if doctors are using the right size of antibiotic hammer, and we found that while we are getting better, we still need to stop swinging sledgehammers at tiny bugs.

Technical Summary

1. Problem Statement

Antimicrobial resistance (AMR) is a critical global health threat. While antimicrobial stewardship (AMS) is fundamental to controlling AMR, evaluating the appropriateness of parenteral antimicrobial use in hospitalized patients remains challenging.

• Complexity of Assessment: Unlike ambulatory care (where inappropriate prescribing is often linked to viral upper respiratory infections), hospital settings involve complex comorbidities and empiric therapy.
• The De-escalation Challenge: Standard practice involves initiating broad-spectrum empiric therapy and "de-escalating" to narrow-spectrum agents once culture results are available. However, there is a lack of quantitative, population-level indicators to assess whether this de-escalation is actually occurring effectively at the facility level.
• Gap in Metrics: Existing metrics often focus on prevalence or resistance rates but fail to compare the appropriateness of prescribing behavior across different healthcare facilities.

2. Methodology

The study employed a retrospective cohort design to develop and validate a new metric, the Susceptibility–Spectrum Discrepancy Index (S2DI).

• Data Source: A database from 67 National Hospital Organization (NHO) hospitals across Japan covering the period 2017–2023.
• Study Population: Patients with confirmed Staphylococcus aureus or Escherichia coli bacteremia.
  • Exclusion Criteria: Cases with multiple bacteria, or patients discharged/died within 7 days of blood culture collection (to ensure sufficient time for de-escalation assessment).
  • Final Cohort: 4,505 S. aureus cases and 9,563 E. coli cases.
• The S2DI Scoring System:
  • Developed by an expert panel of 10 infectious disease physicians.
  • Component A (Susceptibility): Ranked based on the susceptibility profile of the isolated bacterium (3 levels for S. aureus; 5 levels for E. coli).
  • Component B (Spectrum Width): Ranked based on the spectrum width of the prescribed antimicrobial (4 levels for S. aureus; 6 levels for E. coli).
  • Calculation: $S2DI = B - A$.
  • Interpretation: Calculated on Day 7 after treatment initiation. A score closer to zero indicates higher appropriateness (narrow-spectrum agent matched to susceptibility). Higher positive values indicate overuse of broad-spectrum agents.
• Statistical Analysis:
  • Used Linear Mixed-Effects Models with hospital-level random effects to account for inter-facility variability.
  • Fixed Effects: Calendar year, facility size (<400 vs. ≥400 beds), region, department type (Infectious Diseases, Pediatrics, Psychiatry), and patient demographics (sex, age, Charlson Comorbidity Index).
  • Visualization: Used caterpillar plots and snail plots to visualize facility-level random intercepts (BLUPs).

3. Key Contributions

• Novel Metric (S2DI): Introduced the first quantitative index specifically designed to evaluate the appropriateness of parenteral antimicrobial use after de-escalation is theoretically possible (Day 7).
• Facility-Level Benchmarking: Unlike population-level surveillance, S2DI allows for the comparison of stewardship performance across individual hospitals, identifying outliers with suboptimal prescribing practices.
• Temporal Analysis: Demonstrated the utility of the metric in tracking improvements in antimicrobial use over time at a national level.

4. Key Results

• Baseline Appropriateness:
  • Median S2DI was 1 (IQR 0–1) for S. aureus and 2 (IQR 0–3) for E. coli.
  • The non-zero medians indicate that even 7 days post-treatment, many patients were still receiving broader-spectrum antibiotics than necessary.
• Temporal Trends:
  • There was a statistically significant association between later calendar years and more favorable (lower) S2DI scores for both pathogens ($p < 0.001$ for E. coli; $p < 0.002$ for S. aureus), suggesting a gradual improvement in stewardship practices in Japan.
• Patient and Facility Factors:
  • E. coli: Female sex and younger age were associated with more appropriate therapy.
  • S. aureus: Facilities with a dedicated Department of Infectious Diseases showed significantly more favorable S2DI scores ($p = 0.008$).
• Inter-Facility Variability:
  • Significant variation in prescribing behavior existed between hospitals.
  • Random intercept variance was 0.168 for E. coli and 0.053 for S. aureus, confirming that facility-specific factors heavily influence prescribing patterns.
• Clinical Outcomes:
  • No significant association was found between lower S2DI scores (better stewardship) and increased 30-day all-cause mortality, supporting the safety of de-escalation strategies.

5. Significance and Implications

• Policy and Practice: The S2DI provides a simple, actionable tool for hospital administrators and policymakers to monitor and benchmark antimicrobial stewardship programs. It moves beyond simple volume metrics to assess quality of prescribing.
• AMR Countermeasures: The study validates that national action plans are yielding results, as stewardship has improved over the 7-year study period.
• Targeted Intervention: By identifying facilities with consistently high S2DI scores (indicating overuse of broad-spectrum drugs), health authorities can prioritize resources and interventions for specific institutions rather than applying blanket policies.
• Limitations: The metric is currently limited to S. aureus and E. coli bacteremia and relies on retrospective data where disease severity and allergy history were not fully captured. However, given that these two pathogens account for a majority of the AMR burden in Japan, the metric offers a robust starting point for broader evaluation.

In conclusion, the S2DI is a validated, quantitative framework that successfully quantifies the gap between bacterial susceptibility and antibiotic spectrum, offering a critical tool for optimizing parenteral antimicrobial use in hospitalized patients.