How to Use STOPP/START Criteria in Primary Care
STOPP/START criteria should be systematically applied during medication reviews for all patients ≥65 years old, particularly those with polypharmacy (≥5 medications), using either manual review or computerized clinical decision support systems to identify potentially inappropriate medications and prescribing omissions. 1
Application Method
The STOPP/START tool functions as a two-part screening system 2:
- STOPP (Screening Tool of Older Persons' Prescriptions): Identifies 80+ potentially inappropriate medications to avoid or discontinue
- START (Screening Tool to Alert doctors to Right Treatment): Identifies 34 prescribing omissions where indicated medications are missing
Step-by-Step Implementation
1. Target Population Identification
- Apply to all patients aged ≥65 years 2
- Prioritize those with polypharmacy (≥5 drugs) or multimorbidity 3, 4
- Essential at care transitions (hospital discharge, nursing home admission) 1
2. Medication Review Process Integrate STOPP/START into your systematic polypharmacy assessment 1:
- Perform medication reconciliation first
- Apply STOPP criteria to identify inappropriate medications
- Apply START criteria to detect prescribing omissions
- Document findings and create deprescribing plan
3. Most Common PIMs to Screen For
The highest-yield STOPP criteria in primary care 5, 4, 6, 7:
- Benzodiazepines >28 days (most prevalent - found in 25.9% of elderly patients)
- NSAIDs in patients with heart failure, CKD, or hypertension
- Opioids without laxatives (13.8% prevalence)
- Antiplatelet/anticoagulant combinations without clear indication
- Drug duplications (5.7% prevalence, especially benzodiazepines and NSAIDs)
- Anticholinergics (diphenhydramine, oxybutynin, cyclobenzaprine)
- Sulfonylureas in patients with renal impairment
4. Most Common START Omissions
Priority prescribing gaps to address 5, 6:
- Vitamin D and calcium supplements in osteoporosis
- Antidepressants for depression
- Cardiovascular medications (statins in CAD, antiplatelet after stenting)
- Appropriate preventive therapies based on comorbidities
Frequency of Use in Primary Care
The reality is that STOPP/START criteria are underutilized in routine primary care despite their proven value. 8
Current Usage Patterns
- Prevalence of inappropriate prescribing detected: 37.5-89.6% of elderly patients when STOPP/START is applied 5, 4, 6
- Actual implementation rate: Low in daily practice without systematic integration 8
- Clinical impact when used: Independently predicts emergency department visits and hospitalizations at 6 months 3
Barriers to Frequent Use
GPs identify several obstacles 8:
- Time-consuming manual application (80 STOPP + 34 START criteria to review)
- Lack of computerized clinical decision support integration
- Questions about comprehensiveness and relevance to specific patients
- Need for clinical judgment on implicit criteria (drugs without indication, treatment duration)
Facilitators for Implementation
To increase usage frequency, the following strategies are essential 8:
- Computerized clinical decision support systems embedded in electronic health records
- Automated algorithms (RStudio® or similar platforms can screen thousands of patients) 7
- Multidisciplinary collaboration with clinical pharmacists 1
- Targeted application at high-risk transitions (hospital discharge, nursing home placement)
- Education and training programs for primary care teams
Clinical Outcomes
When properly applied, STOPP/START criteria reduce adverse outcomes 3:
- Decreased emergency department visits
- Reduced hospitalizations (independent of age, frailty, and total medication number)
- Lower medication-associated adverse events
Important Caveats
Inter-rater reliability varies significantly 9:
- Explicit criteria (specific drug-disease combinations): 56% positive agreement, kappa 0.56
- Implicit criteria (drugs without indication, excessive duration): Only 16% positive agreement, kappa 0.09
- This means clinical judgment remains essential, particularly for criteria A1 (drugs without indication) and A2 (treatment beyond recommended duration)
Tool limitations 10:
- Developed primarily for European populations
- Low concordance with other tools like Beers criteria (kappa 0.093)
- May require adaptation for specific populations or healthcare systems
- 20 criteria account for 80% of PIMs, suggesting focused screening may be practical 4
Practical Recommendation
Begin with high-yield screening: Focus on the 6-8 most common STOPP criteria (benzodiazepines, NSAIDs, opioids without laxatives, anticholinergics, drug duplications) and 3-4 START criteria (vitamin D/calcium, cardiovascular prevention, antidepressants) for patients with polypharmacy. This captures the majority of clinically significant inappropriate prescribing while remaining time-efficient. 4, 7